Research On Methodology Of Hand-Dorsa Veinrecognition Under Non-contact Imaging Condition

With the economy and information technology developing rapidly, information has been closely related to people's lives, and the requirements of security, such as personal items, venue channels, e-commerce network,are strict. At the same time the damand of individual identity authencation is also becoming more and more urgent. Biometrics, which is a safe and convenient identification method, can satisfy people's demands. Hand-dorsal vein recognition method is an important means of them, whether as a major identification means, or combining with other identification methods, has a very large studying potentiality and practical significance.Hand vein recognition uses the characteristics that red blood cells absorb in the near infrared of specific wavelengths to read vein information, then through a series of image processing technology analyzes the characteristic of hand-dorsal vein for identification method. This paper analyzes the theoretical basis of hand-dorsal vein recognition and research status in and abroad, and compares it with other biometric on advantages and disadvantages. Then the pretreatment, segmentation, refinement, feature extration and match of hand-dorsal vein recognition algorithms are proposed. At last, improving algorithm relevant to each stage of processing has been put forward.Vein image preprocessing contains design,authentication,normalized gray,and image smoothing of hand-dorsal vein image size, normalized gray image and smooth algorithm,and then compares the segmentation effect of different methods,which contain the threshold value method,niblack algorithm and the direction of the valley shape measuring. And it proposes a vein segmentation method based on the 1-d-gray level, then refines skeleton image after segmentation and removing the noise to meet the requirements of hand-dorsal vein image.Then, feature extraction and matching are listed in the paper, introducing unchanged from the recognition method, recognition method based on vein structures, and a vein recognition method aiming at the characteristics of venous lines. The recognition method is carried out in the 100 personal sample databases, obtaining the 99% recognition rate, which is verified to be an effective hand vein recognition method.Finally, this paper puts forward expect and prospect about vein recognition's research work.biometrics recognition; hand vein recognition; pretreatment; vein segmentation; vein recognition...