科研之友
微信
新浪微博
Facebook
分享链接
摘要
In order to solve problems such as traffic jams and poor traffic management, a kind of illegal lane changing monitoring system based oil traffic video is presented. The hardware system is composed of industrial computers and cameras fixed above the road. The software part consists of video processing, background extraction and updating, lane line extraction, vehicle recognition, moving shadow analysis, illegal lane change detection, vehicle tracking and prediction. First, the video collected is transformed into a series of images. Then a background is obtained by using averaging, and the lane lines are extracted by an improved Hough transform. In order to obtain complete and accurate outlines of the moving vehicles, a method of obtaining a vehicle profile combined with background subtraction and edge detection is proposed. A simple method for illegal lane change discrimination is also put forward. A mean-shift algorithm combined with a Kalman filter is used to track and predict the locations of vehicles. Experimental results show that this system satisfies the need for real-time accuracy, and it can be used for intelligent traffic supervison.
- 出版日期2009-11
- 单位天津大学
