轨道扣件缺失检测是铁路日常巡检的一项重要内容，结合现代化铁路对自动化检测技术的实时性和自适应性要求，提出了一种基于机器视觉的轨道扣件缺失实时检测方法.为了应对环境光线的干扰，设计了遮光罩加LED辅助光源的图像采集装置，利用开关型中值滤波和基于图像梯度幅值的改进Canny边缘检测方法，对扣件边缘特征进行自适应图像增强.结合扣件弹条稳定的内外边缘轮廓特征，利用基于曲线特征投影的模板匹配实现了扣件缺失的实时检测.经过实验验证，平均每帧图像的处理时间为245.61ms，平均正确识别率为85.8%，且该方法具有一定的自适应性，最高支持3.82m/s的推行速度，可满足对实际运营线路进行扣件缺失实时检测的需求.
The detection of missing track fasteners is an important part of daily inspection of the railway. Owing to the new requirements of realtime and self adaptation of the automatic detection technology, a method of realtime detection of missing track fasteners based on machine vision is proposed in this paper. In order to deal with the interference of environmental light, the image acquisition device includes hood and lightemitting diode (LED) auxiliary light source. The switching median filtering is used and the Canny edge detection is improved based on image gradient amplitude for image adaptive enhancement of the characteristics of fastener edges. The feature of article fasteners to play in stable boundary and the projection curve characteristics of template matching are combined to realize the realtime detection of missing fasteners. Experimental results showed that the image detection time of every frame is 245.61ms, the average of ecognition accuracy is 85.8%. The highest speed of real time detection of missing fastener of actual operation line is 3.82m/s. These results suggest that the proposed method is adaptive.
［1］LI Q, REN S. A visual detection system for rail surface defects［J］.IEEE Transactions on Systems Man & Cybernetics Part C, 2012, 42(6): 15311542.
［2］钱广春. 基于计算机视觉的铁路扣件缺失快速探测方法研究［D］.上海: 上海交通大学电子信息与电气工程学院, 2011.
［3］王凌, 张冰, 陈锡爱. 基于计算机视觉的钢轨扣件螺母缺失检测系统［J］.计算机工程与设计, 2011, 32(12): 41474150.
WANG Ling, ZHANG Bing, CHEN Xiai. Inspection system for loss of rail fastening nut based on computer vision［J］. Computer Engineering and Design, 2011, 32(12): 41474150.
［4］谢凤英, 吴叶芬, 周世新. 基于互信息的铁路轨枕扣件自动定位算法［J］.中国体视学与图像分析, 2013, 18(2): 145149.
XIE Fengying, WU Yefen, ZHOU Shixin. Railway fastener locating algorithm based on mutual information［J］. Chinese Journal of Stereology and Image Analysis, 2013, 18(2): 145149.
［5］罗建桥, 刘甲甲, 李柏林, 等. 基于局部特征和语义信息的扣件图像检测［J］.计算机应用研究, 2016, 33(8): 25142523.
LUO Jianqiao, LIU Jiajia, LI Bailin, et al. Detection for railway fasteners based on local features and semantic information［J］. Application Research of Computers, 2016, 33(8): 25142523.
［6］ZHANG S, KARIM M A. A new impulse detector for switching median filters［J］. IEEE Signal Processing Letters, 2002, 9(11): 360363.
［7］张小虎, 李由, 李立春, 等. 一种基于梯度方向直方图的直线轮廓提取新方法［J］.光学技术, 2006, 32(6): 824826.
ZHANG Xiaohu, LI You, LI Lichun, et al. A newline boundary detection algorithm based on histogram of gradient’s direction［J］. Optical Technique, 2006, 32(6): 824826.
［8］曾俊. 图像边缘检测技术及其应用研究［D］.武汉: 华中科技大学自动化学院, 2011.
［9］胡铟. 基于单目视觉的运动目标检测与跟踪算法研究［D］. 南京: 南京理工大学计算机科学与工程学院, 2008.
