基于视觉的列车车轮踏面擦伤定位方法

为了实现列车车轮踏面擦伤检测,建立了基于机器视觉的踏面擦伤动态检测系统.对该系统中踏面擦伤可疑图像区域定位技术进行了研究.首先,采用快速中值滤波算法平滑原始图像,然后采用基于Canny算法的踏面区域分割算法提取踏面区域,再根据踏面擦伤图像特征设计了基于踏面边缘线扫描搜索擦伤可疑区域的方法提取擦伤可疑图像区域.实验结果表...

Research on peeling defect location method of wheel tread based on vision

In order to realize peeling defects inspection for the train wheel tread,a defect inspecting system of the train wheel tread based on computer vision was established and suspicious regions of peeling defects location method was investigated.Firstly,the original tread image was smoothed by fast medium filter algorithm.Then the tread region was extracted by tread segmentation technology based on Canny algorithm.Finally,the peeling defect location method based on tread edge lines searching way was used to locate the peeling defect suspicious regions.The experimental results show that the correct rate of peelings location comes to 97.62%,the missing recognition rates of the peelings is 2.38%,false acceptance rate is 3.81%.It can lay the basis for peeling defects recognition.