摘要
提出了一种基于机器视觉的轨距检测方法,该方法采用4个CCD摄像机和2个红色扇形光源构成检测系统。对检测系统进行了定标分析,采用提取分量的方式对图像的目标区域和背景区域进行分割,利用图像差影法去除噪声,应用自适应迭代阈值法对图像进行二值化处理,并通过膨胀和细化算法得到轨道的截面轮廓线。试验结果表明,该方法能有效地实现轨距参数的高精度动态测量,精度可达到0.07mm。
A method based on computer vision for inspecting track gauge is presented in this paper.The inspection system consists of 4 CCD cameras and 2 red sector-lights,the calibration method of inspection system is analyzed.Extracting image component is utilized to segment object region and background region,and difference image method is adopted to eliminate noise in image processing.Adaptive iteration threshold method is applied to get the binary image,and ultimate section profile of track is obtained by using image dilation and thinning algorithm.Experiment results show that this method is effective to measure track gauge value dynamically and accurately with only 0.07 mm maximum erro.
出处
《城市轨道交通研究》
北大核心
2010年第9期73-76,共4页
Urban Mass Transit
基金
上海市科委地方院校能力建设课题(08220511000)
上海高校选拔培养优秀青年教师科研专项基金(gjd08050)
关键词
轨距检测
定标
图像处理
钢轨截面轮廓线
track gauge inspection
calibration
image processing
rail section profile
