摘要
转辙机缺口大小是其健康状态的重要标志之一.针对大多数现有转辙机缺口检测装置需要在表示杆额外增加标记的缺点,以及目前边缘检测技术难以消除图像中噪声的问题,以某铁路公司ZYJ7型与ZD6型转辙机缺口图像为基础,设计了一种基于超像素的快速模糊聚类算法(Superpixel-based Fast Fuzzy C-Means clustering,SFFCM)图像分割的转辙机缺口自动检测方法 .首先,采用SFFCM对缺口图像进行精准分割,保持缺口图像边缘的清晰度和连续性;其次,利用Canny算子对分割后的图像进行边缘提取,获取缺口图像的单像素边缘;最后,利用算法对图像中转辙机缺口两端进行标记,从而实现转辙机缺口的自动检测.实验结果表明,该转辙机缺口检测方法与架构具有一定的有效性.
The size of a switch machine′s gap is an important indication of its health status. Most of the existing detection devices for switch machine gap need to add additional marks on the indicating rod,and the current edge detection technology has trouble eliminating the noise in images. To address these problems, an automatic detection method of switch machine gap based on Superpixel-based Fast Fuzzy C-Means clustering(SFFCM) image segmentation is proposed, which is based on the gap images of ZYJ7 and ZD6 switch machines of a railway company. Firstly, a SFFCM algorithm is used to segment the gap image accurately, while maintaining clear and continuous edge of the gap image. Secondly, the Canny operator is used to extract the edge of the segmented image, thus obtaining the single-pixel edge of the gap image. Finally, the algorithm is used to mark the edges at both ends of the switch machine gap, so as to realize the automatic detection of gaps. The test results show that the proposed detection method and architecture for switch machine gaps are effective.
作者
刘云婷
陈光武
LIU Yunting;CHEN Guangwu(Key Laboratory of Opt-electonic Technology and Intelligent Control Ministry of Education,Lanzhou Jiaotong University,Lanzhou 730070,China;School of Automa-tion&Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China;Gansu Provincial Key Labora-tory of Traffic Information Engineering and Control,Lanzhou 730070,China)
出处
《北京交通大学学报》
CAS
CSCD
北大核心
2022年第2期29-36,共8页
JOURNAL OF BEIJING JIAOTONG UNIVERSITY
基金
国家自然科学基金(61863024)
甘肃省科技引导计划(2020-61-14)。
