| Because of its advantages of large volume,low cost and long-distance transportation,belt conveyor has been widely used as an important transportation tool for bulk materials in coal enterprises.The conveyor belt is an indispensable part of the belt conveyor.In the process of coal production,due to long,high load operation and some unexpected reasons,it is easy to cause longitudinal tear accident of conveyor belt.The fault is concealed and abrupt,and once the fault occurs,the safety of the underground workers and the economic loss of the coal mine can not be remedied.The longitudinal tearing of the conveyer belt is one of the most common accidents in the coal enterprises.Therefore,it is necessary to study a real-time and reliable on-line detection system for the longitudinal tear of the conveyer belt.In view of the problems of low efficiency,complicated detection system,large space and expensive price,this paper presents an on-line visual detection method based on visible light CCD and auxiliary laser line light source.The laser line source is projected on the surface of the conveyor belt to diagnose the fault,on the basis of the laser line shape captured by the camera.Based on this method,an on-line detection system for longitudinal tear of conveyor belts is designed.This paper introduces the principle of longitudinal tearing detection and defect location,and designs the vision detection scheme,including image acquisition,image processing platform,auxiliary light source,a platform of binocular vision detecting system.The image sensor selects the PointGrey CMLN-13S2M-CS CCD type visible light camera,the light source adopts the "one" word line laser line light source,the image processing platform selects the high performance NVIDIA Jetson TK1 development board.According to the characteristics of the longitudinal tearing of the conveyer belt,the image processing flow chart of the binocular vision detection is designed.The related image preprocessing algorithms,image processing and fault diagnosis and fault location algorithms are deeply studied.Considering the characteristics of conveyer belt longitudinal tearing image and the detection methods,combined with the two important factors of reliability and real-time,a comparative analysis is carried out to select the appropriate algorithm.In this paper,two cameras are used to collect images,and the images of high resolution,wide field of vision and large scene are synthesized through real-time image stitching algorithm.FAST corner detection algorithm,Hough transform line detection and circular detection algorithm are used to detect and locate the tear location of longitudinal tearing.This paper designs the embedded processing terminal software and the remote monitoring software on PC.The embedded processing terminal software mainly uses C++and OpenCV to realize the function of longitudinal tearing detection and positioning on the NVIDIA Jetson TK1 development board,and the function of sending data to the host computer software.It mainly includes image acquisition module,detection module,fault location module,data transmission module,interface module and etc.The monitor software is used to realize the functions of user login,receiving data,data management and historical view.In order to test the reliability of detection,we set up a platform for detecting physical objects and analyze the experimental data.The test results show that the accuracy and stability of the detection system are improved on the premise of ensuring real time,which is of great significance for the safety production of coal mining enterprises. |
PDF Full Text Request