Application of computer vision methods in studying the fracturing of mineral raw materials

Machine learning techniques in experimental data processing enable the identification of hidden patterns through the use of nonlinear relationships, providing opportunities for economic gains. This paper presents the results of a study on the application of machine vision methods for automating the analysis of video materials recorded during fracturing of mineral raw materials. In the face of increasing competition in the mining industry, there is a growing need for efficient, high-precision tools. The tool developed for this study is designed to detect defects occurring during the fracturing process and generate regression equations that characterize these defects. A thorough review of existing software libraries for automatic detection of fault contours, as well as their merging and approximation, was conducted. These methods can significantly reduce analysis time and enhance the accuracy of mineral fracture assessments. Special attention is given to the image processing algorithms used, which allow for the effective handling of data captured by video cameras. The paper also presents a block diagram of the tool, outlining its functionality and data processing flow at each stage. The tool was tested on real-world data, demonstrating that the use of machine vision and regression analysis algorithms can reduce costs and improve the overall efficiency of mining operations. It optimizes the rock crushing process by enabling precise equipment adjustments. This tool holds potential value for both academic researchers and professionals in the mining and machine learning fields.