Research On Safety Protection System Of Mine Belt Conveyor Based On Vision

The coal industry is an important basic industry related to the country's economic lifeline and energy security,belt conveyor is the key equipment for coal transportation and play a vital role in coal transportation.However,during the transportation of coal,when the coal falls at the coal drop,it is easy to cause accidents of coal stacking caused by blockage of bulk materials.At the same time,due to the design and performance of the belt conveyor equipment,every rotating part such as the roller and idler may be a pinch point that hurts people.Therefore,it is very dangerous when someone approaches the belt conveyor in operation.With the rapid development of current machine vision technology,in order to make up for these shortcomings,this paper studies and designs a vision-based safety protection system for mining belt conveyors,and the key algorithms(personnel intrusion detection,bulk material detection,and coal stacking detection algorithms)applied in the system in this paper have been studied in depth and necessary improvements,and the final detection algorithms have been verified through multiple sets of experiments.The main research work of this paper is as follows:(1)Aiming at the problems existing in the application of belt conveyors in coal mines,this paper designs the safety protection system scheme of belt conveyors.The contents of the scheme include the overall architecture of the system,the equipments selection of hardware parts,the selection of development software,and the research of visual inspection algorithms.(2)A personal intrusion detection algorithm based on the combination of the improved Vibe algorithm and the improved three-frame difference method is proposed.This paper first aims at the problem of Ghost regions that are prone to occur in the process of detecting targets by traditional Vibe algorithm,and improves the algorithm from two aspects: improving the establishment way of background model and distance determination self-adaption.Then,due to the problems still existing in the improved Vibe algorithm,this paper introduces improved three-frame difference method.And it is combined with the improved Vibe algorithm for the judgment of personnel intrusion detection.Finally,the effectiveness of the algorithm is verified through experiments and analysis.(3)Aiming at the problem that the traditional image target detection algorithm is difficult to distinguish the material targets on the conveyor belt,this paper studies a detection algorithm for bulk materials on the conveyor belt based on deep learning target detection.First,the conveyor belt material target detection data set was produced according to the video images collected on the spot.Then,the YOLOv3 algorithm in the deep learning target detection was selected to be used in the conveyor belt material target detection and tested and verified the images under different sizes,different times,and different positions.Finally,the size of the material target identified by the algorithm is judged,and the detected bulk material target is output.(4)Aiming at the problem that the detection of coal stacking on belt conveyors is still based on contact sensors,this paper studies an algorithm for detecting coal stacking faults through non-contact image processing methods.After cutting out the areas where coal stacking faults are likely to occur in the original image,this paper first uses the gray level difference between coal and background to segment the image and select the material area.Then it performs morphological processing on the image to remove the interference of the dropped coal area.Finally it sets the early warning threshold line and alarm threshold line to detect whether the coal stacking area exceeds these threshold lines,and then it can judge whether a coal stacking fault has occurred in the image.(5)Based on the front research of related theoretical algorithms such as the belt conveyor personnel intrusion detection,bulk material detection and coal stacking detection,the belt conveyor safety protection system is implemented through the VS2019 development platform combined with software such as Halcon and OpenCV vision development tools.At the same time,this paper builds an experimental platform to test the detection effects of related algorithms.Finally,the system is installed and applied on the actual site to test the on-site industrial application effects.This paper has 47 figures,8 tables and 83 references.