Abstract: In order to monitor the working state of the face scraper in real time in production practice, avoid the loss caused by chain deviation, wear and breakage, and improve the intelligence and automation level of the mining process, this paper proposes an online monitoring system for the face scraper based on image processing technologies such as semantic segmentation and edge detection and magnetic flaw detection sensor technology. The results show that, on the premise of meeting the recognition conditions, the recognition rate of the scraper chain can reach 90%, the recognition delay is less than 2 seconds, and the system can have high detection accuracy when the chain of the scraper is stretched, worn or deformed more than 5%. In addition, the ground dispatching command center can receive the video, pictures and voice alarm information of the ground monitoring system of the broken chain and inclined chain of the scraper conveyor in real time.