星下目标的光谱反射特性分析及云层的自动识别

本文着重讨论了遥感数据的实时分类和星载相机快门的自动关启问题.首先建立了计算各类目标景象亮度的物理模型,分析了大气等因素对遥感数据的影响及各类自然目标的光谱反射特性,计算了三十四种目标在九种大气条件下的亮度值.进而提出了识别植被、土壤、水、冰雪和云等五类目标的最佳波段和分类函数及分类流程图,得到了很高的分类精度.然后又特别讨论了云层和所有非云目标(包括混合目标)的分类问题,并用密度计测量了十张MSS卫星图片上的180个目标的亮度值.最后提出了云层开关的实现途径和系统设计的基本思想. 

The Analysis for Spectral Reflectance Characteristics of Targets under Satellites and the Automatic Recognition of Cloud

A method for real-time automatic classification of targets and the automatic control of camera shutter are discussed. The physical model of computing the radiance is presented, the influences of atmosphere is considered and the spectral reflectance characteristics of five kinds of targets is analyzed. The radiance of 34 tar.gets at sorts of atmospheric conditions are computed with the model. The best wavebands, the classification functions and classification flow diagrams for five kinds of targets are determined. The problem of distinguishing cloud and non-cloud is studied. Ten Landsat MSS pictures are measured with the TD504 densimeter. The radiance distribution chart is similar to the result of calculation. How to realize the real-time automatic recognition of cloud on the satellite is also discussed.