Design And Implementation Of Near-infrared Face Recognition System For Treasury

Treasury is an important place for banks to store cash.The security of treasury is significant for banks.In recent years,treasury always has a variety of security issues.It becomes a burning issue to establish an access control system of modernization and informatization.With the development of near-infrared imaging technology and face recognition technology,it is possible to apply the near-infrared face recognition technology to treasury access control.Face recognition has been a developed technology,after several years' development.However,most of its applications are based on visible light.It is sensitive to illumination angle and intensity.The result is that recognition efficiency will be reduced greatly,and photos and videos cannot be identified,and there could be easily happened that a wrong person to recognize his or her face.To the disadvantages of visible-light face recognition,some scholars put forward the near-infrared face recognition technology.Near-infrared face recognition has some distinguishing feature.For example,it cannot be affected by visible light and the imaging of infrared light is stable.The near-infrared face recognition can work normally in the environment of all black or direct sunlight.Some doors of treasury are outside and others are inside.These two illumination conditions are greatly different.Therefore,conventional visible-light face recognition is not a right way for treasury.Therefore,this paper chooses the near-infrared face recognition technology to design and implement treasury system.The main contents of this paper are as follows: First,the present situation in and abroad and various face recognition algorithms are analyzed.In addition,combined with the special field of treasury,the research fields and research contents are confirmed.Second,the near-infrared imaging technology,access control technology and face recognition algorithm is studied,and they are the foundation of system development.Third,according to the real situation of treasury,the system's requirements are analyzed and the system's functional requirements and non-functional requirements are confirmed.Moreover,the functional modules are divided.Then,based on the requirement analysis,the design and implementation of system is done,including system design,database design,all functional modules design and implementation,and the detailed implementation of face recognition algorithm—contourlet transform and SVM.Bes ides,every module is tested and analyzed.Finally,the paper is summarized and the next step of work is prospected.The main achievements of this paper are as follows: The face recognition algorithm based on the near-infrared imaging is implemented,including face detection and face acquisition,denoising and rotation of face image,geometric size normalization of face image,eigenvalue computation and eigenvalue comparison of face image.The business process of system meets the demands of treasury,including communication among devices,doors' interlock,two-people control and so on.This system configures many security measures to ensure the security of treasury,including communication encryption,hardware encryption and hardware binding,two-times authentication and other measures.The system designs two input methods of face registration and face recognition,including credit card and touch screen.Meanwhile,it designs two authentication methods including 1V1 and 1VN.These methods satisfy the individ ualized needs of different treasury.This paper describes the information management of access control,such as the personnel information and records of coming in or out of treasury.This can make administrators keep records and select historical records easily.After a series of tests,the system can meet the security needs and business needs of treasury,and it achieves the anticipated target.