Research On Belt Speed Detection Of Belt Conveyor Based On Machine Vision

Belt conveyor is a kind of continuous transportation equipment in modern production.It has the advantages of long transportation distance,low freight,large volume,convenient loading and unloading,low energy consumption and suitable for bulk transportation,etc.It has been widely used in coal,mines,ports,electric power,metallurgy,chemical industry and other fields.It has become the three main industrial transportation vehicles together with automobiles and trains.In order to ensure the safe operation of the belt conveyor,monitoring of the belt conveyor's slipping,overspeed and other operating conditions,as well as speed control.The belt speed of the belt conveyor needs to be detected.At present,the belt speed detection method of the belt conveyor mainly uses the contact type belt speed detection method of the linear sensor.This method uses the circumference of the roller to calculate the belt speed.However,when the surface of the roller and the conveyor belt adheres to the material,the surface circumference of the roller increases.The accuracy of detection is poor.The roller is in contact with the conveyor belt for a long time,which can easily cause damage to the roller,scratch the conveyor belt,and make reliability worse.This thesis proposes a belt speed detection method of belt conveyor based on machine vision.This method is as follows.Serial images of the surface of the conveyor belt collected by camera in the same time interval are transmitted to the belt speed detector.The belt speed detector uses the belt speed detection algorithm to calculate the speed of belt,and then it sends the calculation result to the host computer.The belt speed detector uses non-contact detection of the belt speed,so it is highly accurate and reliable.A belt speed detection algorithm based on the SURF-ORB image matching algorithm is proposed.Firstly,this algorithm uses SURF-ORB image matching algorithm to achieve feature extraction,feature description and feature matching of every surface image of conveyor belt.What's more,in order to achieve accurate matching,the RANSAC algorithm is used to eliminate mismatched points.Through obtaining the pixel coordinates of the matched point pair,the relative pixel displacements of two adjacent frames are calculated.The mapping relationship between the image coordinate system and the physical coordinate system calculates the actual displacement and belt speed of the conveyor belt.The method of dividing the interest region of the image is used to reduce the matching search range,which improves the computing speed of belt speed detection.A belt speed detector is developed,its hardware and software are designed,and an experimental platform is set up for the belt speed detection system of belt conveyor based on machine vision,and conducted experiments.The result shows that the belt speed detection method proposed in this thesis can achieve the belt speed detection,and it's detection accuracy and reliability are high.The belt speed detection method proposed in this thesis and the developed belt speed detector can be applied to the belt conveyor monitoring system,it can achieve belt speed detection,monitoring of the belt conveyor's slippery,overspeed and other operating conditions,as well as speed control.The method is a non-contact detection method with high accuracy and reliability,and it has a wide application prospect in coal,mines,ports,electric power,metallurgy,chemical industry and other industries.