Design Of Machine Vision Image Acquisition System Based On FPGA

With the development of science and technology,machine vision technology has beenwidely used in military,transportation,aerospace,medical,monitoring,control and other fields.This design is mainly the use of machine vision technology to solve the problem of intelligent vehicle laying waterproof coating.Through the comparison of ARM,ASIC,DSP,FPGA.Considering the advantages of parallelism and assembly line of FPGA.A design scheme of FPGA based machine vision image acquisition system is proposed to obtain the position information of the car.This system uses the Cyclone series EP4CE10E22C8 N of Altera company as the control chip and each module is realized by Verilog HDL language.The whole system hardware module mainly includes: CMOS image acquisition module,SDRAM memory module,image display module,JTAG interface and power module.FPGA implementation of functional module is mainly include: CMOS image sensor configuration module,image acquisition module,image processing module,image storage module,image storage module,image display module and clock management module.And each functional module is simulated by Modelsim.The whole system work flow is to configure the OV7725 CMOS image sensor by FPGA and capture video image of video camera by image acquisition module,then the video image are processed by image processing module.The image processing module include median filter and edge detection.The processed video image is cached by SDRAM.Read the SDRAM data to calculate the image edge points of the average value of the pixel coordinates.The location information of the car is judged by the average value of the pixel coordinates.Finally,the VGA display window is designed to display the video image,the frame rate of the image and the pixel coordinate average of the edge points.Finally,the whole system is debugged,and the result shows that the system uses OV7725 camera image acquisition in 24 MHZ clock.The resolution of the image is 640×480,and the frame rate of image acquisition reaches 25 fps.And be able to accurately calculate the real-time position information of car movement.The system in driven by 100 MHZ clock,through the simulation of image,image acquisition rate of the system can reach 60 fps.In summary,the system performance is stable,basically meet the expected requirements.