论文部分内容阅读
本文描述了一种图像分解处理技术。该技术基于像元反射率是各组分反射率的线性组合这一假设,最初用三个步骤将多波段数据标定为视反射率:(1)最小值减法;(2)波段平均值标准化;(3)反射率平均值均衡化。用实验室或野外波谱数据,解线性混合方程,求出各种成分所占的比例。对于澳大利亚北昆士兰的一个试验场,有两种植被(绿色和干的)及4种矿物(粘土、赤铁矿、针铁矿和石英)用NS001机载扫描仪数据进行了分解,所生成的6种图像上显示出单一的植被和矿物种类在空间上的分布。通过按比例增加每个像元上的矿物所占比例,来消除矿物图中绿色和干的植被的掩蔽效应。分解矿物图用来圈定具潜在经济意义的热液蚀变带。另外,增加了的矿物比例还用于重新计算原始波段的亮度值,产生对识别岩石类型很有用的去植被地质反射率图像。
This article describes an image decomposition processing technique. Based on the assumption that pixel reflectivity is a linear combination of reflectivities of components, the multi-band data is initially calibrated as apparent reflectivity in three steps: (1) minimum subtraction; (2) band average normalization; (3) Average reflectance leveling. Using laboratory or field spectroscopy data, solve linear mixed equations and find the proportions of various components. For a test site in North Queensland, Australia, two vegetation types (green and dry) and four minerals (clay, hematite, goethite and quartz) were decomposed using NS001 airborne scanner data. The resulting Six kinds of images showed the spatial distribution of single vegetation and mineral species. Eliminate the masking effect of green and dry vegetation on mineral maps by scaling up the proportion of minerals on each pixel. Decomposition mineralogy is used to delineate hydrothermal alteration zones of potential economic importance. In addition, the increased proportion of minerals is also used to recalculate the brightness values of the original bands, producing degenerated geologic reflectance images that are useful for identifying rock types.