The Application Of Embedded Face Recognition Technology In The Access Control System

Modern social security equipment is paid attention to by the people day by day, people are not satisfied the security defense system, and biometric identification technology has achieved great development, especially in recent years the development of embedded and facial recognition technology more quickly. Most current face recognition entrance guard system is implemented based on PC, it has the shortcoming of large volume, high costs, for use in some places. But with the development of embedded technology, the use of embedded technology to realize face recognition system become a reality. This article revolves around how to facial recognition technology in the entrance guard system realize the intelligent, miniaturization, launched for small office or home face recognition equipment application and design.This article first elaborated the research of embedded face recognition technology in the application of entrance guard system background and significance; Face recognition related theoretical basis is analyzed in detail, including Adaboost based face detection algorithm and the algorithm of face recognition based on PCA; To complete the construction of the embedded development platform, including selecting ARM processor as the core controller, development platform for Ubuntu12.04 host, establish a cross-compilation environment, etc.; Detailed analysis of this system is used in the hardware driver framework in Linux system, including the V4L2 driver, Framebuffer driver, Input subsystem and RTC driver framework, and transplant the USB camera driver; Then transplanted opencv-2.4.9 to open-source computer vision library into the ARM + Linux platforms, and on the basis of the complete application development of embedded face recognition system, realize the face detection, face training and the basic function such as face recognition, and according to the result of recognition to control electronic lock.Finally, test the basic function of the system in detail and the performance of the system, the test results show that in the case of small system running stability, faster recognition and higher recognition rate.