基于截线法的铁路扣件骨架提取算法

为了能够精确描述铁路扣件端部形状,提高铁路扣件的识别率,以及满足扣件检测的实时性要求,提出了一种基于截线法的铁路扣件骨架提取算法。首先,通过预处理得到原始图像的二值图像及其感兴趣区域;其次,应用不同方向的方位距离变换计算扣件各端部区域的骨架生长点;然后,以骨架生长点为中心沿扣件的弯曲方向分别利用水平线族和竖直线族扫描扣件端部轮廓,得到截线段后,计算其中心点;最后依次连接骨架生长点和所有截线中心点即可得到完整的扣件图像骨架。理论分析与实验结果证明,本文算法对扣件形状的描述能力强,并且可以准确识别扣件形状。对于复杂图像的识别具有较好的效果,在识别缺陷扣件时具有较强的应用优势。算法运行时间较短,可以达到实时性的要求。

Detection Algorithm for Railway Fastener Skeleton Extraction Based on Cut-off Line Method

In order to describe the shape of the end of the railway fasteners accurately,improve the recognition rate of the railway fastener and meet the real-time requirements of fastener detection,an algorithm based on the cut-off line method to extract the skeleton of railway fastener is proposed. Firstly,the binary image of the original image and its region of interest through preprocessing is obtained.Secondly,the azimuth distance transformation of different directions are used to calculate the skeleton growth point at each end of the fastener. Then,the skeleton growth point is taken as the center,and the horizontal lines and the vertical lines are used respectively to scan the end profile of the fastener in the bending direction of the fastener. After the section line is obtained,the center point is calculated.Finally,the skeleton growth point and all the center points of the line segments are connected to obtain the complete fastener image skeleton. The theoretical analysis and experimental results show that the algorithm has strong ability to describe the shape of fastener and can accurately identify the fastener shape. It is effective in recognizing complex images and applicable in identifying defective fastener. The algorithm runs in very short time and complies with real-time requirement.