基于混合像元光谱信息的正交试验设计与交互作用分析

Mixed pixels are not only a major obstacle to the development of quantitative remote sensing. They also seriously affect computer processing in the field of remote sensing. Lambertian surface not only has the spatial distribution with multi-angle and multi-directional characteristics, also has the polarization. This paper aims to the most common mixed pixels-vegetation-soil mixed pixels as an example, analyzed the hyperspectral curves under different polarized angles and different area ratios, and designed an orthogonal experiment scientifically. The result was found that the above factors and their interactions could impact on the hyperspectral characteristics of mixed pixels, which has great theoretical significance and wide application prospect on the use of mixed pixels fusion by polarization and hyperspectral remote sensing, thus achieving target identification, classification and promote the development of quantitative remote sensing.