现代科学仪器Issue(6)：87-90,4.
基于PSO-SVM的重轨标识识别方法研究
Study of Heavy Rail Mark Recognition Method Based on PSO-SVM
摘要
Abstract
Heavy rail mark automatic recognition is crucial for enterprise quality control,aiming at the status relying on manual detection method to observe the character of heavy rail,thought of using machine vision to obtain image for recognition was put forward: after regional locating of the characters of the image,heavy rail recognition was classified using particle swarm optimization algorithm for support vector machine parameter selection method.The experiment result indicated that PSO-SVM algorithm have high forecast accuracy and detection efficiency,the training set accuracy rate reached 99%,the test set accuracy rate reached 83%,the training time was 62.195s,each index was higher than that of GA-SVM.It could be used for on-line detection of the heavy rail mark.关键词
标识识别/区域定位/支持向量机/粒子群算法Key words
Mark Recognition/Recognition Locating/SVM/PSO分类
信息技术与安全科学引用本文复制引用
米曾真,谢志江,刘琴,楚红雨..基于PSO-SVM的重轨标识识别方法研究[J].现代科学仪器,2012,(6):87-90,4.基金项目
国家自然科学基金委员会与中国工程物理研究院联合基金资助 （）
