摘要
高光谱遥感是一门正在兴起的极具发展潜力的应用科学技术 ,能直接快速地识别岩矿组成、丰度及其分布 ;高光谱图像丰富的空间信息也包含一定的地质构造信息。目前 ,光谱分解模型和修正的高斯模型在岩矿信息定量研究中极具潜力 ;对纹理信息的提取侧重于空域的灰度剪切与有效平均梯度相结合及频域空间内的小波包等技术。根据目前高光谱遥感信息的地质应用 ,总结出高光谱地质应用的技术流程。在此基础之上 ,以青藏高原为例 ,阐述了高光谱遥感的地质应用前景。最后 ,对高光谱遥感技术存在的问题进行了一定的评述。
Hyperspectral remote sensing has great potentialities for application in geology. It can identify and extract directly the material of earth surface, especially the altered rock. And it also can quantify the composition and the abundance of material of surface using SUM (spectral unmixing method) and MGM(modified Gaussian method). But the latent texture information of hyperspectral remote sensing image was often neglected. The paper groped the method for extracting and quantity those rich texture information using the clip transformation and effective average gradient (EAG) technique in gray space and the multi-frequency transformation or the wavelet packets technique in frequency space. A processing diagram of geological application of hyperspectral remotely sensing was given. It's potentialities for geological application can be imagined using Tibet Plateau as a test sample. At last, some questions, which must be resolved now or in the future, were discussed.
出处
《国土资源遥感》
CSCD
2000年第3期38-44,共7页
Remote Sensing for Land & Resources
关键词
高光谱遥感
定量化岩矿信息
纹理信息提取
Hyperspectral remote sensing
Identify and extract directly
SUM and MGM
Texture information
Tibet Plateau