Detecting Fine-Grained Airplanes in SAR Images With Sparse Attention-Guided Pyramid and Class-Balanced Data Augmentation
Wei Bao, Jingjing Hu, Meiyu Huang, Yao Xu, Nan Ji, Xueshuang Xiang
Abstract
Airplane detection in synthetic aperture radar (SAR) images has drawn much attention owing to the success of deep learning methods. However, the development of fine-grained airplane detection in SAR images is still in a dilemma due to the small inter-class variance and the large intra-class variance in complex scenes with strong interference from the background. In addition, the class imbalance problem in multi-class fine-grained airplane recognition also significantly limits the direct application of general deep-learning-based airplane detectors. This paper proposes two effective methods to tackle the above two problems, respectively. First, we propose a sparse attention-guided fine-grained pyramid (SA-FP) module to simultaneously sample discriminative local features scattered in multi-scale layers and adaptively aggregate them with fine-grained attention to better classify subordinate-level airplanes with multiple scales. Second, a simple class-balanced copy-paste data augmentation (CC-DA) strategy, which randomly copies an airplane of one category and pastes it onto an image according to the class-wise probability, is proposed for class balance. Finally, extensive experiments on one public dataset and three representative deep-learning-based detection benchmarks are conducted to show the effectiveness and generalization of the two proposed methods. The combination of these two methods based on the Cascade R-CNN benchmark also won fifth place in fine-grained airplane detection in SAR images in the 2021 Gaofen challenge.