论文部分内容阅读
土壤盐渍化是干旱区绿洲稳定与可持续发展面临的主要环境问题之一,因此借助遥感手段及时准确地提取盐渍地信息并掌握其空间分布有着重要的现实意义。本文以渭干河—库车河三角洲绿洲为例,使用Radarsat SAR与Landsat TM影像进行主成分融合,同时与HIS和Brovey变换的融合效果作定量比较,并利用BP神经网络模型,以相同的训练样本分别对融合前后的影像进行分类。结果表明:盐渍地主要分布在绿洲的和沙漠之间的交错带,盐渍地的分布在绿洲内部呈条形状分布,而在绿洲外部呈片状分布,且绿洲外部重度盐渍地交错分布在中轻度盐渍地中;主成份变换融合影像的光谱信息保持性、信息量都优于其它常用的融合方法,且分类精度比单一LANDSAT TM多光谱影像有较大提高,是监测干旱区盐渍地变化的有效手段。
Soil salinization is one of the major environmental problems in the stability and sustainable development of oasis in arid regions. Therefore, it is of great practical significance to extract salinized and timely information and grasp the spatial distribution of salinity information by means of remote sensing. Taking the Weigan-Kuqa River delta oasis as an example, the Radarsat SAR and Landsat TM images are used for principal component fusion, and the fusion effect between HIS and Brovey transform is quantitatively compared. The BP neural network model and the same training Samples were classified before and after the fusion images. The results show that salinized land is mainly distributed in the ecotone between the oasis and the desert. The distribution of saline land is strip-shaped in the oasis and flaky in the outer oasis, and the distribution of salty ground outside the oasis is staggered In moderately salted land, the spectral information retentivity and the amount of information of the main component transformed fusion image are superior to other commonly used fusion methods, and the classification accuracy is greatly improved than the single LANDSAT TM multi-spectral image, Salinization of the effective means of change.