基于图像处理的钢轨表面缺陷检测算法分析

Analysis on Rail Surface Defect Detection Algorithm Based on Image Processing

针对钢轨表面缺陷种类多、产生概率大、人为巡检工作效率低和强度大等问题,设计了一种基于图像处理的钢面缺陷检测方法。通过对360 K980相机采集的钢轨表面图像进行预处理,再利用灰度垂直投影法对预处理后图像进行轨面缺陷区定位提取。基于Canny、Roberts和Sobel算子对轨面ROI区域进行边缘检测,并依据recall和accuracy对仿真结果进行比对,以此筛选出边缘检测效果较好的算子。实验结果表明:本文提出的检测方法实际可行,在轨道状态寿命及维护中有较大的应用价值。

Aiming at the problems of many kinds of rail surface defects,high probability of occurrence,low efficiency and high strength of human inspection,a steel surface defect detection method based on image processing was designed.The rail surface image collected by 360 K980 camera was preprocessed,and the gray vertical projection method was used to locate and extract the rail surface defect area of the preprocessed image.Based on Canny,Roberts and Sobel operators,edge detection of the ROI region of the rail surface is carried out.The simulation results are compared according to recall and accuracy,so as to screen out the operators with better edge detection effect.The experimental results show that the method proposed in this paper is practical and feasible,and has great application value in orbit state life and maintenance.