The Research on the SAR Image Aircraft Object Detection Algorithm Based on Improved YOLOv8s

Be aimed at the problems of target discrete characteristics, structural similarity, background complexity, and difficulty in detecting small and multi-scale targets in SAR (Synthetic Aperture Radar), in this paper, a method of aircraft object detection based on improved YOLOv8s algorithm is proposed. Through multi-level feature enhancement, progressive attention strategy, and lightweight design, the model has better feature expression ability, stronger small object detection performance, and higher detection accuracy. This paper designed a multi-level feature enhancement method, adding the channel attention mechanism in the C2f module in the Backbone stage, replacing conventional convolution with dynamic convolution, and GC-Block module in the Head stage. Through the above improvements, the channel features at the Backbone level, the spatial features at the Neck level, and the comprehensive features at the Head level. At the same time, from the shallow to the deep, from the single attention to the gradual expression strategy of the model. The advantages of small calculation amounts and parallel calculation are that they ensure the real-time performance of the model. The validity of the proposed method is confirmed by experiments performed on the publicly available SAR image dataset SAR-AIRcraft-1.0. The experimental results show that the improved YOLO algorithm is superior to the original detection accuracy and provides an efficient and accurate solution for the field of SAR image aircraft object detection.