基于几何与颜色特征的公交车辆视频检测算

A Bus Video Detection Algorithm Based on Geometry and Color

公交站场的监控管理是城市交通管理的重要组成部分。要实现公交站场自动化监控管理,首先要实现公交车辆的检测。针对公交站场中的公交车相对固定、车辆类型相对单一以及背景复杂、变化大等特点,该文提出了一种基于视频图像处理的检测方法。该算法用公交车车体宽度的几何特征和公交车车牌的颜色特征作为检测依据,通过联合检测公交车的保险杠和黄色牌照来达到检测公交车的目的。该方法在广州市中山八路公交站场智能化监控系统的车辆检测子系统中得到应用,测试结果表明,这是一种实用有效的公交车辆视频检测算法。

The supervised management of bus station is the important part of the traffic management of the city. To achieve intelligent supervised management in the bus station is needed to detecting the bus firstly. Considering the specialty of the bus station： the bus is relatively fixed, the types of vehicle is relatively single, the background is complex and variational, this paper presents a detection method based on the video image processing for the bus. Two characters are used as the capture foundation： one is geometry character denoting the width of the bus, the other is color character denoting the color of plate. This algorithm detects the bus by detecting both the bumper and the yellow plate of bus. This algorithm is used in vehicle detection subsystem of the intelligent supervision system in Zhongshan Balu bus station of Guangzhou. The detection result shows that the proposed algorithm is effective for the bus detection.