| China is an excessive coal production and utilization country,and the normal supply of coal resources enterprises supports the normal development of the industrial economy.As a key transportation equipment in coal mine production,belt conveyors are often under high-load operation.In addition,the mine production environment is complex and harsh.The conveyor belt encounters foreign matter friction,metal tool blockage,and changes in roller structure.Longitudinal tearing and deviation of the conveyor belt are prone to occur.Longitudinal tearing of the conveyor belt will cause the local temperature of the conveyor belt to be too high,increasing the probability of fire.Deviation of the conveyor belt will cause coal piles on the conveyor,which will further cause accidents such as roadway blockage,conveyor belt friction and fracture,and mine fires.Therefore,it is necessary to study and design a monitoring system that can detect the abnormal state of coal mine belt conveyors to ensure the smooth and safe operation of belt conveyors during production operations.The system uses two industrial cameras to form a binocular vision imaging system.Under the action of an auxiliary light source,images of the conveyor belt are synchronously collected,and image data with a complete conveyor belt angle of view is obtained by real-time stitching of the same frame of images.Due to insufficient light and excessive dust in the coal mine,the quality of the collected images of the conveyor belt is poor.This paper designs an image preprocessing scheme to filter the images collected in the coal mine by noise filtering,gray enhancement,and contrast improvement.At the same time,the related algorithms are improved and performance is improved,so that the collected conveyor belt images can meet the subsequent fault detection requirements.For the identification and judgment of the conveyor belt faults,the detection schemes for the longitudinal tearing of the conveyor belt and the faults of the conveyor belt deviation are designed respectively.By identifying and judging the corner features and linear features in the conveyor belt image,the detection of longitudinal tearing faults is realized;By extracting the edge feature of the conveyor image collected by the camera,the coordinate information of the straight line feature is obtained,and the fault detection of the conveyor deviation is realized.And proposed an improved method for the Canny algorithm.After experimental simulation analysis and verification,the optimized algorithm has achieved better image processing results.In the process of realizing the software algorithm of the detection device,firstly build a software development environment based on VS2015 and Open CV3.4.0,and implement the computer to capture and call the programmable camera according to the API function interface;secondly,use the C++ programming language to write image processing algorithms and fault recognition algorithm programs,To identify and judge the fault characteristics in the conveyor belt image;finally,combined with the MFC application framework,Develop and design the host computer interactive interface of the detection device,combine with the database function record,store the operating status data of the conveyor belt,so as to facilitate the analysis of the cause of the failure later,To achieve the purpose of better prevention of conveyor belt failure.Based on the establishment of a simulation experiment platform and the realization of the entire detection system,experimental testing and analysis of the functions of each subsystem have been carried out.The results show that the system can accurately detect the image corners and linear features in the tearing state of the conveyor belt,and realize the transmission Detection with longitudinal tearing;By using the improved algorithm to process the experimental objects,it is shown that after the feature extraction of the improved algorithm,the C/A value and C/B value of the pixel are greatly optimized compared with the original algorithm,and the image can be extracted more accurately.The characteristic pixel information in the system ensures the accurate identification of subsequent faults,and can be accurately detected The position information of the edge of the conveyor belt realizes the accurate identification of the deviation fault.The algorithm running time of the whole detection system is between 25ms-30 ms,The detection system device proposed in this paper can effectively ensure the smoothness of the conveyor running state,and is of great significance in remote control of the conveyor belt and ensuring the economic benefits of the mine. |
