Target Detection Method for SAR Images Based on Feature Fusion Convolutional Neural Network

Target Detection Method for SAR Images Based on Feature Fusion Convolutional Neural Network

For the image target of Synthetic Aperture Radar (SAR), it is more difficult to detect targets in complex background, large scene and more clutter. This paper designs a less layer convolutional neural network (CNN), the complete data validates its feature extraction effect, as a basis for feature extraction networks. In the training dataset, it supplements the target training samples in complex large scene, meanwhile a multi-level convolution feature fusion network is designed to enhance the detection ability of small targets in large scene. After the joint training of the region proposal network (RPN) and the target detection network, a complete model for SAR image target detection in different complex large scenes is obtained. The experimental results show that the proposed method has a good result on SAR image target detection and has an average precision (AP) value of 0.86 in the validation dataset.