Intelligent access control system

Abstract

The invention belongs to the field of intelligent building access control systems, and specifically relates to an intelligent access control system for recognizing faces based on RGB image deep learning. The system includes a face recognition module, an ultrasonic module and an access control switch control module, the access control switch control module is electrically connected with the face recognition module, and the ultrasonic module is electrically connected with the face recognition module. The system solves the power imbalance problem in the face recognition 1:N scheme of access control systems, and improves the multi-view robustness in face recognition. The above scheme adopting the assistance of the ultrasonic module can carry out living body recognition, reduce the calculationamount of the access control system, improve the heat dissipation stability of the system and the heat aging resistance performance of a PCB, prolong the service life and greatly reduce the failure rate and the maintenance cost due to heat generation. A video streaming-based feature extraction scheme is used for various pose face models of faces, so an access control base library can be conveniently logged, and the recognition accuracy is high.

Description

technical field [0001] The invention belongs to the field of intelligent building access control systems, in particular to an intelligent access control system based on RGB image deep learning and face recognition. Background technique [0002] The traditional access control system mainly uses RFID and fingerprint verification; a small number of products use face recognition, but most of them are based on traditional algorithms, such as fisher, eigenface algorithm, etc., not deep learning algorithms; inevitably there are problems such as poor robustness and poor stability. [0003] The face recognition system based on RGB image deep learning is divided into two modes: 1:1 and 1:N in the field of access control. The 1:1 mode is to use non-contact methods such as RFID to obtain the user's ID information, and then use the ID corresponding to the ID information. The information is verified with the video stream information obtained on the spot; and the 1:N mode saves the step of...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a solution to the problem of unbalanced power consumption in the access control 1:N scheme, improve the problem of accuracy drop caused by viewing angle errors in face recognition, and assist living body detection intelligent access control system

Images