改进P-WSVM的矿区积水塌陷地信息提取  

Water subsidence land information extraction in mining area based on improved P-WSVM

作　　者：范忻[1,2] 汪云甲[1,2] 王行风[1,2] 张书建[1,2] 

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221008 [2]江苏省资源环境信息工程重点实验室,江苏徐州221008

出　　处：《测绘科学》2012年第4期82-84,共3页Science of Surveying and Mapping

基　　金：国家环保部公益性行业专项基金项目(200809128);国家自然科学基金项目(50774080;40971275);江苏高校优势学科建设工程资助项目(简称:PAPD)

摘　　要：煤矿区积水塌陷地的动态监测是矿产资源管理的重要方面。针对积水塌陷地的光谱特征和空间结构特征的特点,本文引入非平衡数据集分类思想,提出一种改进的P-WSVM方法。实验表明:基于改进的P-WSVM方法能够快速、准确地提取积水塌陷地信息,并达到良好的效果。It is one important aspect of RS application in mining areas to extract mining subsiding land from RS Images.In this paper,making full use of spectral information and spatial texture feature,water subsided land extraction was realized based on improved P-WSVM method by introducing unbalanced data classification.The conclusion indicated that the P-WSVM based on improved methods could extract the information of water subsidence areas rapidly and accurately,and achieved good results.

关 键 词：P-WSVM 遥感 积水塌陷地 非平衡数据集 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置] P237.9[自动化与计算机技术—控制科学与工程]