A Cluster Analysis Based On Rock Image Feature

The identification and division of different components rock image are of great significance in the field of geological research.It is time-consuming and subjective to identify the rock flakes artificially under the microscope,and the analysis results are difficult to quantify and characterize.Therefore,the use of digital image processing technology to analyze the rock image of has become a hot topic in current research.The identification and division of rock components are the basis for analysis of rock casting thin slice.However,due to the complexity of the image of the rock casting thin slice,it is difficult to obtain the ideal result by applying the image segmentation algorithm to the rock image for component division directly,and it can not meet the requirements of rock image analysis.The usual solution is to combine the image segmentation algorithm with other analysis methods,but it also makes the problem more complicated.In order to divide and identify the components in the rock image,In this paper,weights are used to combine the color features of the rock components with the texture features,followed by the addition of spatial features.The FCM clustering algorithm is used to realize the division and identification of rock components.The specific process is as follows.Firstly,the extraction of the color features and texture features of the rock image are taken as starting points.Selecting the color component of the HSV color space as the color feature,and the contrast of gray-level co-occurrence matrix is selected as the texture feature.This color features and texture features were used to classify the rock image by FCM clustering.The experimental results show that this method can roughly divide the rock components but it is difficult to accurately classify the different rock components with similar texture features and small differences in color characteristics.Subsequently,this paper weights color features and texture features during the process of rock composition using the FCM clustering algorithm.Experiments show that this algorithm can more accurately classify minerals such as sandstone particles,pores,etc.,but there are still some problems with the wrong division of isolated pixels.Finally,based on the above methods,the spatial information features of rock images are added and then used FCM clustering algorithm to divide the rock components in this paper.The experimental results show that the algorithm can not only accurately divide the rock components but also reduce the wrong division of isolated pixels.Through analysis and comparison with other image segmentation algorithms,the rationality and effectiveness of the algorithm for the classification of rock components are verified.