车载GigE Vision实时前方车辆防撞预警系统设计

Design of real-time preceding vehicle collision warning system with mobile GigE Vision

为弥补传统车辆防撞预警系统图像采集速率慢、图像分辨率低等不足,本文设计了一种应用于结构化道路环境的车载GigE Vision实时前方车辆防撞预警系统。利用GigE千兆以太网工业相机提供的高分辨率图像,对汽车前方图像进行检测;利用基于单帧静态图像的测距方法实时计算前车与本车的纵向距离及前车车速;通过安全车距预警模型计算报警距离,对车辆行驶状态进行判别。试验结果表明,系统车辆平均检测率为95.02%,检测时间为20.71ms,并能较准确地进行车距测量,实现及时、可靠的防撞预警。

To compensate the low image capture speed and low image resolution deficiencies of traditional vehicle collision warning system,a novel real-time preceding vehicle collision warning system with mobile GigE Vision is designed for structural road environment.Firstly,the preceding vehicle image is detected by utilizing the high-resolution images provided by GigE Gigabit ethernet industrial cameras.Then,the longitudinal distance and the vehicle speed are measured with the use of the distance measurement method based on single frame static image.Finally,the vehicle driving status is judged through the safety vehicle distance calculated by the proposed warning model.The results show that the average vehicle detection rate of the system is 95.02% and the detection time is 20.71 ms.The system can accurately measure the distance between vehicles and provide timely and reliable collision warning.