Research On Face Recognition System Hardware Based On Arm9

Face Recognition technology covers image processing, pattern classification, machine vision and many other studies. With the rapid development of computer technology, the cost of automatic identification technology with high performance which is included in face recognition system has reduced to an acceptable level. As traditional face recognition system usually consists of giant or high performance computers and technical imaging equipments, they are often applied to huge recognition occasions with their good storage properties. Hitherto, in small occasions, like attendance management, home security system, phone unlocking, these computer won't be used because of their high costs, even if they will performance well. In this paper, a micro face recognition system is raised, and its hardware platform design process is introduced from three aspects: hardware design, hardware debugging and software design.First, hardware designing is presented. It is divided into four parts including power system design, data acquisition module design, classification and recognition module design and input and output module design, and a detailed description of the design process is also given in the paper.Second, system debugging process is presented. What are introduced firstly are about the ADS cross-compiler, H-jtag debug agent and the use of AXD debugging tools, and then the serial port debugging, TVP5147M1 chip initialization, and so on. And the results of these debuggings are also showed in the paper.Which is given finally is the system software design process which includes embedded Linux operating system migration, system main program designing and embedded WEB server building. In the Linux operating system migration section, the cross-compiling environment, U-boot transplantation, the kernel compiling and file system building are briefly described, and the driver development is detailedly presented behind. Acquisition program section includes main framework, TVP5147M1 chip initialization, LCD control subprogram, and the image storing program are introduced too. And finally an embedded WEB server is build for further use.A Samsung S3C2440A processor with high-performance ARM9 core is selected for the system, which ensures high efficiency. By the design, the system can capture a whole frame of image data which brings up the effective protection of data integrity. This hardware platform as a portable face recognition system can also be programed to adapt to the applications of various kinds of image acquisition occasions.