Local Robust Feature Based on FAST Corner Detection

The popular SIFT, SURF and other local features exist computational complexity, poor real-time performance for matching and other follow-up steps. Therefore, a novel visual tracking feature algorithm is proposed based on the FAST corner detection, in order to overcome the impact of noise and outdoor lighting changes in the practical application and, which can quickly match the feature points. Experiments indicate that the proposed tracking feature can ensure the better matching accuracy and robustness than the original visual tracking features, with the lower computational and real-time processing.