Study Of The Vehicle Speed Detection Used In Video Detection System

Traffic information surveillance system based on video image processing is becoming more and more important because of its simple, convenient. And in recent years, a lot of researches focus on it. Vehicle speed detection is a very important part in the traffic information surveillance system.In this paper, camera calibration and image-based matching algorithms are studied. Main contents are included:(1) Camera calibration. It proposes a new method of camera calibration based on the traditional Pin-hole Model. Firstly, one-dimensional algorithm is calculated out with the geometric method, and then extends it to two-dimensional in order to enhance the accuracy of calibration.. Finally the two-dimensional calibration algorithm is simplified for reducing the computational complexity.(2) Image-based matching algorithm. Firstly, the position of template is fixed according to the position of Virtual Loop. Then matches the template with optimal rule in near area of the next frame. Finally, the minimum in template's wave is selected as feature point to calculate out the vehicle's absolute positon in pixel.A practical vehicle speed detection algorithm is designed in this paper, it can detect vehicle's instantaneous speed in real-time. The results of experiment in tunnel show that: it can detect vehicle's instantaneous speed accurately, in practice, the accuracy of measurement is more than 95%. Early applications indicate that: the algorithms can reach the requirements of practical applications, it can be used in tunnel in vehicle speed detection, speed survey etc.