Application of Wavelets to Rock Components Classification

Abstract

Based on digital image feature of granite and sandstone, the spectrum diagram of several ingredients in rock is established by using wavelet theory and Bayes theory. Firstly, the sample set and distribution domain are obtained by rock image gray statistical function, whose minima have been optimized by iterative algorithm. Secondly, the establishment of each rock component bands, which are wavelet decomposition coe‐cient matrices that has been homogenized, is done in according to pyramid principles. Finally, only if the coordination set of all levels wavelet decomposition coe‐cients and corresponding point set is done which is based on the formerly gained sample set, the spectrum diagram of each component containing in granite and sandstone are flnally set up with Bayes theory. The application of wavelet to rock components recognition and classiflcation is based on priori theory, that is to say, once the spectrum of each known rock component are established with the method discussed in this paper that is easy to conflrm the rock components and their distribution in other unknown rock image.