ASTER高光谱影像提取地面人工建筑物信息的应用

Extracting the Information of Concrete/Urban Development from High Spectrum Resolution Image of ASTER

高级星载热辐射热反射探测仪(ASTER)为对地观测提供更高质量的信息源。本文对福州市ASTER影像数据进行了主成分分析、波段运算和自动分类,结果表明其能较好地提取地面的人工建筑物信息。

This paper introduced ASTER image as the sign of a high spectrum resolution era. A comparison was made between high spectrum resolution images and previous common remote sensing images. Great differences between ASTER and other common images such as TM are obvious as far as some technical parameters are concerned. It is clear that high spectrum resolution images could be information resources of higher quality. ASTER images are very useful in various fields, such as glaciology, hydrology, urban expansion, volcanic forecast, evapotranspiration / surface temperation and geology. In this study, an ASTER image, covering the citywide of Fuzhou,is preliminarily disposed during some pre-classification processes such as principal com-ponents analysis and band math. In principal components analysis process, the entire information of the initial images was compressed to six bands of the disposed image,which can well represent initial information and meanwhile can make concrete/urban buildings more distinct.In the process of band math, which is based on the spectrum characteristics of previous image, a common band math formula, B1-B2, is used. B1 and B2 represent Band 1 and Band 2 in the previous image respectively. The value of B1-B2 depends on the type of land cover. Two kinds of values can be seen, the one greater than zero represents buildings and the one smaller than zero represents natural cover such as trees or water bodies. Unsupervised classification method of Isodata is used to extract the cover of buildings. Then the disposed image is used to extract the concrete/urban development information of the research area with a good result. This result can be used in further analyses, such as urban planning, urban environment evaluation with ecological theory of interstice and so on.