The task of instance segmentation in remote sensing images, aiming at performing per-pixel labeling of objects at instance level, is of great importance for various civil applications. Despite previous successes, most existing instance segmentation methods designed for natural images encounter sharp performance degradations when they are directly applied to top-view remote sensing images. Through careful analysis, we observe that the challenges mainly come from the lack of discriminative object features due to severe scale variations, low contrasts, and clustered distributions. In order to address these problems, a novel context aggregation network (CATNet) is proposed to improve the feature extraction process. The proposed model exploits three lightweight plug-and-play modules, namely dense feature pyramid network (DenseFPN), spatial context pyramid (SCP), and hierarchical region of interest extractor (HRoIE), to aggregate global visual context at feature, spatial, and instance domains, respectively. DenseFPN is a multi-scale feature propagation module that establishes more flexible information flows by adopting inter-level residual connections, cross-level dense connections, and feature re-weighting strategy. Leveraging the attention mechanism, SCP further augments the features by aggregating global spatial context into local regions. For each instance, HRoIE adaptively generates RoI features for different downstream tasks. Extensive evaluations of the proposed scheme on iSAID, DIOR, NWPU VHR-10, and HRSID datasets demonstrate that the proposed approach outperforms state-of-the-arts under similar computational costs. Source code and pre-trained models are available at https://github.com/yeliudev/CATNet.

翻译：遥感图像中的实例分割任务旨在对物体进行像素级实例标记，在各类民用应用中具有重要意义。尽管已有众多成功案例，但大多数为自然图像设计的现有实例分割方法在直接应用于俯视遥感图像时，性能会显著下降。通过仔细分析，我们观察到这些挑战主要源于尺度剧烈变化、低对比度和密集分布导致的判别性物体特征缺失。为解决这些问题，我们提出了一种新颖的上下文聚合网络（CATNet）以增强特征提取过程。该模型利用三个轻量级即插即用模块——密集特征金字塔网络（DenseFPN）、空间上下文金字塔（SCP）和分层感兴趣区域提取器（HRoIE），分别在特征域、空间域和实例域聚合全局视觉上下文。DenseFPN是一种多尺度特征传播模块，通过采用层间残差连接、跨层密集连接和特征重加权策略，建立了更灵活的信息流。借助注意力机制，SCP通过将全局空间上下文聚合到局部区域进一步增强特征。对于每个实例，HRoIE自适应地针对不同下游任务生成RoI特征。在iSAID、DIOR、NWPU VHR-10和HRSID数据集上的广泛评估表明，所提方法在相似计算成本下优于现有最先进方法。源代码与预训练模型已开源：https://github.com/yeliudev/CATNet。