Hyperspectral Alteration Information Extraction

1. Background

At present, although theoretically altered minerals (rocks) have very obvious spectral characteristics, due to the insufficient spatial resolution of high (Multi) spectral data, at the same time altered minerals (rocks) it is always a "small category" in the objects studied"
Target, coupled with the interference and mixing of background and disturbance spectral data. As a result, the presentation of the information of the mineralization and alteration in the remote sensing images is always "weak. Therefore, the extraction of remote sensing alteration exception information is still quite difficult. In terms of physical significance, the spectral characteristics of altered minerals (rocks) on the spectral curve are clearly spectral-identifiable, but the alteration information is recognizable in the Multi-dataset of remote sensing images, there are still great uncertainties and difficulties.

2. Main Ideas

(1) Sequential Analysis (2) Gradual Change of geographical phenomena (3) extraction process of Alteration Information (1) data selection (with no Radiation Correction performed )) (2) analysis on the sequence structure of B and hyperspectral bands with a and regression skewness

C. Select a band

By selecting several representative band sequence dolls,
Including (b195, b205), (b148, b216), (b130, b221), (3) information Extraction A, Principal Component Analysis 4, validation and spectral library for Spectral Angle matching (image must be first corrected by radiation) 5. Reference Source

(1) project 863 extraction technology and software development of high (Multi) Spectral Remote Sensing mineral alteration information (2010 second prize of China Nonferrous Metals Industry Science and Technology)

(2) Analysis and Application of band Sequence Structure of hyperspectral data