Abstract: Machine vision technology has positively improved coal mine safety monitoring methods and enhanced equipment automation levels. This article elaborates in detail on the principles of equipment information state perception based on machine vision in different scenarios and systems during the current intelligent construction process of coal mines. It summarizes the practical applications of machine vision perception technology in coal mine safety monitoring, picking recognition, coal rock recognition, positioning navigation, transportation detection, pose detection, and information measurement. The analysis points out that in the future, coal mine machine vision perception technology should deeply explore the understanding needs of mining face machine vision scenes. It is suggested to build a production full field of view monitoring and detection system, and improve the integrated monitoring effect of multiple spatiotemporal, multi-dimensional, and multivariate. It is suggested to improve the video autonomous monitoring and alarm capability, enhance visual guidance capability, and form a unified visual data management method for ground production management and operation systems. The key research should focus on technologies such as simultaneous spatiotemporal measurement of the pose of fully mechanized mining equipment (groups), perception of dynamic changes in the mining environment, full field of view monitoring and autonomous warning for production, and visual guidance and control of coal mining robots. It is pointed out that the coal mine machine vision perception technology still has challenges in explosion-proof or intrinsically safe intelligent vision sensors, efficient methods of visual measurement and analysis, the measurement precision of detection and recognition, and high-quality image annotation. Through the development of visual sensors with edge computing capabilities, a distributed vision measurement scheme is constructed to achieve accurate recognition and measurement of mining information in various complex environments. It can effectively improve the deeper integration and application of machine vision perception technology in the coal industry.