Research On Identification Algorithm Based On Knuckle Prints And Palm Prints

With the development of science and technology,people pay more and more attention to secure identity authentication methods.The traditional identity recognition has gradually been unable to meet people’s needs.Therefore,the biometric recognition technology based on fingerprints and faces have begun to emerge.This kind of single-mode biometric recognition method is safer and more convenient than traditional recognition methods.However,single-modal identity recognition still has the problems of unstable authentication and low security.This article fully considers the advantages that the knuckle pattern has the characteristics of simple and fast recognition speed,the palm print has the characteristics of rich features and high security,and the convenience of capturing images with non-contact devices.The two kinds of biometrics are merged to make up for the shortcomings of single-modal recognition.On the premise of fast speed,the matching accuracy is improved,and the security and stability of recognition are improved.The main work of this paper is as follows:(1)The advantages and disadvantages of non-contact acquisition are analyzed.In the image preprocessing,aiming at the problems of complex background and different light effects in the acquisition of non-contact equipment,the H-channel based on HSV space is used to detect the skin color of the hand to remove the influence of complex background,V-channel based on HSV space extracts the illumination component,and the illumination adaptive correction algorithm is designed to remove the influence of illumination and enrich the application scenarios of non-contact acquisition equipment.(2)In order to improve the effect of knuckleprint recognition,a knuckleprint feature extraction algorithm based on Hessian matrix is proposed.Aiming at the problem of low recognition rate caused by weak knuckleprint features extracted by Hessian matrix,an edge enhancement algorithm is designed to enhance the knuckleprint features extracted by Hessian matrix to improve the effect of knuckleprint feature extraction.Aiming at the problem that the knuckleprint feature matching is susceptible to noise interference,the Hausdorff knuckleprint feature matching algorithm is improved based on the average distance to improve the accuracy of fingerprint matching.(3)In order to improve the accuracy of identity recognition,based on the above research,the palmprint with rich features are introduced,and the decision fusion algorithm of knuckle and palmprint is designed.Aiming at the problem that MB-LBP algorithm can only describe palmprint texture features and cannot describe the change of palmprint edge with palm extension,a palmprint feature extraction algorithm with texture features and edge features is designed based on HOG operator to improve the richness of palmprint features.A palmprint feature matching algorithm based on SVM-KNN cascade classifier is designed to improve the real-time of matching.