Abstract: Chemical enterprises encounter diverse safety risks during production and operation, including tank area filling, high-risk chemical loading/unloading, and chemical equipment maintenance. Recently, to enhance on-site safety, numerous surveillance cameras have been installed within facilities, comprehensively monitoring personnel, devices, and equipment. The advent of artificial intelligence and computer vision technology ushers in video intelligent analysis as a novel safety monitoring approach, which liberates security personnel from tedious manual video surveillance. It concurrently monitors multiple video channels, reducing labor costs and elevating monitoring efficiency. Moreover, devoid of human eye fatigue issues, it enhances monitoring precision. This paper proposes a computer vision-based chemical safety supervision system. Under the man-machine integration paradigm, it expounds how computer vision technology empowers chemical safety monitoring, offering a valuable reference for AI-driven enhancement of chemical industry safety production.