基于感受野注意力优化的SAR 船舶检测方法研究

Abstract

This paper proposes a SAR ship detection method based on a receptive field attention mechanism. By integrating Receptive Field Attention Convolution (RFAConv) into the YOLOv11 model, the method dynamically generates attention weights to enhance the perception of local feature significance. Experimental results demonstrate that the improved model significantly enhances detection accuracy on ship datasets, particularly excelling in the mAP50-95 metric, while maintaining low computational overhead.