5 . analyze the spectral characteristics of typical object . process spectrum especially the vegetable spectrum with differential calculation, normalize processing and differential calculation after normalize processing, the spectral difference of the same object was removed 分析了光譜庫(kù)中典型地物的光譜特性,對(duì)光譜庫(kù)光譜特別是植被光譜做了導(dǎo)數(shù)、歸?化及歸?化后再計(jì)算導(dǎo)數(shù)等三種處理,消除了同一地物因背景、光照差異等因素造成的光譜差異。
5 . analyze the spectral characteristics of typical object . process spectrum especially the vegetable spectrum with differential calculation, normalize processing and differential calculation after normalize processing, the spectral difference of the same object was removed 分析了光譜庫(kù)中典型地物的光譜特性,對(duì)光譜庫(kù)光譜特別是植被光譜做了導(dǎo)數(shù)、歸?化及歸?化后再計(jì)算導(dǎo)數(shù)等三種處理,消除了同一地物因背景、光照差異等因素造成的光譜差異。
thirdly, the paper analyzes the terrain entropy in detail . the primary bug of this algorithm is that discards the relative position information of terrain data and normalized processing . based on this, the kalman filter for optional estimation is offered and the maximization of mutual information is synthesized, the performance of algorithm is improved, the filter divergence is restrained 基于此點(diǎn),論文對(duì)地形熵算法做了詳細(xì)的分析,提出了基于最優(yōu)估計(jì)思想的卡爾曼濾波方法,并綜合使用了交互信息量等方法,改進(jìn)了地形熵算法在tfta中應(yīng)用的性能,取得了較好的效果。