Abstract:The new generation of information technology has emerged as a crucial driving force for industrial and social advancement, and its profound integration with mines has become an inevitable trend for the innovative development of mining. Machine vision technology, as one of the significant representatives of information technology, plays a vital role in enhancing the perception capabilities of mines, reducing safety risks, and achieving fewer personnel in hazardous areas, digitization of production indicators, and prompt information feedback. This paper systematically reviews the basic concepts, principles, typical algorithms, and development history of machine vision technology. Around the main production elements involved in mine production, it elaborates on the latest research progress and application status of machine vision technology in the mining field from the four aspects of materials, machinery, personnel, and environment, and analyzes the main characteristics of the current research and application. The research findings indicate that the application of machine vision technology in mines confronts issues such as the difficulty in obtaining basic data, no substantial breakthroughs in algorithms and computing power, and poor stability in complex environment detection. In light of the current development status of machine vision technology and the practical problems encountered, future breakthroughs can be sought in the directions of small sample detection, 3D vision detection, and multi-data fusion detection. Small sample technology reduces the cost of data collection through efficient learning approaches; multi-data fusion technology enhances the adaptability of detection by integrating multi-information; 3D vision technology strengthens the detection capabilities in complex three-dimensional spaces and improves the detection accuracy. Applying the aforementioned cutting-edge technologies to visual detection in the complex mining environment can address the application limitations of traditional visual detection methods under challenges such as scarce basic data and variable environments.