火烧迹地信息遥感快速提取方法研究

THE EXTRACTING OF FIRE SCARS FROM TM IMAGE

利用福建省福州地区ETM图像,在分析火烧迹地光谱特征的基础上,研究了火烧迹地信息遥感快速提取方法。文中设置了4种提取方案,即监督分类法、NDVI参与的监督分类法、NDVI差值法及谱间关系法,最后,对提取结果进行了精度评价。

Based on an analysis of the spectral characteristics of the fire burned areas in ETM image, the authors have studied the methods for rapid extraction of the burned areas. In this paper, four methods were set up. They are the supervised classification method, the supervised classification method with the participation of NDVI, the supervised classification method of the combination image of the NDVI difference between two - period image and the first two principal components （NDVI difference method for short） , and the method based on relations of the spectral characteristics （spectral relationship method）. The results of the extraction are evaluated in this paper.