Research On Safety Monitoring Of Belt Conveyor Based On Deep Learning

Belt conveyor uses belt and traction parts to transport materials.It is often used in industrial environment such as mines or iron and steel plants to transport materials such as coal,metallurgy and cement.Generally,the environment of belt conveyor is relatively poor,including dark light,large dust,narrow road,odor,temperature,moisture and other characteristics of transportation materials,and narrow pedestrian walkways on both sides.When workers are taking care of the patrol inspection,the idler,roller and other facilities of belt conveyor system are in an open state during operation,which is likely to happen with these parts It is necessary to touch,which not only poses a potential threat to workers' lives,but also causes safety problems related to production such as equipment outage.Therefore,the safety of belt conveyor can not be ignored.This paper analyzes and summarizes the belt conveyor transportation accidents in recent years,deeply explores and studies the graphics,image processing and the current intelligent factory and intelligent video monitoring technology,through artificial intelligence technology,automatic control technology,real-time monitoring technology,to ensure the operation,status and personnel monitoring of the belt conveyor system.The main work of this paper is as follows:(1)In this paper,according to the specific situation of the application of image processing and other related algorithms,the design of custom regulatory area.The custom supervision module designed in this paper is the premise to judge the position of personnel.The supervision module area includes red locking linkage control emergency stop area,yellow warning area and black safety area.(2)Aiming at the problem of factory safety monitoring,the data set of this environment is established by building the behavior of employees on both sides of the conveyor Road,collecting a large number of images of different actions,and a set of object extraction and detection method based on yolov3 algorithm is designed.Aiming at the problem of small amount of data in the actual environment,the methods of data enhancement and learning migration are adopted to achieve a small number of annotation sets Ensure the validity of the algorithm model.The yolov3 algorithm is tested and analyzed.(3)Based on the supervision area module and personnel detection module of the first two chapters,the overall structure of the safety monitoring system based on deep learning is designed.The coordinates of the system are determined by yolov3 algorithm,and the positions of its employees are determined,so as to trigger early warning or linkage control according to the results.On the client server platform,through the combination of software and hardware,the human-computer interaction function module is designed to achieve the functional requirements of the on-site safety production intelligent monitoring system.