摘要
【目的】针对已有水体信息提取方法无法准确识别干旱区湖泊水体和非水体边界信息的缺点,提出一种新的水体信息提取方法,以期高精度地提取水体信息。【方法】以我国最大的干旱区内陆淡水湖———博斯腾湖及其周边地区为研究对象,以研究区Landsat ETM+遥感影像图为主要数据源,以大量的湖岸线GPS定位数据为检验依据,提出了多波段DN值比较法(CBDN),并对比分析了CBDN与常用的单波段阀值法、NDWI、MNDWI、谱间关系法和NDWI3等方法的水体信息提取效果,对各种方法提取结果的精度进行评价。【结果】CBDN法不仅能准确地提取水体信息,而且该法提取的水体与非水体之间的边界线与实际湖岸线非常接近,水体与非水体信息提取的总体精度最高,为99.29%。【结论】CBDN方法优于其他已有的常用水体信息提取方法,该法操作过程较简单、快捷方便,有一定的推广价值。
【Objective】This study proposed a new water body information extracting method for accurate extraction of water information from inland lakes.【Method】The largest inland freshwater lakeBosten Lake and surrounding area in China was investigated based on Landsat ETM+ data and lake shoreline GPS data.Comparison of multiple bands digital numbers (CBDN) method was proposed and compared with single band threshold method,NDWI,MNDWI,spectral relationship method and NDWI-3.【Result】CBDN method accurately extracted water body information and the extracted boundaries between water body and nonwater body were very close to GPS data with a total precision of 99.29%.【Conclusion】CBDN was the best method to extract water body information for inland lakes.CBDN is simple and convenient and it can be widely used.
出处
《西北农林科技大学学报(自然科学版)》
CSCD
北大核心
2013年第12期227-234,共8页
Journal of Northwest A&F University(Natural Science Edition)
基金
国家自然科学基金项目(41161007)
新疆自然科学基金项目(2011211A034)
新疆干旱区湖泊环境与资源实验室开放基金项目(XJDX0909-2010-12)
新疆维吾尔自治区高校科研计划青年教师启动基金项目(XJEDU2010S33)