Research On Detecting Methods For Abnormal Status Of Mine Conveyor Belt Based On Computer Vision

Coal is the basic industry of China’s economic development,which plays an important role in social development and economic construction.As the main equipment of coal transportation,mining conveyor belt will bring great economic loss to the enterprise once it breaks down,and even endanger the safety of miners.The detection of mining conveyor belt abnormal situation has become an important content of information management and monitoring in coal enterprises.The detection method based on computer vision has the advantages of low cost,good reusability and high accuracy,which is the first choice of many scholars.But at present,there is not a relatively complete and effective detection method based on computer vision for mining conveyor belt abnormalities.To this end,this paper proposes a set of computer vision-based mining belt abnormality detection method.This method mainly conducts a comprehensive and in-depth study on the tearing and deviation of the conveyor belt:1.For small conveyor belt tear detection,the existing detection methods lack of early warning and low detection accuracy.For this reason,this paper puts forward a method of tear detection for small conveyor belt.In this method,the "one" shaped line laser is projected on the lower surface of the conveyor belt,and the CCD industrial camera is used to collect the image online;the improved Canny algorithm combined with the slope difference method is used to extract the slope difference feature of the image,and the tearing trend of the conveyor belt is judged early according to the slope difference;and the torn part is detected by the target detection algorithm based on YOLO(you only look once).The experimental results show that this method can not only early warn the tearing trend of conveyor belt,but also accurately detect the tearing position.The average detection speed is 35 fps,and the accuracy rate is 94.6%.In this paper,the deep learning method is combined with the traditional linear segment detection method to detect the tearing of large conveyor belt.Firstly,YOLO algorithm based on deep learning is used to detect the supporting shafts on both sides of the conveyor belt and then determine the tearing detection area;secondly,the improved LSD algorithm is used to detect the straight line segments on both sides of the conveyor belt,and the points on the same vertical coordinate on the straight line segments on both sides of the conveyor belt are selected as the bandwidth length.When the belt is torn,the belt width will obviously narrow.According to the detected bandwidth change,we can judge whether the conveyor belt is torn.The experimental results show that the detection method can accurately measure the belt width and judge whether the belt is torn.2.In this paper,a computer vision based method is proposed to detect the deviation of the conveyor belt.In this method,firstly,the conveyor belt region in the image is extracted and preprocessed by the example segmentation algorithm based on the convolution neural network;secondly,the edge contour information of the conveyor belt region is extracted by the improved Canny algorithm;finally,the edge straight line feature of the conveyor belt is extracted by the improved LSD straight line segment detection algorithm,and the relative position of the center line and the slope of the edge straight line is determined Check whether the broken conveyor belt deviates.The experimental results show that the method can detect the straight line of the conveyor belt edge accurately and judge the deviation,which verifies the effectiveness of the method.