Research On Temperature Anomaly Detection And Early Warning Of Ore Conveyor Roller Based On Infrared Image Recognition

With the implementation of the intelligent and intelligent construction of green mines and mines in our country,the monitoring and maintenance of all kinds of production equipment in open-pit mining areas are becoming more and more intelligent,and the ore conveyor belt is an indispensable transport equipment for transporting ore.the daily monitoring of its related components is one of the keys to ensure the safety of production in open-pit mining areas.The anomaly detection method based on image recognition has gradually become one of the main means of temperature anomaly detection of ore conveyor roller because of its high fault recognition rate,fast recognition speed,no contact and high safety.This paper makes an in-depth study on the temperature anomaly detection of ore conveyor roller in open-pit mine based on depth learning and infrared image.The main contents of this paper include:(1)Research on infrared image preprocessing and data set construction of ore conveyor roller.Through on-site collection and manual labeling,the recognition degree and related scene information of the idler are increased,and the state of the idler is further subdivided.At the same time,in order to solve the problem of uneven distribution of data samples in the training process,the traditional data amplification method and mosaic data enhancement method are used to expand the data set to further enhance the data characteristics of roller and background information.(2)Research on temperature anomaly detection model of ore conveyor roller based on machine vision.In order to solve the problem that the daily monitoring of roller in open-pit mining area is not intelligent and the current manual inspection method is difficult to meet the needs of mine intelligent production,a temperature anomaly detection model of ore conveyor roller is constructed.In the feature extraction module,the detection model uses the improved GhostNet backbone feature extraction network to reduce the cost of feature extraction.In the feature fusion module,the SPP-Net module is used to optimize the PaNet feature fusion network,increase the receptive field of the model,simplify the model structure through deep separable convolution blocks,reduce the amount of calculation and the number of parameters of the model,and improve the learning ability of the model through LeakyReLU activation function.As a result,the real-time and accurate detection of conveyor belt idler in mining area is realized.(3)the early warning model of abnormal temperature of ore conveyor roller based on infrared image.In view of the low accuracy of edge data recognition by the ore conveyor roller detection model,the temperature of the infrared image of the roller is calibrated,and the early warning model of abnormal temperature of the roller is constructed.Based on the infrared image of the idler,the model calculates the temperature of the idler through the three-channel value of the image,which is used to judge the temperature state of the idler,which provides reliable temperature data for the daily monitoring of the ore conveyor roller in the open-pit mining area.the accurate measurement of the temperature of the ore conveyor belt idler and the daily detection of the state of the idler are realized.Through the analysis of the experimental results,it can be seen that the detection method of the ore conveyor roller proposed in this paper can better locate the position of the idler,and the abnormal early warning model can calibrate the temperature of the infrared image of the idler.Make a corresponding early warning to the idler with abnormal temperature.Finally,this method is applied to the experimental test,and the results show that this method can realize the real-time monitoring and early warning of the idler through the infrared video taken by the infrared camera.