Comparison of Deep Learning-based SAR Target Detection Algorithms

In order to explore the performance of different types of deep learning target detection algorithms in SAR image detection, this paper uses the MSAR dataset consisting of four types of targets, namely, aircraft, oil tanks, bridges and ships, from Haisi-1 and Gaofen-3 satellite images, and conducts experiments using two types of deep learning target detection, SSD and Faster R-CNN, to conduct a comparative analysis of the detection accuracy, and the experimental The results show that Faster R-CNN detection accuracy is higher.