Design Of Real-time Image Acquisition And Preprocessing System Based On FPGA

The automobile safety problem has been concerned by people all the time.The automobile collision avoidance system based on video can analyze the surrounding environment of the vehicle to judge the possible dangerous situation,and then give a timely warning,so as to improve the safety of driving.However,because of the huge amount of video image data,the traditional serial processor uses software to process the image,which makes the system’s real-time performance difficult to meet the actual requirements.Therefore,a real-time image acquisition and preprocessing system based on FPGA is designed in this paper.As the front end of the vehicle anti-collision system,the image preprocessing is moved to FPGA,which can effectively improve the real-time performance of the whole system.The hardware architecture of the system built in this paper includes image acquisition by OV5640 image sensor,image data cache by SDRAM,image preprocessing by FPGA and LCD display display.On this basis,the image acquisition and display functions are completed at first.Then,from the perspective of real time,the image processing flow is determined to be image grayscale,median filtering algorithm for image filtering and Sobel operator for edge detection,so as to extract the edge contour of the object in the image more quickly.In order to improve the speed of image processing,the original serial image algorithm is processed in parallel by pipelining design.Finally,after the circuit design is completed on FPGA,it is built into a complete system together with OV5640,SDRAM and LCD.The simulation software Model Sim and logic analyzer Signal Tap are used for testing and analysis.The experimental results show that for a frame of 800×480 resolution image pretreatment,the processing time of the system is about 8ms,which has a great improvement in real-time,and has a good processing effect,in line with the design expectation.