New Approach to Apply Texture Features in Minerals Identification in Petrographic Thin Sections Using ANNs

Identification of minerals in petrographic thinsections using intelligent methods is very complex andchallenging task which, mineralogists and computer scientists arefaced with it. Textural features have very important role toidentify minerals, and undoubtedly without using these features,recognition minerals in thin sections yield to many missclassification results. Thin sections have been studied applyingplane-polarized and cross-polarized lights. In this paper, in orderto extract textural features of minerals in thin section, cooccurrencematrix is used, and six features as Entropy,Homogeneity, Energy, Correlation and Maximum Probabilityare extracted from each image. Then, ANNs are used foridentifying in complex situation and experimental results haveshown that using textural features in mineral identification,significant improve classification result in petrographic thinsections.

Keywords:

Texture Analysis , Mineral Identification , Thin Texture Analysis , Mineral Identification , Thin