地理空间信息2011，Vol.9Issue(3)：3-5,8,4.
高光谱影像混合像元分解比较研究
Comparative Study of Mixed Pixel Unmixing in Hyperspectral Image
刘涛 1严海英 1于曼竹2
作者信息
- 1. 新疆维吾尔自治区第二测绘院,新疆乌鲁木齐830002
- 2. 武汉大学遥感信息工程学院,湖北武汉430079
- 折叠
摘要
Abstract
Spectral unmixing of remote sensing imagery as a remote sensing classification and cartographic method has its unique advantages. In this paper, Hyperion remote sensing images of Fukang region was used, in ENVI/IDL software running environment, respectively, utilizing the hourglass algorithm method, SMACC algorithm and volume method for endmember extraction. A comparative analysis of these methods to choose the actual extraction of the end of the hourglass algorithm as the most terminal endmember had been carried out. On this basis, respectively using the fully constrained least square method and OSP algorithm for mixed-pixel decomposition, all unmixing methods had been counted and compared. The results showed that in the area, the endmember extracted from hourglass algorithms had higher accuracy, and the result of all constrained least squares was satisfied to the ground fact.关键词
混合像元/端元/混合像元分解Key words
mixed pixel, pixel unmixing, endmenber分类
天文与地球科学引用本文复制引用
刘涛,严海英,于曼竹..高光谱影像混合像元分解比较研究[J].地理空间信息,2011,9(3):3-5,8,4.
