首页 | 本学科首页   官方微博 | 高级检索  
     

图像特征识别方法研究
引用本文:门蓬涛,张秀彬,张峰,孙志旻,吴炯.图像特征识别方法研究[J].微计算机信息,2004,20(5):103-105.
作者姓名:门蓬涛  张秀彬  张峰  孙志旻  吴炯
作者单位:上海交通大学
摘    要:图像特征识别的方法及其技术实现系当前模式识别研究领域中最为热门的研究课题之一。本文针对NMI(归一化转动惯量)特征识别、不变矩特征识别和比例特征识别三种图像特征识别方法.通过实验分析了该三种识别方法的缩放不变性、旋转不变性、平移不变性以及不同物体之间的特征差异。实验数据显示NMI特征识别方法具有最佳的识别效果和最快的处理速度。

关 键 词:特征识别  NMI特征  不变矩特征  比例特征  RST不  变性
文章编号:1008-05270(2004)05-0103-02
修稿时间:2003年11月2日

Study on Methods about Disposal and Recognition of Image
Men,Pengtao Zhang,Xiubin Zhang,Feng Sun,Zhimin Wu,Jiong.Study on Methods about Disposal and Recognition of Image[J].Control & Automation,2004,20(5):103-105.
Authors:Men  Pengtao Zhang  Xiubin Zhang  Feng Sun  Zhimin Wu  Jiong
Abstract:The method of Image feature recognition and the technical realization c onstitute one of the most popular study tasks in the field of pattern recognitio n. Focusing on the three Feature Recognition methods of NMI (Normalized Moment o f Inertia) feature recognition, Moment invariant feature recognition and Scaling feature recognition, and through experiment, the thesis analyzes the scaling in variability, rotation invariability, translation invariability in the three meth ods and the feature difference of various objects. The experimental data show th at the recognition method of NMI feature is the best of the three in recognition effect and processing speed.
Keywords:Feature Recognition  NMI feature  Moment invariant feature  Scaling fea ture  RST Invariant feature
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号