Rock Image Pore Identification Based on Fuzzy C-Means Clustering and Neural Networks

Abstract

In order to realize the recognition automation of rock section pore images, a method combined Fuzzy C-Means clustering with BP neural network is proposed to recognize the pore of rock images. Firstly, Fuzzy C-Means clustering as segmentation algorithm are applied to the rock images and they are divided into two types, then using the BP neural network training and classification recognition. It is shown that the trained BP neural network can accurately identify the effective porosity in the casting image, and lay a good foundation for practical applications.