Study On Key Technologies Of Visual Sorting Of Potassium Feldspar

With the advancement of semiconductor manufacturing technology,the design technology for Field Programmable Gate Array(FPGA)devices has also made rapid progress.FPGA is different from the traditional processor architecture.It uses hardware description language to program.After a series of designs,it can realize the logic function similar to the underlying hardware circuit.In recent years,the production technology of image sensor also develops rapidly.Compared with CCD image sensor,CMOS image sensor has the advantages of low cost,low power consumption and fast speed.In this paper,a potassium feldspar recognition and sorting system based on FPGA chip is designed,which uses FPGA chip as the main control chip,CMOS image sensor as the image acquisition chip,and USB interface for serial communication,making use of the characteristics of strong flexibility and parallelism of FPGA.The research of this topic can realize the intelligent separation of potassium feldspar,liberate the concentrator from the harsh environment,reduce the loss of water resources,reduce air pollution,and has high economic and social value.The research content of this topic is divided into three parts:image acquisition,image transmission and image processing and recognition.Hardware system implementation is based on Altium Designer 15.0 software,mainly including power circuit,image sensor related circuit,FPGA related circuit,SDRAM image cache circuit,serial communication circuit,etc.,and then through Xilinx company's ISE14.7 software to write the corresponding hardware chip driver.Image processing and recognition mainly includes image preprocessing,image feature extraction and image sorting and recognition.Image preprocessing and image feature extraction are realized in FPGA by Verilog language,and image sorting and recognition algorithm is realized in upper computer.This paper mainly analyzes the hardware system of image acquisition,studies the timing drive of CMOS image sensor,uses SDRAM chip to cache image,and the recognition result is uploaded to the upper computer display through USB interface communication.According to the sorting algorithm of K-feldspar,the system extracts the color features and texture features of the image.Finally,using the image data to train the logistic regression model,the final sorting of K-feldspar is completed.