Study On The Key Technology Of Long-distance Incoordinate Iris Acquisition Device

The realization of iris acquisition by automatic intelligent devices will greatlypromote the popularity of iris recognition. And the proportion of iris recognition devicesin monitor and other service market will consequently scales up. However, most of theexisting iris capture devices are developed in the cooperate type, which means that, tocapture the desired recognition image, the workers must operate with a certain gestureand at a certain distance. Undoubtedly, the promotion of iris recognition equipmentswill be greatly constrained. Then, a more functional and convenient iris capture deviceis needed in the market.In this paper binocular visual positioning, console control, face recognition, andautomatic zoom algorithms are designed, according to which, a kind of long-distancenoncooperative iris acquisition system is developed based on VC platform. Firstly,using a mature algorithm to calibrate the camera, and image registration is realized byan improved SUSAN corner point achieve arithmetic. Then binocular visual positioningis realized by the principles of affine transformation. Secondly, face detection is realizedby the AdaBoost and some new algorithms, which include particle filter based on facecolor model and ellipse restriction, and a pattern matching based on the image Humoment used to distinguish users. Thirdly, auto zoom is realized by a improvedhill-climbing algorithm and a multi-direction acquire edge information algorithm, andbinocular visual positioning help to accelerate automatic zoom. Finally, to achieve thefunctions of database management, user interface, substep debugging, the applicationprogram framework based on multi-thread technology is developed by VC.Following are the main contributions of this dissertation. First of all, the systemrealized the remote acquisition of iris image. Besides, this system realized the gatheringof iris image without users’ active cooperation. Furthermore, the system has strongapplicability for it could achieve user’s iris image or infrared iris image under thecondition of natural light. At last, the system could coordinate with iris recognitionsystem for authentication...