Preceding Vehicles Detection Based on Vehicle Features

In order to enhance the real-time and robustness of driver assistance system, an algorithm of preceding vehicles detection based on vehicle features is proposed in this paper. According to the characteristics of lane marks and vehicles, such methods as adaptive threshold based on sampling, adaptive double-threshold and the smallest univalve segment assimilating nucleus (SUSAN) algorithm based on adaptive threshold in different contrast ratio are introduced to extract the information which is needed during detection of front vehicles. Experiment results indicate that this algorithm is feasible for different road conditions, and can realize the detection of the preceding vehicles with robustness and efficiency.