Research On Visual Inspection System Of Conveyor Belt Tearing Based On Cloud

Belt conveyors are widely used in the transportation of bulk materials in coal mining enterprises because of their low cost,large load,and long-distance transportation.As the core component of the belt conveyor,the conveyor belt not only bears the traction and load of the belt conveyor,but also bears the responsibility to ensure the smooth completion of the task in the continuous conveying of scattered materials.Because the belt conveyor is in a high-load working environment for a long time,there will always be the danger of sudden tearing of the conveyor belt,and the tearing of the conveyor belt is mostly the idler roller,the roller failure friction,the metal object jamming and piercing,Gangue scratches and penetrates the tape.Once a tearing failure occurs,if it is not discovered and stopped in time,the entire conveyor belt will be torn,and the frame,reducer,motor and other related equipment will be damaged;in severe cases,the local temperature of the conveyor belt will increase,thereby The fire caused a major safety accident and brought huge economic losses to the entire coal mine.Therefore,it is necessary to study a real-time,reliable,safe and intelligent on-line detection system for conveyor belt tearing.On the basis of summarizing the current research status at home and abroad,this paper designed a set of practical and effective cloud-based visual detection system for conveyor belt tearing,aiming at the existing problems such as low detection efficiency,high delay,chaotic data management,large space occupied,and low automation and intelligence.The system combines the traditional fault diagnosis of conveyor belt tearing with the Internet,cloud technology and machine vision,and establishes a complete set of diagnosis system from image data acquisition,tear fault analysis,evaluation and diagnosis,information data management,and finally to the maintenance of belt conveyor.The experimental results show that the system has ideal diagnostic effect,and can meet the real-time detection requirements,and has high practical application value.Firstly,this article introduces the principle of longitudinal tear detection in detail,and designs a visual inspection scheme according to actual inspection requirements,and builds a cloud-based visual online inspection system platform.In terms of hardware,it mainly selects and arranges industrial cameras,light source systems,PC devices,and cloud servers;in terms of software,it mainly designs the overall design of client software modules,cloud service systems and data acquisition modules.The camera used is a line laser assisted binocular stereo camera,which is based on point cloud data for processing;the image data processing platform uses a KJD127high-performance industrial computer;the cloud server uses Huawei's CS6 flexible cloud server.Secondly,the three-dimensional point cloud data of the conveyor belt collected by the camera is characterized by high density,large amount of data,high redundancy and inevitably containing noise points.If the original point cloud is not processed to some extent,it will cause huge waste of resources and time,and also affect the accuracy of detection of conveyor belt tear.In view of this problem,this paper carried out an in-depth study on the preprocessing algorithm of image data.First,the collected point cloud data was de-noised by combining statistical filtering and bilateral filtering smoothing processing method.Then,the de-noised point cloud data was simplified by using K adjacent point algorithm.It provides the conditions for subsequent extraction of the characteristics of conveyor belt tearing,3D reconstruction and quantitative processing.Thirdly,through the observation and contrast belt longitudinal tearing failure and the longitudinal tear of point cloud data collected fault structure,puts forward the clustering algorithm and threshold segmentation algorithm for clustering and selection of tear characteristics,through the algorithm can successfully to tear feature extraction,put some outside longitudinal tearing defects(such as local protuberance,fold,become warped skin,burrs,etc.)to give out.At the same time,because there are a lot of image processing API in Open CV,feature extraction based on image is more convenient.This paper also introduces the method of Open CV contour search to extract the tear feature points and draw them out.Then,based on PCA principal component analysis,normal vector solving and other algorithms,the extracted tear feature points were quantitatively calculated,and the length,maximum width and Angle of the conveyor belt tear were obtained.Finally,in order to more visually see the three-dimensional surface of the conveyor belt crack morphology and other important information,this paper adopts a combination of Open GL and Qt method to carry out three-dimensional real-time display of the conveyor belt image.Then,a cloud-based Visual detection system for conveyor belt tearing is designed through programming.The system adopts a combination of Microsoft Visual Studio 2010,Open GL and Qt to develop the platform.It mainly includes three modules: image data acquisition module,client software detection module and remote cloud module.The image data acquisition module is to collect the image data in real time through the binocular stereo camera,and then transmit the collected data to the image data processing module.The image data processing module is the client software module,which is responsible for data processing,algorithm diagnosis,historical data query,real-time display of fault diagnosis results and real-time transmission of the processing results to the cloud database.The remote cloud module classifies and saves the diagnostic results so as to check the running status of the conveyor belt in real time and conduct secondary data processing,etc.At the same time,the remote expert access and mobile APP interface are reserved in the system,and the remote expert diagnosis system can be established in the later stage to realize the intelligent diagnosis on the cloud server.Finally,this article summarizes the research work this time,and looks forward to the future development of related research.