On-line Detection And Analysis Of Ore Particle Size Based On Image Processing

The particle size information of broken ore has always been a key data index in mineral processing,and it is an important criterion for evaluating the quality of crushing.It is very important to optimize the process parameters of mineral processing equipment in beneficiation production and realize mineralization automation.The guiding significance,through the detection and analysis of the ore dressing ore size,can timely understand the current working state of the crusher,so as to adjust the size of the crusher discharge port in time according to the actual situation,improve the working efficiency and crushing precision of the crusher beneficiation.In the current traditional mineral processing,the detection of ore particle size is mainly achieved by artificial screening and sedimentation detection methods.These detection methods have long detection time,the distribution information of the feedback ore particle size is lagging,and the influence of human interference is large,leading to detection.The accuracy of the granularity is relatively low and the efficiency is not high.In view of the above reasons,this paper proposes a method of applying machine vision in particle size detection--online ore particle size detection based on image processing.This method introduces machine vision technology into the ore particle size detection,and uses industrial camera to the beneficiation workshop conveyor belt.The ore particle image on the realtime acquisition,real-time acquisition of the current ore particle size distribution information,and then feedback to the crusher in time to optimize the control of the size of the discharge port and other parameters,eliminating the error of manual detection,improving the efficiency and accuracy of the beneficiation Sex,reducing energy consumption.Aiming at the actual production situation of the Qi Dashan iron ore crushing workshop of Anshan Iron and Steel Group Mining Company and the field investigation of the site environment,an online detection method of ore particle size based on image processing technology was proposed.Firstly,the industrial camera is used to collect the ore image,the image noise is eliminated by the bilateral filtering algorithm and the binary image is used for distance transformation and reconstruction to perform seed extraction,which prepares for the watershed segmentation and reduces the phenomenon of over-segmentation.Secondly,the watershed segmentation algorithm based on morphological gradient reconstruction combined with Canny edge detection algorithm(threshold image segmentation algorithm)combines the segmentation edge line obtained by Canny edge detection algorithm with the segmentation line of improved watershed algorithm to achieve effective segmentation of ore particles..Then,using the hit and miss transformation method to count the number of particle size grades in the ore image,the calibration of the ore particles and the calculation of the ore particle size are realized,and the particle size distribution information of the broken ore on the conveyor belt is obtained.Finally,a BP neural network algorithm based on weight correction is proposed to predict the depth distribution of ore particle size,which provides a guarantee for further improving the accuracy of ore particle size detection.Through the above pretreatment analysis and on-site artificial screening comparison,the prediction accuracy meets the actual situation.The on-site demand has effectively improved the crushing capacity and achieved the production goal of multi-grinding and grinding.