Abstract:

In order to improve the efficiency of vehicle recognition, this paper puts forward a fusion recognition algorithm based on a variety of local features of the vehicles. The algorithm adopts the optimized Principal Component Analysis (PCA) to extract, analyze, and fuse the local features of the vehicles so as to realize the recognition of vehicles. The results of the experiments indicate that the algorithm can effectively reduce the influence of the shadows of vehicles and the surrounding environment, and improve the recognition efficiency under the circumstance of partial blocking of a vehicle, and that the algorithm has pretty good accuracy and stability.

Key words: vehicle detection;feature region;PCA;gray balance