EM-YOLO: A Fine-grained Recognition Model for Aircraft Targets in SAR Images

Fine-grained recognition of aircraft targets in SAR images is a challenging task due to the complexity of target appearance, detail differences and sparse information. Improving the recognition accuracy has important academic and practical significance for improving the target detection effect. In order to solve the problems of sparse features, missing information, large intra-class difference and small inter-class difference in the finegrained recognition task of aircraft targets in SAR images, this paper designs ASCPA-EMF network, and constructs a finegrained recognition model named EM-YOLO combined with YOLOv8n model. The experimental results show that EM-YOLO performs well in the fine-grained recognition task of SAR image aircraft and significantly improves the recognition accuracy.