基于NDWI分割与面向对象的水体信息提取  

Water Body Information Extraction Based on NDWI Segmentation and Object-oriented Method

作　　者：童李霞[1,2] 燕琴[3] 骆成凤[2] 杜英坤[2,4] 

机构地区：[1]山东科技大学,山东青岛266590 [2]中国测绘科学研究院,北京100830 [3]国家测绘地理信息局,北京100830 [4]山东农业大学,山东泰安271018

出　　处：《地理空间信息》2017年第5期57-59,共3页Geospatial Information

基　　金：中国测绘科学研究院基本科研业务资助项目(7771512)

摘　　要：基于OLI融合影像,运用面向对象的目标提取方法对龙羊峡水体信息进行了提取。先将NDWI作为一个波段叠加到融合后的多波段影像中参与多尺度分割,再结合隶属度函数分类方法,选定NDWI的阈值范围提取水体信息,最后与基于像元的NDWI水体提取方法进行了比较。结果表明,两种方法受冰雪影响均较小,但面向对象的NDWI水体提取方法受山体阴影和薄云的影响较小,能完全消除基于像元的NDWI水体提取方法出现的"零星水体"现象。Taking OLI images as the data source, this paper used the object-oriented object extraction method to extract water body information in Longyangxia. NDWI as a band was stacked in the merged multi-band images to participate multiresolution segmentation. And then, combined with membership function classification method, the paper chose the threshold range of NDWI, and extracted the water body information. Finally, this paper compared this method with the NDWI water body extraction method based on pixel. The results show that the two methods are affected by the snow less, and the object-oriented NDWI water body information extraction method is less affected by the mountain shadow and thin clouds, which can completely eliminate the Sporadic Water caused by the NDWI water body information extraction method based on pixel.

关 键 词：NDWI 面向对象 水体提取 OLI影像 

分 类 号：P237[天文地球—摄影测量与遥感]