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

基于改进邻域均值的全色影像水域提取
引用本文:金飞,王番,刘智,芮杰.基于改进邻域均值的全色影像水域提取[J].测绘科学,2013,38(3):125-127.
作者姓名:金飞  王番  刘智  芮杰
作者单位:信息工程大学,郑州,450052
摘    要:利用传统邻域均值法对全色影像水域进行提取时,所提取的水域边界精确度严重依赖于搜索窗口的大小。本文通过分析水域灰度特征以及提取边界不精确的原因,对方法进行针对性的改进:依据邻域各像素点与种子点的灰度差值分别对其赋予不同的灰度值,以处理后的邻域均值作为特征进行水域提取。实验结果表明:基于改进的邻域均值方法对遥感影像水域有较好的提取结果,提取精度较之改进前有明显改善。

关 键 词:邻域均值  聚类分析  区域生长  人机交互

Water area extraction from remote sensing image based on improved means of neighborhood
JIN Fei,WANG Fan,LIU Zhi,RUI Jie.Water area extraction from remote sensing image based on improved means of neighborhood[J].Science of Surveying and Mapping,2013,38(3):125-127.
Authors:JIN Fei  WANG Fan  LIU Zhi  RUI Jie
Institution:(Information Engineering University,Zhengzhou 450052,China)
Abstract:For water-body extraction in panchromatic remote sensing images,the traditional solutions by means of neighborhood have a strong uncertainty of accuracy in segmentation result of boundary,as different sizes of searching windows are used.In this paper,an alternative means of neighborhood was proposed by analyzing the intensity features of water-body and the uncertainty of accuracy in boundary extraction.It assigned different values of intensity to the pixels in a neighborhood according to their difference of intensity between the pixel and the seed point,and used the processed means value of neighborhood as feature for water extraction.The experimentation results showed that the solution based on alternative means of neighborhood would have an effective result with higher accuracy in water-body extraction.
Keywords:means of neighborhood  cluster analysis  region growing  interactive processing
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