Research On Mica Ore Pre-concentration System Based On Machine Vision

As one of the main raw materials of electrical equipment,the separation requirements of mica raw ore are relatively strict.At present,in industrial production,the separation methods of mica raw ore with large chip diameter are mostly manual separation,air separation and gravity separation.The traditional manual method is time-consuming,laborious and inefficient,while the air separation and gravity separation methods can not effectively identify the internal defects of mica sheet.Therefore,this paper designs and implements a mica ore pre-concentration system based on machine vision technology.Firstly,according to the analysis of mica characteristics and industry status,the selection basis and indicators are determined with reference to the principle of color selection,and a research platform is built to select appropriate auxiliary devices to form the overall framework.Then,combined with the actual situation,the hardware circuit and the processing program of basic image data are designed to make the data flow and image display normal,which is convenient for subsequent image processing.In addition,the sorting mechanism is studied,and the hardware program is compiled according to the communication protocol to separate pure high-quality mica flakes(concentrate)and defective mica(tailings).Combined with the above work,the design of dispersion and transmission mechanism is completed,and the mica image processing algorithm is mainly discussed.A variety of algorithms are compared from three aspects: image pre-processing,contour recognition and positioning and defect detection.Finally,the processing flow of gray processing,bilateral filtering,binarization,morphological operation and edge detection is determined.The image processing results can accurately identify the internal stain defects of mica and output the position information of mica.After completing the overall design,1000 mica raw material samples were sorted and tested.The final screening efficiency of defective mica(tailings)was 87%,and the screening efficiency of high-quality mica(concentrate)was 81.78%.Through the analysis of the separation experimental results,the system designed in this paper can basically meet the needs of mica raw ore pre-concentration,and the operation is simple.It replaces the process of manual visual recognition and manual separation,saves manpower and increases the automation degree of mica pre-concentration.