轨道扣件缺失的机器视觉快速检测方法
发布时间：2018-12-11 19:59
【摘要】：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.
[Abstract]:Rail fastener missing detection is an important part of railway routine inspection. According to the real-time and adaptive requirements of modern railway automatic detection technology, a real-time detection method based on machine vision for rail fastener missing is proposed. In order to deal with the interference of environmental light, an image acquisition device with mask and LED auxiliary light source is designed. The method of edge detection based on the image gradient is improved by using the switching median filter and the improved Canny edge detection method based on the gradient amplitude of the image. The edge feature of fastener is enhanced by adaptive image enhancement. Combined with the inner and outer edge contour features of fastener elastic strip, the template matching based on curve feature projection is used to realize the real-time detection of fastener missing. The experimental results show that the average image processing time is 245.61msand the average correct recognition rate is 85.8ms.The method is self-adaptive and has the highest speed to support the implementation of 3.82m/s. It can meet the need of real-time detection of missing fastener.
【作者单位】： 兰州交通大学自动化与电气工程学院;兰州工业学院电子信息工程学院;
【基金】：国家自然科学基金项目(61663022,61461023) 教育部创新团队发展计划(IRT_16R36) 甘肃省高原信息工程及控制重点实验室开放课题基金(20161105)资助
【分类号】：TP391.41;U216.3
本文编号：2373136
[Abstract]:Rail fastener missing detection is an important part of railway routine inspection. According to the real-time and adaptive requirements of modern railway automatic detection technology, a real-time detection method based on machine vision for rail fastener missing is proposed. In order to deal with the interference of environmental light, an image acquisition device with mask and LED auxiliary light source is designed. The method of edge detection based on the image gradient is improved by using the switching median filter and the improved Canny edge detection method based on the gradient amplitude of the image. The edge feature of fastener is enhanced by adaptive image enhancement. Combined with the inner and outer edge contour features of fastener elastic strip, the template matching based on curve feature projection is used to realize the real-time detection of fastener missing. The experimental results show that the average image processing time is 245.61msand the average correct recognition rate is 85.8ms.The method is self-adaptive and has the highest speed to support the implementation of 3.82m/s. It can meet the need of real-time detection of missing fastener.
【作者单位】： 兰州交通大学自动化与电气工程学院;兰州工业学院电子信息工程学院;
【基金】：国家自然科学基金项目(61663022,61461023) 教育部创新团队发展计划(IRT_16R36) 甘肃省高原信息工程及控制重点实验室开放课题基金(20161105)资助
【分类号】：TP391.41;U216.3
【相似文献】
相关硕士学位论文 前6条
1 王昆;基于机器视觉的轨道几何特征提取算法研究及实现[D];上海工程技术大学;2016年
2 殷超;基于机器视觉的轨道缺陷识别方法研究[D];兰州交通大学;2016年
3 丁政开;基于机器视觉的钢轨表面缺陷检测技术研究[D];北京交通大学;2017年
4 孙丽丽;机器视觉与铁路货车超限监测应用研究[D];西南交通大学;2005年
5 李冬;基于机器视觉的钢轨磨耗检测系统研究[D];南京航空航天大学;2013年
6 柴华;基于机器视觉的嵌入式铁路异物侵限检测系统研究[D];北京交通大学;2015年
，本文编号：2373136
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2373136.html
