A Method of LVQ Network to Detect Vehicle Based on Morphology

In this paper, the main research is to recognize and classify vehicle on the traffic road in the hi-resolution satellite image. Because the satellite image contains the vehicle characters pixel set and the alike vehicle characters pixel set. So, there are a series of pattern recognition methods from image preprocessing to the LVQ artificial neural network works. The application of morphology is the main preprocessing technology in this paper. After image is morphologically preprocessed, it makes the pixel matrix more clear and differentiable, and is good for the next step, which is to calculate the preprocessed image texture. According to the gray level and the texture, a vector is built as input of LVQ Artificial Neural Network system; the trained LVQ artificial neural network will give a satisfying outcome. Finally, with these processing, the result images are given. The white vehicle and the black vehicle are obvious. Although there are some errors in the results, by contrast with the origin images; the results still give a good effort of this method.