基于图像处理技术的转辙机实时在线监测研究

介绍了一种通过对转辙机表示杆末端边缘位置图像的监测来反映道岔尖轨密贴状态的方法，以及对道岔尖轨密贴状况监测所采用的图像检测微小位移的基本方法、原理和步骤。对采集的图像使用中值滤波和高斯滤波相结合的预处理方法滤除噪声，通过迭代阈值分割、小面积删除、图像边缘检测得到表示杆位置特征，并采用Hough变换完成坐标提取，从而得出表示杆位移具体数值，并通过实际监测测量分析系统对转辙机尖轨密贴状态的监测精度和应用中问题。

Online Monitoring Based on Image Processing Technology for Switch Machine

A railway switch close gap monitoring system based on image processing techniques was proposed to detect the close-up status of the end of indication lever. The basic structure, working principles and procedure of this system for monitoring tiny displacement were presented. The collected image was preprocessed by Median filter and Gauss filter to get rid of the noise. The character of the image was accomplished by iterative threshold division, deleting small area and image edge detection, and the coordinates of the character was gotten by Hough algorithm. Though such steps the numerical displacement of indication lever was gotten. Finally, the precision of monitoring system analyzed by actual operation. Some application questions were found, corresponding recommendations of this system were presented.