A Study On The Blurred Iris Images Pre-processing Algorithms

As a kind of biometric features,iris has great practical value in information security domain, since it has many desirable properties, such as unique, stable, non-intrusive, and so on, then caused scientific research institutions and enterprises at home and abroad interest, and spent a great deal of human, material and financial resources to research this identification technology. So far, this technology has been applied to the banks, airport security domain widely.This thesis summarizes iris recognition research status in domestic and foreign, analyzes the existing iris recognition system shortcomings, which points out direction in iris recognition system improvement. In addition, introduces the main steps of iris recognition system briefly. Through the analyses of the iris localization algorithms based on circular template, aiming at the shortcomings and insufficiency of existing localization algorithms, a novel iris localization algorithm based on point and line is used in this thesis. Experimental results demonstrate that the improved method can enhance the localization accuracy effectively both in clear or slightly defocused images compared with the original algorithm.According to influence that the defocused degeneration process to the iris image, this thesis mainly studies the application of spatial domain method, frequency domain method and 2D wavelet method on assessment the iris image clarity, and has evaluated by experiments the performance of evaluation functions being used to describe the corresponding relation between the image clarity and the eye shooting distance, and analyses comparatively the several the performance of clarity evaluation functions, which provides reference for selection of the clarity assessment function.The existing iris reader usually requires human eye imaging in smaller depth of filed, making many users adjust position repeatedly when acquisition the iris image, this condition is not conducive to the use of users. Aiming at this problem, this thesis, under the condition of the human eye deviating the focus of the camera, studies and establishes the corresponding relationship between the iris image clarity and the shooting distance between the eye and lens, and gives the maximal moving range of human eyes in the process of shooting under the premise of the clarity of iris image meeting the performance of iris recognition system, and develops the practical system in the basis of this theory. By giving distance prompt information guides users to shoot iris image in the correct position, which facilitates the use of users.