Design of an identity recognition system based on infinite radio frequency identification technology RFID
chrispeng [ad_1]
introduction
The future warfare is mainly information warfare. To adapt to this new combat modality, the identity of soldiers on the battlefield in the 21st century should also be digitized, invisible, and secured through information networks. Biometric identification technology is a science that uses the inherent physiological or behavioral characteristics of the human body to identify and authenticate. It is safer, more reliable and effective than traditional methods based on passwords and identification numbers, and is attracting more and more attention. And applied in various fields. Military ID cards, also known as /soldier cards 0 or /life cards 0, as early as the American Civil War, the U.S. military began to issue such small metal cards as identity markers carried by soldiers. In order to meet the needs of modern high-tech local warfare, using the characteristics of biometric identification technology, based on infinite radio frequency identification technology (Radio Frequency Ident i ficati on, referred to as RFID), human body information such as fingerprints, facial images and DNA, etc., through information fusion technology It is stored in the background database, and the key information for retrieval is loaded into the RFID card, namely /soldier card 0. It is used as the only basis for the accurate identification of soldiers, and the personal identification system/battlefield is formed through the computer terminal and the network database server.
At present, identification systems based on various biometric identification technologies are becoming more and more popular and gradually applied in various fields. However, because the information expression and feature description of a single biometric identification have certain limitations, for this reason, the use of information fusion technology, through the fusion of fingerprints, facial images, and DNA and other information fusion processing personal multi-modal biometric identification system, not only It strengthens the security of identity recognition, and improves its accuracy, and will fill the research gaps in related fields of our army similar to /soldier card 0 or /soldier card 0.
1 System overall design
At present, identification systems based on various biometric identification technologies are gradually being applied in various fields. However, whether large-capacity human body information such as fingerprints, facial images, and DNA needs to be stored in the background database in the battlefield personal identification system 0 must be determined according to application requirements. For this reason, the /soldier card 0, as an identity marker carried by soldiers, should have the requirements of low price, waterproof, antimagnetic, high temperature resistance, and long service life. In view of the characteristics of biometric information, we mainly collect and process some biometrics. At the same time, according to the requirements of the specific combat environment and the overall design, different contents and functions have been expanded, such as multi-biological information fusion and GPS positioning.
The personal multi-modal biometric identification system is mainly through the computer terminal. First, fingerprints, facial images, DNA, etc., and text information are collected and input into the microprocessor. After information fusion, it is written into the back-end individual biometric information database management system through the network. , And can be associated with other military information, political information, medical care, clothing, wages, housing and other supply relationship information related to military personnel. At the same time, a unique ID serial number is generated, and the RFID reader and its biometric auxiliary information are written into the RFID card. Then, only need to use the RFID reader to read the ID and biometric auxiliary information during identification, and after inputting it into the computer terminal, it can be matched and inquired in the individual biometric information database via the wireless communication network or local area network. Figure 1:
Among them, the computer terminal can use a microcomputer based on the Windows.net operating system, with an internal RFID reader, and an extended interface to connect to the biometric collector. Except for DNA, it can currently collect, read and write other characteristics in real time. The information in the RFI D card needs to be encrypted while being written. Finally, a database of individual biometric information is established. The following mainly focuses on the collection of biometrics, the fusion of multi-modal biometric information and the design of the RFID system in detail.
2 Collection design of biometrics
2.1 Biometric recognition technology
Biometric identification technology refers to the science of personal identification and identification by using the inherent physiological or behavioral characteristics of the human body through the combination of computers, optical acoustics and biostatistics principles. People usually collectively refer to physiological characteristics and behavioral characteristics as biological characteristics. Physiological characteristics are innate and are mostly congenital; behavioral characteristics are dictated by habit and are mostly acquired. However, for any human physiological or behavioral characteristics to achieve identity recognition, the following conditions must be met:
First, universality: that is, everyone must have this characteristic;
Second, uniqueness: that is, the characteristics of any two people are different;
Third, collectability: that is, features can be measured;
Fourth, stability: that is, the characteristics will not change over a period of time.
At the same time, in actual application, other factors should be considered: real-time and accuracy, etc. Commonly used biological characteristics include: fingerprints, palm prints, human face, iris, retina, DNA, hand shape, signature, voice, gait, etc. Compared with traditional identification methods based on passwords and identification numbers, biometric identification technology is not easy to forget or lose; it has good anti-counterfeiting performance and is not easy to forge or be stolen; it can be carried around 0 and can be used anytime and anywhere; low price and high ease of use , Security and confidentiality and many other advantages. The following introduces the biometric technology used in the system.
2.2 Fingerprint recognition technology
Fingerprint recognition technology mainly includes: reading fingerprint images, feature extraction, and pattern matching. First, the image of the human fingerprint is read by the fingerprint reader, and then the original image is preprocessed to make it clearer. Next, feature extraction is performed. The fingerprint feature extraction algorithm consists of the following three steps: (Figure 2)
(a) Reading fingerprint images; (b) Region positioning; (c) Ridge feature graphics; (d) Minutiae extraction
Direction field estimation: that is, estimate the direction field of the input image.Determine the available area
Ridge feature extraction: extract and refine the ridge feature
Detail detection and post-processing: that is, extract details from the refined ridge feature map, and determine the ridge feature parameter coordinates, direction angle and the relationship with the ridge.
The fingerprint recognition technology is used to build a fingerprint database, and the data points called /minutia 0 are found from the fingerprint, that is, the coordinate position of the branch, termination or circle of those fingerprint lines. Some algorithms combine the node and direction information to generate more data, which indicates the relationship between each node, and some algorithms process the entire fingerprint image. Compare the templates of the two fingerprints, calculate their similarity, and finally get the matching result of the two fingerprints, so as to achieve the purpose of identifying the individual’s identity (Figure 3).
The fingerprint recognition system realizes recognition by comparing the input fingerprint with the fingerprint in the database. Therefore, it requires the database to be large enough. At the same time, the experiment shows that the input sensor of the fingerprint recognition system is not effective for about 4% of people’s fingerprints. Provide fingerprint images of sufficient high quality for identification, including scars on the finger skin, bandages, calluses, dry skin, dry humidity, sick skin, old skin, particularly smooth skin, narrow fingers and input Contamination of the sensor will affect the fingerprint recognition effect.
2.3 Collection of face signals
The principle of the face recognition system is shown in Figure 4. First, the face image is captured by a sensor such as a CCD camera; secondly, the image quality is improved by preprocessing; then the face is located according to the face detection and the face image is set to a predefined size; feature extraction is used to extract effectively Features are used to reduce the dimensionality of the original pattern space, and the classifier makes decision classification based on the features. Finally, the detected face to be recognized is compared and matched with the known face in the database to obtain the recognition result.
Face recognition methods mainly include: method based on geometric features of the side face, method of front face feature, hybrid method of front face feature and side face feature, template matching method, principal component analysis method, isodensity line map method, Multi-template correlation method, template matching method based on neural network, etc.
2.4 Other basic personal information
DNA recognition uses the uniqueness and permanence of the DNA molecular structure in human cells to further identify individual identities, so as to make up for the shortcomings of fingerprint and face recognition technology and improve the accuracy of system recognition. However, compared with fingerprints and face information, the DNA sequence of every cell and tissue of the human body is the same. The accuracy of this identification method is better than any other biometric identification method, so it is widely used in crime detection. Its main problem is the user’s ethical issues and actual acceptability. DNA pattern recognition must be carried out in the laboratory. It is difficult to achieve real-time and anti-interference. Time-consuming is another problem, which limits the use of DNA recognition technology; In addition, certain special diseases may also change the structure of human DNA, and the system cannot identify such people. Because the DNA recognition cycle is too long, the mapping of the sequence takes one month at the fastest. Therefore, its real-time performance, anti-interference ability, and the structural changes of human DNA caused by certain diseases are the main problems that affect the actual application of DNA recognition to the recognition system.
3 Fusion of multi-modal biometric information
Multimodal fusion is the information fusion of multiple biometric indicators (I nd ica -tors). This type of system integrates the evidence scores provided by multiple biometric sources to make more accurate and rapid decisions. The information integrated by the multi-modal biometric identification system can come from one or more biometric indicators. The general multi-modal single biometric identification system refers to the improvement of the system by integrating multiple types of evidence provided by a biological feature. In the multi-modal biometric identification system (Mult-ibi ome-tric System), fingerprints are generally used Multi-modal information fusion, so as to meet the two-way requirements of data volume and accuracy.However, in order to enter
To improve accuracy in one step, a method of fusing multiple biological characteristics can be used to provide accurate identification under special circumstances.
Generally speaking, fusion can be carried out at any of the three levels of pattern recognition, namely data layer fusion, feature layer fusion and decision-making layer fusion. The current research on biometric data fusion is mainly focused on the decision-making level. Different individual features are processed independently, and then matched to obtain a matching score. Finally, a certain fusion algorithm is used to synthesize the result. There are many kinds of fusion algorithms that can be used, such as majority voting, addition and multiplication rules, K-NN classifiers, SVM, Bayesian decision-making, support vector machines, decision trees and other artificial intelligence algorithms.
3.1 Fusion of feature layers
Feature layer fusion is that the input data is processed by the front-end to obtain its feature description vector for each biological feature. Different feature vector sets use different methods to form a new high-dimensional feature vector, and use this high-dimensional feature vector to represent multiple The fusion of biological characteristics. The fusion of the feature layer is more efficient than the fusion of the first two layers.
3.2 Fusion of matching layers
The input of the fusion module of the matching layer is the output scores of the matching modules of several biometric authentication systems. The fusion of the matching layer is performed for these inputs. Among the three fusion methods, the fusion of matching layers is the most common. This is because the fusion of the matching layer not only has relatively small implementation difficulty, but also integrates the respective information of several features. One of the important aspects of the matching layer fusion process is to normalize the scores obtained by different systems. After normalization, the scores obtained by different systems are mapped to an N-dimensional space, and then all points are classified in this space.
3.3 Integration of decision-making levels
Different single features can be processed independently, and then matched to obtain matching scores. Finally, through the process of decision fusion, multiple matching results are integrated through a certain fusion algorithm to obtain the final result. Fusion at the decision-making level is relatively simple, and the amount of information available is relatively small. Since the input of the decision-making layer is already the logical output of a single biometric authentication, the fusion of the decision-making layer can be divided into two forms:
(1) OR rule: In this kind of system, if the user is rejected by the subsystem H1, the subsystem H2 will verify the user again, and if it passes, it will be determined as the real user.
(2) AND rule: In this system, users can only be recognized as real users when they are accepted by subsystem H1 and subsystem H2 at the same time.
In specific applications, each fusion method has its own advantages and disadvantages. Feature layer fusion is more efficient, but it is not easy to integrate a practical single biometric authentication system. Matching layer fusion not only realizes the fusion of information energy algorithms, it is not very difficult to realize, and it has good application value. Although the level of information fusion is relatively small in the decision-making layer, it can be used in some integrated systems.
The current research on biometric data fusion is also mainly focused on decision-making research, but it is not enough to conduct research at the decision-making stage, because the role and impact of the correlation between features are ignored in the processing process. At the same time, the main Focusing on the discussion of fusion algorithms, and ignoring more considerations of biological characteristics, the fusion of the data layer and the feature layer is still needed.
4 Design of identification workstation
4.1 RFID system
4.1.1 Overview of RFID technology
RFID radio frequency identification is a non-contact automatic identification technology, which automatically recognizes the target object and obtains related data through radio frequency signals, and the identification work does not require human intervention. Compared with traditional barcodes, it has the advantages of non-contact, rewritable, fast scanning, good penetration, high security, large data storage capacity, as well as waterproof, antimagnetic, high temperature resistance, long service life, long reading distance, etc. It can adapt to the complex environmental requirements of modern high-tech local wars. The most basic RFID system consists of three parts:
Tag: Also known as electronic tag or smart tag, it is composed of coupling components and chips. Each chip has a unique electronic code, and the information that can identify the target is stored in the chip.
Reader: It is composed of transmitter, receiver, control module and transceiver. The transceiver is connected with a control computer or a programmable logic controller (PLC) to read (write) tag information and is designed to be handheld.
Antenna: wirelessly transmit radio frequency signals between the tag and the reader.
4.1.2 The working frequency of RFID and its application
The working frequency of the RFID system is one of its most important parameters. Radio frequency is a public resource. In the course of use, if different application systems overlap in time, space and frequency, mutual interference may occur. The degree of mutual interference between systems is related to the degree of overlap of radio signals and signal strength. Therefore, the rules of radio applications are the basic principles that various application systems must follow.
At present, there are three well-known international organizations that formulate RI FD standards: ISO, EPC globa l led by the United States, and Ubi qu i tous ID C enter (abbreviated as U ID) in Japan. Commonly used RFID international standards mainly include ISO/IEC18000 standard (including 7 parts, involving 125 KH z, 13.56MH z, 433MHz, 860) 960MH z, 2.45GH z and other frequency bands).
Among them, because the high-frequency tag with a working frequency of 13.56MHz is easy to be made into a card shape, the standard adopted by the second-generation electronic ID card is the IS014443 TYPE B protocol. This kind of label is one of the current mainstream applications. It generally adopts a passive solution, the reading distance is generally about 20 cm, and the farthest can reach 1.5M, which is suitable for ordinary/soldier card 0. The active recognition distance is not less than 9m, the actual sensing area can be set according to the specific situation, and the recognition distance can be adjusted from 1 to 10m. Reliable recognition of high-speed moving targets (people, vehicles, objects)[100 Km /h].[100Km/h的高速移动目标(人、车、物)。
4.2 Establishment of individual biometric information database
Through the acquisition system integrating the RFID reader, the soldier’s ID and biometric information are written into the card and sent to the computer terminal and then stored in the server. The computer processes the personal biological characteristics and enters it into the database management system of the server, and integrates other information. Finally, based on the database platform, a complete database related to biological information is established. Through the query function of the database, personal information in wartime and peacetime can be carried out. manage. At the same time, you can use the database shared by the information network to build a data warehouse, use artificial intelligence methods for data mining, and use effective data information, thereby maximizing the utilization rate of the individual biometric information database, which is beneficial to scientific research and Management decision.
4.3 Extension of other functions
According to the needs of the search and rescue of the wounded on the battlefield, or the needs of other special combat environments, the system can adopt an active solution, select an ultra-high frequency RFID system, build a wireless communication network, and expand the GPS positioning function by a computer terminal. Mainly include: used in active RFID tags and read-write modules, the thickness and cost of such cards will increase compared with passive RFID. Secondly, use Bluetooth and other wireless communication modules to connect the computer terminal with the GPS positioning system, and send the position information in real time within the effective range, and the host will choose to confirm the tracking target for positioning. Thus, the search and rescue work and command work for the wounded on the battlefield are realized.
5 Outlook
Studies at home and abroad have shown that the use of fingerprint and face image, as well as fingerprint, face image and voice combination and other biometric identification technology for identity recognition, to improve the accuracy of recognition, has been widely used in airport and port security systems, identity authentication systems , Financial field, access control system, attendance system, floating population management, e-commerce, e-government, etc. However, because the identification system that integrates fingerprints, face images, and DNA maps is based on infinite radio frequency identification technology, the multi-modal biometric information data storage capacity is large, the calculation speed is slow, and the storage capacity of active and passive radio frequency cards and wireless The limited reading and writing speed and distance make it difficult to adapt to the actual needs of modern high-tech local wars and future wars, which has become the main bottleneck restricting the wide application of multi-modal biometric identification systems. I believe that with the continuous innovation of follow-up research and the rapid development of related hardware technologies, multi-modal biometric recognition technology will eventually be more widely used and play a more important role in personal identification on the battlefield.
[ad_2]
