Research On Face Recognition Access Control System Based On Android System

With the rapid development of Android system and face recognition technology,intelligent access control system is more and more widely used in intelligent buildings,intelligent residential areas,offices and hotels.It is becoming an important part of security protection.In this paper,a face recognition access control system based on Android is designed,which combines face recognition technology with identity recognition technology on mobile phone.It realizes the user's face recognition authentication on Android mobile phone to get the two-dimensional code from the server,and uses the acquired two-dimensional code to open the door of smart access control terminal.In the aspect of face detection,this paper analyses the advantages and disadvantages of two face detection algorithms based on skin color segmentation and Adaboost.By combination the two algorithms it proposes a face detection algorithm based on multi-feature combination.First,the simple and fast features of the face detection algorithm based on skin color segmentation can be used to quickly lock the approximate face area in the image to be detected,and then to this.The region uses the Adaboost algorithm to extract the Haar feature to detect and determine the final face location.Experiments show that the multi-feature combined face detection algorithm greatly improves the efficiency of Adaboost algorithm.It can detect face in real time,and the accuracy of face detection is also greatly improved.In face recognition,considering that the traditional PCA algorithm requires the training samples to conform to the Gauss distribution,and the face images collected in reality often do not conform to the Gauss distribution because of the influence of illumination,expression and posture,which will lead to the low recognition rate of the samples.Therefore,this paper improves the traditional PCA algorithm.On the one hand,the image with the same attitude is partitioned into the same matrix,which makes the sample closer to the Gauss distribution.On the other hand,the traditional PCA algorithm is more vulnerable to the interference of illumination and affects the final recognition efficiency.It is found that in the PCA method,the first three largest principal components in the feature space are most affected by illumination.In this paper,the first three principal components are added with a coefficient whose weight is less than 1 to reduce the influence of illumination on recognition.Experiments show that the improved algorithm can effectively solve the above problems,and the recognition rate is significantly improved.The entrance guard system designed in this paper has the advantages of double authentication.On the one hand,users can authenticate face recognition on Android mobile phone.On the other hand,when the authentication is successful,the mobile phone will get the two-dimensional code sent by the server as the "soft key" to open the door.Dual validation greatly increases the security of access control system.This access control system provides a new design scheme for intelligent access control system.