Fingerprint recognition uses physical sensors to obtain fingerprint images.

#Fingerprint recognition is to classify and compare the fingerprints of the identification objects for identification. As one of the biometric identification technologies, fingerprint recognition technology has gradually matured in the new century and entered the field of human production and life. (Recommended Learning: Web front -end video tutorial )

## This fingerprint mode

Fingerprint is formed by the end of the human finger by bumpy skin Patterns, fingerprints are already formed before human beings are born, and the shape of fingerprints will not change as the individual grows. It only changes to an obvious degree, and everyone's fingerprints are different, and can be well distinguished in many detailed descriptions. , there are three basic shapes of fingerprint patterns: whorl, arch and loop.

There are many feature points in fingerprints. Feature points provide confirmation information of fingerprint uniqueness. This is the basis for fingerprint identification. It is divided into overall features and local features. The overall features include core points (located in The progressive center of the fingerprint pattern), triangular point (located at the first bifurcation point or breakpoint starting from the core point, or at the convergence, isolated point, turning point of the two lines, or pointing to these singular points), the number of patterns ( The number of fingerprint lines);

Local features are the detailed features of fingerprints. The direction, curvature, and location of nodes at feature points are all important indicators for distinguishing different fingerprints.

The fingerprint recognition process is divided into two secondary processes and is divided into four parts. Two secondary processes are the fingerprinting and cross-checking processes.

The fingerprint recording process consists of four parts: fingerprint collection, fingerprint preprocessing, fingerprint inspection and fingerprint template collection.

The fingerprint comparison process also includes four parts: fingerprint collection, fingerprint preprocessing, fingerprint feature comparison and matching. In both processes, pre-processing of fingerprint images exists, but the values ​​of fingerprint images and the values ​​of fingerprint features seem to have the same name, but their intrinsic algorithms and properties are completely different.

In the process of introducing fingerprints, fingerprint images are obtained more frequently, while the algorithm of the single value extraction part pays more attention to the process of analyzing and obtaining some feature values.

The first step in fingerprint identification is to obtain fingerprint images. Currently, there are many ways to obtain fingerprint images, including optical fingerprint acquisition technology, capacitive sensor fingerprint acquisition, temperature sensing fingerprint acquisition technology, and ultrasonic Fingerprint collection technology and electromagnetic wave fingerprint collection technology perform preprocessing after obtaining the image, and realize preprocessing steps such as grayscale transformation, segmentation, equalization, enhancement, and refinement of the image.

First of all, the fingerprint must be separated from the entire pattern. The grayscale of the background image and the fingerprint distribution map are different, which determines the difference in intensity between the two. The fingerprint can be separated from the background using the concept of gradient. The picture is well separated;

Equalization is an important step in preprocessing. During extraction, the pixel distribution points in different areas of the fingerprint image obtained are different depending on the environment. Equalization is to combine different The regionally distributed pixels are averagely divided to obtain an image with balanced brightness distribution;

In order to facilitate feature extraction, the image after several steps of processing must be intelligently enhanced. Daugmann implemented the Gabor wavelet approximation method to make the fingerprint The lines of the image are clearer, that is, the white parts are whiter, the black parts are darker, and the edge distribution of the lines is smoother.

The above is the detailed content of Fingerprint recognition uses physical sensors to obtain fingerprint images.. For more information, please follow other related articles on the PHP Chinese website!