Current location - Education and Training Encyclopedia - Graduation thesis - Main methods of extracting mineralization and alteration information in remote sensing image processing
Main methods of extracting mineralization and alteration information in remote sensing image processing
Remote sensing image enhancement technology is to distinguish the spectral information of different types of geological bodies, expand the difference of information through different image enhancement technologies, and obtain the required geological information.

There are differences in mineral types, structures and colors between the altered rocks formed by the alteration of surrounding rocks near the mine and the surrounding normal rocks, which leads to differences in the reflection spectral characteristics of rocks, and the spectral anomalies of specific altered rocks appear in some specific spectral bands.

Band ratio method: Band ratio method is based on the principle of algebraic operation. When the differences between bands are similar but the slopes are different, the ratio of reflection band to absorption band is used to enhance the spectral differences between various lithology! Suppress the influence of terrain and show the dynamics.

The scope of.

Principal component analysis (PCA) is a widely used method to extract rock alteration information. This method concentrates and compresses image data. It concentrates the highly relevant information of each band in a multi-spectral image into several bands, and ensures that the information of these bands is irrelevant as far as possible, that is, several comprehensive bands represent the multi-band original image! Reduce the amount of data processed.

Spectral angle mapping (SAM): Also known as spectral angle matching method. Refers to the standard spectrum measured in the laboratory or the average spectrum of known points extracted from the image. Calculate the generalized included angle between the vector of each pixel in the image (taking the spectral response of the pixel in n bands as the vector of n-dimensional space) and the reference spectral vector. Determine the similarity between spectra according to the size of the included angle! In order to identify ground objects.

The purpose.

Correspondence analysis (R-Q) type factor analysis: a multivariate statistical method that reduces some complicated factors (samples or variables) to several comprehensive factors. Through a series of coordinate rotation transformation, it can be between n variables! Extract several main factors. Reflect the main information of n variables! Also called dimensionality reduction analysis.

Others include G-s projection method, mixed pixel decomposition method, MPH technology and so on.

You can look for better domestic reviews first, then search for foreign reviews according to the English keywords inside, then look at several typical and classic documents, and then search for relatively new related materials to see the development trend, which should be more understandable.