Moving Vehicle Detection In Vision Navigation Of Intelligent Vehicle

This paper mainly studied the front moving vehicle detection in vision navigation of intelligent vehicle.First, the video data is gained from the driving car while the camera installed at the front of the car which is toward the front. These video sequences contain different transportation condition and visual circumstance.Secondly, the system is mainly constituted of two parts: the segment of the road and the front vehicle detection based on symmetry. The segment of the road is mainly for further vehicle detection. First, we calculate the best threshold value by which the image is thresholding. Then, we adopt the line grow algorithm to segment the road district. The detection the road is completed. The moving vehicle detection is mainly realized by whether ROI district which is inside the road district has a symmetry characteristic. If it has, we think this ROI contains a moving vehicle. At the end, I realize the system by the Visual C++ 6.0.At the different weather, this system have completed the front moving vehicle detection. The experiment results show that the algorithm is effective and robust in detecting moving vehicles in complicated background.The main contribution of this thesis is summarized as follows:1. In the road partition, this paper put forward a kind of road segment based on line-grow. This method is effective to segment the road district.2. Because the video data is collected by the driving car, the front moving vehicle will appear obvious symmetry character. According to these, this paper puts forward to a vehicle detection algorithm based on the road segment and symmetry detection.