基于像元二分模型改进的水体指数提取方法研究  

Improved Water Body Index Extraction Method Based on Dimidiate Pixel Model

作　　者：丁鹏飞 刘汉湖[1] DING Pengfei;LIU Hanhu(Key Laboratory of Land and Resources Ministry for Geoscience Spatial Information Technology,Chengdu University of Technology,Chengdu 610059,China)

机构地区：[1]成都理工大学国土资源部地学空间信息技术重点实验室,成都610059

出　　处：《河南科学》2020年第10期1625-1632,共8页Henan Science

基　　金：国家自然科学基金(41102225)。

摘　　要：湖泊是生态圈的重要组成,湖泊面积变化与人类活动和气候变化具有高度敏感性,研究湖泊变化可以帮助我们了解生态环境变化.近年来遥感技术的进步为湖泊面积提取提供了更好的方法,但是由于空间分辨率的影响,混合像元的存在制约水体分类与提取的精度.针对低分辨率影像单个像元内地物复杂,难以选择端元等问题,选用Landsat8 OLI数据,利用改进归一化水体指数基础上的像元二分模型对纳木错等湖泊进行面积提取,分别利用线性分解模型和高分辨率影像提取的水体对结果进行精度评价.研究发现改进像元二分模型提取的水体面积与高分辨影像高度相关,且Kappa系数更高,说明该方法提取水体信息更精确,特别是对水陆模糊边界区域提取精度效果较好.Lakes are an important part of the ecosystem,and the lake area changes are highly sensitive to human activities and climate change.Studying lake changes can help us understand ecological changes.Recent advances in remote sensing technology have provided better methods for lake area extraction,but due to the spatial resolution,the presence of mixed pixels limits the accuracy of water classification and extraction.In order to solve the problem of complex inland features of single image elements in low-resolution images and difficulty in selecting end elements,this paper selects Landsat8 OLI data and uses the dimidiate pixel model based on the improved normalized water body index to extract the area of lakes such as Namucuo,and evaluates the accuracy of the results by using the linear decomposition model and the water body extracted from high-resolution images,respectively.It is found that the water body area extracted by the improved dimidiate pixel model is highly correlated with the high-resolution images,and the Kappa coefficient is higher,indicating that the method is more accurate in extracting water body information,especially for the fuzzy boundary region of land and water.

关 键 词：像元二分模型 线性分解 MNDWI 水体信息提取 

分 类 号：P237[天文地球—摄影测量与遥感] P343.3[天文地球—测绘科学与技术]