基于机器视觉的公铁联运车导引落轨检测系统

为了解决目前公铁联运车落轨时需要人工辅助,效率较低的问题,基于机器视觉技术,设计并实现了公铁联运车导引落轨检测系统;系统利用安装在转向架附近车架下的摄像头获取车辆与钢轨的图像信息,采用边缘检测和Hough变换结合的算法来检测地面标志线与车辆钢轮的相对位置,计算出车辆钢轮与钢轨之间的偏移距离和角度;测量结果通过工业无线网络传送到驾驶室里的平板显示器上,辅助司机调整车辆位置完成精准落轨;现场试验表明:导引落轨检测系统能够实时有效测量钢轨与钢轮的距离,检测误差不超过±3 mm,且结果更新时间低于200ms,满足转向架落轨精度和实时性要求.

Guiding wheel dropped detection system for rail-road vehicle based on machine vision

In order to solve the problem that current dropping the steel wheel of rail-road vehicle needs manual assistance and the efficiency is low, a guiding wheel dropped detection system for rail-road vehicle was designed and implemented based on machine vision. The system made use of the camera mounted near the bogie to obtain the image information of the vehicle and the rail, and combined the algorithms of the Edge detection and Hough transform to detect the relative position of the ground marking line and the steel wheel of the vehicle,accordingly to calculate the distance and angle between the rail and the steel wheel of the vehicle finally. The measurement results were transmitted to the flat-panel monitor. in the cab through the industrial wireless network to assist the driver adjusting the position of the vehicle to complete the accurate landing. Experiments show that the system can effectively measure the distance between the rail and the steel wheel in real time. The measurement error is less than ±3mm, and the update time of the result is less than 200ms, which meets the requirements of the accuracy and real-time of the bogie dropping.