Design And Implementation Of Coal Gangue Linear Array Imaging And Sorting System

With the advancement of optical imaging technology and computing technology,the price of imaging and computing equipment has been greatly reduced,and machine vision technology has been greatly improved and widely used.In this paper,the online sorting of coal and gangue is realized by using optical imaging technology and image processing technology,and an online image acquisition and processing system for coal and gangue is designed.The research of this subject can improve the automation level of coal gangue sorting,improve the working environment of mine workers,reduce the consumption of water resources,improve the quality of clean coal,reduce air pollution and it has certain application value.The main components of the coal and gangue imaging and sorting system are divided into three parts:image acquisition,image transmission and image processing.The research mainly includes optical circuit,CCD timing driving circuit,signal conversion circuit,image buffer transmission circuit,and DSP image processing and so on.The linear array CCD is used as a photosensitive element,FPGA is used as the master chip to complete the imaging part of the timing drive,dual-RAM achieves seamless image transmission by ping-pang operation,DSP is used to achieve image processing,and they play respective advantages to complete the overall system functions.On the basis of clarifying the overall system architecture,the hardware circuit such as CCD image acquisition circuit,emitter follower circuit,analog-to-digital conversion circuit,FPGA minimum system circuit,SRAM ping-pang buffer circuit and DSP floor circuit and the whole system power supply circuit are designed.The Verilog HDL language is used to write linear array CCD timing driver,AD timing driver,AD configuration program and SRAM storage program.The C language is used to implement image processing procedures and classifier program.The C++language is used to write the host computer image recognition results display program.The internal structure and imaging principle of CCD are analyzed,and the dual-sampling processing method of CCD timing and output signal is studied emphatically.A two-piece SRAM ping-pang cached image transmission circuit is designed,which utilizes two SRAM time-sharing operations to realize real-time seamless transmission of coal and gangue images between FPGA and DSP.Commonly the using of coal and waste rock image recognition is based on its gray value and gray variance recognition,this paper adds texture features.Firstly,according to the difference of coal and gangue image,the gray level co-occurrence matrix is used to calculate the entropy,the second moment energy value,the contrast and the autocorrelation texture.Finally,fisher classifier trained by sample data is used to complete the coal and gangue final identification.According to the image de-noising,segmentation and other image processing algorithms and classifier algorithms,the program working in the DSP is written.