Research on vehicle tracking algorithm using Contourlet transform

Vehicle tracking is a very important part of ITS and become a hot research field at home and abroad. In this paper, according to the multi-scales and multi-resolution analysis of Contourlet transform, put forward a video vehicle tracking algorithm based on the Contourlet transform. At first, according to the fixed characteristics of traffic video monitoring camera. we get the ideal vehicle target and the circum-rectangle with the method of reducing background, then, using Contourlet transform for the area of vehicle, we get the son of spectrum data in video vehicles. At last, realize video vehicle target tracking based on the vehicle image similarity measure and linear forecasting. Experiments show that this algorithm with higher tracking accuracy and better real-time.