At present, due to its long development time and rapid development speed in China, fingerprint identification has mature products in access control, attendance, safe, identity authentication and other aspects after years of market promotion and application. At the same time, due to its lower cost than other identification technologies, it is easier to promote and be accepted by users.



Development of fingerprint identification civil market



In the 1990s, fingerprint identification technology rose in China. At that time, its application was limited to criminal investigation. It can improve the detection rate of public security organs and save a lot of costs. However, fingerprint identification technology belongs to the use of "unknown subject" in criminal investigation, and "comparison" needs to be processed by mainframe, so the cost of price and time is relatively large. Therefore, fingerprint identification is often only used as an auxiliary means of criminal investigation, which is not widely used and the market share is very limited.



Through the continuous promotion and application of fingerprint identification technology in the market, fingerprint identification enterprises found that the real market of fingerprint identification technology should be the civil market. Because the civil fingerprint identification technology is "know the main body", the "comparison" speed is faster and the accuracy is high. With more and more electronic equipment, such as PC, ATM, access control system, etc., it is entering our daily life.



From the trend, the civil fingerprint identification technology will eventually replace the security and convenience of the identification code and password, to prevent unauthorized access. Because the civil application of fingerprint identification technology is easier to popularize than criminal investigation application, and the market capacity is larger, it has the foundation of large-scale promotion. In western countries, fingerprint identification technology has entered the stage of large-scale civil use. Los Angeles in the United States used the world's first fingerprint identification system for relief payments in 1990; SFNB, the world's first virtual bank, also implemented a security project based on fingerprint identification technology to enhance the security of transactions.



Because of its mature technology and low cost, fingerprint identification has begun to be used for civil use. In the future, the domestic biometrics market will be more than 10 billion yuan, among which security industry is one of the most important application fields, and the market space is very large.



Decomposition of fingerprint identification technology



As we all know, people's fingerprints have the characteristics of uniqueness and stability, that is, each person's fingerprint is unique, there is no same fingerprint between the two, and each person's fingerprint is quite fixed, and generally will not change with the change of age and health status, so people's identity can be identified by fingerprint. At present, there are many automatic fingerprint identification systems at home and abroad, but their structures are similar, including fingerprint collection and classification, fingerprint detail matching and fingerprint compression and storage.



1. Fingerprint collection and classification



This is the first step in the operation of AFIS. Through optical or CMOS fingerprint acquisition instrument, the living fingerprint image is input into the system, and the image is segmented. On the premise of keeping the useful fingerprint information basically complete, some redundant image information is cut out to produce a relatively small fingerprint. The image is enhanced to reduce noise, enhance the contrast of ridge and Valley, and improve the image quality. Then the features of the image are extracted and the direction array is generated. According to the ridge and valley flow direction of the fingerprint, the fingerprint can be divided into five categories or smaller categories, such as sharp arch, arch, left ring, right ring and vortex. The main purpose of fingerprint matching is to speed up the fingerprint matching process and reduce the capacity of fingerprint database.



2. Fingerprint collection



After the fingerprint identification equipment is connected normally, the fingerprint can be registered and entered. In the process of reader's fingerprint collection, the reader's fingerprint needs to be input twice. The fingerprint collected for the first time is compared with the fingerprint collected for the second time. If successful, the system will save it normally and add it to the fingerprint identification system.



If the fingerprint collected is unqualified, the system will give an audio prompt. If the identification is not qualified, the save function cannot be used and needs to be collected again. In order to control the quality of the collected reader fingerprint information before saving to the database. To judge the quality of collected fingerprints, the fingerprint information collected for the second time is compared with the fingerprint template collected for the first time in a single ratio of 1:1, so as to ensure the quality of fingerprint collection and avoid storing illegal and unqualified fingerprint information into the database.



3. Fingerprint detail matching



This is the core of automatic fingerprint identification system. Generally, biokey algorithm is used, which is a fast and accurate 1:1 and 1:n fingerprint identification algorithm. When using biokey for fingerprint identification, 2000 ~ 6000 fingerprints can be easily completed in the test under 1 ~ 5S without pre classification of names, pins, etc. all tests are carried out in Pentium iii900mhz128mb memory environment).



4. Fingerprint compression and storage



In order to save storage space, fingerprint image must be compressed. At present, JPEG, WSQ and EZW are the most common compression algorithms for fingerprint image data. The image compression algorithm based on adaptive scalar quantization and wavelet decomposition based on biokey WSQ is adopted. When the algorithm is used for fingerprint image compression, the information of fingerprint detail feature points can be kept as far as possible, and the effect of decompression on the accuracy of fingerprint detail feature extraction is small. When the fingerprint image is compressed by WSQ algorithm with large compression ratio, the location of minutiae points and related information of restored and decompressed fingerprint image are better, which has little influence on the fingerprint recognition to be carried out later. The WSQ compression ratio is 1:20 or 1:15, that is, a fingerprint image can be compressed to 6 ~ 10K, which is very suitable for this compression ratio