受电弓滑板磨耗检测图像处理算法的研究及实现

Research and Realization of Image Processing Algorithm in Pantograph Slide Abrasion Detection

由于图像检测受电弓滑板磨耗能实现在线自动检测,且检测精度较高,而该检测方法的关键是对采集到的受电弓图片进行图像处理,因此研究图像处理算法具有非常重要的意义。检测系统采用4个CCD摄像机实现受电弓前后滑板的初始图像采集,然后将获取到的图片传输并存储到检测主机上,通过后期的图像处理,实现滑板磨耗的自动检测,并输出数据报表和磨耗曲线等。图像处理算法主要包括图像预处理、边缘检测和连接、导线及滑板边缘定位等。其中,边缘检测和定位算法是关键,也是决定系统检测精度的重要因素。主要对Canny边缘检测、形态学边缘检测和小波边缘检测进行实验验证,并比较其边缘提取效果。结果表明,Canny边缘检测精度更高,而小波边缘检测虽然定位更准确,但精度较低。因此,最终选择Canny边缘检测,结合边缘连接进行处理,并给出了检测结果图。

As detecting the pantograph slide abrasion based on image processing can realize the online and automatic inspection, and the image processing after collecting the picture of the pantograph is the crucial part of the system, therefore researching the image processing algorithm is of great importance. The system adopt 4 CCD cameras to collect the original images of the front and back slide of the pantograph, then the images are transmitted to the main detection PC and stored there. Through post-processing of the pictures, the system can inspect the slide abrasion automatically and output the data report and abrasion curve. The image processing algorithms mainly include image preprocessing, edge detecting and linking,contact wire and slide edge location and so on. There into, the edge detecting and location algorithm are the key and they are signifi dant factors which determine precision of the inspection system. This paper emphasizes on the experimental validation and effect of the edge distilling of three main edge detecting methods which are Canny edge detecting, morphology edge detecting and wavelet detecting. The result shows that the measuring precision of the Canny edge detecting is higher than that of the wavelet edge detecting which has higher edge location accuracy. So the Canny edge detecting combined edge linking is adopted finally. At last,the paper gives the detecting result picture.