SAR Image Target Recognition Based on Hu Invariant Moments and SVM

Target recognition is a key step in the application of SAR images, but because of the existing of speckles in SAR images, targets can not be recognized well by using traditional methods. According to the advantages of invariant moments extraction and support vector machine (SVM) classification, an efficient method of SAR image target recognition is proposed. First, image preprocessing is performed by using wavelet transform. Second, seven Hu moments, which have the properties of rotation invariability, translation invariability and scaling invariability, are extracted as feature vectors and are normalized. Then a SVM classifier is designed and trained by using normalized feature vectors. Finally, the testing sets of SAR images are recognized by trained SVM. In the experiments of recognizing planes and tanks, better results have been obtained with this new method.