遥感图像因子分析的环境意义

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。

THE ENVIRONMENT MEANING OF FACTOR ANALYSIS OF REMOTE SENSING IMAGES

In this paper, experiments and theory analysis were made. The remote sensing images of 6 geographical environment were processed with factor analysis algorithm, the results were stated in table 1. Results introduced the relationship of image factors and the bases of Tasseled Cap space Generally, the first factor vector most similar to SB and second factor vector most similar to GV, as for plain agricultural environments (environment 5 in the table), inverse re-sults appeared.To analyze the experimental results, we developed a kind of theory to explore these results with physical geographical theory.In first work, the unequivalentity of the response to geographical landscapes of the sensor has been discovered, with using the supposing that ground geographical feature is 2-di-mension zero mean stationary stochastic processes and the effect of IFOV's convolution integral. It shows that the signal of primary environment landscape factor would been increased and secondary environment landscape factor would relatively been decreased under the observation of the sensor. The gain factor is sinwx sinwy/wxwy, where wx and wy are angular frequency of space, and X, Y are defined as the sweeping direction of the sensor and the moving direction of the sensor, respectively.In second work, a concept of resolution of a environment landscape, was developed dependent on the discovery. A method that equivalent to SVD transformation of image processing has been developed, it would used to process the environment information of remote sensing images.