The Research And Design Of CMOS Video Image Processing Algorithm Based On FPGA

Currently,In the field of video image processing,not only requires the video images which collect by the system to have high frame rate,high quality and fast transmission,but also requires the system to have high-speed and real-time characteristics when processing video images.Because the FPGA hardware has the excellent features of high-speed parallel processing of signals.Thus,this paper which makes use of the advantages of FPGA hardware parallel processing,high-speed acquisition of image data from CMOS image sensors,and USB high-speed data transmission capabilities to collect video images,preprocess video images,detect edges of video images,and extract features of video images,the transmission of video image data and the design of upper monitor on the PC are studied.An experimental system for processing of CMOS video images based on FPGA is developed.The main research work is as follows:1.Based on the parallel acceleration of FPGA hardware,Combining the characteristics of high-speed video capture of CMOS sensors,the advantages of USB real-time data transmission,and the corresponding system hardware platform is built.2.Aiming at the shortcomings of traditional software methods such as the edge detection and feature extraction processing speed of video images.Firstly,the Canny edge detection algorithm of CMOS video images is studied by virtue of the parallel acceleration of FPGA hardware,and then the feature extraction algorithm based on FAST is studied.Simulation and experimental results show that under the same conditions,using FPGA hardware method compared to traditional software methods,the quality of video image edge detection and feature extraction is not only not reduced,and the processing speed of the entire video image has been greatly improved.3.Aiming at the shortcomings of traditional software methods such as edge detection and feature extraction processing of video images,the Canny-based edge detection and FAST-based feature extraction are performed on CMOS video images by virtue of the parallel acceleration of FPGA hardware.The experimental results show that under the same conditions,using FPGA hardware method compared to traditional software methods,the quality of video image edge detection and feature extraction is not only not reduced,and the overall video image processing speed has greatly improved.4.In order to be able to conveniently observe and collect the processing results of FPGA-based CMOS video images.The USB logic function under the main control of FPGA,the firmware program under the USB interface and the design of the USB device driver under the PC platform are studied,the upper moniter software based on MFC was written.The software can capture and save the currently displayed image data.5.Finally,video image acquisition and transmission tests,video image edge detection tests,and video image feature extraction tests are performed on the entire system.The whole test shows that the system not only can perform real-time image acquisition,but also can perform real-time image edge detection and image real-time feature extraction.It has good practical prospect.