A New Access Control System Based On Multi-information Fusion

With the rapidly development of the information society,people's living standards continue to improve at the same time,the masses of their own safety awareness is also increasing.As an important part of the security system,the security performance of the access control system has become the focus of attention.Nowadays,digital intelligent access control system has been widely concerned,the wildly application including of password unlocking,identification card unlocking and so on,but these unlocking system contain some disadvantages: passwords are easy to steal and identity cards are easily to lost.And the biological characteristics of the human body is unique,if regard the human biological characteristics as the access control system to unlock the way,it can further enhance the security of the access control system and make up for the shortcomings of traditional access control system.In this paper,the design of the new access control system is the result of hand gesture recognition and face recognition through the information fusion,in order to complete the identification of the identity.For the gesture recognition system,this paper proposes a Mean Shift algorithm to improve the gesture tracking,and then proposed the gesture segmentation method that combining HSV and YCbCr color space,and compare the segmentation effect with the two color spaces using alone.then proposed that Combined with the Hu moment features,finger number and convexity and compactness in hand image for feature extraction.Finally,through the template matching algorithm,the gesture recognition.For the face recognition system,this paper adopts the method based on Haar feature for face detection,the face to locate eye position,then introduces the basic theory of LBP algorithm,and put forward the improved LBP descriptor to extract the facial features,finally,this paper introduces the application of PCA algorithm in face recognition.and proposed a face recognition algorithm based on improved LBP algorithm and PCA algorithm through the characteristics of PCA algorithm.This method of dimension reduction by PCA on the extraction of LBP descriptor to obtain effective face recognition.For the fusion of the two subsystems,considering the access control system,we select decision template matching method to fusion the gesture recognition system and the face recognition system for the judgment,and then according to the needs,choose the weighted voting method and rules for final decision fusion,so the output results to control the access control system.At present,there are many single biometric access control system,but the recognition rate is often unsatisfactory,through the integration human biometrics to complete the recognition,not only greatly improves the security of access control system,but also improve the recognition rate of the system and make the access control system more reliable and secure.In this paper,we have compare the recognition rate through the traditional recognition system with our method,that this system can get a better rate,the actual demonstration experiment in complex background,and the system also shows a variety of digital gesture better,the recognition rate is up to 94%.For the face recognition system who has proposed in this paper,through the experiments compare this method with other face recognition methods based on FERET and ORL face database,the experimental results show that the face recognition algorithm is proposed in this paper has better stability,shorter training time and higher recognition rate.Therefore,the gesture and face recognition system proposed in this paper can be applied to the design of access control system.