Research On Application Of Palm Vein Recognition Technology In The Valuables Logistics

Obviously, the valuables were lost, damaged and impersonator in logistics, whicheasily caused huge property losses and compensation disputes.Traditional authenticationmethods such as ID, text messaging, phone numbers, job number, user name, passwordswere easily lost, forgotten, copied and at theft risks, so these methods had been unableto meet the needs of valuables logistics. However, palm vein characteristics have uniqueadvantages such as stability, uniqueness, rich information, difficulty to be copied andstolen, gathering methods are easily accepted. Therefore, this paper applies palm veinrecognition technology, a new authentication method, to supervise and manage clients,couriers, the warehouse administrators in the valuables logistics. When the valuableswere lost, damaged and impersonator, it guarantees responsibilities to individuals.This paper do experiments on self-built palm vein image database, mainly analysisand discuss around the subspace which is a kind of palm vein feature extractionalgorithm. We propose a method based on principal component analysis(PCA) andFISHER linear discriminant(FLD), FLD is the best classification feature extraction,Through the PCA dimension reduction method, we can overcome the difficulty whenusing the FLD method alone, the within class scatter matrix is singular. In addition, putforward the improvement method, in the recognition stage, PCA features and finallyextracted FLD features were fused by addition, this method achieved better recognitionresults.In order to extract nonlinear optimal discriminant features and address the problemof small sample size, this paper presents a new palm vein feature extraction methodbased on kernel principal component analysis(KPCA) and fisher linear discriminant(FLD). First with KPCA reduce the dimension of the image, and then use FLD toclassify features, finally calculate the Euclidean distance between feature vectors tomatch. The experiments results show that, compared with the traditional 2DFLD andPCA+FLD, improved PCA+FLD, this method can achieve a higher recognition rate of96%, the recognition time is shorter, run faster, which meets the application needs of thevaluables logistics.Finally, write a MATLAB GUI user interface and complete from obtaining palmvein image to the matching results. The simulation results show that, this palm veinrecognition system is safe and reliable. Apply it to valuables logistics, not only of greattheoretical significance, but also has broad application prospects.