基于LandsatTM数据的济宁市水体识别方法研究  

Identification of Water Bodies in Jining City Based on LandsatTM Dataset

作　　者：李晶[1] 李恩来[1] 苗慧玲[1] 余洋[1] 王宇鹏[1] 王培俊[1] 

机构地区：[1]中国矿业大学(北京)土地复垦与生态重建研究所,北京100083

出　　处：《环境科学与技术》2013年第9期175-179,186,共6页Environmental Science & Technology

基　　金：国家科技支撑计划课题(2012BAC04B03)

摘　　要：文章基于济宁市2009年LANDSAT TM影像,根据影像各波段间的相关性和地物在各影像上的灰度差异,分别采用单波段阈值法、多波段谱间关系法、改进的水体指数法(MNDWI)及HIS空间水体提取模型法对济宁市水体信息进行识别提取。在此基础上,结合归一化指数法和HIS空间水体提取模型,提出了一种综合性的水体信息提取方法,并对各方法提取结果进行了精度评价。结果表明,已有方法对济宁市水体识别存在一定的缺陷,而综合法集成了归一化指数法和HIS空间水体模型的优点,很好地弥补了已有方法的不足,对济宁市水体信息进行了较好的提取,尤其是细小水体,同时也很好地消除了建筑物、阴影、水生植被等的影响,提高了济宁市水体信息识别的精度,为更好地定量评价济宁市地表水体时空演变规律等提供帮助。According to correlation of inter bands and gray difference of features in the image of Landsat dataset of Jining City in 2009, water body information was picked up by methods including single band threshold method, method of normalized water index （MNDWI）, b2＋b3〉b4＋b5 band relations method and HIS space water extraction method. A comprehensive method of extracting water body information was proposed by the combination of MNDWI and HIS space water extraction model, and the extraction results are accuracy evaluated. Results revealed that the methods have some shortcomings on water body identification of the City, while the comprehensive method has a good water extraction effect by integrated advantage of MNDWI and HIS, especially on small water body, meanwhile it has a better elimination on the effect of building, shadows and near-water vegetation, which would improve the recognition accuracy to a large extent, and provide better help for the quantitative evaluation of temporal and spatial evolution rule of surface water of the City.

关 键 词：TM影像 阈值 水体识别 HIS空间水体提取模型 水体指数法 

分 类 号：X87[环境科学与工程—环境工程]