主成分变换和彩色变换在TM图像信息提取中的应用——以苏州市为例

本文讨论了IHS彩色变换和主成分变换在苏州市TM多波段图像信息提取研究中的应用效果。研究表明,地物在多波段数据的彩色合成图像上具有较为稳定、特征的色调信息。利用苏州市水体、植被、城镇三大类地物的色调特征对TM_3(红)、TM_4(绿)、TM_5(蓝)彩色合成图像的色调成分H作了伪彩色密度分割,准确地反映了它们的分布状况。在主成分图像的IHS反变换彩色合成图像上,主成分的信息内容得到了定量、形象直观的反映,苏州市新旧城区的光谱差异、该市现状以及现代城市规划特点得到了清晰的显现。水稻、菜地、山地之间的光谱差异也得到明显的增强。主成分变换和彩色变换相结合,为多波段数据信息提取提供了一种新的方法。

THE COMBINED USE OF PRINCIPAL COMPONENT TRANS FORM AND IHS COLOR TRANSFORM FOR SPECTRAL SIGNATURE EXTRACTION FROM TM DATA

This paper shows the application of IHS color transform and principal component transform for information extraction from TM image in Suzhou study area. It is shown that lan-duse types have characteristic hues on color composites of multispectral data. Based on the hue analysis of water body, vegetation and urban area on TMs(R), TM4(G), TM5(B) color composite, the H component is color-sliced, the spatial distribution of the three major landuse types is distinctly demonstrated. On the inverse IHS color transform composite of principal components (PCs), the information content of PCs is quantitatively and vividly displayed, the spectral difference between the old urban area and the newly developed urban area, the urban status and the feature of contemporary city planning are conspicuously demonstrated. The spectral difference among rice, vegetable and hill forest is also effectively enhanced. The combined use of principal component transform and IHS color transform provides a powerful information extraction tool for multispectral data processing.