The Design Of The Distributed Attendance System With Fingerprint Identification

In recent years, attendance system has got great attention and comprehensive application.More and more enterprises, schools etc, began to record attendance of employees, students'attendance. But the ordinary attendance system is not satisfactory, because it exists some shortcomings. It cannot very well for fairness and accuracy.It broke the original intention of checking attendance system. Therefore, this paper explores a distributed attendance system with fingerprint identification instead of the ordinary attendance system. It avoided the problems of integration. More and more fingerprint identification algorithms used in the attendance system. It adopted the most convenient,the quickest and novel algorithm to improve the speed of fingerprint identification.Thereby, the system perfected the fingerprint indentification attendance system.This paper mainly completed four parts design: the hardware part design of the system, the algorithm design of the fingerprint identification system ,the design of distributed database system and the design of the management system with fingerprint identification. Among them, the thesis is focused on: the algorithm design of the fingerprint identification system.The thesis accomplished the design of the fingerprint identification system in Visual c + + 6.0 development platform ,including the preprocessing of the fingerprint image, the feature extraction of the fingerprint image and the matching algorithms of the fingerprint image.Based on frequently-used many kinds of algorithms, according to advantages and disadvantages of the algorithms, comparing and selecting the optimal algorithm for programming, to realize the corresponding function. The innovation points of the thesis is using the fractal dimension in the fractal geometry - Box-counting dimension, came to the segmentation of the fingerprint image. In the image processing, the box-counting dimension as one of the fractal dimension reflects the regularity and complexity of the image block texture and grayscale distribution. Box-counting dimension is the gradient on unit area, it utilization the different gradient value of the foreground and background of the fingerprint image, namely the different box-counting dimension to distinguish the prospects and background of the fingerprint image, this method is simple and easy to implement.