论文部分内容阅读
本文讨论了当前遥感图像自动识别与分类中存在的一些问题,提出由地形、地质、植被数据辅助执行的土壤类型模糊分类方法,在土壤数据分类和遥感图像识别中,定量地引入了土壤发生分类学的概念,并赋予成土因素对分类结果的影响作用以新的含义,从而把地学相关分析与自动识别分类有机地结合了起来。这对消除自动分类中光谱特征相互混淆的现象及提高分类精度具有一定的意义。
In this paper, some problems existing in the current automatic recognition and classification of remote sensing images are discussed. A fuzzy classification method based on topography, geology and vegetation data is proposed. Soil taxonomy and remote sensing image recognition are used to quantitatively introduce soil classification The concept of learning and the influence of soil-forming factors on the classification results are given a new meaning, which organically combines the geo-related analysis and the automatic identification classification. This is of some significance to eliminate the mutual confusion of spectral features in automatic classification and to improve the classification accuracy.