Binocular Stereovision Based On Vehicles Distance Measurement Algorithm Research

In this paper a measuring method of leading vehicles safe distance is presented based on the rapid increasing of the private car in our nation, the high traffic accident occurring rate still in high position and the traffic safety problems increasingly outstanding. Firstly systemically studied on the theory of binocular stereovision and the inside and outside parameters calibration methods, and combing the traffic applications, designed the calibrating scheme. The parameters that obtained according to the scheme could meet the demands of measuring the vehicle distance.In allusion to the needs of the vehicle safe distance measurement using binocular stereovision system, the vehicle number plate was chose as measuring region that represents the front vehicle. In this thesis an orientation algorithm of the vehicle brand and number-plate locating algorithm using color image were designed. And a center of the number-plate coordinates based on number-plate area gray characteristic was obtained. Then, according to the left and right number-plate center points, the 3D information of the vehicle were obtained through using three-dimensional reconstruction equations. Finally, the distance of the vehicle from the cameras was computed. Experimental results demonstrated that the accuracy of this program can satisfy the need of distance measurement for vehicle safety in highway transportation system, so it is an effective algorithm for the measurement of the front vehicle in transportation supervision.