Research On Ore Particle Size Image Detection Technology

The research topic is derived from the actual production.The ore particle size detection equipment used in the field has the disadvantages of high cost,and the sample is broken again during the detection process,resulting in distortion of the detection result.The ore image processing algorithm is used to obtain the ore particle size distribution.The advantages and disadvantages of ore image processing filtering and segmentation of different algorithms are studied.An ore image segmentation algorithm based on genetic algorithm is proposed.Then the watershed algorithm is used to process the complete boundary of the ore particles.A staged treatment of ore particles is proposed.Image,a method of achieving ore particle segmentation.Test verification was carried out.The error between the test result and the image analysis result is less than 5%.The specific research process is as follows:Firstly,comparing the experimental results,the bilateral filtering method is used to denoise the ore image.Since the collected ore image is affected by multiple fac-tors,the image itself carries some noise,which will cause interference when extract-ing information from the image.Therefore,it is necessary to filter the collected ore particle image.This paper uses various filtering methods to image the ore.The pro-cessing is carried out.By comparing the experimental results,this paper selects the bilateral filtering to process the image.This method can not only remove the noise on the image,but also not destroy the information of the ore particles in the image.Secondly,the bilaterally filtered image is processed by a threshold segmentation algorithm.Different from the general threshold segmentation algorithm,this paper introduces the genetic algorithm in the commonly used threshold segmentation algo-rithm.The basic idea of using the genetic algorithm to segment the ore image to be processed in this paper is to divide the pixels of the image into two types of images according to the gray value of the image by using the threshold value M.One type is the foreground image,that is,the ore particle image G₁ to be analyzed.The other type is the background image G₂.The image of the ore particle in the foreground is composed of pixels whose gray value is between 0 and M,and the image G₂ is composed of pixels whose gray value is between M+1 and L-1（L is the gray level of the image）.Since the unevenness of the surface of the ore particles causes uneven distribution of the gray scale of the surface,the image obtained by threshold division is binarized.The binarized image is segmented using a watershed algorithm.The watershed algorithm can obtain continuous and complete boundary information,which can facilitate the next step of marking.Thirdly,compare with the test data to verify the accuracy of the image pro-cessing detection granularity.On the basis of the segmentation process,the relevant calibration theory is used to calibrate the se gmented ore particle region,and the ore particle size is obtained through correlation operation.The data is distributed,and then the image processing data is compared with the artificial screening test data,and the maximum error of the two is 5%,and the accuracy of the image proce ssing detection granularity is verified.Finally,a real-time detection system for ore particle size images is designed.According to the characteristics of the processing flow of ore particle image,the ore particle size detection system is designed,and the related user operation interface for ore image processing is designed,and the functions contained in the user operation interface are explained.