SAR Image Target Recognition Algorithm Based On Convolutional Neural Network

Abstract

SAR has high application value in both military and civil fields because of its unique working characteristics, and because of the inevitable existence of coherent speckle noise in SAR images, the existence of noise has a great impact on the recognition processing of images later, therefore, this paper firstly performs noise suppression on SAR images, and then constructs convolutional neural network to fully learn the feature information of images through the network to classify them recognition. For the published MSTAR dataset, the method in this paper achieves better recognition results in both three and ten classes of MSTAR datasets.