摘要
鉴于煤矿安全事故时有发生,利用签到系统准确掌握井下人员出入情况,对煤矿安全生产与救援有着重要的意义。将基于人脸识别的签到系统用于煤矿,遇到光线昏暗、人脸易附着黑色粉尘等因素影响,传统的基于PCA(PrincipalComponent Analysis)的人脸识别算法检测率低。为了解决该问题,论文提出了一种基于KL变换(Karhunen-Loeve Trans-form)和TAN分类器(Tree-Augmented Naive Bayesian network)相结合的人脸识别方法。该算法通过KL变换使特征点更突出,通过TAN分类器使匹配结果更准确。仿真研究结果表明:该算法既减小了计算复杂度,又提高了人脸识别率。
The coalmine accident happens sometimes. In order to be convenient to rescue,it＇s significance to know the accurate number of the miners in coalmine or outside. When the traditional face recognition system was used in coal mine, the system meets new problems, such as black, hazy face etc. The detection rate based on PCA （Principal Component Analysis） of traditional face recognition algorithm is low. Aiming at this issue,put forward a face recognition algorithm based on the combination of KL transform （Karhunen-Loeve Trans- form） and TAN classifier （Tree-Augmented Naive Bayesian network）. The algorithm through the KL transform makes feature point more outstanding, through the TAN classifier makes matching result more accurate. Simulation shows that this algorithm not only reduces the computational complexity, but also improves the human face recognition rate.
出处
《计算机技术与发展》
2012年第7期171-173,共3页
Computer Technology and Development
基金
"211工程"三期创新人才培养计划建设项目(S-09108)
