摘要
针对传统人脸识别技术识别率不高,稳定性不强等缺点,本文以Smart210开发板为核心,提出一种嵌入式人脸识别系统。通过采用AdaBoost算法进行图像人脸检测,将二维小波变换法与PCA降维算法相结合,以提取图像中的人脸特征。通过PC机训练人脸图像样本,获得特征子空间文件后发送到嵌入式系统内,这种方式可以有效解决图像处理数据巨大而造成处理效率低的问题。最后,通过对本文设计的嵌入式人脸识别系统的识别率和运行速度等性能指标进行测试。测试结果表明:该系统功能完备、界面友好、使用方便、识别效果良好。
In view of the shortcomings of traditional face recognition technology,such as low recognition rate and poor stability,this paper proposes an embedded face recognition system based on smart210 development board.By using AdaBoost algorithm to detect the face of the image,the two-dimensional wavelet transform and PCA dimension reduction algorithm are combined to extract the face features in the image.The face image samples are trained by PC,the feature subspace files are obtained and sent to the embedded system.This method can effectively solve the problem that the image processing data is huge and the processing efficiency is low.Finally,the performance indexes of the embedded face recognition system designed in this paper,such as recognition rate and running speed,are tested.The test results show that the system has complete functions,friendly interface,convenient use and good recognition effect.
作者
张婷
ZHANG Ting(Department of Information Technology,Anhui Grain Engineering Vocational College,Hefei 230011,Anhui,China)
出处
《贵阳学院学报(自然科学版)》
2020年第1期82-86,共5页
Journal of Guiyang University:Natural Sciences
关键词
嵌入式操作系统
人脸识别
二维小波变换
PCA降维算法
embedded operating system
face recognition
2D wavelet transform
PCA dimension reduction algorithm