| Belt conveyor is the most common transport machine adopted in coal mine for a long distance. But during the running process, conveyor belts are often ripped longitudinally because of the high-speed operation and internal structure of conveyor belt. Once the accident happened, the entire conveyor belt will be ripped which cause serious financial losses. In order to solve poor reliability and low efficiency of the device, the author studies on-line detection system for longitudinal rip of conveyor belt which based on machine vision technique. By means of collecting images in real time, the images could be preprocessed, divided, extracted and recognized by machine vision system. Then the results are transformed into controlling signal, which could control the actuator to meet the needs of industrial detection.This paper mainly about:(1) By analyzing current longitudinal rip technique home and abroad, considering the most ideal needs of longitudinal tearing detection device and the advantages of machine vision technique, the author proposes an on-line detection system for longitudinal rip of conveyor belt which based on machine vision technique.(2) The author elaborates the operating principle and the general structure of on-line detection system for longitudinal rip of conveyor belt which based on machine vision technique. Considering the systematic function requirements and special underground environment, the software and the hardware needed by system are selected and designed optimally. So the software and hardware platform of system are established.(3) Directed at the feature of the image of longitudinal rip of conveyor belt, this paper does a study on systematic image processing and recognizing. Then the crack image is extracted, characteristic quantity of crack outline is considered as alarm threshold. Therefore, systematic alarm is achieved.(4) Considering the complicated underground situation and how to improve the function of systematic detection, the author selects appropriate image preprocessing when the image is processed and recognized. Then Canny edge detection is improved. By comparing the experimental results, the effect of systematic detection is improved.(5) In the aspect of methods of software implementation, this study uses the dynamic collection of systematic image realized by API function which is provided by CCD manufacturer at first. Then the image of longitudinal rip of conveyor belt is processed and recognized by using OpenCV machine vision. Combined with database technology, on-line detection parameter and alarm data are managed. Finally, all the functions are gathered under MFC framework, friendly interface is formed.There are some innovations in this study, at first, an non-contact on-line detection system for longitudinal rip of conveyor belt is proposed; secondly, crack image on the surface of conveyor belt could be processing and recognizing. The needs of on-line detection and alarm are satisfied. Systematic real-time is improved. At the same time, it also has other advantages, such as low cost, simple structure, long working time, high accuracy, friendly interface etc. A new thought is proposed for longitudinal rip of conveyor belt detection. According to analysing the local experimental results, the effect of on-line detection is achieved the expected goal. |
PDF Full Text Request