Design Of Real_Time Video Signal Processing System Based On FPGA

It has close relationship between the real-time of real-time video signal processing system and auto-focus technology. As the development of electronic technology and digital image processing technology, the direction of auto-focus is based on digital image processing technology. Auto-focus based on high frequency component was born. Video signal format conversion and implementation are also focused in present. This paper descript an integrated camera that completes the auto-focus function, and devise a system of real-time video signal processing, its content can be summarized as the follows:(1)Auto-focus is divided into two parts:focus adjustment and focus detection. Focus adjustment contains zoom tracking and focus searching. Zoom tracking:The focus motor rotates corresponding number of steps according to the zoom tracking curve when the zoom motor changes the zoom of the images. Focus searching:Focus automatically again according to the search algorithm after zoom tracking. The main work of focus detection is determining whether restart auto-focus and the condition of auto-focus. (2)Improves the gray scale differential algorithm and Increase the number of adjacent pixels comparison to increase the absolute value of focus evaluation function and improve the accuracy of the auto-focus.(3)Describe an improved auto-focus searching algorithm. Its main features can be described as follow:step is dominated by the reciprocal of the slope of the focusing evaluation function curve, and the best focus point is fixed when compare the focusing evaluation function value two times. This will not only raise the speed of auto-focus, avoid falling into local maximum, but also resist the failure of focus because of the quick change of the vidio signal.(4)Analysis and demonstrate a8051IPcore and design it based on FPGA. Its every module is realized by VHDL language and simulated by Modelsim。(5)Motor-driven and focus searching code are write by C language and compiled in Keil C environment. At last, we Generate.coe file and download it into805IIP core to complete the zoom tracking and focus searching.(6)Process video signals at the end of the system which includes color-space conversion, frame rate ascension and image magnification and realize them by VHDL language。(7)We get the hardware structure of the system after software description. Hardware structure contains the connection among the FPGA board, video signal acquisition board and lens.The whole video signal processing system is tested and verified on XUPV5-LX110T FPGA board. It has the function of focus when zoom finish and its execution time of auto-focus is less than1second. The resolution of the video signal is elevated from736*576up to1024*768.The frame of the video signal is elevated from12Hz up to60Hz.The format of the video signal is changed from8-bit YCbCr into24-bit RGB. This system can be expanded, such as the image compression encoding/decoding, local storage and Ethernet transport etc.