激光与光电子学进展, 2016, 53 (8): 080104, 网络出版: 2016-08-11
基于陆地成像仪影像和主成分分析的水体信息提取——以鄱阳湖区为例
Water Body Information Extraction Based on Operational Land Imager Imagery and Principal Component Analysis: A Case Study of Poyang Lake Region
海洋光学 陆地成像仪影像 主成分水体指数 水体提取 鄱阳湖 oceanic optics operational land imager imagery principal components water index water extraction Poyang Lake
摘要
利用卫星影像快速准确地提取地球表面水体分布一直是一个重要的研究课题,其对水体灾害监测、水资源利用等具有重要意义。利用陆地成像仪(OLI)遥感影像,通过对图像进行辐射定标、大气校正获得影像的地表反射率值;然后通过分析典型水体指数的构建方法及地物的主分量空间特征,构建了一种主成分水体指数(PCWI)。以鄱阳湖区为研究区,选取枯水期两个不同时间的OLI影像,利用PCWI进行了水体提取,提取总体精度分别为95.92%、95.52%。与其他已有的5种典型水体指数的水体信息提取效果对比显示,PCWI对两期OLI影像水体提取的总体效果最好。综上表明,利用PCWI进行水体提取是可行且有效的。
Abstract
The use of satellite images to quickly and accurately extract the distribution of the earth′s surface water body has been an important research topic all the time, which has important significance for water disaster monitoring, water resource utilization, etc. The operational land imager (OLI) remote sensing imagery is used to obtain the surface reflectance values of the images by radiometric calibration and atmospheric correction. Then principal components water index (PCWI) is constructed by analyzing the construction method of the typical water index and the principal component space features of ground objects. Taking Poyang Lake as the research area, two different time OLI images are selected during the dry season to extract water body information by using PCWI. The overall accuracy is 95.92% and 95.52% respectively. Compared with the water body extraction effects of other existing five typical water index, the overall effect of water extraction based on PCWI for the two different time OLI images is the best. In conclusion, PCWI of water extraction is feasible and effective.
周小莉, 郭加伟, 刘锟铭. 基于陆地成像仪影像和主成分分析的水体信息提取——以鄱阳湖区为例[J]. 激光与光电子学进展, 2016, 53(8): 080104. Zhou Xiaoli, Guo Jiawei, Liu Kunming. Water Body Information Extraction Based on Operational Land Imager Imagery and Principal Component Analysis: A Case Study of Poyang Lake Region[J]. Laser & Optoelectronics Progress, 2016, 53(8): 080104.