Design Of Video Acquisition And Processing System Based On FPGA And USB3.0

In view of the high requirements of video acquisition and processing system applied in intelligent monitoring and security,medical imaging and automatic driving,it is particularly important to design a real-time and accurate video acquisition and processing system for extracting edge information.With the continuous innovation and rapid development in the field of video acquisition and processing technology,more and more information is contained in the image,and the requirements for the accuracy of the information obtained from the image and the clarity of the image are also higher and higher,but the existing video acquisition and processing system can not simultaneously meet the real-time and accurate extraction of edge information needs.Therefore,this paper designs a video acquisition and processing system based on FPGA and USB3.0.The main work is as follows:(1)According to the video acquisition and processing system’s requirements on real-time performance and video image preprocessing hardware,determine the overall design of the whole system,analyze the performance and advantages of the current mainstream device models,select the devices that meet the needs of the system,and complete the circuit construction of each device.(2)The image processing algorithm is realized based on FPGA,and the advantages of using FPGA hardware to realize image processing algorithm compared with PC software are described.The image processing algorithms included in the study include gray image extraction,Sobel edge detection,median filtering,binary corrosion and expansion,Laplacian sharpening and histogram equalization algorithm.Among them,the design for Sobel edge detection removes the traditional DDR memory module to save the time of memory reading and reduce the overall system delay.For histogram balancing design,statistical results and cumulative results share one RAM,saving 30%of hardware resources.(3)Design and implement video acquisition system based on FPGA and USB3.0,including: logic programming based on FPGA,USB3.0 programming and PC application software design.The system can stably collect 1280*720@30fps video,data transmission bandwidth up to 240MB/s.(4)In the video acquisition system,a preprocessing module is designed to realize Sobel edge detection.According to the shortcomings of the system results,the preprocessing system is continuously optimized,and the median filter and Laplacian sharpening algorithm are added to filter out the fine noise,highlight the edge and details of the video image,reflect more edge information,and further improve the processing efficiency and accuracy of the video processing system.At the same time,the image processing algorithm is realized by the way of FPGA hardware architecture pipeline,which improves the real-time performance of the whole system and achieves the desired effect.In conclusion,this paper realizes image preprocessing and real-time video transmission based on the proposed hardware architecture,which can not only accurately extract the edge information of video images,but also ensure the real-time performance of the system,verifying the feasibility of the system,which has a wide range of application values in intelligent monitoring and security,medical imaging and automatic driving.