A Logistics Vehicle Classification Method Based Gabor Wavelet Feature Tracking

A new logistics transport classification method for target tracking based on Gabor features is proposed in this paper. To improve computation speed and effect of Gabor wavelet feature tracking, methods of extract initialization feature points was studied. Provided a new method—edge surrounding, which detect important edge automatically based on edge detection and edge surrounding to find Gabor wavelet feature point. Firstly it detects edge to get area of initialization points, and calculate distributing density of feature points, and looks length as estimate of distributing; it needn't optimize position and angle parameter so that we decrease number of parameter, and can improve computer speed of wavelet and effect. Experiment results indicate that edge detection can get effecter target feature shorter computer time than that method, and reduces whole track operation mean time and the slows time.