基于引导滤波和积分投影算法的轨道扣件定位

Track Fastener Positioning Based on Guided Filtering and Integral Projection Algorithm

精确定位是实现轨道扣件缺陷计算机自动检测的基础,为此提出了一种改进引导滤波去噪和灰度积分投影结合模板匹配算法的轨道扣件定位方法。首先,通过具有良好边缘保持能力的改进引导滤波算法对轨道扣件图像进行去噪;其次,利用改进Canny算法在Opencv平台对扣件图像进行边缘检测,实现轨道扣件图像边缘检测的自适应性;再次,采用灰度积分投影算法结合先验知识粗定位扣件区域;最后,通过模板匹配算法精确定位轨道扣件。仿真实验表明:所采用的算法具有较好的定位能力,可以准确地定位轨道扣件区域,为进一步的扣件识别提供了可靠的基础。

Precise positioning is the basis for automatic computer detection of track fastener defects.This study proposed a method of track fastener positioningbased on improved guided filter de-noising and gray integral projection combined with template matching algorithm.Firstly,the track fastener images were de-noised by an improved guided filtering algorithm with good edge retention;Then,using the advanced Canny algorithm on the Opencv platform to quickly achieved the adaptive edge detection of the track fastener image;Thirdly,gray integral projection algorithm combined with prior knowledge is used to locate the coarse fastener area;Finally,the track fastener were precisely located by the template matching algorithm.This simulation results show that the proposed algorithm has good positioning ability and can accurately locate the track fastener region,which provides a reliable basis for further fastener identification.