Research Of The Main Light Extraction Methods Of Palmprint Images In The Natural Environment

With the development of science and technology, biometric technology in all spheres of life more widely used. Biometric authentication technology compared to traditional terms, a higher uniqueness, stability, universality and other advantages. Palmprint recognition technology and biometric identification technology is an emerging field of technology. Because each person’s unique palm palm, fixity, stability, universality, can be collected. But also has abundant palm print information and other features, which makes palmprint recognition technology can be generally accepted. Palmprint recognition technology has become a major research focus in the field of biometrics. The main line is rich in palm palm print information, compared to the mastoid palm pattern, triangulation point, wrinkles and other information palmprint mainline more stable, easier to extract features. So the main palm authentication can be extracted as a feature to identify the palmprint recognition technology in personal information.The vast majority of existing palmprint recognition technolog are under a particular light background, fixed position palm contact or non-contact method to collect palm prints.The paper is to study under the natural light environment. the palm is not fixed position non-contact method to collect palmprint image Line extraction. Specific research results are as follows:(1)Extraction of the palm region of interest. Improved seeds palmprint image positioning algorithm to obtain an image containing the finger area. Then use Sobel operator for image filtering and then locate and merging both targeting methods. Then calculated according to the squareness of the image geometric features and aspect ratio value to determine whether there is a finger into the region in the region to be extracted. And then by analyzing the image of horizontal and vertical integration interception palmprint gray area of interest.(2)The preprocess of palmprint image, including grayscale method, with one method based on the horizontal and vertical integration of local gray gray automatically corrects the brightness to make the image more evenly distributed, median filtering and sharpening Sobel filtering operation to enhance the difference palm edge ridge, class discrimination binary method. Finally useing gray levels based on the integral value of the image denoising.(3)Extraction of palmprint mainline.First, using the principle of median filtering the binary image de-noising. Then connect the main line of fracture phenomena using mathematical morphology operation. Next refining the main line, and finally connecting domain denoising extracted three main effective.