Vehicles detection using sensor fusion

The sensor fusion method is more robust than the method by using a single sensor. So sensor fusion is effective to vehicle recognition at complex scene. But, if each sensing data processed individually for most stage, recognition performance is not always good. In this paper, we propose the extensible and generalized fusion method. First, fusion vector which is combined between image sensor data and laser radar data at primitive level is prepared. We regard fusion vector as sensing data by one robust sensor. Next, fusion vector is compared with discriminated dictionary. We report efficiency of our method at complex scene which recognition error tend to occur by using a single sensor.