一种快速轮对踏面光截曲线图像分割方法  

Fast wheelsets tread light-sectioning curve image segmentation method

作　　者：赵勇[1] 

机构地区：[1]长安大学工程机械学院机械系道路施工技术与装备教育部重点实验室,陕西西安710064

出　　处：《电子技术应用》2011年第3期130-132,136,共4页Application of Electronic Technique

基　　金：中央高校基本科研业务费专项资金资助;长安大学基础研究支持计划专项基金资助(CHD2009JC102)

摘　　要：为准确地实现轮对轮缘磨耗的测量,提出了将二维最大相关准则的阈值分割法应用于轮对踏面光截曲线图像的分割。引入量子粒子群优化算法代替穷尽搜索获取阈值向量。通过对实际的踏面光截曲线图象分割表明,该方法是一种实用、有效的踏面光截曲线图像分割法。In order to measure the wear of train wheelsets flange accurately, an image thresholding method named two dimensional maximum correlation criterion was proposed to do wheelsets tread light-sectioning curve image segmentation. The threshold vector was obtained by using quantum particle swarm optimization algorithm. Through the examples of segmenting the wheelsets tread light-sectioning curve images, the proposed method has been proved to be a practical and efficient wheelsets tread light-sectioning curve image segmentation method.

关 键 词：图像分割 量子粒子群优化 最大相关准则 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]