无砟轨道板表面裂缝的红外热成像检测方法  

Infrared thermal imaging detection method for surface cracks of ballastless track slab

作　　者：马跃坤 李再帏[1] 赵彦旭 路宏遥 何越磊[1] MA Yuekun;LI Zaiwei;ZHAO Yanxu;LU Hongyao;HE Yuelei(School of Urban Rail Transportation,Shanghai University of Engineering Science,Shanghai 201620,China;China Railway 21st Bureau Group Co.,Ltd.,Lanzhou 730070,China)

机构地区：[1]上海工程技术大学城市轨道交通学院,上海201620 [2]中铁二十一局集团有限公司,甘肃兰州730070

出　　处：《铁道科学与工程学报》2022年第3期579-587,共9页Journal of Railway Science and Engineering

基　　金：国家自然科学基金资助项目(52178430,51978393);甘肃省科技计划资助项目(19ZD2FA001);中国铁建科技研发计划资助项目(2019-B08)。

摘　　要：纵连无砟轨道板作为无砟轨道主要结构类型之一,表面裂缝是轨道板最常见的结构病害类型,也是高铁运维重点作业内容之一。实现非接触快速的红外热成像检测裂缝病害方法对于提高作业精度和效率具有显著作用。针对红外热成像仪采集的图像数据存在的边缘模糊、对比度低等问题,提出基于NSCT变换(Non-subsampled Contourlet Transform)的多尺度积阈值红外图像增强算法。基于相位一致性原理,采用形态学处理、k-means聚类手段提取出红外图像中的完整裂缝区域,并建立基于红外热像图中完整裂缝区域的裂缝检测方法。通过现场病害检测实例对所提算法的有效性进行验证。研究结果表明:采用基于NSCT变换的多尺度积阈值红外图像增强算法,可以有效改善轨道板裂缝红外图像中的噪声、边缘模糊、对比度低的问题,强化图像中裂缝边缘细节;通过相位一致性原理可以获取红外热像图中的裂缝区域,利用形态学处理方法可有效去除图像中的孤立噪声和伪边缘区域,结合红外热像图的温度和温度梯度信息通过k-means聚类算法可对裂缝区域细化,提取出完整的裂缝区域图像;算法对轨道板表面裂缝的检测准确率达90%以上,可以有效地实现表面裂缝病害的检测,建议在高铁无砟轨道运维中采用红外热成像方法对纵连轨道板的表面裂缝进行检测。Longitudinal ballastless track slab is one of the main structural types of ballastless track. Surface cracks are the most common structural disease type of track slab, and it is also one of the key operations of highspeed rail operation and maintenance. The realization of a non-contact and fast infrared thermal imaging method for detecting cracks and diseases has significant value for improving the accuracy and efficiency of the operation.Aiming at the problems of blurred edges and low contrast in the image data collected by infrared thermal imaging cameras, a multi-scale product threshold infrared image enhancement algorithm based on NSCT transform(Nonsubsampled Contourlet Transform) was proposed. Based on the principle of phase congruency, morphological processing and k-means clustering were used to extract the complete crack area in the infrared image. A crack detection method based on the complete crack area in the infrared thermal image was established. The effectiveness of the proposed algorithm was verified by field disease detection examples. The results are as follows. The use of a multi-scale product threshold infrared image enhancement algorithm based on NSCT transform can effectively improve the noise, edge blur, and low contrast in the infrared image of track slab cracks,and strengthen the edge details of the cracks in the image. The crack area in the infrared thermal image can be obtained through the principle of phase congruency. The isolated noise and false edge area in the image can be effectively removed by the morphological processing method. Combining the temperature and temperature gradient information of the infrared thermal image, the crack area can be refined through the k-means clustering algorithm. The complete crack area image can be extracted. the algorithm can detect surface cracks on track slabs with an accuracy of more than 90%, which can effectively detect surface cracks. It is recommended to use infrared thermal imaging to detect surface cracks on longitudinally connected track sl

关 键 词：轨道板 裂缝检测 红外热像图 NSCT变换 相位一致性 

分 类 号：U213.2[交通运输工程—道路与铁道工程]