针对受电弓滑板裂纹故障将严重危及行车安全的问题,提出一种基于模糊熵和Hough变换的滑板裂纹检测方法。首先依据像素邻域内的灰度分布,提出一种基于区间二型模糊熵的边缘检测方法,可以获得主体特征增强的滑板边缘图像;然后,采用连通域方法去除孤立噪声点,获得主要包含边界线、接缝、螺钉和裂纹四类图形元素的滑板边缘图像;在此基础上,采用Hough变换分析各类图形元素在参数空间的特征分布,从而提出一种基于极角约束Hough变换的裂纹提取方法,通过有效地排除非裂纹图形元素的特征点,最终实现滑板裂纹的自动检测。仿真实验表明,本文的边缘检测方法能够获得线性特征增强的滑板边缘图像,为后续的Hough变换提供有利基础;本文的裂纹提取方法能够准确实现滑板裂纹的识别和定位。
In order to deal with the problem that faults of pantograph slide cracks may impair traffic safety,a novel method of pantograph slide crack detection was put forward on the basis of fuzzy entropy and Hough transform.According to the gray distribution of pixels in the neighborhood,the edge detection method based on the entropy of interval type 2 fuzzy sets was forward to obtain the slide edge image with enhanced features;then,the connected domain method was used to remove isolated noise points to obtain the slide edge image mainly including four types of graphic elements,i.e.,boundary lines,joints,rivets and cracks;on this basis,the characteristics distribution of the various types of graphic elements in the parameter space was analyzed by using Hough transform,thus,the method to extract slide cracks based on Hough transform under constraint of polar angles was proposed;through effectively excluding feature points of non-crack graphic elements,automatic slide crack detection was realized finally.The simulation results indicate as follows:The proposed edge detection method can obtain slide edge images with enhanced linear features,thus laying a foundation beneficial to subsequent Hough transform;The proposed crack extraction method can be applied to accurately identify and locate slide cracks.