首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于区域增长的图像跟踪算法的研究
引用本文:薛雪,刘泽平,丁艳.基于区域增长的图像跟踪算法的研究[J].光学技术,2005,31(1):152-154.
作者姓名:薛雪  刘泽平  丁艳
作者单位:北京理工大学,机电工程学院,北京,100081
摘    要:为了提高序列图像的跟踪精度,提出了基于区域增长的特征量提取方法。这种方法可捕获图像中所有单连通域,并能准确地提取其特征量,然后依据所提取的特征量设计分类器,实现对图像中所有单连通域进行分类识别并加以跟踪的目的。此算法有效地解决了一般传统识别算法难以区分目标和其近邻区域的干扰,而导致跟踪目标特征量提取不准确的问题。在目标特征提取识别算法的基础上,还提出了阈值预测分割算法,并应用在序列图像的跟踪中,取得了较好效果。

关 键 词:图像跟踪  特征提取  区域增长  阈值分割
文章编号:1002-1582(2005)01-0152-03
修稿时间:2004年6月1日

Study on image tracking base on region growing
XUE Xue,LIU Ze-ping,DING Yan.Study on image tracking base on region growing[J].Optical Technique,2005,31(1):152-154.
Authors:XUE Xue  LIU Ze-ping  DING Yan
Abstract:In order to improve the precision of image tracking, a new recognition arithmetic used in region growing is proposed. The method can capture the single connected component in image and extract it's feature. According to the feature, the method of categories to recognize entire single connected components is design. The recognition arithmetic solves inexact extracted feature, which is brought on by disturbance of background image near the object in ordinary recognition arithmetic. Base on the recognition arithmetic, a segmentation method by forecasting threshold is proposed. Using this segmentation method in image tracking, satisfying effect is gotten.
Keywords:image tracking  feature extraction  region growing  threshold segmentation  
本文献已被 CNKI 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