Face The Real-time Detection System Based On Fpga + Arm Research

With the rapid development of image processing technology, visual inspection technology asone of today’s hottest technical areas has been applied to the aerospace, industrial,transportation, home, communications and other fields. It has the merit of non-contact andnon-offensive which are unable to compare by other technologies. Since most of the currentimage detection technology developed on the pc platform has the disadvantages of bulkinessand poor portability, It is not suitable on some special occasions. The embedded systems withthe advantages of small size, low power consumption and low cost is just to make up for thedisadvantages of PC platform, Therefore, design an image detecting system based onembedded system has great practical value and broader development prospects.The paper designed a face detection system based on embedded system, and realized theimaging, processing, examination, result demonstration and other functions. Since the presentface image inspecting technology uses the Adaboost algorithm generally, The paper firstintroduced the Haar characteristics, the integral image, sorter as well as the cascade sorterAdaboost algorithm in detail, then using a combination FPGA+ARM as the hardwareplatform of system which significantly improved the processing speed of the system, andcompatibility, meanwhile, carried on the optimized processing of Image acquisition. Throughthe ping-pong operation to the buffer of image, the image acquisition efficiency and thetimeliness of image processing was greatly improved. By transplants the embedded Linuxoperation tradition which has the merits of open source, small size and scalable to the ARMembedded system, Strengthened the processing complex situation ability of system. Generally,the design of image processing software was difficult, High costs, heavy and longdevelopment cycles. But with the emergence of OpenCV, has made the significant contribution for the imagery processing technology widespread application. The OpenCV asan open source computer vision library developed by enthusiasts, Contains many imageprocessing and computer vision algorithms, effectively improved the develop efficiency, Thisarticle detailed elaborated transplant process of OpenCV in ARM, and designed OpenCVbased face detection procedures.Finally, the image acquisition hardware platform was build and the software procedure wasdesigned, the experimental results proved the correctness and stability of the system.