摘要
由于茂密的植被覆盖,西南岩溶区在遥感图像上反映的大部分是植被覆盖层的光谱信息,而直接的岩性光谱信息很弱,因而利用遥感图像识别岩性,尤其是碳酸盐岩岩性的难度很大。以广西灌江流域为例,在植被茂密的南方岩溶区利用多源遥感数据,如ETM、SPOT、ASTER数据进行碳酸盐岩的计算机自动岩性识别。结果表明:多源遥感数据的岩性识别效果远大于单一类型的遥感数据;未进行融合的遥感数据的分类效果好于融合后的数据;利用高分辨率遥感数据的纹理图像参与分类,有助于提高分类的精度。最终,采用SPOT的4个多光谱波段、ASTER的14个波段、TM的6个波段3种遥感数据共24个波段加上4个SPOT纹理图像和3个ASTER可见光波段的纹理图像共31个波段组合进行岩性的自动识别分类,取得了82.01%的自动识别总精度。
The difficulty in geological mapping by remote sensing is lithological discrimination,especially the carbonate area covered by deep flourishing vegetation in southwest China karst area.Multi-source remote sensing data,such as TM,SPOT and ASTER,for lithological discrimination is evaluated.The results indicate that the accuracy of lithological discrimination using single RS data is not so high as 69.36% for ASTER,64.37% for TM and 54.41% for SPOT.When more sorts of RS data used for classification,higher accuracy is obtained.Except for spectral information,the inclusion of variogram texture images may improve the classification accuracy.In the case of 4 SPOT spectral bands and 4 texture images,6 TM spectral bands,14 ASTER spectral bands and 3 texture images extract from 3 VNIR spectral bands for classification,with final overall classification accuracy of 82.01%.
出处
《桂林工学院学报》
北大核心
2010年第1期41-46,共6页
Journal of Guilin University of Technology
基金
中国地质调查局地质大调查项目(200310400043)
中国地质科学院岩溶地质研究所所控基金项目(200706)
关键词
碳酸盐岩
岩性识别
遥感
lithological discrimination
remote sensing
image classification
karst area
