With the rapid development of industrial intelligence, there is a growing demand for automated safety supervision and vehicle management on construction sites. Traditional manual management methods, plagued by low efficiency and high error rates, are struggling to meet the safety and management requirements of modern industrial scenarios. To address these challenges, this study proposes an intelligent management system based on computer vision. The system encompasses worker identity and attendance management, vehicle management (including license plate recognition and vehicle type classification), as well as personnel and worksite monitoring. By integrating key technologies such as edge detection, image classification/segmentation, and scene semantic recognition, the system achieves efficient and precise information extraction and intelligent analysis. Compared to conventional methods, this system not only improves recognition accuracy but also significantly enhances the real-time and adaptive capabilities of environmental monitoring. Experimental results indicate that the system achieves a worker identity recognition accuracy of 98.5% and a license plate recognition accuracy of 97.2%. Additionally, it enables real-time monitoring of the worksite environment, providing an efficient and intelligent solution for industrial safety management.