Study of Two-Dimensional Motion Estimation Technique in Vehicle Tracking

摘要

It is very important to achieve reliable vehicle tracking in ITS application such as accident detection. In order to achieve reliable vehicle tracking we designed the dedicated algorithm to traffic images. But the most difficult problem associated with the dedicated algorithm is vector displacement estimation. So we present three criteria to evaluate different displacement vector estimation algorithms. As a result, we can evaluate the convergence time and the capability of different algorithms to resist noise and detect velocity, which is critical for practical application. After groups of experiments had in practical environment, the algorithm were able to track vehicle ate a high success rate of 90%-95%. The algorithm does not assume only physical models of vehicles.