A new method for image recognition

An efficient target recognition method for remote sensing image is proposed in this paper, which is based on the moment invariant, the self-organizing map neural network and the support vector machine. Firstly, seven Hu's invariant moments are extracted as feature vectors, then a self-organizing map neural network is used to cluster feature vectors. Finally, a support vector machine is applied to recognize targets of planes and ships on binary remote sensing images. The test show that the recognition results of the new method are better, and the training time is shorter than those by using the support vector machines only.