Research On The Design Of Embedded Face Recognition System Based On ARM

With the rapid development of networking and informatization,the demand for security and convenience in various occasions increases.Face recognition is widely researched and applied in many fields because of its unique advantages.Especially,in the trend of mobile Internet,the popularity of various embedded intelligent devices has brought opportunities for the research and application of embedded face recognition system.Firstly,based on the TQ210 embedded development board,this paper builds a hardware and software platform.Based on the Linux operating system,the Qt library and the OpenCV library are compiled and transplanted for the host and target machine respectively,and the Qt integrated development environment is configured.Ultimately,a reasonable and effective system development platform is formed,simulating on the host machine and running on the target machine.Secondly,according to the logical order of face recognition system,the function modules of the face recognition system are researched and implemented,including the image acquisition module based on V4L2,the face detection module based on AdaBoost,the face image preprocessing module and the face recognition module based on LBP feature.Then,according to the requirement analysis of the embedded face recognition system,the general design for the system and the graphic user interface design are completed.And two key parts of the system,face detection and face recognition,are tested and analyzed by the designed test face database.Finally,from the correct recognition rate and the matching success rate of the embedded face recognition system,an optimization method of face recognition reliability based on LBP feature is proposed.On the one hand,the system searches for the optimal threshold and the minimum number of members' registered samples to ensure the correct recognition rate of the system.On the other hand,the matching success rate is improved continuously through the automatic incremental training of the system.This paper completed the embedded face recognition system based on ARM with relatively complete function and high reliability.It provides ideas for the design and development of the embedded face recognition system for the practical application.At the same time,it lays a good platform for the follow-up related research and application.