基于ETM+数据的大型水生植物信息分类波段选择方法研究

METHODOLOGICAL STUDY OF INFORMATION EXTRACTION OF MACROPHYTES USING ETM+ REMOTE SENSING IMAGE

以乌梁素海湿地为研究对象,对基于ETM+数据的乌梁素海湿地大型水生植物信息分类的一般方法进行研究。首先分析了研究区ETM+数据各波段的相关系数、特征值、不同波段像元值与像元数的关系,得出TM3、TM4、TM5相关性较小,包含信息量较大。采用Brovey变换将TM3、TM4、TM5与Pan波段融合,然后利用最大似然法对融合图像进行分类,结果表明,分类后影像分类精度为89.72%,影像Kappa系数为0.8363,分类效果已经达到分类精度要求。

The common methods of extracting information of macrophytes in Lake Wuliangsuhai using ETM＋ remote sensing image were discussed in this paper.According to the analysis of eigenvalue and relativity,different band pixel value and the relationship between the number of pixels obtained TM3,TM4,TM5,it was found that bands of TM3,TM4 and TM5 had small relativity and contain ample information.Bands of TM3,TM4 and TM5 were merged with Pan bands using Brovey transform.Then the combination image of fusion remote sensing images was classified by method of maximum likelihood.The post-classification accuracy and the Kappa coefficient was computed to be 89.72% and 0.8363 according to the classification.The classification results have been achieved classification precision.