基于面部识别的新型智能门禁系统设计与实现

Design and implementation of a new intelligent access control system based on facial recognition

数据采集利用OpenMV等信息采集模块对人脸特征信息进行提取、收集,从而建立系统图像库。利用Facenet算法计算面部特征向量,该算法用于处理相同面部下不同角度、姿态形成的差异。具体通过计算特征向量L获取一个阈值,从而实现了面部检测的量化计算,提高了面部检测算法MTCNN的准确度,智能门禁系统启动时采集到的面部特征信息进行特征向量计算,计算结果将作为门禁系统是否解锁的依据,该门禁系统在传统生物识别的基础上,运用算法模型提高了安全性、准确性。

Data acquisition is to use OpenMV and other information acquisition modules to extract and collect face image information and face feature information to establish the system image library.At the same time,the data comparison module is activated to enter the current information into the image library for comparison and feature comparison pretreatment.This process will save a lot of image processing time and sacrifice less accuracy.Facenet algorithm was used to obtain the feature vector of facial information.The algorithm is used to deal with the differences of different angles and postures under the same face.In particular,a threshold value is obtained by calculating the feature vector L,so as to realize the quantitative calculation of face detection and enhance the accuracy of face detection algorithm MTCNN.