基于聚类及霍夫变换的轮对廓形提取算法研究  

Research on algorithm of wheelset profile extraction based on clusteringand Huff transform

作　　者：王洪峥 从光涛 邢宗义[2] 张永[2] WANG Hongzheng;CONG Guangtao;XING Zongyi;ZHANG Yong(Operation Headquarters,Guangzhou Metro Group Co.Ltd.,Guangzhou 510335,China;School of Auotomation,Nanjing University of Science&Technology,Nanjing 210094,China)

机构地区：[1]广州地铁集团有限公司运营事业总部,广州510335 [2]南京理工大学自动化学院,南京210094

出　　处：《铁路计算机应用》2020年第6期12-17,24,共7页Railway Computer Application

基　　金：国家重点研发计划课题(2017YFB1201201)。

摘　　要：针对轮对尺寸在线检测中廓形计算复杂的问题,提出一种新的轮对廓形提取算法。总结在线检测系统中的6种典型轮对廓形,采用DBSCAN算法对廓形数据进行聚类,并采用改进霍夫变换确定廓形有效区域;采用曲率熵与最小二乘法进行曲线拟合,获取轮缘基准点;通过坐标旋转与平移等操作实现完整的轮对廓形提取,为轮对尺寸计算提供支持。通过现场试验验证算法可行性,并与人工检测结果进行对比,结果表明:在线检测系统精度优于人工检测,验证该算法能够有效地实现轮对廓形提取。Aiming at the problem that it is complex to determine the wheelset pro file in the online wheelset pro files detection system,a novle algorithm on wheelset profile extraction is proposed.Firstly,six types of typical wheelset pro files are analyzed.Then,the DBSCAN method is used to culster the pro file data and an improved Huff transform is employed to determine the effective area of the pro file.Secondly,a curve fit is performed to obtain the reference point of wheelset rim based on the curvature entropy and the least square methods.Thirdly,the complete wheel pro file can be acquired after the operations of coordinate rotation and translation.Finally,the feasibility of the proposed algorithm is veri fied through field experiments.The results indicate that the precision of wheel pro files acquired from the online detection system is higher than that of the geometrical pro files measured manually and the proposed algorithm can satisfy the requirement of pro file calculation of the online detection system.

关 键 词：轮对廓形 霍夫变换 密度聚类 

分 类 号：U270.331.1[机械工程—车辆工程] TP39[交通运输工程—载运工具运用工程]