摘要
获取道砟颗粒廓形是研究有砟轨道道床力学性能及开展仿真分析的基础。针对传统图像处理中利用边缘检测算法获得道砟二维廓形信息的不足,提出追踪提取算法。采用数码相机和计算机技术搭建图像采集平台,通过颜色空间转化、二值化处理、边界追踪和数据优化等步骤实现了道砟二维廓形的提取。算法特点在于,提取的道砟颗粒二维廓形数据为有序的点列,廓形曲线连续且闭合良好,检测出来的边缘是单像素宽,通过使用游标卡尺测量证明实际廓形的最大粒径在考虑测量误差及精度的条件下二者数值基本相同。同时,基于前期学者对棱角指数的算法对本文提取廓形进行相应计算并与实测值比较,数值基本吻合,从而说明本算法在一定误差及精度条件下的有效性与可靠性。
Extracting 2D contour of railway ballast is fundamental in the study and simulation analysis of ballast bed. Due to the shortage of traditional edge detection algorithms that are used to obtain 2D contour of ballast, the tracing and extracting algorithm is presented. Digital camera and computer are used to accomplish hardware construction, and the transformation of images' color space, binary processing, boundary tracing, data optimization and other procedures are employed to extract 2D contour of railway ballast. Features of the algorithm are that the extracted 2D contour data is a series of ordered dots, and the contour not only has good continuity and closeness, but also ensures single pixel wide of edge. The maximum particle size of real ballast is measured by Vernier caliper and compared with the maximum particle size of 2D contour. The two values are mostly equal to each other. At the same time, the extracted profile in this paper is calculated based on the algorithm of priori scholars on the angular exponent and compared with the measurement. The consistency of the values verifies the validity and reliability of this algorithm with allowable error and accuracy.
引文
[1] 张徐,赵春发,翟婉明.铁路碎石道砟静态压碎行为数值模拟[J].西南交通大学学报,2015,50(1):137-143.
[2] Mcdowell G R, M. L U. Discrete element modelling of railway ballast under monotonic and cyclic triaxial loading[J]. Géotechnique, 2010,60(6):459-467.
[3] 井国庆,封坤,高亮,等.循环荷载作用下道砟破碎老化的离散元仿[J].西南交通大学学报,2012,47(2):187-191.
[4] 付招兴.高速铁路道砟形态特征量化方法及其应用研究[D].成都:西南交通大学,2013.
[5] Ferellec J F, Mcdowell G R. A method to model realistic particle shape and inertia in DEM[J]. Granular Matter, 2010,12(5):459-467.
[6] 丁秀丽,李耀旭,王新.基于数字图像的土石混合体力学性质的颗粒流模拟[J].2010,29(3):477-484.
[7] Wu Fenghe, Zhang Xiaofeng, Shi Fazhong, et al. Research on 3D reconstruction method based on single image data[J]. China Mechanical Engineering, 2007,18(17):2071-2075.
[8] Wu Fenghe. A study on contour extraction method in computer vision measurement technology[J]. Acta Metrologica Sinica, 2007,28(1):18-22.
[9] 高朝阳,张太发,曲亚男.图像边缘检测研究进展[J].科技导报,2010,28(20):112-117.
[10] 李艳荻,徐熙平,钟岩.特征弦约束随机Hough变换在椭圆检测中的应用[J].仪器仪表学报,2017,38(1):50-56.
[11] 谢林海,刘相滨,何昭青.基于最佳阈值和轮廓提取的边缘检测方法[J].计算机与现代化,2007(10):117-118.
[12] 张慧,刘伟军.一种基于小波的轮廓特征提取算法[J].中国图象图形学报,2005,10(7):828-833.
[13] 杨帆.数字图像处理及应用:Matlab版[M].北京:化学工业出版社,2013.
[14] 许莉,王敏,温月.基于Lab颜色空间的运动目标检测[J].华中科技大学学报(自然科学版),2013,41(S1):219-222.
[15] 古晶.RGB到Lab颜色空间转换[J].广东印刷,2009(5):11-14.
[16] 万波,王泉,潘蓉,等.基于CIELab颜色空间的矢量彩色图像放大算法[J].华中科技大学学报(自然科学版),2008,36(8):17-20.
[17] laviewpbt.颜色空间系列2: RGB和CIELAB颜色空间的转换及优化算法[EB/OL].(2013-02-02)[2017-06-15]. http://www.cnblogs.com/Imageshop/archive/2013/02/02/2889897.html.
[18] Klotz L D. Automated Measurement of Aggregate Indices of Shape[J]. Particulate Science & Technology, 1997,15(1):13-35.
[19] Rao C, Tutumluer E, Kim I T. Quantification of coarse aggregate angularity based on image analysis[J]. Journal of the Transportation Research Board, 2002,1787(1):117-124.
[20] Masad E, Al-Rousan T, Button J, et al. Test Methods for Characterizing Aggregate Shape, Texture and Angularity[R]. Washington, D.C: National Cooperative Highway Research Program, 2005.