重视网站商务通,中小型网站建设与管理 唐军民,seo课程总结怎么写,广州app开发软件Remote Sensing Object Detection Meets Deep Learning#xff1a; A Meta-review of Challenges and Advances 遥感目标检测与深度学习的相遇#xff1a;挑战与进展的元综述 论文链接
0.论文摘要和作者信息
摘要
遥感目标检测#xff08;RSOD#xff09;是遥感领域最基…Remote Sensing Object Detection Meets Deep Learning A Meta-review of Challenges and Advances 遥感目标检测与深度学习的相遇挑战与进展的元综述 论文链接
0.论文摘要和作者信息
摘要
遥感目标检测RSOD是遥感领域最基础和最具挑战性的任务之一长期以来一直受到人们的关注。近年来深度学习技术展示了强大的特征表示能力并导致了RSOD技术发展的巨大飞跃。在这个技术快速发展的时代这篇综述旨在全面回顾基于深度学习的RSOD方法的最新成就。这篇综述涵盖了300多篇论文。我们确定了RSOD中的五个主要挑战包括多尺度目标检测、旋转目标检测、弱目标检测、微小目标检测和有限监督下的目标检测并以分层划分的方式系统地回顾了相应的方法。我们还回顾了RSOD领域广泛使用的基准数据集和评估指标以及RSOD的应用场景。为进一步推进RSOD的研究提供了未来的研究方向。
索引术语-目标检测、遥感图像、深度学习、技术演进
作者信息
Xiangrong Zhang, Tianyang Zhang, Guanchun Wang, Peng Zhu, Xu Tang, and Licheng Jiao are with the School of Artificial Intelligence, Xidian University, Xi’an 710071, China (e-mail: xrzhangmail.xidian.edu.cn). Xiuping Jia is with the School of Engineering and Information Technology, University of New South Wales, Canberra, ACT 2612, Australia.
1.研究背景
随着地球观测技术的快速进步遥感卫星如Google Earth[1]、WordWide-3[2]和Gaofen系列卫星[3]-[5]在空间、时间和光谱分辨率方面取得了显著提高现在可以获得大量遥感图像。受益于可用RSI的急剧增加人类进入了一个遥感大数据时代RSI的自动解译成为一个活跃的产量挑战性课题[6]-[8]。
RSOD旨在确定给定RSI中是否存在感兴趣的目标并返回每个预测目标的类别和位置。本调查中的术语“目标”是指人造或高度结构化的目标如飞机、车辆和船舶而不是非结构化的场景目标如土地、天空和草地。作为RSI自动解释的基石RSOD受到了极大的关注。
一般来说RSI是在具有不同地面采样距离GSD的头顶视点拍摄的并且覆盖了地球表面的广泛区域。结果地理空间目标在规模上表现出更显著的多样性角度和外观。基于RSI中地理空间目标的特点我们总结了RSOD在以下五个方面的主要挑战
1巨大的尺度变化。一方面不同类别的目标通常存在巨大的比例变化如图1(b)所示车辆可以小至10个像素区域而飞机可以比车辆大20倍。另一方面类别内目标也表现出广泛的尺度。因此检测模型需要处理大规模和小规模目标。
2任意取向。独特的头顶视点导致地理空间目标通常以任意方向分布如图1c所示。这种旋转目标检测任务加剧了RSOD的挑战使得检测器能够感知方向非常重要。
3弱特征响应。通常RSI包含复杂的上下文和大量的背景噪声。如图1a所示一些车辆被阴影遮挡并且周围的背景噪声往往具有与车辆相似的外观。这种复杂的干扰可能会淹没感兴趣的目标并恶化它们的特征表示这导致感兴趣的目标被呈现为弱的特征响应[9]。
4微小目标。如图1d所示微小目标往往表现出极小的尺度和有限的外观信息导致质量差的特征表示。此外当前流行的检测范式不可避免地削弱甚至丢弃了微小目标的表示[10]。微小目标检测中的这些问题给现有的检测方法带来了新的困难。
5昂贵的注释。地理空间目标在尺度和角度方面的复杂特征以及细粒度注释所需的专家知识[11]使得RSI的精确框级注释成为一项耗时耗力的任务。然而当前基于深度学习的检测器严重依赖于丰富的标记良好的数据来达到性能饱和。因此在缺乏足够监督信息的情况下有效的RSOD方法仍然具有挑战性。 图1。遥感图像的典型示例。a复杂的上下文和大量的背景噪声导致目标的特征响应较弱。b类别间和类别内目标都存在巨大的尺度差异。c目标以任意方向分布。d微小目标往往表现出极小的尺度。
为了应对这些挑战在过去二十年中出现了许多RSOD方法。早期研究人员采用模板匹配[12]-[14]和先验知识[15]-[17]进行遥感场景中的目标检测。这些早期的方法更多地依赖于手工制作的模板或先验知识导致结果不稳定。
后来机器学习方法[18]–[21]已经成为RSOD的主流它将目标检测视为一项分类任务。具体地机器学习模型首先从输入图像中搜索一组目标建议并提取这些目标建议的纹理、上下文和其他特征。然后它采用独立的分类器来识别这些目标建议中的目标类别。然而来自机器学习方法的基于浅层学习的特征显著限制了目标的表示尤其是在更具挑战性的场景中。此外基于机器学习的目标检测方法不能以端到端的方式训练这在遥感大数据时代不再适用。
最近深度学习技术[22]已经从海量数据中展示了强大的特征表示能力计算机视觉中最先进的检测器[23]-[26]实现了与人类相媲美的目标检测能力[27]。利用深度学习技术的先进进展各种基于深度学习的方法已经主导了RSOD并导致了检测性能的显著突破。与传统方法相比深度神经网络架构可以提取高级语义特征并获得更鲁棒的目标特征表示。此外高效的端到端训练方式和自动化的特征提取方式使得基于深度学习的目标检测方法更适合遥感大数据时代的RSOD。
随着RSOD的流行近年来发表了许多地理空间目标检测调查[9]、[28]-[34]。例如Cheng等人[29]回顾了RSOD的早期发展。韩等人[9]重点研究了RSI中的小目标和弱目标检测。在[30]中作者回顾了飞机检测方法。李等人[31]根据各种改进策略对遥感界基于深度学习的探测器进行了彻底的调查。此外一些工作[28], [33], [34]主要集中在发布RSOD的新基准数据集并简要回顾了遥感领域的目标检测方法。与以往的工作相比本调查基于地理空间目标的特征全面分析了RSOD中的主要挑战并根据这些挑战对基于深度学习的遥感目标检测器进行了系统的分类和总结。此外这项工作回顾了300多篇关于RSOD的论文从而进行了更全面和系统的调查。
图2示出了本综述中目标检测方法的分类。根据RSOD中的主要挑战我们将当前基于深度学习的RSOD方法分为五大类多尺度目标检测、旋转目标检测、弱目标检测、微小目标检测和有限监督下的目标检测。在每个类别中我们根据针对特定类别挑战设计的改进策略或学习范式进一步总结子类别。对于多尺度目标检测我们主要回顾了三种广泛使用的方法数据增强策略、多尺度特征表示和高质量多尺度锚点生成。关于旋转目标检测我们主要关注旋转检测框表示和旋转不敏感特征学习。对于弱目标检测我们将其分为两类背景噪声抑制和相关上下文挖掘。对于微小目标检测我们将其细化为三个流判别特征提取、超分辨率重建和改进的检测度量。根据学习范式我们将有限监督下的目标检测分为弱监督目标检测、半监督目标检测和少样本目标检测。值得注意的是每个子类别中仍然有详细的划分如图2中的圆角矩形所示。这种层次划分提供了对现有方法的系统回顾和总结。它有助于研究人员更全面地了解RSOD并促进进一步的进展这是本综述的主要目的。 图2。本文综述了基于深度学习的RSOD方法的结构化分类。采用层次划分来详细描述每个子类别。
综上所述本次审查的主要贡献如下
•我们根据地理空间目标的特征全面分析了RSOD中的主要挑战包括巨大的尺度变化、任意的方向、弱的特征响应、微小的目标和昂贵的注释。 •我们系统地总结了遥感界中基于深度学习的目标检测器并根据他们的动机。 •我们对RSOD的未来研究方向进行了前瞻性讨论以激励RSOD的进一步发展。
2.多尺度目标检测
由于RSIs之间不同的空间分辨率巨大的尺度变化是RSOD中众所周知的具有挑战性的问题并严重降低了检测性能。如图3所示我们展示了DOTAv2.0数据集中每个类别的目标像素区域的分布[33]。显然不同类别之间的比例差异很大其中小型车辆可能仅包含小于10个像素区域而机场超过 1 0 5 10^5 105个像素区域。更糟糕的是巨大的类别内尺度变化进一步加剧了多尺度目标检测的困难。为了解决巨大的尺度变化问题目前的研究主要分为数据扩充、多尺度特征表示和多尺度锚点生成。图4给出了多尺度目标检测方法的简要概述。 图3.DOTAV2.0数据集中每个类别的比例变化类别的简称可参考[33]。类别间和类别内都存在巨大的尺度差异。
图4。多尺度目标检测方法的简要总结。
A.数据扩充
数据扩充是一种简单但广泛应用的增加数据集多样性的方法。对于多尺度目标检测中的尺度变化问题图像缩放是一种直接有效的增强方法。赵等人[35]将多尺度图像金字塔馈送到多个网络中并融合这些网络的输出特征以生成多尺度特征表示。在[36]中Azimi等人提出了一种组合图像级联和特征金字塔网络来提取各种尺度上的目标特征。虽然图像金字塔可以有效地提高对多尺度目标的检测性能但严重增加了推理时间和计算复杂度。为了解决这个问题Shamsolmoali等人[37]设计了一种轻量级图像金字塔模块LIPM。所提出的LIPM接收多个下采样图像以生成多尺度特征图并将输出的多尺度特征图与来自主干的相应尺度特征图融合。此外一些现代数据增强方法例如Moscia和Stitcher[38]在多尺度目标检测中也显示出显著的有效性尤其是对于小目标[39]–[41]。
B.多尺度特征表示
RSOD的早期研究通常利用主干的最后一个单一特征图来检测目标如图5a所示。然而这种单尺度特征图预测限制了检测器处理具有宽尺度范围的目标[42]–[44]。因此多尺度特征表示方法被提出并成为RSOD中巨大目标尺度变化问题的有效解决方案。目前的多尺度特征表示方法主要分为三股多尺度特征集成、金字塔特征层次和特征金字塔网络。 图5。单尺度特征表示和多尺度特征表示的六种范例。a单尺度特征表示。b多尺度特征集成。c金字塔特征层次。d特征金字塔网络。(e)自上而下和自下而上。(f)跨尺度特征平衡。
1多尺度特征集成卷积神经网络CNN通常采用深度层次结构不同层次的特征具有不同的特性。浅层特征通常包含细粒度特征例如目标的点、边缘和纹理并提供详细的空间位置信息这更适合于目标定位。相反来自较高层的特征显示出更强的语义信息并呈现用于目标分类的判别信息。为了组合来自不同层的信息并生成多尺度表示一些研究人员引入了多层特征集成方法将来自多层的特征集成到单个特征图中并在该重建的特征图上执行检测[45]–[52]。图5(b)描绘了多层特征集成方法的结构。
Zhang等[48]设计了一种分层鲁棒CNN通过融合三个不同层的多尺度卷积特征提取分层空间语义信息并引入多个全连接层来增强网络的旋转和缩放鲁棒性。考虑到多层特征之间的不同范数Lin等人[49]在集成之前对每个特征应用L2归一化以保持网络训练阶段的稳定性。与以前在卷积层级别的多尺度特征集成不同Zheng等人[51]设计了HyBlock来构建层内级别的多尺度特征表示。
HyBlock采用具有锥体感受野的可分离卷积来学习超尺度特征缓解了RSOD中的尺度变化问题。
2金字塔特征层次金字塔特征层次背后的关键洞察力是不同层中的特征可以编码来自不同尺度的目标信息。例如小目标更可能出现在浅层中而大目标往往存在于深层中。因此金字塔特征层次结构采用多层特征进行独立预测以检测具有宽比例范围的目标如图5c所示。SSD[53]是金字塔特征层次的典型代表在自然场景[54]-[56]和遥感场景[57]-[63]中都有广泛的扩展应用。
为了提高小型车辆的检测性能Liang[60]等人在SSD中添加了一个额外的缩放分支该分支由反卷积模块和平均池化层组成。参考SSD中的分层回归层Wang等人[58]介绍了尺度不变回归层SIRLs其中采用三个孤立的回归层来捕获全尺度目标的信息。在SIRLs的基础上引入了一种新的特定尺度联合损耗来加速网络收敛。在[64]中李等人提出了在RPN和检测子网络中引入分层选择性过滤层的HSF-Net。具体地分层选择性滤波层采用具有不同核大小例如1 × 1、3 × 3和5 × 5的三个卷积层来获得多个感受野特征这有利于多尺度船舶检测。
3特征金字塔网络Feature Pyramid Networks金字塔特征层次方法使用独立的多级特征进行检测忽略了不同级别特征之间的互补信息导致低级别特征的语义信息较弱。为了解决这个问题林等人[65]提出了特征金字塔网络FPN。如图5d所示FPN引入了一种自上而下的路径将丰富的语义信息从高层特征转移到浅层特征导致所有级别的丰富语义特征请参考[65]中的详细信息。由于FPN在多尺度目标检测方面的显著改进FPN及其扩展[66]–[68]在多尺度特征表示中起着主导作用。
考虑到地理空间目标如桥梁、港口和机场的极端纵横比Hou等人[69]提出了一种非对称特征金字塔网络AFPN。AFPN采用非对称卷积块来增强关于十字形骨架的特征表示并提高大纵横比目标的性能。Zhang等[70]设计了一种拉普拉斯特征金字塔网络LFPN将高频信息注入到多尺度金字塔特征表示中这对于精确的目标检测是有用的但被以前的工作所忽略。在[71]中Zhang等人引入了高分辨率特征金字塔网络HRFPN以充分利用高分辨率特征表示从而实现精确和鲁棒的SAR船舶检测。此外一些研究人员将新颖的特征融合模块[72], [73]、注意力机制[74]–[77]或膨胀卷积层[78], [79]集成到FPN中以进一步获得更具区分性的多尺度特征表示。
FPN引入了一种自上而下的路径将高层语义信息传递到浅层而低层空间信息在主干网中长距离传播后仍然在顶层丢失。利用这个问题傅等人[80]提出了一种特征融合架构FFA该架构将辅助的自下而上路径集成到FPN结构中以通过短路径将低级空间信息传输到顶层特征如图5e所示。FFA确保检测器提取具有丰富语义和详细空间信息的多尺度特征金字塔。同样在[81], [82]中作者引入了一种双向FPN该FPN通过可学习参数学习不同级别特征的重要性并通过迭代的自上而下和自下而上的路径融合多级别特征。
与上述顺序增强途径[80]不同一些研究[83]-[94]采用了跨层次的特征融合方式。如图5f所示跨级特征融合方法充分收集所有级别的特征以自适应地获得平衡的特征图。程等人[83]利用特征级联操作实现跨尺度特征融合。考虑到来自不同层次的特征对特征融合应该有不同的贡献Fu等人[84]提出了基于级别的注意力以学习每个级别特征的独特贡献。由于transformer结构强大的全局信息提取能力一些工作[88]、[89]引入了transformer结构来集成和细化多级特征。在[90]中Chen等人提出了一种级联注意网络其中引入了位置监督来增强多级特征的语义信息。
C.多尺度锚生成
除了数据扩充和多尺度特征表示方法之外多尺度锚点生成还可以解决RSOD中巨大的目标尺度变化问题。由于自然场景和遥感场景中目标尺度范围的差异一些研究[95]-[104]修改了常见目标检测中的锚点设置以更好地覆盖地理空间目标的尺度。
Guo等[95]在检测器中注入了具有更多尺度和纵横比的额外锚点用于多尺度目标检测。Dong等[98]根据训练集中目标尺度的统计量设计了更合适的锚定尺度。Qiu等[99]将原始的方形RoI特征扩展为垂直、方形和水平RoI特征并融合这些RoI特征以更灵活的方式表示不同纵横比的目标。上述方法遵循固定的锚点设置而当前的研究[100]-[104]试图在训练阶段动态学习锚点。考虑到不同类别之间的纵横比变化Hou等人[100]设计了一种新的自适应纵横比锚SARA来自适应地学习每个类别的适当纵横比。SARA将可学习的类别纵横比值嵌入到回归分支中以利用位置回归损失的梯度自适应地更新每个类别的纵横比。受GA-RPN[105]的启发一些研究人员[102]-[104]在检测器中引入了轻量级子网络以自适应地学习锚点的位置和形状信息。
3.旋转目标检测
目标的任意方向是RSOD中的另一个主要挑战。由于RSI中的目标是从鸟瞰图中获取的它们表现出任意方向的属性因此在一般目标检测中广泛使用的水平检测框HBB表示不足以准确定位旋转目标。因此许多研究人员将注意力集中在地理空间目标的任意方向属性上这可以概括为旋转目标表示和旋转不变特征学习。图6中描绘了旋转目标检测方法的简要概述。 图6.旋转目标检测方法的简要总结。
A.旋转目标表示
旋转目标表示对于RSOD避免冗余背景和获得精确检测结果至关重要。最近的旋转目标表示方法主要可以概括为几类五参数表示[107]–[116]、八参数表示[117]–[126]、角度分类表示[106], [127]、[129]、高斯分布表示[130]–[133]和其他[134]–[144]。
1五参数最流行的解决方案是用五参数方法 ( x , y , w , h , θ ) (x, y, w, h, θ) (x,y,w,h,θ)表示目标这只是在HBB[107]-[115]上增加了一个额外的旋转角度参数θ。角度范围的定义在这种方法中起着至关重要的作用其中导出了两种定义。一些研究[107]-[112]将θ定义为与x轴的锐角并将角度范围限制为90°如图7(a)所示。作为最具代表性的工作Yang等人[107]遵循五参数方法来检测旋转目标并设计了一个IoU感知损失函数来解决旋转角度的边界不连续性问题。另一组[113]-[116]将θ称为x轴和长边之间的角度其范围为180°如图7(b)所示。丁等人[114]通过五参数方法回归旋转角度并将水平区域的特征转换为旋转区域以方便旋转目标检测。
图7。旋转目标的五参数表示和八参数表示方法的可视化[106]。
2八参数与五参数方法不同八参数方法[117]-[126]解决的是通过直接回归四个顶点 { ( a x , a y ) , ( b x , b y ) , ( c x , c y ) , ( d x , d y ) } \{(a_x, a_y), (b_x, b_y), (c_x, c_y), (d_x, d_y)\} {(ax,ay),(bx,by),(cx,cy),(dx,dy)}来表示旋转的目标如图7c所示。Xia等[117]首先采用了用于旋转目标表示的八参数方法该方法通过在训练过程中最小化每个顶点与地面真实坐标之间的差异来直接监督检测模型。然而这些顶点的序列顺序对于八参数方法避免不稳定的训练是必不可少的。如图8所示直观地从红色虚线箭头回归目标是一条更容易的路线但实际过程遵循红色实线箭头这造成了模型训练的困难。为此钱等人[119][121]提出了一种调制损失函数该函数计算不同排序顺序下的损失并选择最小情况进行学习有效地提高了检测性能。 图8。五参数法和八参数法的边界不连续性挑战[119], [121]。
3角度分类为了从源头解决图8中描述的问题许多研究人员[106], [127] [129]通过将角度预测问题转化为角度分类任务绕过了回归的边界挑战。Yang等[106]提出了用于旋转目标检测的第一种角度分类方法该方法将连续角度转换为离散角度并用新颖的圆形平滑标签训练模型。然而角度分类头[106]引入了附加参数并降低了检测器的效率。为了克服这一点杨等人[129]用一个密集编码的标签改进了[106]确保了模型的准确性和效率。
4高斯分布虽然上述方法取得了有希望的进展但它们没有考虑实际检测性能和优化度量之间的不对准。最近一系列工作[130]–[133]旨在通过用高斯分布表示旋转目标来处理这一挑战如图9所示。具体地这些方法将旋转的目标转换为2D高斯分布 N ( μ , Σ ) N(μ, Σ) N(μ,Σ)如下所示 图9。旋转目标的高斯分布表示方法的可视化[130]。 其中R表示旋转矩阵Λ表示
特征值的对角矩阵。利用等式1中的高斯分布表示两个旋转目标之间的IoU可以简化为两个分布之间的距离估计。此外高斯分布表示舍弃了角边界的定义有效地解决了角边界问题。杨等人[130]提出了一种新的高斯瓦瑟斯坦距离GWD度量来测量分布之间的距离该度量通过有效地近似旋转IoU来实现显著的性能。基于此Yang等人[131]引入了Kullback-Leibler散度KLD度量来增强其尺度不变性。
5其他一些研究人员通过其他方法解决旋转目标表示如基于分割的[134]-[136]和基于关键点的[137]-[144]。基于分割的方法中具有代表性的是Mask OBB[134]它在每个水平建议上部署分割方法以获得像素级目标区域并产生最小外部矩形作为旋转的边界框。另一方面魏等人[142]对旋转目标采用了基于关键点的表示该表示定位目标中心并利用一对中线来表示整个目标。此外Yang等[145]提出了第一个水平框标注监督的旋转目标检测器该检测器采用两种不同视图的自监督学习来预测旋转目标的角度。
B.旋转不变特征学习
旋转不变特征指示特征在任何旋转变换下保持一致。因此目标的旋转不变特征学习是解决旋转目标检测中任意方向问题的一个重要研究领域。为此许多研究人员提出了一系列学习目标旋转不变性的方法[146]–[157]显著改善了RSI中的旋转目标检测。
Cheng等人[146]提出了第一个旋转不变目标检测器通过使用rotationinsensitive特征来精确识别目标该检测器强制目标的特征在不同旋转角度下保持一致。后来程等人[148]、[149]采用旋转不变和fisher判别正则化器来鼓励检测器学习旋转不变和判别特征。在[150][151]中Wu等人分析了傅立叶域极坐标下目标的旋转不变性并设计了空间频率通道特征提取模块来获得旋转不变性特征。考虑到轴对齐卷积特征和旋转目标之间的未对准Han等人[156]提出了一种定向检测模块该模块采用一种新的对齐卷积操作来学习方向信息。在[155]中Han等人进一步设计了一个旋转等方差检测器来显式编码旋转等方差和旋转不变性。此外一些研究人员[80], [157]用一系列预定义的旋转锚扩展了RPN以应对地理空间目标的任意方向特征。
我们在表I中总结了里程碑旋转目标检测方法的检测性能。 表I 旋转目标检测方法在具有旋转注释的DOTAV1.0数据集上的检测性能。
4.弱目标检测
RSI中感兴趣的目标通常嵌入在具有复杂目标空间模式和大量背景噪声的复杂场景中。复杂的上下文和背景噪声严重损害了感兴趣目标的特征表示导致对感兴趣目标的特征响应较弱。因此许多现有的工作集中在改进感兴趣目标的特征表示上这可以分为两个流抑制背景噪声和挖掘相关的上下文信息。在图10中给出了弱目标检测方法的简要概述。 图10。弱目标检测方法的简要总结。
A.抑制背景噪声
这类方法旨在通过弱化背景区域的响应来加强特征图中目标区域的弱响应。主要可以分为两类内隐学习和外显监督。
1内隐学习内隐学习方法在检测器中采用精心设计的模块在训练阶段自适应地学习重要特征并抑制冗余特征从而减少背景噪声干扰。
在机器学习中降维可以有效地学习紧凑的特征表示抑制不相关的特征。利用上述性质叶等人[158]提出了一种特征过滤模块通过连续的瓶颈层捕获低维特征图以过滤背景噪声干扰。受人类视觉感知选择性聚焦的启发注意机制被提出并得到了大量研究[159]-[161]。注意机制在网络学习阶段重新分配特征重要性以增强重要特征并抑制冗余信息。因此注意力机制也在RSOD中被广泛引入以解决背景噪声干扰问题[57], [162]、[170]。在[162]中Huang等人强调了补丁-补丁依赖性对RSOD的重要性并设计了一种新的非局部感知金字塔注意NP-Attention。NP-Attention学习空间多尺度非局部依赖性和通道依赖性以使检测器能够专注于目标区域而不是背景。考虑到SAR图像中陆地区域的强散射干扰Sun等[163]提出了一种船舶注意力模块以突出船舶的特征表示减少来自陆地区域的虚警。此外为RSOD设计的一系列注意力机制例如空间洗牌组增强注意力[165]、多尺度空间和通道注意力[166]、离散小波多尺度注意力[167]等。已经证明了它们在抑制背景噪声方面的有效性。
2显式监督与隐式学习方法不同显式监督方法采用辅助显著性监督信息来显式引导检测器突出前景区域并弱化背景。
李等人[171]采用区域对比度法获得显著图并构建显著特征通过融合多尺度特征图与显著图来构建金字塔。在[172]中Lei等人用显著性检测方法[173]提取显著性图并提出显著性重建网络。显著性重建网络利用显著性图作为像素级监督来指导检测器的训练以加强特征图中的显著性区域。上述显著性检测方法通常是无监督的并且生成的显著性图可能包含非目标区域如图11(b)所示从而向检测器提供不准确的引导。因此后来的工作[107]、[134]、[174]-[180]将框级注释转化为目标级显著性引导信息如图11c所示以生成更准确的显著性监督。杨等人[107]设计了一个像素注意力网络该网络采用目标级显著性监督来增强目标线索并削弱背景信息。在[175]中Zhang等人提出了FoRDet以更简洁的方式利用目标级显著性监督。具体地所提出的FoRDet利用粗糙阶段中前景区域的预测在框级注释下监督来增强精细阶段中前景区域的特征表示。 图11.(a)输入图像。b显著性检测方法生成的显著性图[173]。c目标级显著性图。
B.挖掘相关上下文信息
上下文信息通常是指目标与周围环境或场景之间的空间和语义关系。该上下文信息可以为无法清楚区分的目标提供辅助特征表示。因此挖掘上下文信息可以有效地解决RSOD中的弱特征响应问题。根据上下文信息的类别现有方法主要分为局部和全局上下文信息挖掘。
1局部上下文信息挖掘局部上下文信息是指目标与其周围环境在视觉信息和空间分布上的相关性[147]、[181]-[187]。Zhang等人[181]通过将原始区域建议缩放为三种不同的大小来生成多个局部上下文区域并提出了一种上下文双向增强模块来融合局部上下文特征和目标特征。上下文感知卷积神经网络CA-CNN[182]采用上下文RoI挖掘层来提取目标周围的上下文信息。首先通过合并围绕目标的一系列过滤建议来生成目标的上下文RoI然后与目标RoI融合作为用于分类和回归的最终目标特征表示。在[183]中Ma等人利用门控递归单元GRU将目标特征与局部上下文信息融合从而获得目标的更具鉴别性的特征表示。图卷积网络GCN最近在目标-目标关系推理方面表现出了更好的性能。因此田等人[184][185]构建了空间和语义图来建模和学习目标之间的上下文关系。
2全局上下文信息挖掘全局上下文信息利用目标和场景之间的关联[188]-[195]例如车辆通常位于道路上船舶通常出现在海上。Chen等人[188]通过RoI-Align操作从全局图像特征中提取场景上下文信息并将其与目标级RoI特征融合以加强目标和场景之间的关系。Liu等人[192]设计了一种场景辅助检测头在场景级监督下利用场景上下文信息。场景辅助检测头将预测出的场景向量嵌入到分类分支中实现目标级特征与场景级上下文信息的融合。在[193]中陶等人提出了一种场景上下文驱动的车辆检测方法。具体来说引入预训练的场景分类器将每个图像块分类为三个场景类别然后采用特定场景的车辆检测器来获得初步检测结果最后利用场景上下文信息进一步优化检测结果。
考虑到局部和全局上下文信息的互补性Zhang等人[196]提出了一种CAD-Net来挖掘局部和全局上下文信息。CAD-Net采用金字塔局部上下文网络来学习目标级局部上下文信息并设计了全局上下文网络来提取场景级全局上下文信息。在[103]中Teng等人提出了一种GLNet来收集从全局到局部的上下文信息从而实现RSI的鲁棒和准确的检测器。此外一些研究[197]–[199]还引入了ASPP[200]或RFB模块[54]来利用本地和全球上下文信息。
5.微小目标检测
RSI的典型地面采样距离GSD为1-3米这意味着即使是大型目标例如飞机、船舶和储罐也只能占据小于16 × 16像素。此外即使在GSD为0.25 m的高分辨率RSI中尺寸为 3 × 1.5 m 2 3×1.5m^2 3×1.5m2的车辆也仅覆盖72个像素12 × 6。RSI中微小目标的普遍存在进一步增加了RSOD的难度。目前关于微小目标检测的研究主要分为判别特征学习、基于超分辨率的方法和改进的检测度量。图12简要总结了微小目标检测方法。 图12。微小目标检测方法的简要总结。
A.判别特征学习
微小目标极小的尺度小于16 × 16像素使其表现出有限的外观信息这对探测器学习微小目标的特征提出了严峻的挑战。为了解决上述问题许多研究人员专注于提高微小目标的判别特征学习能力[201]–[208]。
由于微小目标主要存在于浅层特征中缺乏高层语义信息[65]一些文献[201]–[203]引入自上而下的结构将高层语义信息融合到浅层特征中以加强微小目标的语义信息。考虑到微小目标的有限外观信息一些研究[204]–[208]通过自注意机制或扩张卷积建立微小目标与周围上下文信息之间的联系以增强微小目标的特征辨别能力。值得注意的是前面提到的一些关于多尺度特征学习和上下文信息挖掘的研究也证明了在微小目标检测中的显著有效性。
B.基于超分辨率的方法
极小的尺度是微小目标检测的关键问题因此提高图像的分辨率是提高微小目标检测性能的直观解决方案。一些方法[209]-[212]采用超分辨率策略作为检测流水线的预处理步骤以扩大输入图像的分辨率。例如Rabbi等人[211]强调了边缘信息对于微小目标检测的重要性并提出了一种边缘增强的超分辨率生成对抗网络GAN来生成具有详细边缘信息的视觉上令人愉悦的高分辨率RSI。吴等人[212]开发了一种微小目标的点到区域检测框架。点到区域框架首先通过关键点预测获得建议区域然后采用多任务GAN对建议区域执行超分辨率并检测这些建议区域中的微小目标。然而超分辨率生成的高分辨率图像给检测流水线带来了额外的计算复杂性。利用这个问题[213]和[214]在特征级采用超分辨率策略来获取微小目标的判别特征表示并有效地节省计算资源。
C.改进的微小目标检测度量
与前两种类型的方法不同最近的高级工作[10]、[215]-[222]断言当前流行的检测范例不适合微小目标检测并且不可避免地阻碍微小目标检测性能。Pang等人。[215]认为现代探测器中过度的下采样操作导致特征图上微小目标的丢失并提出了一种缩小和放大结构来放大特征图。在[218]中Yan等人调整标签分配中的IoU阈值增加微小目标的正分配锚点有利于微小目标的学习。董等人[219]设计了Sig-NMS以减少传统非最大抑制NMS中大型和中型目标对微小目标的抑制。
在[10]中Xu等人指出IoU度量不适用于微小目标检测。如图13所示IoU度量对轻微的位置偏移敏感。此外基于IoU的标签分配存在严重的尺度不平衡问题其中微小的目标往往被分配的阳性样本不足。为了解决这些问题Xu等人[10]设计了一种归一化的Wasserstein距离NWD来代替IoU度量。NWD将微小目标建模为2D高斯分布并利用高斯分布之间的归一化Wasserstein距离来表示微小目标之间的位置关系详见[10]。与IoU度量相比所提出的NWD度量对位置偏差是平滑的并且具有尺度平衡的特性如图13b所示。在[222]Xu等人进一步提出了用于微小目标检测的感受野距离RFLA并实现了最先进的性能。 图13。aIoU偏差曲线和bNWD偏差曲线之间的比较[10]。详见[10]。
6.有限监督下的目标检测
近年来RSI中广泛使用的基于深度学习的检测器严重依赖于具有高质量注释的大规模数据集来实现最先进的性能。然而收集大量标记良好的数据是相当昂贵和耗时的例如一个边界框注释将花费大约10秒这导致了RSOD中数据受限或注释受限的场景[11]。这种缺乏足够的监督信息严重降低了检测性能。为了解决这个问题研究人员在有限的监督下探索了RSOD中的各种任务。我们将以往的研究总结为三种主要类型弱监督目标检测、半监督目标检测和少样本目标检测。图14提供有限监督下的目标检测方法综述。 图14。有限监督下的目标检测方法的简要总结。
A.弱监督目标检测
与全监督目标检测相比弱监督目标检测WSOD只包含弱监督信息。形式上WSOD由训练数据集 D t r a i n { ( X i , y i ) } i 1 I D_{train} \{(X_i, y_i)\}^I_{i1} Dtrain{(Xi,yi)}i1I组成其中 X i { x 1 , . . . , x m i } X_i \{x_1, ..., x_{m_i} \} Xi{x1,...,xmi}是训练样本的集合称为bag m i m_i mi是bag中训练样本的总数 y i y_i yi是 X i X_i Xi的弱监督信息例如图像级标签[223]或点级标签[224]。将图像级监督有效地转移到目标级标签是WSOD的关键挑战[225]。
韩等人[226]引入深度玻尔兹曼机来学习目标的高级特征并提出了一种基于贝叶斯原理的遥感WSOD弱监督学习框架。李等人[227]利用场景对之间的互信息来学习判别卷积权重并采用多尺度类别激活图来定位地理空间目标。
在WSDDN[228]显著性能的激励下提出了一系列遥感WSOD方法[229]、[241]。如图15所示当前WSOD方法的范例通常由两个步骤组成其首先构建多实例学习模型(MIL)以找到对图像分类任务有贡献的建议作为伪标签然后利用它们来训练检测器。姚等人[229]引入了一种动态课程学习策略其中检测器通过从易到难的训练过程逐步提高检测性能。冯等人[231]设计了一种渐进的上下文实例细化方法该方法通过利用周围的上下文信息来抑制低质量的目标部分并突出整个目标。Wang等[233]将空间和外观关系图引入WSOD传播高质量的标签信息以挖掘更多可能的目标。在[240]中Feng等人认为现有的遥感WSOD方法忽略了地理空间目标的任意方向导致了旋转敏感的目标探测器。为了解决这个问题冯等人[240]提出了一种RINet它通过采用旋转不变学习和多实例挖掘为WSOD带来了旋转不变但多样的特征学习。 图15。最近WSOD方法的两步范例[229]–[241]。
我们在表II中总结了里程碑WSOD方法的性能其中采用正确定位度量CorLoc[242]来评估定位性能。 B.半监督目标检测
半监督目标检测SSOD通常仅包含一小部分不超过50%的良好标记样本 D l a b e l e d { ( x i , y i ) } i 1 I l a b e l e d D_{labeled} \{(x_i, y_i)\}^{I_{labeled}}_{i1} Dlabeled{(xi,yi)}i1Ilabeled难以构建可靠的监督检测器并且具有大量未标记样本 D u n l a b e l e d { ( x j ) } j 1 I u n l a b e l e d D_{unlabeled} \{(x_j )\}^{I_{unlabeled}}_{j1} Dunlabeled{(xj)}j1Iunlabeled。SSOD旨在通过从大量未标记样本中学习潜在信息来提高稀缺监督信息下的检测性能。
侯等人[243]提出了一种用于半监督SAR船舶检测的SCLANet。SCLANet采用标记和未标记样本之间的对抗性学习来利用未标记样本信息并对未标记样本采用一致性学习来增强网络的鲁棒性。伪标签生成机制也是半监督目标检测的广泛使用的方法[244]-[248]典型范例如图16所示。首先使用从scare标记样本中学习的预训练检测器来预测未标记样本然后选择置信度分数较高的伪标签作为可信部分最后用标记和伪标记样本重新训练模型。Wu等人[246]提出了一种自定进度的课程学习该学习遵循“从易到难”的方案来选择更可靠的伪标签。钟等人[245]采用主动学习策略其中高分预测由专家手动调整以获得精炼的伪标签。陈等人[247]采用师生相互学习来充分利用未标记的样本并迭代生成更高质量的伪标签。 图16.SSOD中伪标签生成机制的流水线。
此外一些研究[249]–[253]致力于弱半监督目标检测其中未标记样本被弱注释样本取代。杜等人[251][252]采用大量imagelevel标记样本来提高稀缺检测框级标记样本下的SAR车辆检测性能。陈等人[253]采用了一小部分像素级标记样本和大量的检测框级标记样本来提高标签稀缺实例分割的性能。
C.少样本目标检测
少样本目标检测FSOD是指仅用有限数量不超过30个的样本检测新的类别。通常FSOD包含一个具有丰富样本的基类数据集 D b a s e { ( x i , y i ) , y i ∈ C b a s e } i 1 I b a s e D_{base} \{(x_i, y_i) , y_i ∈ C_{base}\}^{I_{base}}_{i1} Dbase{(xi,yi),yi∈Cbase}i1Ibase和一个只有K-样本样本的新类数据集 D n o v e l { ( x j , y j ) , y j ∈ C n o v e l } j 1 C n o v e l ∗ K D_{novel} \{(x_j , y_j ) , y_j ∈ C_{novel}\}^{C_{novel}∗K}_{j1} Dnovel{(xj,yj),yj∈Cnovel}j1Cnovel∗K。请注意 C b a s e C_{base} Cbase和 C n o v e l C_{novel} Cnovel是脱节的。如图17所示典型的FSOD范例由两阶段训练流水线组成其中基础训练阶段利用丰富的基础类样本建立先验知识并且少数样本微调阶段利用先验知识来促进少数样本新概念的学习。遥感FSOD的研究主要集中在元学习方法[254]-[259]和迁移学习方法[260]-[269]。 图17.FSOD的两阶段训练管道。
基于元学习的方法通过模拟一系列少样本学习任务来获取任务级知识并将这些知识推广到新类的少样本学习中。李等人[255]首次将元学习用于遥感FSOD仅用1至10个标记样本就获得了令人满意的检测性能。后来遥感界开发了一系列基于元学习的少样本探测器[254]–[259]。例如Cheng等人[254]提出了一种原型CNN通过学习特定于类的原型来为遥感FSOD生成更好的前景建议和类感知RoI特征。王等人[258]提出了一种元度量训练范式使少样本学习者具有灵活的可扩展性以快速适应少样本的新颖任务。
基于迁移学习的方法旨在将从丰富的注释数据中学习到的公共知识微调到少量的新数据并且通常包括基本训练阶段和少量的微调阶段。黄等人[266]提出了一种平衡微调策略以缓解新颖类样本和基类样本之间的数量不平衡问题。周等人[265]在微调阶段引入了建议级对比度学习以在少样本场景中学习更鲁棒的特征表示。与基于元学习的方法相比基于迁移学习的方法具有更简单和记忆有效的训练范式。
7.数据集和评估指标
A.数据集介绍和选择
数据集在RSI中目标检测的整个发展过程中发挥了不可或缺的作用。一方面数据集作为探测器性能评估和比较的共同点。另一方面数据集推动研究人员解决RSOD领域越来越具有挑战性的问题。在过去十年中发布了几个具有不同属性的数据集以促进RSOD的发展如表三所示。在本节中我们主要介绍10个广泛使用的具有特定特征的数据集。 表III RSOD领域广泛使用的数据集的比较。HBB和OBB分别指水平边界框和定向边界框。*代表平均图像宽度。
NWPU VHR-10[18]。该数据集是一个多类地理空间目标检测数据集。它包含十个类别的3,775个HBB注释实例飞机、轮船、储罐、棒球场、网球场、篮球场、跑道、港口、桥梁和车辆。有800张非常高分辨率的RSI包括来自谷歌地球的715张彩色图像和来自Vaihingen数据的85张泛锐化彩色红外图像。图像分辨率范围为0.5至2米。
VEDAI[272]。VEDAI是一个细粒度的车辆检测数据集包含五个细粒度的车辆类别露营车、轿车、皮卡、拖拉机、卡车和货车。VEDAI数据集中有1,210张图像和3,700个实例每个图像的大小为1,024 × 1,024。小区域和车辆的任意方向是VEDAI数据集中的主要挑战。
UCAS-AOD[274]。UCAS-AOD数据集包括910幅图像和6029个目标其中600幅图像中包含3210架飞机310幅图像中包含2819辆车辆。所有图像均从Google Earth获取图像大小约为1,000 × 1,000。
HRSC[276]。HRSC数据集广泛用于任意方向的船舶检测由1,070幅图像和2,976个带有OBB注释的实例组成。这些图像取自谷歌地球包含近海和近岸场景。图像尺寸从300 × 300到1500 × 900不等图像分辨率从2米到0.4米不等。
SSDD[277]。SSDD是第一个用于SAR图像船舶检测的开放数据集包含1,160幅SAR图像和2,456艘船舶。SSDD数据集中的SAR图像是从不同的传感器收集的分辨率从1米到15米具有不同的偏振HH、VV、VH和HV。随后作者将SSDD数据集进一步细化和丰富为三种不同类型以满足当前SAR船舶检测的研究[286]。
xView[2]。xView数据集是ROSD中最大的公开可用数据集之一在60个细粒度类中拥有大约100万个标记目标。与其他RSOD数据集相比xView数据集中的图像是从WorldView-3在0.3 m地面样本距离处收集的提供了更高分辨率的图像。此外xView数据集覆盖了超过1,400平方公里的地球表面这导致了更高的多样性。
DOTA[117]。DOTA是一个大规模数据集由188,282个用HBB和OBB注释的目标组成。所有目标分为15类飞机、轮船、储罐、棒球场、网球场、游泳池、地面跑道、港口、桥梁、大型车辆、小型车辆、直升机、环形交叉路口、足球场和篮球场。该数据集中的图像收集自谷歌地球、JL-1卫星和GF-2卫星空间分辨率为0.1至1米。最近最新的DOTAv2.0[33]已经公开其中包含18个类别的170多万个目标。
DIOR[28]。DIOR是光学RSI的目标检测数据集。该数据集中有23,463幅光学图像空间分辨率为0.5至30米。数据集中的目标总数为192,472个所有目标都用HBB标记。目标类别如下飞机、机场、棒球场、篮球场、桥梁、烟囱、大坝、高速公路服务区、高速公路收费站、海港、高尔夫球场、地面田径场、立交桥、船舶、体育场、储罐、网球场、火车站、车辆、风车。
FAIR1M[34]。FAIR1M是一个更具挑战性的数据集用于RSI中的细粒度目标检测包括5个类别和37个子类别。有超过40,000张图像和超过100万个目标由定向边界框注释。这些图像是从多个平台获取的分辨率为0.3米至0.8米分布在不同的国家和地区。细粒度的类别、大量的目标、大范围的尺寸和方向以及多样化的场景使FAIR1M更具挑战性。
SODA-A[284]。SODA-A是最近发布的数据集专为RSI中的微小目标检测而设计。该数据集由2,510幅平均图像大小为4,761 × 2,777的图像和800,203个带有OBB注释的目标组成。所有目标根据其面积范围分为四个子集即极小、相对小、一般小和正常。该数据集中有九个类别包括飞机、直升机、小型车辆、大型车辆、船舶、集装箱、储罐、游泳池和风车。
上述综述表明早期发表的数据集通常样本有限。例如NWPU VHR10[18]仅包含10个类别和3,651个实例而UCAC-AOD[274]由2个类别和6,029个实例组成。近年来研究人员不仅引入了海量数据和细粒度级别的目标还收集了来自多传感器、各种分辨率和不同场景的数据例如DOTA[117]、DIOR[28]、FAIR1M [34]以满足RSOD的实际应用。图18描绘了不同RSOD数据集的典型样本。 图18。不同RSOD数据集的可视化。不同的分辨率、海量的实例、多传感器图像和细粒度的类别是RSOD数据集的典型特征。
我们还在表IV中提供了数据集选择指南以帮助研究人员为不同的挑战和场景选择合适的数据集和方法。值得注意的是只有数据集的imagelevel注释可用于弱监督场景。至于少样本监督场景每个场景只有K-shot框级注释样本新类其中K设置为 { 3 , 5 , 10 , 20 , 30 } \{3,5,10,20,30\} {3,5,10,20,30}。 表IV RSOD针对不同挑战和场景的数据集选择指南。
B.评价指标
除了数据集评估指标也同样重要。通常推理速度和检测精度是评估检测器性能的两个常用指标。
每秒帧数FPS是推理速度评估的标准指标表示检测器每秒可以检测到的图像数量。值得注意的是图像大小和硬件设备都会影响推理速度。
平均精度AP是检测准确度最常用的指标。给定测试图像i设 { ( b i , c i , p i ) } i 1 N \{(b_i, c_i, p_i)\}^N_{i1} {(bi,ci,pi)}i1N表示预测检测其中 b i b_i bi是预测框 c i c_i ci是预测标签 p i p_i pi是置信度分数。设 { b j g t , c j g t } j 1 M \{b^{gt}_j , c^{gt}_j\}^M_{j1} {bjgt,cjgt}j1M指的是测试图像 I I I上的真实注释其中 b j g t b^{gt}_j bjgt是地面实况框 c j g t c^{gt}_j cjgt是真实类别。如果预测检测 ( b i , c i , p i ) (b_i, c_i, p_i) (bi,ci,pi)满足以下两个标准则将其分配为真实注释 b j g t , c j g t b^{gt}_j , c^{gt}_j bjgt,cjgt的真阳性(TP)
•置信度得分 p i p_i pi大于置信度阈值 t t t并且预测的标签与地面真实标签 c j g t c^{gt}_j cjgt相同。
•预测框 b i b_i bi和真实框 b j g t b^{gt}_j bjgt之间的IoU大于IoU阈值 ε ε ε。IoU的计算方法如下 其中 a r e a ( b i ∩ b j g t ) area(b_i ∩ b^{gt}_j ) area(bi∩bjgt)和 a r e a ( b i ∪ b j g t ) area(b_i ∪ b^{gt}_j ) area(bi∪bjgt)代表预测框和真实框的交集和并集面积。
否则认为是假阳性(FP)。值得注意的是根据上述标准多个预测检测可能匹配相同的真实注释但只有具有最高置信度分数的预测检测被分配为TP其余的是FPs[287].。
基于TP和FP检测精度P和召回率R可以计算为等式3和等式4. 其中FN表示假阴性的数量。精确度测量预测检测的真阳性的分数召回率测量正确检测的阳性的分数。然而上述两个评估指标仅反映了检测性能的单一方面。
考虑到精确度和召回率AP提供了检测性能的综合评估并为每个类别单独计算。对于给定的类别根据每次召回时最大精度的检测绘制精度/召回曲线PRCAP总结PRC的形状[287]。对于多类目标检测采用所有类的AP值的平均值称为mAP来评估整体检测精度。
早期研究主要采用基于固定IoU的AP度量即 A P 50 AP_{50} AP50[18]、[28]、[117]其中IoU阈值 ε ε ε被给出为0.5。该低IoU阈值表现出对检测框偏差的高容限并且不能满足高定位精度要求。后来一些工作[130], [131], [284]引入了一种新的评估度量称为 A P 50 : 95 AP_{50:95} AP50:95它对10个IoU阈值上的AP进行平均从0.5到0.95间隔为0.05。 A P 50 : 95 AP_{50:95} AP50:95考虑了更高的IoU阈值并鼓励更准确的定位。
AP作为RSOD中评估度量的基石针对不同的具体任务有各种扩展。在少样本学习场景中 A P n o v e l AP_{novel} APnovel和 A P b a s e AP_{base} APbase是评估少样本检测器性能的两个关键指标其中 A P n o v e l AP_{novel} APnovel和 A P b a s e AP_{base} APbase分别表示新类和基类上的检测性能。一个优秀的少样本检测器应该在新类中实现令人满意的性能并避免基类中的性能下降[269]。在遥感目标的增量检测中采用 A P o l d AP_{old} APold和 A P i n c AP_{inc} APinc来评估旧类和增量类在不同增量任务上的性能。此外调和平均值也是增量目标检测的重要评估度量[288]它提供了旧类和增量类的综合性能评估如等式5所述 8.应用
深度学习技术为RSOD注入了重大创新导致了一种从大量RSIs中自动识别感兴趣目标的有效方法。因此RSOD方法已被应用于丰富多样的实践场景中这些实践场景极大地支持了可持续发展目标SDGs的实施和社会的改善[289]–[291]如图19所示。 图19.RSOD的广泛应用为实施可持续发展目标和改善社会做出了重大贡献。a震后灾害评估中倒塌建筑物的检测。b用于精确农业的玉米植物检测。c-d可持续城市和社区的建筑和车辆检测。(e)减缓气候变化的太阳能光伏探测。(f)沿海岸探测垃圾以保护海洋。(g)检测非洲哺乳动物用于野生动物监测。(h)森林生态系统保护的单树检测。
A.灾害管理
自然灾害对人类生命财产安全构成严重威胁。快速准确地了解灾害影响和破坏程度对于灾害管理至关重要。RSOD方法可以从受灾地区的鸟瞰图中准确识别地面目标为灾害管理提供了一种新的潜力[292]–[296]。Guan等人[293]提出了一种新颖的实例分割模型用于在复杂环境中准确检测火灾可应用于森林火灾灾害响应。Ma等[295]设计了震后倒塌建筑评估的实时检测方法。
B.精准农业
随着前所未有且仍在增长的人口确保农业生产是养活不断增长的人口的根本障碍。RSOD有能力监测作物生长和估计粮食产量促进精准农业的进一步进步[297]-[302]。庞等人[298]将RSI用于早季玉米检测并实现了出苗率的准确估计。Chen等[302]设计了一种自动草莓花检测系统用于监测草莓田的生长周期。
C.可持续城市和社区
现在全球一半的人口居住在城市而且这一人口在未来几十年还将继续增长。可持续城市和社区是现代城市发展的目标RSOD可以在其中产生重大影响。例如建筑和车辆检测[303]、[306]可以帮助估计人口密度分布和交通统计为城市发展规划提供建议。基础设施分布检测[307]可以辅助城市环境中的灾害评估和预警。
D.气候行动
持续的气候变化迫使人类面临气候危机的艰巨挑战。一些研究人员[308]–[310]采用目标检测方法自动绘制苔原冰楔多边形以记录和分析气候变暖对北极地区的影响。此外RSOD可以统计太阳能电池板和风力涡轮机的数量和空间分布[311]、[314]有助于减少温室气体排放。
E.海洋养护
海洋覆盖了地球表面的近四分之三30多亿人依赖海洋和海岸的多样生命。海洋因污染而逐渐恶化RSOD可以为海洋保护提供强有力的支持[315]。一些工作将检测方法应用于沿海垃圾检测[316]、海上漂浮塑料检测[317]、深海碎片检测[318]等。另一个重要的应用是船舶检测[135][136]它可以帮助监控非法捕鱼活动。
F.野生动物监测
在各个层面都观察到了全球生物多样性的丧失目标检测与RSI相结合为野生动物保护提供了一个新的视角[319]–[323]。Delplanque等人[322]采用基于深度学习的检测器对非洲哺乳动物进行多物种检测和识别。Kellenberger等[323]设计了一个弱监督野生动物检测框架只需要图像级标签即可识别野生动物。
G.森林生态系统保护
森林生态系统在生态保护、气候调节、碳循环等方面发挥着重要作用。了解树木的状况对于森林生态系统保护至关重要[324]-[328]。萨福诺娃等人[326]分析了检测到的树木树冠的形状、纹理和颜色以确定其损害阶段为评估森林健康提供了一种更有效的方法。萨尼-穆罕默德等人[328]利用实例分割方法绘制直立枯树这对于森林生态系统管理和保护至关重要。
9.未来方向
除了本次调查中提到的五个RSOD研究课题外该领域还有很多工作要做。因此我们对未来方向进行了前瞻性的讨论以进一步改进和增强遥感场景中的探测器。
A.大比例尺遥感影像统一检测框架
受益于遥感技术的发展可以很容易地获得高分辨率的大规模RSI例如超过10,000 × 10,000像素。但受限于GPU内存目前主流的RSOD方法在大规模RSI中未能直接进行目标检测而是采用滑动窗口策略主要包括滑动窗口裁剪、补丁预测和结果合并。一方面与统一检测框架相比这种滑动窗口框架需要复杂的数据预处理和后处理。另一方面目标通常占据RSI的小区域海量背景的无效计算导致计算时间和内存消耗的增加。一些研究[215], [329], [330]提出了一种从粗到细的检测框架用于大规模RSI中的目标检测。该框架首先通过过滤掉无意义区域来定位感兴趣区域然后从这些过滤后的区域中实现准确检测。
B.利用多模态遥感图像进行检测
受传感器成像机理的限制基于单模态RSI的检测器往往存在检测性能偏差在实际应用中难以满足[331]。相反来自不同传感器的多模态RSI有其特点。例如高光谱图像包含高光谱分辨率和细粒度光谱特征SAR图像提供丰富的纹理信息光学图像表现出高空间分辨率和丰富的细节信息。多模态RSIs的集成处理可以提高场景的解释能力获得对地理空间目标更客观、更全面的理解[332]-[334]为进一步提高RSOD的检测性能提供了可能性。
C.遥感图像中的域自适应目标检测
由于遥感卫星传感器、分辨率和波段的多样性以及天气条件、季节和地理空间区域的影响[6]从不同卫星收集的RSI通常来自相似但不相同的分布。这种分布差异也称为畴隙严重限制了检测器的泛化性能。最近关于域自适应目标检测的研究[335]–[338]已经提出解决域间隙问题。然而这些研究仅关注单模态中的域自适应检测器而跨模态域自适应目标检测例如从光学图像到SAR图像[339], [340]是一个更具挑战性和值得研究的课题。
D.遥感目标的增量检测
现实世界的环境是动态和开放的类别的数量随着时间的推移而变化。然而主流检测器在遇到新类别时需要新旧数据来重新训练模型导致计算成本较高。最近增量学习被认为是解决这一问题最有希望的方法它可以仅用新数据就能学习新知识而不忘记旧知识[341]。增量学习在遥感界已有初步探索[342]-[345]。例如Chen等人[342]将知识蒸馏集成到FPN和检测头中以在保持旧概念的同时学习新概念。增量式RSOD仍需要更深入的研究以满足实际应用中的动态学习任务。
E.遥感场景的自监督预训练模型
当前的RSOD方法总是用ImageNet[346]预训练的权重初始化。然而在自然场景和遥感场景之间存在不可避免的域差距这可能限制了RSOD的性能。近年来自监督预训练方法受到了广泛的关注并在自然场景中的分类和下游任务中表现出了优异的性能。受益于遥感技术的快速进步丰富的遥感数据[347], [348]也为自监督预训练提供了充足的数据支持。一些研究人员[349]-[353]已经初步证明了遥感预训练在代表性下游任务上的有效性。因此探索基于多源遥感数据的自监督预训练模型值得进一步研究。
F.紧凑和高效的目标检测架构
大多数现有的机载和星载卫星都需要将遥感数据发回进行解译从而导致额外的资源开销。因此有必要研究用于机载和星载平台的紧凑高效的探测器以减少数据传输中的资源消耗。利用这一需求一些研究人员通过模型设计[285]、[354]、[355]、网络剪枝[356]、[357]和知识蒸馏[358]-[360]提出了轻量级检测器。然而这些探测器仍然严重依赖高性能GPU无法部署在机载和星载卫星上。因此为有限资源场景设计紧凑高效的目标检测架构仍然具有挑战性。
10.总结
目标检测一直是遥感界一个基础但具有挑战性的研究课题。由于深度学习技术的快速发展RSOD在过去十年中受到了相当大的关注并取得了显著的成就。在这篇综述中我们对RSOD中现有的基于深度学习的方法进行了系统的综述和总结。首先我们根据地理空间目标的特点总结了RSOD中的五个主要挑战并将方法分为五个流多尺度目标检测、旋转目标检测、弱目标检测、微小目标检测和有限监督目标检测。然后我们采用了系统的层次划分对每个类别中的方法进行了回顾和总结。接下来我们介绍了RSOD领域的典型基准数据集、评估指标和实际应用。最后考虑到现有RSOD方法的局限性我们讨论了一些有希望的进一步研究方向。 鉴于RSOD技术的高速发展我们相信这项调查可以帮助研究人员更全面地了解该领域的主要主题并找到未来研究的潜在方向。
11.参考文献 [1] N. Gorelick, M. Hancher, M. Dixon, S. Ilyushchenko, D. Thau, and R. Moore, “Google earth engine: Planetary-scale geospatial analysis for everyone,” Remote Sens. Environ., vol. 202, pp. 18–27, 2017. [2] D. Lam, R. Kuzma, K. McGee, S. Dooley, M. Laielli, M. Klaric, Y. Bulatov, and B. McCord, “xview: Objects in context in overhead imagery,” 2018. [Online]. Available: http://arxiv.org/abs/1802.07856 [3] Z. Li, H. Shen, H. Li, G. Xia, P. Gamba, and L. Zhang, “Multi-feature combined cloud and cloud shadow detection in gaofen-1 wide field of view imagery,” Remote Sens. Environ., vol. 191, pp. 342–358, 2017. [4] S. Zhang, R. Wu, K. Xu, J. Wang, and W. Sun, “R-cnn-based ship detection from high resolution remote sensing imagery,” Remote Sens., vol. 11, no. 6, p. 631, 2019. [5] Y. Wang, C. Wang, H. Zhang, Y. Dong, and S. Wei, “Automatic ship detection based on retinanet using multi-resolution gaofen-3 imagery,” Remote Sens., vol. 11, no. 5, p. 531, 2019. [6] X. X. Zhu, D. Tuia, L. Mou, G.-S. Xia, L. Zhang, F. Xu, and F. Fraundorfer, “Deep learning in remote sensing: A comprehensive review and list of resources,” IEEE Geosci. Remote Sens. Mag., vol. 5, no. 4, pp. 8–36, 2017. [7] L. Zhang, L. Zhang, and B. Du, “Deep learning for remote sensing data: A technical tutorial on the state of the art,” IEEE Geosci. Remote Sens. Mag., vol. 4, no. 2, pp. 22–40, 2016. [8] L. Zhang and L. Zhang, “Artificial intelligence for remote sensing data analysis: A review of challenges and opportunities,” IEEE Geosci. Remote Sens. Mag., vol. 10, no. 2, pp. 270–294, 2022. [9] W. Han, J. Chen, L. Wang, R. Feng, F. Li, L. Wu, T. Tian, and J. Yan, “Methods for small, weak object detection in optical high-resolution remote sensing images: A survey of advances and challenges,” IEEE Geosci. Remote Sens. Mag., vol. 9, no. 4, pp. 8–34, 2021. [10] C. Xu, J. Wang, W. Yang, H. Yu, L. Yu, and G.-S. Xia, “Detecting tiny objects in aerial images: A normalized wasserstein distance and a new benchmark,” ISPRS J. Photogrammetry Remote Sens., vol. 190, pp. 79–93, 2022. [11] J. Yue, L. Fang, P. Ghamisi, W. Xie, J. Li, J. Chanussot, and A. Plaza, “Optical remote sensing image understanding with weak supervision: Concepts, methods, and perspectives,” IEEE Geosci. Remote Sens. Mag., vol. 10, no. 2, pp. 250–269, 2022. [12] C. Xu and H. Duan, “Artificial bee colony (abc) optimized edge potential function (epf) approach to target recognition for low-altitude aircraft,” Pattern Recognit. Lett., vol. 31, no. 13, pp. 1759–1772, 2010. [13] X. Sun, H. Wang, and K. Fu, “Automatic detection of geospatial objects using taxonomic semantics,” IEEE Geosci. Remote Sens. Lett., vol. 7, no. 1, pp. 23–27, 2010. [14] Y. Lin, H. He, Z. Yin, and F. Chen, “Rotation-invariant object detection in remote sensing images based on radial-gradient angle,” IEEE Geosci. Remote Sens. Lett., vol. 12, no. 4, pp. 746–750, 2015. [15] H. Moon, R. Chellappa, and A. Rosenfeld, “Performance analysis of a simple vehicle detection algorithm,” Image Vis. Comput., vol. 20, no. 1, pp. 1–13, 2002. [16] S. Leninisha and K. Vani, “Water flow based geometric active deformable model for road network,” ISPRS J. Photogrammetry Remote Sens., vol. 102, pp. 140–147, 2015. [17] D. Chaudhuri and A. Samal, “An automatic bridge detection technique for multispectral images,” IEEE Trans. Geosci. Remote Sens., vol. 46, no. 9, pp. 2720–2727, 2008. [18] G. Cheng, J. Han, P. Zhou, and L. Guo, “Multi-class geospatial object detection and geographic image classification based on collection of part detectors,” ISPRS J. Photogrammetry Remote Sens., vol. 98, pp. 119–132, 2014. [19] L. Zhang, L. Zhang, D. Tao, and X. Huang, “Sparse transfer manifold embedding for hyperspectral target detection,” IEEE Trans. Geosci. Remote Sens., vol. 52, no. 2, pp. 1030–1043, 2013. [20] J. Han, P. Zhou, D. Zhang, G. Cheng, L. Guo, Z. Liu, S. Bu, and J. Wu, “Efficient, simultaneous detection of multi-class geospatial targets based on visual saliency modeling and discriminative learning of sparse coding,” ISPRS J. Photogrammetry Remote Sens., vol. 89, pp. 37–48, 2014. [21] H. Sun, X. Sun, H. Wang, Y. Li, and X. Li, “Automatic target detection in high-resolution remote sensing images using spatial sparse coding bag-of-words model,” IEEE Geosci. Remote Sens. Lett., vol. 9, no. 1, pp. 109–113, 2011. [22] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” nature, vol. 521, no. 7553, pp. 436–444, 2015. [23] S. Ren, K. He, R. B. Girshick, and J. Sun, “Faster R-CNN: towards realtime object detection with region proposal networks,” in Proc. Annu. Conf. Neural Inf. Process. Syst, 2015, pp. 91–99. [24] J. Redmon and A. Farhadi, “YOLO9000: better, faster, stronger,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2017, pp. 6517–6525. [25] T. Lin, P. Goyal, R. B. Girshick, K. He, and P. Doll ́ ar, “Focal loss for dense object detection,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2017, pp. 2999–3007. [26] Z. Tian, C. Shen, H. Chen, and T. He, “FCOS: fully convolutional onestage object detection,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2019, pp. 9626–9635. [27] L. Liu, W. Ouyang, X. Wang, P. W. Fieguth, J. Chen, X. Liu, and M. Pietik ̈ ainen, “Deep learning for generic object detection: A survey,” Int. J. Comput. Vis., vol. 128, no. 2, pp. 261–318, 2020. [28] K. Li, G. Wan, G. Cheng, L. Meng, and J. Han, “Object detection in optical remote sensing images: A survey and a new benchmark,” ISPRS J. Photogrammetry Remote Sens., vol. 159, pp. 296–307, 2020. [29] G. Cheng and J. Han, “A survey on object detection in optical remote sensing images,” ISPRS J. Photogrammetry Remote Sens., vol. 117, pp. 11–28, 2016. [30] U. Alganci, M. Soydas, and E. Sertel, “Comparative research on deep learning approaches for airplane detection from very high-resolution satellite images,” Remote Sens., vol. 12, no. 3, p. 458, 2020. [31] Z. Li, Y. Wang, N. Zhang, Y. Zhang, Z. Zhao, D. Xu, G. Ben, and Y. Gao, “Deep learning-based object detection techniques for remote sensing images: A survey,” Remote Sens., vol. 14, no. 10, p. 2385, 2022. [32] J. Kang, S. Tariq, H. Oh, and S. S. Woo, “A survey of deep learningbased object detection methods and datasets for overhead imagery,” IEEE Access, vol. 10, pp. 20 118–20 134, 2022. [33] J. Ding, N. Xue, G. Xia, X. Bai, W. Yang, M. Y. Yang, S. J. Belongie, J. Luo, M. Datcu, M. Pelillo, and L. Zhang, “Object detection in aerial images: A large-scale benchmark and challenges,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 44, no. 11, pp. 7778–7796, 2022. [34] X. Sun, P. Wang, Z. Yan, F. Xu, R. Wang, W. Diao, J. Chen, J. Li, Y. Feng, T. Xu, M. Weinmann, S. Hinz, C. Wang, and K. Fu, “Fair1m: A benchmark dataset for fine-grained object recognition in highresolution remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 184, pp. 116–130, 2022. [35] W. Zhao, W. Ma, L. Jiao, P. Chen, S. Yang, and B. Hou, “Multi-scale image block-level F-CNN for remote sensing images object detection,” IEEE Access, vol. 7, pp. 43 607–43 621, 2019. [36] S. M. Azimi, E. Vig, R. Bahmanyar, M. K ̈ orner, and P. Reinartz, “Towards multi-class object detection in unconstrained remote sensing imagery,” in Asian Conference on Computer Vision, vol. 11363, 2018, pp. 150–165. [37] P. Shamsolmoali, M. Zareapoor, J. Chanussot, H. Zhou, and J. Yang, “Rotation equivariant feature image pyramid network for object detection in optical remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3112481 [38] Y. Chen, P. Zhang, Z. Li, Y. Li, X. Zhang, G. Meng, S. Xiang, J. Sun, and J. Jia, “Stitcher: Feedback-driven data provider for object detection,” 2020. [Online]. Available: https://arxiv.org/abs/2004.12432 [39] X. Xu, X. Zhang, and T. Zhang, “Lite-yolov5: A lightweight deep learning detector for on-board ship detection in large-scene sentinel-1 SAR images,” Remote Sens., vol. 14, no. 4, p. 1018, 2022. [Online]. Available: https://doi.org/10.3390/rs14041018 [40] N. Su, Z. Huang, Y. Yan, C. Zhao, and S. Zhou, “Detect larger at once: Large-area remote-sensing image arbitrary-oriented ship detection,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3144485 [41] B. Zhao, Y. Wu, X. Guan, L. Gao, and B. Zhang, “An improved aggregated-mosaic method for the sparse object detection of remote sensing imagery,” Remote Sens., vol. 13, no. 13, p. 2602, 2021. [42] X. Han, Y. Zhong, and L. Zhang, “An efficient and robust integrated geospatial object detection framework for high spatial resolution remote sensing imagery,” Remote Sens., vol. 9, no. 7, p. 666, 2017. [43] Y. Long, Y. Gong, Z. Xiao, and Q. Liu, “Accurate object localization in remote sensing images based on convolutional neural networks,” IEEE Trans. Geosci. Remote Sens., vol. 55, no. 5, pp. 2486–2498, 2017. [44] Y. Zhong, X. Han, and L. Zhang, “Multi-class geospatial object detection based on a position-sensitive balancing framework for high spatial resolution remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 138, pp. 281–294, 2018. [45] P. Ding, Y. Zhang, W.-J. Deng, P. Jia, and A. Kuijper, “A light and faster regional convolutional neural network for object detection in optical remote sensing images,” ISPRS J. Photogrammetry Remote Sens., vol. 141, pp. 208–218, 2018. [46] W. Liu, L. Ma, and H. Chen, “Arbitrary-oriented ship detection framework in optical remote-sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 15, no. 6, pp. 937–941, 2018. [47] W. Liu, L. Ma, J. Wang, and H. Chen, “Detection of multiclass objects in optical remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 16, no. 5, pp. 791–795, 2019. [48] Y. Zhang, Y. Yuan, Y. Feng, and X. Lu, “Hierarchical and robust convolutional neural network for very high-resolution remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 8, pp. 5535–5548, 2019. [49] Z. Lin, K. Ji, X. Leng, and G. Kuang, “Squeeze and excitation rank faster R-CNN for ship detection in SAR images,” IEEE Geosci. Remote Sens. Lett., vol. 16, no. 5, pp. 751–755, 2019. [50] Z. Deng, H. Sun, S. Zhou, J. Zhao, L. Lei, and H. Zou, “Multi-scale object detection in remote sensing imagery with convolutional neural networks,” ISPRS J. Photogrammetry Remote Sens., vol. 145, pp. 3–22, 2018. [51] Z. Zheng, Y. Zhong, A. Ma, X. Han, J. Zhao, Y. Liu, and L. Zhang, “Hynet: Hyper-scale object detection network framework for multiple spatial resolution remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 166, pp. 1–14, 2020. [52] Y. Ren, C. Zhu, and S. Xiao, “Deformable faster r-cnn with aggregating multi-layer features for partially occluded object detection in optical remote sensing images,” Remote Sens., vol. 10, no. 9, p. 1470, 2018. [53] W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y. Fu, and A. C. Berg, “Ssd: Single shot multibox detector,” in in Proc. Euro. Conf. Comput. Vis. Springer, 2016, pp. 21–37. [54] S. Liu, D. Huang, and Y. Wang, “Receptive field block net for accurate and fast object detection,” in in Proc. Euro. Conf. Comput. Vis., 2018, pp. 385–400. [55] Z. Shen, Z. Liu, J. Li, Y.-G. Jiang, Y. Chen, and X. Xue, “Dsod: Learning deeply supervised object detectors from scratch,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2017, pp. 1919–1927. [56] Z. Zhang, S. Qiao, C. Xie, W. Shen, B. Wang, and A. L. Yuille, “Singleshot object detection with enriched semantics,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 5813–5821. [57] X. Lu, J. Ji, Z. Xing, and Q. Miao, “Attention and feature fusion SSD for remote sensing object detection,” IEEE Trans. Instrum. Meas., vol. 70, pp. 1–9, 2021. [58] G. Wang, Y. Zhuang, H. Chen, X. Liu, T. Zhang, L. Li, S. Dong, and Q. Sang, “Fsod-net: Full-scale object detection from optical remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 118, 2022. [59] B. Hou, Z. Ren, W. Zhao, Q. Wu, and L. Jiao, “Object detection in high-resolution panchromatic images using deep models and spatial template matching,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 2, pp. 956–970, 2020. [60] X. Liang, J. Zhang, L. Zhuo, Y. Li, and Q. Tian, “Small object detection in unmanned aerial vehicle images using feature fusion and scalingbased single shot detector with spatial context analysis,” IEEE Trans. Circuits Syst. Video Technol., vol. 30, no. 6, pp. 1758–1770, 2020. [61] Z. Wang, L. Du, J. Mao, B. Liu, and D. Yang, “Sar target detection based on ssd with data augmentation and transfer learning,” IEEE Geosci. Remote Sens. Lett., vol. 16, no. 1, pp. 150–154, 2018. [62] S. Bao, X. Zhong, R. Zhu, X. Zhang, Z. Li, and M. Li, “Single shot anchor refinement network for oriented object detection in optical remote sensing imagery,” IEEE Access, vol. 7, pp. 87 150–87 161, 2019. [63] T. Xu, X. Sun, W. Diao, L. Zhao, K. Fu, and H. Wang, “ASSD: feature aligned single-shot detection for multiscale objects in aerial imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–17, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3089170 [64] Q. Li, L. Mou, Q. Liu, Y. Wang, and X. X. Zhu, “Hsf-net: Multiscale deep feature embedding for ship detection in optical remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 12, pp. 71477161, 2018. [65] T.-Y. Lin, P. Doll ́ ar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2017, pp. 2117–2125. [66] S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path aggregation network for instance segmentation,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR), Jun. 2018, pp. 8759–8768. [67] J. Pang, K. Chen, J. Shi, H. Feng, W. Ouyang, and D. Lin, “Libra RCNN: towards balanced learning for object detection,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR), Jun. 2019, pp. 821830. [68] M. Tan, R. Pang, and Q. V. Le, “Efficientdet: Scalable and efficient object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2020, pp. 10 781–10 790. [69] L. Hou, K. Lu, and J. Xue, “Refined one-stage oriented object detection method for remote sensing images,” IEEE Trans. Image Process., vol. 31, pp. 1545–1558, 2022. [70] W. Zhang, L. Jiao, Y. Li, Z. Huang, and H. Wang, “Laplacian feature pyramid network for object detection in VHR optical remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3072488 [71] S. Wei, H. Su, J. Ming, C. Wang, M. Yan, D. Kumar, J. Shi, and X. Zhang, “Precise and robust ship detection for high-resolution SAR imagery based on hr-sdnet,” Remote Sens., vol. 12, no. 1, p. 167, 2020. [72] G. Cheng, M. He, H. Hong, X. Yao, X. Qian, and L. Guo, “Guiding clean features for object detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [73] J. Jiao, Y. Zhang, H. Sun, X. Yang, X. Gao, W. Hong, K. Fu, and X. Sun, “A densely connected end-to-end neural network for multiscale and multiscene SAR ship detection,” IEEE Access, vol. 6, pp. 20 88120 892, 2018. [74] Q. Guo, H. Wang, and F. Xu, “Scattering enhanced attention pyramid network for aircraft detection in SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 9, pp. 7570–7587, 2021. [75] Y. Li, Q. Huang, X. Pei, L. Jiao, and R. Shang, “Radet: Refine feature pyramid network and multi-layer attention network for arbitraryoriented object detection of remote sensing images,” Remote Sens., vol. 12, no. 3, p. 389, 2020. [76] L. Shi, L. Kuang, X. Xu, B. Pan, and Z. Shi, “Canet: Centerness-aware network for object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–13, 2022. [77] R. Yang, Z. Pan, X. Jia, L. Zhang, and Y. Deng, “A novel cnn-based detector for ship detection based on rotatable bounding box in SAR images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 1938–1958, 2021. [78] Y. Zhao, L. Zhao, B. Xiong, and G. Kuang, “Attention receptive pyramid network for ship detection in SAR images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 13, pp. 2738–2756, 2020. [79] X. Yang, X. Zhang, N. Wang, and X. Gao, “A robust one-stage detector for multiscale ship detection with complex background in massive SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–12, 2022. [80] K. Fu, Z. Chang, Y. Zhang, G. Xu, K. Zhang, and X. Sun, “Rotationaware and multi-scale convolutional neural network for object detection in remote sensing images,” ISPRS J. Photogrammetry Remote Sens., vol. 161, pp. 294–308, 2020. [81] W. Huang, G. Li, B. Jin, Q. Chen, J. Yin, and L. Huang, “Scenario context-aware-based bidirectional feature pyramid network for remote sensing target detection,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS. 2021.3135935 [82] V. Chalavadi, J. Prudviraj, R. Datla, C. S. Babu, and K. M. C, “msodanet: A network for multi-scale object detection in aerial images using hierarchical dilated convolutions,” Pattern Recognit., vol. 126, p. 108548, 2022. [83] G. Cheng, Y. Si, H. Hong, X. Yao, and L. Guo, “Cross-scale feature fusion for object detection in optical remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 18, no. 3, pp. 431–435, 2021. [84] J. Fu, X. Sun, Z. Wang, and K. Fu, “An anchor-free method based on feature balancing and refinement network for multiscale ship detection in SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 2, pp. 1331–1344, 2021. [85] Y. Liu, Q. Li, Y. Yuan, Q. Du, and Q. Wang, “Abnet: Adaptive balanced network for multiscale object detection in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3133956 [86] H. Guo, X. Yang, N. Wang, B. Song, and X. Gao, “A rotational libra R-CNN method for ship detection,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 8, pp. 5772–5781, 2020. [87] T. Zhang, Y. Zhuang, G. Wang, S. Dong, H. Chen, and L. Li, “Multiscale semantic fusion-guided fractal convolutional object detection network for optical remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–20, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3108476 [88] Y. Zheng, P. Sun, Z. Zhou, W. Xu, and Q. Ren, “Adt-det: Adaptive dynamic refined single-stage transformer detector for arbitrary-oriented object detection in satellite optical imagery,” Remote Sens., vol. 13, no. 13, p. 2623, 2021. [89] Z. Wei, D. Liang, D. Zhang, L. Zhang, Q. Geng, M. Wei, and H. Zhou, “Learning calibrated-guidance for object detection in aerial images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 27212733, 2022. [90] L. Chen, C. Liu, F. Chang, S. Li, and Z. Nie, “Adaptive multi-level feature fusion and attention-based network for arbitrary-oriented object detection in remote sensing imagery,” Neurocomputing, vol. 451, pp. 67–80, 2021. [91] X. Sun, P. Wang, C. Wang, Y. Liu, and K. Fu, “Pbnet: Part-based convolutional neural network for complex composite object detection in remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 173, pp. 50–65, 2021. [92] T. Zhang, X. Zhang, C. Liu, J. Shi, S. Wei, I. Ahmad, X. Zhan, Y. Zhou, D. Pan, J. Li, and H. Su, “Balance learning for ship detection from synthetic aperture radar remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 182, pp. 190–207, 2021. [93] T. Zhang, X. Zhang, and X. Ke, “Quad-fpn: A novel quad feature pyramid network for sar ship detection,” Remote Sens., vol. 13, no. 14, p. 2771, 2021. [94] J. Song, L. Miao, Q. Ming, Z. Zhou, and Y. Dong, “Fine-grained object detection in remote sensing images via adaptive label assignment and refined-balanced feature pyramid network,” IEEE J. Sel. Top. Appl. Earth Obs. Remote. Sens., vol. 16, pp. 71–82, 2023. [95] W. Guo, W. Yang, H. Zhang, and G. Hua, “Geospatial object detection in high resolution satellite images based on multi-scale convolutional neural network,” Remote Sens., vol. 10, no. 1, p. 131, 2018. [96] S. Zhang, G. He, H. Chen, N. Jing, and Q. Wang, “Scale adaptive proposal network for object detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 16, no. 6, pp. 864–868, 2019. [97] C. Li, C. Xu, Z. Cui, D. Wang, T. Zhang, and J. Yang, “Featureattentioned object detection in remote sensing imagery,” in Proc. IEEE Int. Conf. Image Process. Conf. (ICIP). IEEE, 2019, pp. 3886–3890. [98] Z. Dong, M. Wang, Y. Wang, Y. Zhu, and Z. Zhang, “Object detection in high resolution remote sensing imagery based on convolutional neural networks with suitable object scale features,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 3, pp. 2104–2114, 2020. [99] H. Qiu, H. Li, Q. Wu, F. Meng, K. N. Ngan, and H. Shi, “A2rmnet: Adaptively aspect ratio multi-scale network for object detection in remote sensing images,” Remote Sens., vol. 11, no. 13, p. 1594, 2019. [100] J. Hou, X. Zhu, and X. Yin, “Self-adaptive aspect ratio anchor for oriented object detection in remote sensing images,” Remote Sens., vol. 13, no. 7, p. 1318, 2021. [101] N. Mo, L. Yan, R. Zhu, and H. Xie, “Class-specific anchor based and context-guided multi-class object detection in high resolution remote sensing imagery with a convolutional neural network,” Remote Sens., vol. 11, no. 3, p. 272, 2019. [102] Z. Tian, R. Zhan, J. Hu, W. Wang, Z. He, and Z. Zhuang, “Generating anchor boxes based on attention mechanism for object detection in remote sensing images,” Remote Sens., vol. 12, no. 15, p. 2416, 2020. [103] Z. Teng, Y. Duan, Y. Liu, B. Zhang, and J. Fan, “Global to local: Clip-lstm-based object detection from remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–13, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3064840 [104] Y. Yu, H. Guan, D. Li, T. Gu, E. Tang, and A. Li, “Orientation guided anchoring for geospatial object detection from remote sensing imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 160, pp. 67–82, 2020. [105] J. Wang, K. Chen, S. Yang, C. C. Loy, and D. Lin, “Region proposal by guided anchoring,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2019, pp. 2965–2974. [106] X. Yang and J. Yan, “On the arbitrary-oriented object detection: Classification based approaches revisited,” Int. J. Comput. Vis., vol. 130, no. 5, pp. 1340–1365, 2022. [107] X. Yang, J. Yang, J. Yan, Y. Zhang, T. Zhang, Z. Guo, X. Sun, and K. Fu, “Scrdet: Towards more robust detection for small, cluttered and rotated objects,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2019, pp. 8232–8241. [108] X. Yang, J. Yan, Z. Feng, and T. He, “R3det: Refined single-stage detector with feature refinement for rotating object,” in Pro. AAAI Conf. Artific. Intell., vol. 35, no. 4, 2021, pp. 3163–3171. [109] X. Yang, H. Sun, K. Fu, J. Yang, X. Sun, M. Yan, and Z. Guo, “Automatic ship detection in remote sensing images from google earth of complex scenes based on multiscale rotation dense feature pyramid networks,” Remote Sens., vol. 10, no. 1, p. 132, 2018. [110] X. Yang, H. Sun, X. Sun, M. Yan, Z. Guo, and K. Fu, “Position detection and direction prediction for arbitrary-oriented ships via multitask rotation region convolutional neural network,” IEEE Access, vol. 6, pp. 50 839–50 849, 2018. [111] Q. Ming, L. Miao, Z. Zhou, and Y. Dong, “Cfc-net: A critical feature capturing network for arbitrary-oriented object detection in remotesensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3095186 [112] Q. Ming, Z. Zhou, L. Miao, H. Zhang, and L. Li, “Dynamic anchor learning for arbitrary-oriented object detection,” in Pro. AAAI Conf. Artific. Intell., 2021, pp. 2355–2363. [113] Y. Zhu, J. Du, and X. Wu, “Adaptive period embedding for representing oriented objects in aerial images,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 10, pp. 7247–7257, 2020. [114] J. Ding, N. Xue, Y. Long, G.-S. Xia, and Q. Lu, “Learning roi transformer for oriented object detection in aerial images,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2019, pp. 28492858. [115] Q. An, Z. Pan, L. Liu, and H. You, “Drbox-v2: An improved detector with rotatable boxes for target detection in SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 11, pp. 8333–8349, 2019. [116] Q. Li, L. Mou, Q. Xu, Y. Zhang, and X. X. Zhu, “R3-net: A deep network for multi-oriented vehicle detection in aerial images and videos,” 2018. [Online]. Available: http://arxiv.org/abs/1808.05560 [117] G. Xia, X. Bai, J. Ding, Z. Zhu, S. J. Belongie, J. Luo, M. Datcu, M. Pelillo, and L. Zhang, “DOTA: A large-scale dataset for object detection in aerial images,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 3974–3983. [118] Y. Liu, S. Zhang, L. Jin, L. Xie, Y. Wu, and Z. Wang, “Omnidirectional scene text detection with sequential-free box discretization,” arXiv preprint arXiv:1906.02371, 2019. [119] W. Qian, X. Yang, S. Peng, J. Yan, and Y. Guo, “Learning modulated loss for rotated object detection,” in in Proc. AAAI Conf. Artific. Intell., vol. 35, no. 3, 2021, pp. 2458–2466. [120] Y. Xu, M. Fu, Q. Wang, Y. Wang, K. Chen, G.-S. Xia, and X. Bai, “Gliding vertex on the horizontal bounding box for multi-oriented object detection,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 43, no. 4, pp. 1452–1459, 2021. [121] W. Qian, X. Yang, S. Peng, X. Zhang, and J. Yan, “Rsdet: Pointbased modulated loss for more accurate rotated object detection,” IEEE Trans. Circuits Syst. Video Technol., vol. 32, no. 11, pp. 7869–7879, 2022. [122] J. Luo, Y. Hu, and J. Li, “Surround-net: A multi-branch arbitraryoriented detector for remote sensing,” Remote Sens., vol. 14, no. 7, p. 1751, 2022. [123] Q. Song, F. Yang, L. Yang, C. Liu, M. Hu, and L. Xia, “Learning pointguided localization for detection in remote sensing images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 1084–1094, 2021. [124] X. Xie, G. Cheng, J. Wang, X. Yao, and J. Han, “Oriented r-cnn for object detection,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2021, pp. 3520–3529. [125] Y. Yao, G. Cheng, G. Wang, S. Li, P. Zhou, X. Xie, and J. Han, “On improving bounding box representations for oriented object detection,” IEEE Trans. Geosci. Remote Sens., vol. 61, pp. 1–11, 2023. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3231340 [126] Q. Ming, L. Miao, Z. Zhou, X. Yang, and Y. Dong, “Optimization for arbitrary-oriented object detection via representation invariance loss,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3115110 [127] X. Yang and J. Yan, “Arbitrary-oriented object detection with circular smooth label,” in Proc. Euro. Conf. Comput. Vis. Springer, 2020, pp. 677–694. [128] X. Yang, L. Hou, Y. Zhou, W. Wang, and J. Yan, “Dense label encoding for boundary discontinuity free rotation detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2021, pp. 15 819–15 829. [129] J. Wang, F. Li, and H. Bi, “Gaussian focal loss: Learning distribution polarized angle prediction for rotated object detection in aerial images,” IEEE Trans. Geosci. Remote Sens., 2022. [130] X. Yang, J. Yan, Q. Ming, W. Wang, X. Zhang, and Q. Tian, “Rethinking rotated object detection with gaussian wasserstein distance loss,” in Proc. Int. Conf. Machine Learn, 2021, pp. 11 830–11 841. [131] X. Yang, X. Yang, J. Yang, Q. Ming, W. Wang, Q. Tian, and J. Yan, “Learning high-precision bounding box for rotated object detection via kullback-leibler divergence,” vol. 34, pp. 18 381–18 394, 2021. [132] X. Yang, Y. Zhou, G. Zhang, J. Yang, W. Wang, J. Yan, X. Zhang, and Q. Tian, “The kfiou loss for rotated object detection,” arXiv preprint arXiv:2201.12558, 2022. [133] X. Yang, G. Zhang, X. Yang, Y. Zhou, W. Wang, J. Tang, T. He, and J. Yan, “Detecting rotated objects as gaussian distributions and its 3-d generalization,” IEEE Trans. Pattern Anal. Mach. Intell., 2022. [Online]. Available: https://doi.org/10.1109/TPAMI.2022.3197152. [134] J. Wang, J. Ding, H. Guo, W. Cheng, T. Pan, and W. Yang, “Mask obb: A semantic attention-based mask oriented bounding box representation for multi-category object detection in aerial images,” Remote Sens., vol. 11, no. 24, p. 2930, 2019. [135] X. Zhang, G. Wang, P. Zhu, T. Zhang, C. Li, and L. Jiao, “Grs-det: An anchor-free rotation ship detector based on gaussian-mask in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 4, pp. 3518–3531, 2020. [136] Y. Yang, X. Tang, Y. Cheung, X. Zhang, F. Liu, J. Ma, and L. Jiao, “Ar2det: An accurate and real-time rotational one-stage ship detector in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3092433 [137] F. Zhang, X. Wang, S. Zhou, Y. Wang, and Y. Hou, “Arbitraryoriented ship detection through center-head point extraction,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2021. [138] J. Yi, P. Wu, B. Liu, Q. Huang, H. Qu, and D. Metaxas, “Oriented object detection in aerial images with box boundary-aware vectors,” in Proc. IEEE Winter Conf. Appl. Comput. Vis. (WACV), 2021, pp. 2150–2159. [139] Z. Xiao, L. Qian, W. Shao, X. Tan, and K. Wang, “Axis learning for orientated objects detection in aerial images,” Remote Sens., vol. 12, no. 6, p. 908, 2020. [140] X. He, S. Ma, L. He, L. Ru, and C. Wang, “Learning rotated inscribed ellipse for oriented object detection in remote sensing images,” Remote Sens., vol. 13, no. 18, p. 3622, 2021. [141] K. Fu, Z. Chang, Y. Zhang, and X. Sun, “Point-based estimator for arbitrary-oriented object detection in aerial images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 5, pp. 4370–4387, 2020. [142] H. Wei, Y. Zhang, Z. Chang, H. Li, H. Wang, and X. Sun, “Oriented objects as pairs of middle lines,” ISPRS J. Photogrammetry Remote Sens., vol. 169, pp. 268–279, 2020. [143] L. Zhou, H. Wei, H. Li, W. Zhao, Y. Zhang, and Y. Zhang, “Arbitraryoriented object detection in remote sensing images based on polar coordinates,” IEEE Access, vol. 8, pp. 223 373–223 384, 2020. [144] X. Zheng, W. Zhang, L. Huan, J. Gong, and H. Zhang, “Apronet: Detecting objects with precise orientation from aerial images,” ISPRS J. Photogrammetry Remote Sens., vol. 181, pp. 99–112, 2021. [145] X. Yang, G. Zhang, W. Li, X. Wang, Y. Zhou, and J. Yan, “H2rbox: Horizontal box annotation is all you need for oriented object detection,” 2022. [Online]. Available: https://doi.org/10.48550/arXiv.2210.06742 [146] G. Cheng, P. Zhou, and J. Han, “Learning rotation-invariant convolutional neural networks for object detection in vhr optical remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 12, pp. 74057415, 2016. [147] K. Li, G. Cheng, S. Bu, and X. You, “Rotation-insensitive and contextaugmented object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 4, pp. 2337–2348, 2017. [148] G. Cheng, P. Zhou, and J. Han, “Rifd-cnn: Rotation-invariant and fisher discriminative convolutional neural networks for object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 2884–2893. [149] G. Cheng, J. Han, P. Zhou, and D. Xu, “Learning rotation-invariant and fisher discriminative convolutional neural networks for object detection,” IEEE Trans. Image Process., vol. 28, no. 1, pp. 265–278, 2019. [150] X. Wu, D. Hong, J. Tian, J. Chanussot, W. Li, and R. Tao, “Orsim detector: A novel object detection framework in optical remote sensing imagery using spatial-frequency channel features,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 7, pp. 5146–5158, 2019. [151] X. Wu, D. Hong, J. Chanussot, Y. Xu, R. Tao, and Y. Wang, “Fourierbased rotation-invariant feature boosting: An efficient framework for geospatial object detection,” IEEE Geosci. Remote Sens. Lett., vol. 17, no. 2, pp. 302–306, 2019. [152] G. Wang, X. Wang, B. Fan, and C. Pan, “Feature extraction by rotationinvariant matrix representation for object detection in aerial image,” IEEE Geosci. Remote Sens. Lett., vol. 14, no. 6, pp. 851–855, 2017. [153] X. Wu, D. Hong, P. Ghamisi, W. Li, and R. Tao, “Msri-ccf: Multi-scale and rotation-insensitive convolutional channel features for geospatial object detection,” Remote Sens., vol. 10, no. 12, p. 1990, 2018. [154] M. Zand, A. Etemad, and M. Greenspan, “Oriented bounding boxes for small and freely rotated objects,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–15, 2021. [155] J. Han, J. Ding, N. Xue, and G.-S. Xia, “Redet: A rotation-equivariant detector for aerial object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2021, pp. 2786–2795. [156] J. Han, J. Ding, J. Li, and G. Xia, “Align deep features for oriented object detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–11, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3062048 [157] X. Yao, H. Shen, X. Feng, G. Cheng, and J. Han, “R2ipoints: Pursuing rotation-insensitive point representation for aerial object detection,” IEEE Trans. Geosci. Remote Sens., 2022. [158] X. Ye, F. Xiong, J. Lu, J. Zhou, and Y. Qian, “R3-net: Feature fusion and filtration network for object detection in optical remote sensing images,” Remote Sens., vol. 12, no. 24, p. 4027, 2020. [159] J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 7132–7141. [160] X. Li, W. Wang, X. Hu, and J. Yang, “Selective kernel networks,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2019, pp. 510–519. [161] J. Fu, J. Liu, H. Tian, Y. Li, Y. Bao, Z. Fang, and H. Lu, “Dual attention network for scene segmentation,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), Jun. 2019, pp. 3146–3154. [162] Z. Huang, W. Li, X. Xia, X. Wu, Z. Cai, and R. Tao, “A novel nonlocal-aware pyramid and multiscale multitask refinement detector for object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–20, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3059450 [163] Y. Sun, X. Sun, Z. Wang, and K. Fu, “Oriented ship detection based on strong scattering points network in large-scale SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–18, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3130117 [164] W. Ma, N. Li, H. Zhu, L. Jiao, X. Tang, Y. Guo, and B. Hou, “Feature split-merge-enhancement network for remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–17, 2022. [165] Z. Cui, X. Wang, N. Liu, Z. Cao, and J. Yang, “Ship detection in largescale SAR images via spatial shuffle-group enhance attention,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 1, pp. 379–391, 2021. [166] J. Chen, L. Wan, J. Zhu, G. Xu, and M. Deng, “Multi-scale spatial and channel-wise attention for improving object detection in remote sensing imagery,” IEEE Geosci. Remote Sens. Lett., vol. 17, no. 4, pp. 681–685, 2020. [167] J. Bai, J. Ren, Y. Yang, Z. Xiao, W. Yu, V. Havyarimana, and L. Jiao, “Object detection in large-scale remote-sensing images based on time-frequency analysis and feature optimization,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3119344 [168] J. Hu, X. Zhi, S. Jiang, H. Tang, W. Zhang, and L. Bruzzone, “Supervised multi-scale attention-guided ship detection in optical remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS. 2022.3206306 [169] Y. Guo, X. Tong, X. Xu, S. Liu, Y. Feng, and H. Xie, “An anchor-free network with density map and attention mechanism for multiscale object detection in aerial images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3207178 [170] D. Yu and S. Ji, “A new spatial-oriented object detection framework for remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3127232 [171] C. Li, B. Luo, H. Hong, X. Su, Y. Wang, J. Liu, C. Wang, J. Zhang, and L. Wei, “Object detection based on global-local saliency constraint in aerial images,” Remote Sens., vol. 12, no. 9, p. 1435, 2020. [172] J. Lei, X. Luo, L. Fang, M. Wang, and Y. Gu, “Region-enhanced convolutional neural network for object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 8, pp. 56935702, 2020. [173] Y. Yuan, C. Li, J. Kim, W. Cai, and D. D. Feng, “Reversion correction and regularized random walk ranking for saliency detection,” IEEE Trans. Image Process., vol. 27, no. 3, pp. 1311–1322, 2018. [174] C. Xu, C. Li, Z. Cui, T. Zhang, and J. Yang, “Hierarchical semantic propagation for object detection in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 6, pp. 4353–4364, 2020. [175] T. Zhang, X. Zhang, P. Zhu, P. Chen, X. Tang, C. Li, and L. Jiao, “Foreground refinement network for rotated object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–13, 2021. [176] J. Wang, W. Yang, H. Li, H. Zhang, and G. Xia, “Learning center probability map for detecting objects in aerial images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 5, pp. 4307–4323, 2021. [177] Z. Fang, J. Ren, H. Sun, S. Marshall, J. Han, and H. Zhao, “Safdet: A semi-anchor-free detector for effective detection of oriented objects in aerial images,” Remote Sens., vol. 12, no. 19, p. 3225, 2020. [178] Z. Ren, Y. Tang, Z. He, L. Tian, Y. Yang, and W. Zhang, “Ship detection in high-resolution optical remote sensing images aided by saliency information,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3173610 [179] H. Qu, L. Shen, W. Guo, and J. Wang, “Ships detection in SAR images based on anchor-free model with mask guidance features,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 666–675, 2022. [180] S. Liu, L. Zhang, H. Lu, and Y. He, “Center-boundary dual attention for oriented object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3069056 [181] J. Zhang, C. Xie, X. Xu, Z. Shi, and B. Pan, “A contextual bidirectional enhancement method for remote sensing image object detection,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 13, pp. 4518–4531, 2020. [182] Y. Gong, Z. Xiao, X. Tan, H. Sui, C. Xu, H. Duan, and D. Li, “Contextaware convolutional neural network for object detection in VHR remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 1, pp. 34–44, 2020. [183] W. Ma, Q. Guo, Y. Wu, W. Zhao, X. Zhang, and L. Jiao, “A novel multi-model decision fusion network for object detection in remote sensing images,” Remote Sens., vol. 11, no. 7, p. 737, 2019. [184] S. Tian, L. Kang, X. Xing, Z. Li, L. Zhao, C. Fan, and Y. Zhang, “Siamese graph embedding network for object detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 18, no. 4, pp. 602–606, 2021. [185] S. Tian, L. Kang, X. Xing, J. Tian, C. Fan, and Y. Zhang, “A relationaugmented embedded graph attention network for remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–18, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3073269 [186] Y. Wu, K. Zhang, J. Wang, Y. Wang, Q. Wang, and Q. Li, “Cdd-net: A context-driven detection network for multiclass object detection,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2020.3042465 [187] Y. Han, J. Liao, T. Lu, T. Pu, and Z. Peng, “Kcpnet: Knowledge-driven context perception networks for ship detection in infrared imagery,” IEEE Trans. Geosci. Remote Sens., vol. 61, pp. 1–19, 2023. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3233401 [188] C. Chen, W. Gong, Y. Chen, and W. Li, “Object detection in remote sensing images based on a scene-contextual feature pyramid network,” Remote Sens., vol. 11, no. 3, p. 339, 2019. [189] Z. Wu, B. Hou, B. Ren, Z. Ren, S. Wang, and L. Jiao, “A deep detection network based on interaction of instance segmentation and object detection for SAR images,” Remote Sens., vol. 13, no. 13, p. 2582, 2021. [190] Y. Wu, K. Zhang, J. Wang, Y. Wang, Q. Wang, and X. Li, “Gcwnet: A global context-weaving network for object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–12, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3155899 [191] G. Shi, J. Zhang, J. Liu, C. Zhang, C. Zhou, and S. Yang, “Global context-augmented objection detection in VHR optical remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 12, pp. 10 60410 617, 2021. [192] J. Liu, S. Li, C. Zhou, X. Cao, Y. Gao, and B. Wang, “Sraf-net: A scene-relevant anchor-free object detection network in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3124959 [193] C. Tao, L. Mi, Y. Li, J. Qi, Y. Xiao, and J. Zhang, “Scene contextdriven vehicle detection in high-resolution aerial images,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 10, pp. 7339–7351, 2019. [194] K. Zhang, Y. Wu, J. Wang, Y. Wang, and Q. Wang, “Semantic contextaware network for multiscale object detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3067313 [195] M. Wang, Q. Li, Y. Gu, L. Fang, and X. X. Zhu, “Scaf-net: Scene context attention-based fusion network for vehicle detection in aerial imagery,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3107281 [196] G. Zhang, S. Lu, and W. Zhang, “Cad-net: A context-aware detection network for objects in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 12, pp. 10 015–10 024, 2019. [197] E. Liu, Y. Zheng, B. Pan, X. Xu, and Z. Shi, “Dcl-net: Augmenting the capability of classification and localization for remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 9, pp. 79337944, 2021. [198] Y. Feng, W. Diao, X. Sun, M. Yan, and X. Gao, “Towards automated ship detection and category recognition from high-resolution aerial images,” Remote Sens., vol. 11, no. 16, p. 1901, 2019. [199] P. Wang, X. Sun, W. Diao, and K. Fu, “FMSSD: feature-merged singleshot detection for multiscale objects in large-scale remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 5, pp. 33773390, 2020. [200] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 40, no. 4, pp. 834–848, 2018. [201] Y. Bai, R. Li, S. Gou, C. Zhang, Y. Chen, and Z. Zheng, “Crossconnected bidirectional pyramid network for infrared small-dim target detection,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3145577 [202] Y. Li, Q. Huang, X. Pei, Y. Chen, L. Jiao, and R. Shang, “Crosslayer attention network for small object detection in remote sensing imagery,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 2148–2161, 2021. [203] H. Gong, T. Mu, Q. Li, H. Dai, C. Li, Z. He, W. Wang, F. Han, A. Tuniyazi, H. Li, X. Lang, Z. Li, and B. Wang, “Swin-transformerenabled yolov5 with attention mechanism for small object detection on satellite images,” Remote Sens., vol. 14, no. 12, p. 2861, 2022. [204] J. Qu, C. Su, Z. Zhang, and A. Razi, “Dilated convolution and feature fusion SSD network for small object detection in remote sensing images,” IEEE Access, vol. 8, pp. 82 832–82 843, 2020. [205] T. Ma, Z. Yang, J. Wang, S. Sun, X. Ren, and U. Ahmad, “Infrared small target detection network with generate label and feature mapping,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3140432 [206] W. Han, A. Kuerban, Y. Yang, Z. Huang, B. Liu, and J. Gao, “Multi-vision network for accurate and real-time small object detection in optical remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https: //doi.org/10.1109/LGRS.2020.3044422 [207] Q. Hou, Z. Wang, F. Tan, Y. Zhao, H. Zheng, and W. Zhang, “Ristdnet: Robust infrared small target detection network,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3050828 [208] X. Lu, Y. Zhang, Y. Yuan, and Y. Feng, “Gated and axis-concentrated localization network for remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 1, pp. 179–192, 2020. [209] L. Courtrai, M. Pham, and S. Lef evre, “Small object detection in remote sensing images based on super-resolution with auxiliary generative adversarial networks,” Remote Sens., vol. 12, no. 19, p. 3152, 2020. [210] S. M. A. Bashir and Y. Wang, “Small object detection in remote sensing images with residual feature aggregation-based super-resolution and object detector network,” Remote Sens., vol. 13, no. 9, p. 1854, 2021. [211] J. Rabbi, N. Ray, M. Schubert, S. Chowdhury, and D. Chao, “Smallobject detection in remote sensing images with end-to-end edgeenhanced GAN and object detector network,” Remote Sens., vol. 12, no. 9, p. 1432, 2020. [212] J. Wu and S. Xu, “From point to region: Accurate and efficient hierarchical small object detection in low-resolution remote sensing images,” Remote Sens., vol. 13, no. 13, p. 2620, 2021. [213] J. Li, Z. Zhang, Y. Tian, Y. Xu, Y. Wen, and S. Wang, “Target-guided feature super-resolution for vehicle detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3112172 [214] J. Chen, K. Chen, H. Chen, Z. Zou, and Z. Shi, “A degraded reconstruction enhancement-based method for tiny ship detection in remote sensing images with a new large-scale dataset,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3180894 [215] J. Pang, C. Li, J. Shi, Z. Xu, and H. Feng, “R2-cnn: Fast tiny object detection in large-scale remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 8, pp. 5512–5524, 2019. [216] J. Wu, Z. Pan, B. Lei, and Y. Hu, “Fsanet: Feature-and-spatial-aligned network for tiny object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–17, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3205052 [217] M. Pham, L. Courtrai, C. Friguet, S. Lef evre, and A. Baussard, “Yolofine: One-stage detector of small objects under various backgrounds in remote sensing images,” Remote Sens., vol. 12, no. 15, p. 2501, 2020. [218] J. Yan, H. Wang, M. Yan, W. Diao, X. Sun, and H. Li, “Iouadaptive deformable R-CNN: make full use of iou for multi-class object detection in remote sensing imagery,” Remote Sens., vol. 11, no. 3, p. 286, 2019. [219] R. Dong, D. Xu, J. Zhao, L. Jiao, and J. An, “Sig-nms-based faster R-CNN combining transfer learning for small target detection in VHR optical remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 11, pp. 8534–8545, 2019. [220] Z. Shu, X. Hu, and J. Sun, “Center-point-guided proposal generation for detection of small and dense buildings in aerial imagery,” IEEE Geosci. Remote Sens. Lett., vol. 15, no. 7, pp. 1100–1104, 2018. [221] C. Xu, J. Wang, W. Yang, and L. Yu, “Dot distance for tiny object detection in aerial images,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. Workshops. IEEE, 2021, pp. 1192–1201. [222] C. Xu, J. Wang, W. Yang, H. Yu, L. Yu, and G. Xia, “RFLA: gaussian receptive field based label assignment for tiny object detection,” pp. 526–543, 2022. [223] F. Zhang, B. Du, L. Zhang, and M. Xu, “Weakly supervised learning based on coupled convolutional neural networks for aircraft detection,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 9, pp. 5553–5563, 2016. [224] Y. Li, B. He, F. Melgani, and T. Long, “Point-based weakly supervised learning for object detection in high spatial resolution remote sensing images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 5361–5371, 2021. [225] D. Zhang, J. Han, G. Cheng, Z. Liu, S. Bu, and L. Guo, “Weakly supervised learning for target detection in remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 12, no. 4, pp. 701–705, 2015. [226] J. Han, D. Zhang, G. Cheng, L. Guo, and J. Ren, “Object detection in optical remote sensing images based on weakly supervised learning and high-level feature learning,” IEEE Trans. Geosci. Remote Sens., vol. 53, no. 6, pp. 3325–3337, 2015. [227] Y. Li, Y. Zhang, X. Huang, and A. L. Yuille, “Deep networks under scene-level supervision for multi-class geospatial object detection from remote sensing images,” ISPRS J. Photogrammetry Remote Sens., vol. 146, pp. 182–196, 2018. [228] H. Bilen and A. Vedaldi, “Weakly supervised deep detection networks,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 2846–2854. [229] X. Yao, X. Feng, J. Han, G. Cheng, and L. Guo, “Automatic weakly supervised object detection from high spatial resolution remote sensing images via dynamic curriculum learning,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 1, pp. 675–685, 2021. [230] H. Wang, H. Li, W. Qian, W. Diao, L. Zhao, J. Zhang, and D. Zhang, “Dynamic pseudo-label generation for weakly supervised object detection in remote sensing images,” Remote Sens., vol. 13, no. 8, p. 1461, 2021. [231] X. Feng, J. Han, X. Yao, and G. Cheng, “Progressive contextual instance refinement for weakly supervised object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 11, pp. 8002–8012, 2020. [232] P. Shamsolmoali, J. Chanussot, M. Zareapoor, H. Zhou, and J. Yang, “Multipatch feature pyramid network for weakly supervised object detection in optical remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–13, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3106442 [233] B. Wang, Y. Zhao, and X. Li, “Multiple instance graph learning for weakly supervised remote sensing object detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–12, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3123231 [234] X. Feng, J. Han, X. Yao, and G. Cheng, “Tcanet: Triple contextaware network for weakly supervised object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 8, pp. 69466955, 2021. [235] X. Feng, X. Yao, G. Cheng, J. Han, and J. Han, “Saenet: Self-supervised adversarial and equivariant network for weakly supervised object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–11, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3105575 [236] X. Qian, Y. Huo, G. Cheng, X. Yao, K. Li, H. Ren, and W. Wang, “Incorporating the completeness and difficulty of proposals into weakly supervised object detection in remote sensing images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 1902–1911, 2022. [237] W. Qian, Z. Yan, Z. Zhu, and W. Yin, “Weakly supervised part-based method for combined object detection in remote sensing imagery,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 50245036, 2022. [238] S. Chen, D. Shao, X. Shu, C. Zhang, and J. Wang, “Fcc-net: A fullcoverage collaborative network for weakly supervised remote sensing object detection,” Electronics, vol. 9, no. 9, p. 1356, 2020. [239] G. Cheng, X. Xie, W. Chen, X. Feng, X. Yao, and J. Han, “Selfguided proposal generation for weakly supervised object detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–11, 2022. [240] X. Feng, X. Yao, G. Cheng, and J. Han, “Weakly supervised rotationinvariant aerial object detection network,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR). IEEE, 2022, pp. 14 12614 135. [241] G. Wang, X. Zhang, Z. Peng, X. Jia, X. Tang, and L. Jiao, “Mol: Towards accurate weakly supervised remote sensing object detection via multi-view noisy learning,” ISPRS J. Photogrammetry Remote Sens., vol. 196, pp. 457–470, 2023. [242] T. Deselaers, B. Alexe, and V. Ferrari, “Weakly supervised localization and learning with generic knowledge,” Int. J. Comput. Vis., vol. 100, no. 3, pp. 275–293, 2012. [243] B. Hou, Z. Wu, B. Ren, Z. Li, X. Guo, S. Wang, and L. Jiao, “A neural network based on consistency learning and adversarial learning for semisupervised synthetic aperture radar ship detection,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [244] Z. Song, J. Yang, D. Zhang, S. Wang, and Z. Li, “Semi-supervised dim and small infrared ship detection network based on haar wavelet,” IEEE Access, vol. 9, pp. 29 686–29 695, 2021. [245] Y. Zhong, Z. Zheng, A. Ma, X. Lu, and L. Zhang, “COLOR: cycling, offline learning, and online representation framework for airport and airplane detection using GF-2 satellite images,” IEEE Trans. Geosci. Remote Sens., vol. 58, no. 12, pp. 8438–8449, 2020. [246] Y. Wu, W. Zhao, R. Zhang, and F. Jiang, “Amr-net: Arbitrary-oriented ship detection using attention module, multi-scale feature fusion and rotation pseudo-label,” IEEE Access, vol. 9, pp. 68 208–68 222, 2021. [247] S. Chen, R. Zhan, W. Wang, and J. Zhang, “Domain adaptation for semi-supervised ship detection in SAR images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3171789 [248] Z. Zhang, Z. Feng, and S. Yang, “Semi-supervised object detection framework with object first mixup for remote sensing images,” in International Geoscience and Remote Sensing Symposium. IEEE, 2021, pp. 2596–2599. [249] B. Xue and N. Tong, “DIOD: fast and efficient weakly semi-supervised deep complex ISAR object detection,” IEEE Trans. Cybern., vol. 49, no. 11, pp. 3991–4003, 2019. [250] L. Liao, L. Du, and Y. Guo, “Semi-supervised SAR target detection based on an improved faster R-CNN,” Remote Sens., vol. 14, no. 1, p. 143, 2022. [Online]. Available: https://doi.org/10.3390/rs14010143 [251] Y. Du, L. Du, Y. Guo, and Y. Shi, “Semisupervised sar ship detection network via scene characteristic learning,” IEEE Trans. Geosci. Remote Sens., vol. 61, pp. 1–17, 2023. [Online]. Available: https://doi.org/10.1109/TGRS.2023.3235859 [252] D. Wei, Y. Du, L. Du, and L. Li, “Target detection network for SAR images based on semi-supervised learning and attention mechanism,” Remote Sens., vol. 13, no. 14, p. 2686, 2021. [253] L. Chen, Y. Fu, S. You, and H. Liu, “Efficient hybrid supervision for instance segmentation in aerial images,” Remote Sens., vol. 13, no. 2, p. 252, 2021. [254] G. Cheng, B. Yan, P. Shi, K. Li, X. Yao, L. Guo, and J. Han, “Prototype-cnn for few-shot object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–10, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3078507 [255] X. Li, J. Deng, and Y. Fang, “Few-shot object detection on remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3051383 [256] L. Li, X. Yao, G. Cheng, M. Xu, J. Han, and J. Han, “Solo-tocollaborative dual-attention network for one-shot object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–11, 2022. [Online]. Available: https://doi.org/10.1109/TGRS. 2021.3091003 [257] H. Zhang, X. Zhang, G. Meng, C. Guo, and Z. Jiang, “Few-shot multiclass ship detection in remote sensing images using attention feature map and multi-relation detector,” Remote Sens., vol. 14, no. 12, p. 2790, 2022. [258] B. Wang, Z. Wang, X. Sun, H. Wang, and K. Fu, “Dmmlnet: Deep metametric learning for few-shot geographic object segmentation in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–18, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3116672 [259] J. Li, Y. Tian, Y. Xu, X. Hu, Z. Zhang, H. Wang, and Y. Xiao, “Mmrcnn: Toward few-shot object detection in remote sensing images with meta memory,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–14, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3228612 [260] Z. Zhao, P. Tang, L. Zhao, and Z. Zhang, “Few-shot object detection of remote sensing images via two-stage fine-tuning,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2021.3116858 [261] Y. Zhou, H. Hu, J. Zhao, H. Zhu, R. Yao, and W. Du, “Few-shot object detection via context-aware aggregation for remote sensing images,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022.3171257 [262] Y. Wang, C. Xu, C. Liu, and Z. Li, “Context information refinement for few-shot object detection in remote sensing images,” Remote Sens., vol. 14, no. 14, p. 3255, 2022. [263] Z. Zhou, S. Li, W. Guo, and Y. Gu, “Few-shot aircraft detection in satellite videos based on feature scale selection pyramid and proposal contrastive learning,” Remote Sens., vol. 14, no. 18, p. 4581, 2022. [264] S. Chen, J. Zhang, R. Zhan, R. Zhu, and W. Wang, “Few shot object detection for SAR images via feature enhancement and dynamic relationship modeling,” Remote Sens., vol. 14, no. 15, p. 3669, 2022. [265] S. Liu, Y. You, H. Su, G. Meng, W. Yang, and F. Liu, “Few-shot object detection in remote sensing image interpretation: Opportunities and challenges,” Remote Sens., vol. 14, no. 18, p. 4435, 2022. [266] X. Huang, B. He, M. Tong, D. Wang, and C. He, “Few-shot object detection on remote sensing images via shared attention module and balanced fine-tuning strategy,” Remote Sens., vol. 13, no. 19, p. 3816, 2021. [267] Z. Xiao, J. Qi, W. Xue, and P. Zhong, “Few-shot object detection with self-adaptive attention network for remote sensing images,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 4854–4865, 2021. [268] S. Wolf, J. Meier, L. Sommer, and J. Beyerer, “Double head predictor based few-shot object detection for aerial imagery,” in Proc. IEEE Int. Conf. Comput. Vis. Workshops. IEEE, 2021, pp. 721–731. [269] T. Zhang, X. Zhang, P. Zhu, X. Jia, X. Tang, and L. Jiao, “Generalized few-shot object detection in remote sensing images,” ISPRS J. Photogrammetry Remote Sens., vol. 195, pp. 353–364, 2023. [270] G. Heitz and D. Koller, “Learning spatial context: Using stuff to find things,” in Proc. Euro. Conf. Comput. Vis., vol. 5302. Springer, 2008, pp. 30–43. [271] C. Benedek, X. Descombes, and J. Zerubia, “Building development monitoring in multitemporal remotely sensed image pairs with stochastic birth-death dynamics,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 34, no. 1, pp. 33–50, 2012. [272] S. Razakarivony and F. Jurie, “Vehicle detection in aerial imagery : A small target detection benchmark,” J. Vis. Commun. Image Represent., vol. 34, pp. 187–203, 2016. [273] K. Liu and G. M ́ attyus, “Fast multiclass vehicle detection on aerial images,” IEEE Geosci. Remote Sens. Lett., vol. 12, no. 9, pp. 19381942, 2015. [274] H. Zhu, X. Chen, W. Dai, K. Fu, Q. Ye, and J. Jiao, “Orientation robust object detection in aerial images using deep convolutional neural network,” in Proc. IEEE Int. Conf. Image Process., 2015, pp. 37353739. [275] T. N. Mundhenk, G. Konjevod, W. A. Sakla, and K. Boakye, “A large contextual dataset for classification, detection and counting of cars with deep learning,” in Proc. Euro. Conf. Comput. Vis., B. Leibe, J. Matas, N. Sebe, and M. Welling, Eds., vol. 9907, 2016, pp. 785–800. [276] Z. Liu, H. Wang, L. Weng, and Y. Yang, “Ship rotated bounding box space for ship extraction from high-resolution optical satellite images with complex backgrounds,” IEEE Geosci. Remote Sens. Lett., vol. 13, no. 8, pp. 1074–1078, 2016. [277] J. Li, C. Qu, and J. Shao, “Ship detection in sar images based on an improved faster r-cnn,” in SAR in Big Data Era: Models, Methods and Applications (BIGSARDATA). IEEE, 2017, pp. 1–6. [278] Z. Zou and Z. Shi, “Random access memories: A new paradigm for target detection in high resolution aerial remote sensing images,” IEEE Trans. Image Process., vol. 27, no. 3, pp. 1100–1111, 2018. [279] X. Sun, Z. Wang, Y. Sun, W. Diao, Y. Zhang, and K. Fu, “Air-sarship1.0: High-resolution sar ship detection dataset,” Journal of Radars, vol. 8, no. 6, pp. 852–863, 2019. [280] W. Yu, G. Cheng, M. Wang, Y. Yao, X. Xie, X. Yao, and J. Han, “Mar20: A benchmark for military aircraft recognition in remote sensing images,” National Remote Sensing Bulletin, pp. 1–11, 2022. [281] K. Chen, M. Wu, J. Liu, and C. Zhang, “FGSD: A dataset for fine-grained ship detection in high resolution satellite images,” 2020. [Online]. Available: https://arxiv.org/abs/2003.06832 [282] Y. Han, X. Yang, T. Pu, and Z. Peng, “Fine-grained recognition for oriented ship against complex scenes in optical remote sensing images,” IEEE Trans. Geosci. Remote. Sens., vol. 60, pp. 1–18, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3123666 [283] J. Wang, W. Yang, H. Guo, R. Zhang, and G. Xia, “Tiny object detection in aerial images,” in Proc. Int. Conf. Pattern Recognit. IEEE, 2020, pp. 3791–3798. [284] G. Cheng, X. Yuan, X. Yao, K. Yan, Q. Zeng, and J. Han, “Towards large-scale small object detection: Survey and benchmarks,” 2022. [Online]. Available: https://doi.org/10.48550/arXiv.2207.14096 [285] T. Zhang, X. Zhang, J. Shi, and S. Wei, “Hyperli-net: A hyper-light deep learning network for high-accurate and high-speed ship detection from synthetic aperture radar imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 167, pp. 123–153, 2020. [286] T. Zhang, X. Zhang, J. Li, X. Xu, B. Wang, X. Zhan, Y. Xu, X. Ke, T. Zeng, H. Su, I. Ahmad, D. Pan, C. Liu, Y. Zhou, J. Shi, and S. Wei, “SAR ship detection dataset (SSDD): official release and comprehensive data analysis,” Remote Sens., vol. 13, no. 18, p. 3690, 2021. [287] M. Everingham, L. V. Gool, C. K. I. Williams, J. M. Winn, and A. Zisserman, “The pascal visual object classes (VOC) challenge,” Int. J. Comput. Vis., vol. 88, no. 2, pp. 303–338, 2010. [288] A. G. Menezes, G. de Moura, C. Alves, and A. C. P. L. F. de Carvalho, “Continual object detection: A review of definitions, strategies, and challenges,” Neural Networks, vol. 161, pp. 476–493, 2023. [289] C. Persello, J. D. Wegner, R. H ̈ ansch, D. Tuia, P. Ghamisi, M. Koeva, and G. Camps-Valls, “Deep learning and earth observation to support the sustainable development goals: Current approaches, open challenges, and future opportunities,” IEEE Geosci. Remote Sens. Mag., vol. 10, no. 2, pp. 172–200, 2022. [290] T. Hoeser, F. Bachofer, and C. Kuenzer, “Object detection and image segmentation with deep learning on earth observation data: A review—part ii: Applications,” Remote Sen., vol. 12, no. 18, p. 3053, 2020. [291] L. Ma, Y. Liu, X. Zhang, Y. Ye, G. Yin, and B. A. Johnson, “Deep learning in remote sensing applications: A meta-analysis and review,” ISPRS J. Photogrammetry Remote Sens., vol. 152, pp. 166–177, 2019. [292] P. Barmpoutis, P. Papaioannou, K. Dimitropoulos, and N. Grammalidis, “A review on early forest fire detection systems using optical remote sensing,” Sensors, vol. 20, no. 22, p. 6442, 2020. [293] Z. Guan, X. Miao, Y. Mu, Q. Sun, Q. Ye, and D. Gao, “Forest fire segmentation from aerial imagery data using an improved instance segmentation model,” Remote. Sens., vol. 14, no. 13, p. 3159, 2022. [294] Z. Zheng, Y. Zhong, J. Wang, A. Ma, and L. Zhang, “Building damage assessment for rapid disaster response with a deep objectbased semantic change detection framework: From natural disasters to man-made disasters,” Remote Sens. Environ., vol. 265, p. 112636, 2021. [295] H. Ma, Y. Liu, Y. Ren, and J. Yu, “Detection of collapsed buildings in post-earthquake remote sensing images based on the improved yolov3,” Remote. Sens., vol. 12, no. 1, p. 44, 2020. [296] Y. Pi, N. D. Nath, and A. H. Behzadan, “Convolutional neural networks for object detection in aerial imagery for disaster response and recovery,” Adv. Eng. Informatics, vol. 43, p. 101009, 2020. [297] M. Weiss, F. Jacob, and G. Duveiller, “Remote sensing for agricultural applications: A meta-review,” Remote Sens. Environ., vol. 236, p. 111402, 2020. [298] Y. Pang, Y. Shi, S. Gao, F. Jiang, A. N. V. Sivakumar, L. Thompson, J. D. Luck, and C. Liu, “Improved crop row detection with deep neural network for early-season maize stand count in UAV imagery,” Comput. Electron. Agric., vol. 178, p. 105766, 2020. [299] C. Mota-Delfin, G. de Jes ́ us L ́ opez-Cante ̃ ns, I. L. L. Cruz, E. Romantchik-Kriuchkova, and J. C. Olguı ́n-Rojas, “Detection and counting of corn plants in the presence of weeds with convolutional neural networks,” Remote Sens., vol. 14, no. 19, p. 4892, 2022. [300] L. P. Osco, M. d. S. de Arruda, D. N. Gon ̧ calves, A. Dias, J. Batistoti, M. de Souza, F. D. G. Gomes, A. P. M. Ramos, L. A. de Castro Jorge, V. Liesenberg et al., “A cnn approach to simultaneously count plants and detect plantation-rows from uav imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 174, pp. 1–17, 2021. [301] M. M. Anuar, A. A. Halin, T. Perumal, and B. Kalantar, “Aerial imagery paddy seedlings inspection using deep learning,” Remote Sens., vol. 14, no. 2, p. 274, 2022. [302] Y. Chen, W. S. Lee, H. Gan, N. Peres, C. W. Fraisse, Y. Zhang, and Y. He, “Strawberry yield prediction based on a deep neural network using high-resolution aerial orthoimages,” Remote Sens., vol. 11, no. 13, p. 1584, 2019. [303] W. Zhao, C. Persello, and A. Stein, “Building outline delineation: From aerial images to polygons with an improved end-to-end learning framework,” ISPRS J. Photogrammetry Remote Sens., vol. 175, pp. 119–131, 2021. [304] Z. Li, J. D. Wegner, and A. Lucchi, “Topological map extraction from overhead images,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2019, pp. 1715–1724. [305] L. Mou and X. X. Zhu, “Vehicle instance segmentation from aerial image and video using a multitask learning residual fully convolutional network,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 11, pp. 66996711, 2018. [306] J. Zhang, X. Zhang, Z. Huang, X. Cheng, J. Feng, and L. Jiao, “Bidirectional multiple object tracking based on trajectory criteria in satellite videos,” IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–14, 2023. [307] H. Kim and Y. Ham, “Participatory sensing-based geospatial localization of distant objects for disaster preparedness in urban built environments,” Automation in Construction, vol. 107, p. 102960, 2019. [308] M. A. E. Bhuiyan, C. Witharana, and A. K. Liljedahl, “Use of very high spatial resolution commercial satellite imagery and deep learning to automatically map ice-wedge polygons across tundra vegetation types,” J. Imaging, vol. 6, no. 12, p. 137, 2020. [309] W. Zhang, A. K. Liljedahl, M. Kanevskiy, H. E. Epstein, B. M. Jones, M. T. Jorgenson, and K. Kent, “Transferability of the deep learning mask R-CNN model for automated mapping of ice-wedge polygons in high-resolution satellite and UAV images,” Remote Sens., vol. 12, no. 7, p. 1085, 2020. [310] C. Witharana, M. A. E. Bhuiyan, A. K. Liljedahl, M. Kanevskiy, M. T. Jorgenson, B. M. Jones, R. Daanen, H. E. Epstein, C. G. Griffin, K. Kent, and M. K. W. Jones, “An object-based approach for mapping tundra ice-wedge polygon troughs from very high spatial resolution optical satellite imagery,” Remote Sens., vol. 13, no. 4, p. 558, 2021. [311] J. Yu, Z. Wang, A. Majumdar, and R. Rajagopal, “Deepsolar: A machine learning framework to efficiently construct a solar deployment database in the united states,” Joule, vol. 2, no. 12, pp. 2605–2617, 2018. [312] J. M. Malof, K. Bradbury, L. M. Collins, and R. G. Newell, “Automatic detection of solar photovoltaic arrays in high resolution aerial imagery,” Applied energy, vol. 183, pp. 229–240, 2016. [313] W. Zhang, G. Wang, J. Qi, G. Wang, and T. Zhang, “Research on the extraction of wind turbine all over the china based on domestic satellite remote sensing data,” in International Geoscience and Remote Sensing Symposium. IEEE, 2021, pp. 4167–4170. [314] W. Hu, T. Feldman, Y. J. Ou, N. Tarn, B. Ye, Y. Xu, J. M. Malof, and K. Bradbury, “Wind turbine detection with synthetic overhead imagery,” in International Geoscience and Remote Sensing Symposium. IEEE, 2021, pp. 4908–4911. [315] T. Jia, Z. Kapelan, R. de Vries, P. Vriend, E. C. Peereboom, I. Okkerman, and R. Taormina, “Deep learning for detecting macroplastic litter in water bodies: A review,” Water Research, vol. 231, p. 119632, 2023. [316] C. Martin, Q. Zhang, D. Zhai, X. Zhang, and C. M. Duarte, “Enabling a large-scale assessment of litter along saudi arabian red sea shores by combining drones and machine learning,” Environmental Pollution, vol. 277, p. 116730, 2021. [317] K. Themistocleous, C. Papoutsa, S. C. Michaelides, and D. G. Hadjimitsis, “Investigating detection of floating plastic litter from space using sentinel-2 imagery,” Remote. Sens., vol. 12, no. 16, p. 2648, 2020. [318] B. Xue, B. Huang, W. Wei, G. Chen, H. Li, N. Zhao, and H. Zhang, “An efficient deep-sea debris detection method using deep neural networks,” IEEE J. Sel. Top. Appl. Earth Obs. Remote. Sens., vol. 14, pp. 12 34812 360, 2021. [319] J. Peng, D. Wang, X. Liao, Q. Shao, Z. Sun, H. Yue, and H. Ye, “Wild animal survey using uas imagery and deep learning: modified faster rcnn for kiang detection in tibetan plateau,” ISPRS J. Photogrammetry Remote Sens., vol. 169, pp. 364–376, 2020. [320] N. Rey, M. Volpi, S. Joost, and D. Tuia, “Detecting animals in african savanna with uavs and the crowds,” Remote Sens. Environ., vol. 200, pp. 341–351, 2017. [321] B. Kellenberger, D. Marcos, and D. Tuia, “Detecting mammals in uav images: Best practices to address a substantially imbalanced dataset with deep learning,” Remote Sens. Environ., vol. 216, pp. 139–153, 2018. [322] A. Delplanque, S. Foucher, P. Lejeune, J. Linchant, and J. Th ́ eau, “Multispecies detection and identification of african mammals in aerial imagery using convolutional neural networks,” Remote Sensing in Ecology and Conservation, vol. 8, no. 2, pp. 166–179, 2022. [323] D. Wang, Q. Shao, and H. Yue, “Surveying wild animals from satellites, manned aircraft and unmanned aerial systems (uass): A review,” Remote Sen., vol. 11, no. 11, p. 1308, 2019. [324] T. Kattenborn, J. Leitloff, F. Schiefer, and S. Hinz, “Review on convolutional neural networks (cnn) in vegetation remote sensing,” ISPRS J. Photogrammetry Remote Sens., vol. 173, pp. 24–49, 2021. [325] T. Dong, Y. Shen, J. Zhang, Y. Ye, and J. Fan, “Progressive cascaded convolutional neural networks for single tree detection with google earth imagery,” Remote. Sens., vol. 11, no. 15, p. 1786, 2019. [326] A. Safonova, S. Tabik, D. Alcaraz-Segura, A. Rubtsov, Y. Maglinets, and F. Herrera, “Detection of fir trees (Abies sibirica) damaged by the bark beetle in unmanned aerial vehicle images with deep learning,” Remote. Sens., vol. 11, no. 6, p. 643, 2019. [327] Z. Hao, L. Lin, C. J. Post, E. A. Mikhailova, M. Li, Y. Chen, K. Yu, and J. Liu, “Automated tree-crown and height detection in a young forest plantation using mask region-based convolutional neural network (mask r-cnn),” ISPRS J. Photogrammetry Remote Sens., vol. 178, pp. 112–123, 2021. [328] A. Sani-Mohammed, W. Yao, and M. Heurich, “Instance segmentation of standing dead trees in dense forest from aerial imagery using deep learning,” ISPRS O. J. Photogrammetry Remote Sens., vol. 6, p. 100024, 2022. [329] A. V. Etten, “You only look twice: Rapid multi-scale object detection in satellite imagery,” 2018. [Online]. Available: http: //arxiv.org/abs/1805.09512 [330] Q. Lin, J. Zhao, G. Fu, and Z. Yuan, “Crpn-sfnet: A high-performance object detector on large-scale remote sensing images,” IEEE Trans. Neural Networks Learn. Syst., vol. 33, no. 1, pp. 416–429, 2022. [331] D. Hong, L. Gao, N. Yokoya, J. Yao, J. Chanussot, Q. Du, and B. Zhang, “More diverse means better: Multimodal deep learning meets remote-sensing imagery classification,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 5, pp. 4340–4354, 2021. [332] D. Hong, N. Yokoya, G.-S. Xia, J. Chanussot, and X. X. Zhu, “Xmodalnet: A semi-supervised deep cross-modal network for classifica-tion of remote sensing data,” ISPRS J. Photogrammetry Remote Sens., vol. 167, pp. 12–23, 2020. [333] M. Segal-Rozenhaimer, A. Li, K. Das, and V. Chirayath, “Cloud detection algorithm for multi-modal satellite imagery using convolutional neural-networks (cnn),” Remote Sens. Environ., vol. 237, p. 111446, 2020. [334] Y. Shendryk, Y. Rist, C. Ticehurst, and P. Thorburn, “Deep learning for multi-modal classification of cloud, shadow and land cover scenes in planetscope and sentinel-2 imagery,” ISPRS J. Photogrammetry Remote Sens., vol. 157, pp. 124–136, 2019. [335] Y. Shi, L. Du, and Y. Guo, “Unsupervised domain adaptation for SAR target detection,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 6372–6385, 2021. [336] Y. Zhu, X. Sun, W. Diao, H. Li, and K. Fu, “Rfa-net: Reconstructed feature alignment network for domain adaptation object detection in remote sensing imagery,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 5689–5703, 2022. [337] T. Xu, X. Sun, W. Diao, L. Zhao, K. Fu, and H. Wang, “Fada: Feature aligned domain adaptive object detection in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [338] Y. Koga, H. Miyazaki, and R. Shibasaki, “A method for vehicle detection in high-resolution satellite images that uses a region-based object detector and unsupervised domain adaptation,” Remote Sens., vol. 12, no. 3, p. 575, 2020. [339] Y. Shi, L. Du, Y. Guo, and Y. Du, “Unsupervised domain adaptation based on progressive transfer for ship detection: From optical to SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–17, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3185298 [340] P. Zhang, H. Xu, T. Tian, P. Gao, L. Li, T. Zhao, N. Zhang, and J. Tian, “Sefepnet: Scale expansion and feature enhancement pyramid network for SAR aircraft detection with small sample dataset,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 15, pp. 3365–3375, 2022. [341] S. Dang, Z. Cao, Z. Cui, Y. Pi, and N. Liu, “Open set incremental learning for automatic target recognition,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 7, pp. 4445–4456, 2019. [342] J. Chen, S. Wang, L. Chen, H. Cai, and Y. Qian, “Incremental detection of remote sensing objects with feature pyramid and knowledge distillation,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–13, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2020.3042554 [343] X. Chen, J. Jiang, Z. Li, H. Qi, Q. Li, J. Liu, L. Zheng, M. Liu, and Y. Deng, “An online continual object detector on VHR remote sensing images with class imbalance,” Eng. Appl. Artif. Intell., vol. 117, no. Part, p. 105549, 2023. [344] J. Li, X. Sun, W. Diao, P. Wang, Y. Feng, X. Lu, and G. Xu, “Class-incremental learning network for small objects enhancing of semantic segmentation in aerial imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–20, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3124303 [345] W. Liu, X. Nie, B. Zhang, and X. Sun, “Incremental learning with open-set recognition for remote sensing image scene classification,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3173995 [346] J. Deng, W. Dong, R. Socher, L. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR), 2009, pp. 248–255. [347] Y. Long, G. Xia, S. Li, W. Yang, M. Y. Yang, X. X. Zhu, L. Zhang, and D. Li, “On creating benchmark dataset for aerial image interpretation: Reviews, guidances, and million-aid,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 4205–4230, 2021. [348] G. A. Christie, N. Fendley, J. Wilson, and R. Mukherjee, “Functional map of the world,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 6172–6180. [349] D. Wang, J. Zhang, B. Du, G.-S. Xia, and D. Tao, “An empirical study of remote sensing pretraining,” IEEE Trans. Geosci. Remote Sens., pp. 1–1, 2022. [Online]. Available: https: //doi.org/10.1109/TGRS.2022.3176603 [350] W. Li, K. Chen, H. Chen, and Z. Shi, “Geographical knowledgedriven representation learning for remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–16, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3115569 [351] X. Sun, P. Wang, W. Lu, Z. Zhu, X. Lu, Q. He, J. Li, X. Rong, Z. Yang, H. Chang, Q. He, G. Yang, R. Wang, J. Lu, and K. Fu, “Ringmo: A remote sensing foundation model with masked image modeling,” IEEE Trans. Geosci. Remote Sens., pp. 1–1, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3194732 [352] A. Fuller, K. Millard, and J. R. Green, “Satvit: Pretraining transformers for earth observation,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022. [Online]. Available: https://doi.org/10.1109/LGRS.2022. 3201489 [353] D. Wang, Q. Zhang, Y. Xu, J. Zhang, B. Du, D. Tao, and L. Zhang, “Advancing plain vision transformer towards remote sensing foundation model,” IEEE Trans. Geosci. Remote Sens., pp. 1–1, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3222818 [354] T. Zhang and X. Zhang, “Shipdenet-20: An only 20 convolution layers and 1-mb lightweight SAR ship detector,” IEEE Geosci. Remote Sens. Lett., vol. 18, no. 7, pp. 1234–1238, 2021. [355] T. Zhang, X. Zhang, J. Shi, and S. Wei, “Depthwise separable convolution neural network for high-speed SAR ship detection,” Remote Sens., vol. 11, no. 21, p. 2483, 2019. [356] Z. Wang, L. Du, and Y. Li, “Boosting lightweight cnns through network pruning and knowledge distillation for SAR target recognition,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 8386–8397, 2021. [357] S. Chen, R. Zhan, W. Wang, and J. Zhang, “Learning slimming SAR ship object detector through network pruning and knowledge distillation,” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 14, pp. 1267–1282, 2021. [358] Y. Zhang, Z. Yan, X. Sun, W. Diao, K. Fu, and L. Wang, “Learning efficient and accurate detectors with dynamic knowledge distillation in remote sensing imagery,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–19, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2021.3130443 [359] Y. Yang, X. Sun, W. Diao, H. Li, Y. Wu, X. Li, and K. Fu, “Adaptive knowledge distillation for lightweight remote sensing object detectors optimizing,” IEEE Trans. Geosci. Remote Sens., vol. 60, pp. 1–15, 2022. [Online]. Available: https://doi.org/10.1109/TGRS.2022.3175213 [360] C. Li, G. Cheng, G. Wang, P. Zhou, and J. Han, “Instance-aware distillation for efficient object detection in remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 61, pp. 1–11, 2023. [Online]. Available: https://doi.org/10.1109/TGRS.2023.3238801
