The Design Of Face Recognition Access Control System Based On Davinci Platform

With the rapid development of science and technology,intelligent security systems have been widely used in many areas of society.As an important form of it,the access control system is also increasing people's attention.The traditional access control system,such as passwords,card access control system,has deficiencies in terms of convenience,user-friendly and safety.In recent years,with the continuous development of biometric technology and embedded technology,an access control system technology associated with them has become a hot research direction.An access control system based on face recognition and Davinci platform is presented in this thesis.Firstly,the overall design of the access control system is introduced in this thesis.And the applied technology related to it is researched in detail.The hardware configuration and functions related to the realization of Davinci platform are expounded exhaustively.Then software architecture of the core chip DM6446 and method of algorithm packaging are studied deeply.Secondly,the face detection algorithm based on Adaboost is analyzed in depth,and it is applied in the access control system.The experimental results show that the algorithm can adapt to Davinci platform very well,and have a high detection rate and speed.In this thesis,some common pre-processing method for face image are studied and analyzed.A weighted average method is used to gray the detected face images,the eyes' location will be got in the image by using the Adaboost algorithm.Images are rotated,cropped and scaled according to the coordinates of the eyes' center points.And the histogram equalization and median filtering methods are used for further processing to the scaled image.The experimental results show that after the preprocessing we can get the normailized face images without noises.Thirdly,the face recognition algorithm is studied and analyzed in this thesis.On the basis of deeply studying the LRC algorithm,we improve it by using the singular value decomposition method to solve the self-correlation problem between each sample in the sample training stage.The experimental results show that comparing with the traditional LRC algorithm,the recognition rate of the ameliorated algorithm is significantly improved.Lastly,QT is installed in Ubuntu operating system.And then transplant it to the Davinci platform.The communication between DM6446 and F28027 is achieved by I2 C agreement in Davinci platform.User interface is written in QT programming environment,and running on the platform to realize the real time communication.The access control system program is written in QT programming environment,and running on the platform to realize video capture real time display,user registration,user identification,F28027 output frequency control and returning to results in real time.