论文部分内容阅读
本文是一份实验和理论分析报告。六种地理环境类型的遥感图像用因子分析算法作了处理。并介绍了图像因素和Tasseled Cap空间基之间的关系。一般来说第一因素矢量和SB最相似,而第二因素矢量和GV最相似。但在平原农业环境出现相反的结果。 为分析实验的成果,我们试图从理论上进行探讨。假设地面地理特征是一个二维零均值的平稳随机过程,IFOV卷积效应使初级环境景观因子信号被增强,而二级环境因子相对地在传感器的观察下是递降的。 在第二步工作中,发展了一种基于环境景观学概念的方法,它相当于图像处理的奇异值分解(SVD变换),它将用来处理遥感图像的环境信息。
This article is an experimental and theoretical analysis report. Remote sensing images of the six geo-environment types were processed using a factor analysis algorithm. It also describes the relationship between image elements and Tasseled Cap space bases. In general, the first factor vector is most similar to SB, while the second factor vector is most similar to GV. However, the opposite result occurred in the agrarian plain. In order to analyze the experimental results, we try to explore the theory. Assuming that the geography is a stationary two-dimensional zero-mean random process, the IFOV convolution effect enhances the signal of the primary environmental landscape factor, while the secondary environmental factor decreases relatively with the observation of the sensor. In the second step, a method based on the concept of environmental landscape has been developed, which is equivalent to singular value decomposition (SVD) of image processing, which will be used to process the environmental information of remote sensing images.