ILLUMINATION ROBUST VEHICLE DETECTION USING KNOWLEDGE BASED MULTI FEATURES EXTRACTION

Vehicle detection is important in many advanced driver assistance systems. In this paper, we propose a illumination robust forward vehicle detection. The proposed method involves three major components: First, we detect a main lanes and define an adaptive region of interest (ROI) to remove outlier and reduce complexity. Second, we extract vehicle candidates combing multi features of vehicle. Finally, we utilize the distinct property of taillights color to achieve illumination invariant vehicle detection. Experimental results show satisfactory performance with an average detection rate of 96% under various illumination conditions.