摘要
针对传统扣件检测方法式效率低、可靠性差,不能满足现代铁路检修的需要,提出了一种基于计算机视觉的扣件缺失自动检测方法。在对灰度图像进行Canny边缘检测处理后采用十字交叉定位法对扣件位置进行定位,得到120×200像素的扣件区域,并提取扣件图像的20个边缘特征值;最后,利用模糊C均值聚类算法对这两类的特征量进行聚类分析,通过计算待诊断对象与标准模式的隶属度实现对扣件状态的分类。应用验证表明:采用的图像处理方法和识别分类算法能够有效检出轨道扣件缺失,检测速度快,鲁棒性好,检出率达96%。
The traditional fastener detection methods are inefficient and unreliable,can not meet the needs of the modern railway maintenance.This paper proposes a vision-based technique for detecting rail fastening automatically.First,a criss-crossing localization method was proposed to position the fastener for the canny edge processing gray images,and the edge characteristic information of fastener was extracted.Finally,fuzzy C-means clustering algorithm was used to cluster the extracted features,fastener missing detection can be realized by calculating the membership between the unknown samples and the standard modes of fastener.The experiment showed that this image processing and classifying algorithm can realize the automatic detection of missing fastener effectively;the detection rate is above 96%.
作者
杨樊
陈建政
吴梦
YANG Fan, CHEN Jian-zheng, WU Meng (Traction Power State Key Laboratory, Southwest Jiaotong University, Chengdu 610031, China )
出处
《电脑知识与技术》
2014年第4期2367-2370,共4页
Computer Knowledge and Technology
基金
国家科技支撑计划(2009BAG12A01)