Research And Implement Based On FPGA In Infrared Image Preprocessing

With the rapid development of computer, sensor technology, more and more high-resolution visual images, image processing systems integration, reliability have become increasingly demanding. Because the image processing speed of data processing, real-time requirements increasingly high volume of image processing systems, price, portability requirements become more stringent, the traditional use of DSP and ASIC can not meet the needs of the preliminary work, image processing large amount of data transmission speed is high, require high processing speed, FPGA hardware parallel processing structure is very suitable for this operation.For infrared image preprocessing algorithm characteristics and requirements, this paper considers the FPGA logic resources, computing resources, and development tool support, selects the combination of the hardware system of the host computer, the USB 2.0 controller and FPGA architecture as the platform of the experiment and research, designs the research program of image pre-processing modules with the FPGA core.Firstly, we learn and research the basic algorithm of infrared image pre-processing, edge detection, compare the effect of a variety of edge detection algorithm, chose Sobel edge detection algorithm which is nor sensitive to noise for FPGA implementation as the content. Secondly, we combine with FPGA features, research FPGA digital systems design and implementation based on Verilog-HDL language and master the basic skills of a specific digital image processing circuit design through high-level hardware description language. Meanwhile we use Verilog-HDL language, program on FPGA, achieve Sobel edge detection of digital circuits processing module on the FPGA, and attain a good result. We research the communication between FPGA and the USB controller based on the structure and features of the USB controller, study the host computer communications-driven approach, program the USB controller firmware, the application and the relevant driven.The results of the paper experiment show that, the Sobel edge detection algorithm and image pre-processing system accomplished on FPGA not only improve image processing speed while still are able to smooth the noise and provide accurate edge information.