Biometric technology includes iris recognition, retina recognition, face recognition, voice recognition and fingerprint recognition [2]. Among them, fingerprint identification technology is the most mature and widely used biometric technology. Everyone's skin lines, including fingerprints, patterns, breakpoints and intersections are different, which means that these fingerprint features are unique and will remain unchanged for life. Relying on this uniqueness and stability, we can associate a person with his fingerprints, and by comparing his fingerprints with pre-stored fingerprints, we can verify his true identity.
Fingerprint identification system realizes fast and accurate identity authentication by collecting, analyzing and comparing fingerprint features. The block diagram of the fingerprint identification system is shown in figure 1.
After the fingerprint image is collected by the fingerprint collector, it can be recognized and processed by the computer. The quality of fingerprint image will directly affect the accuracy of identification and the processing speed of fingerprint identification system, so fingerprint acquisition technology is one of the key technologies of fingerprint identification system. This paper focuses on the analysis and comparison of different fingerprint acquisition technologies and their performance.
1 fingerprint collection technology
Fingerprint has a small surface area and is often worn by fingers in daily life, so it is a very complicated job to obtain high-quality fingerprint detail images. The fingerprint collection technologies currently used mainly include optical fingerprint collection technology, semiconductor fingerprint collection technology and ultrasonic fingerprint collection technology.
1. 1 optical fingerprint image acquisition technology
Optical fingerprint collection technology is the oldest and most widely used fingerprint collection technology at present. Optical fingerprint acquisition equipment started in 197 1 year, and its principle is total reflection of light (FTIR). Light shines on the glass surface with fingerprints, and the reflected light is obtained by CCD. The amount of reflected light depends on the depth of ridges and valleys of the fingerprint pressed on the glass surface, as well as the grease and moisture between the skin and the glass. After the light hits the valley through the glass, it is totally reflected at the interface between the glass and the air, and the light is reflected to the CCD, while the light directed to the ridge is not totally reflected, but absorbed or diffusely reflected to other places by the interface between the ridge and the glass, thus forming a fingerprint image on the CCD.
Optical acquisition equipment has many advantages: it can stand the test of long-term practical application, can withstand a certain degree of temperature change, has good stability and relatively low cost, and can provide images with a resolution of 500dpi.
Optical acquisition equipment also has shortcomings, mainly image size and potential fingerprints. The drum must be big enough to get a good image. Finger pressing on the platen leaves a latent fingerprint, which reduces the quality of fingerprint image. A serious latent fingerprint will cause two fingerprints to overlap. In addition, the coating (film) and CCD array on the platen will be lost over time, and the accuracy will be reduced.
With the innovation of optical equipment technology, the volume of optical fingerprint collection equipment is also decreasing. Now the sensor can be installed in a 6x3x6 inch box. In the near future, the smaller equipment will be 3x 1X 1 inch. These advances have benefited from the development of various optical technologies. For example, optical fiber bundles can be used to obtain fingerprint images. The optical fiber beam irradiates the fingerprint surface vertically, illuminating the fingerprint and detecting the reflected light. Another solution is to mount a surface containing a matrix of microprisms on an elastic plane. When a finger is pressed on this surface, the surface of microprism changes due to the different pressures of fingerprint ridges and valleys, and these changes are reflected by the reflection of prism light.
The U.are.U series optical fingerprint collector introduced by DigitaIPersona[4] company in the United States is a widely used optical fingerprint collector at present, which is mainly used to confirm the identity of users when logging into computer windows system. It integrates the precision optical system, LED light source and CMOS camera to work together, has the characteristics of three-dimensional living body, can accept fingerprint input from all directions, even if it rotates 180 degrees, it is one of the safest optical fingerprint identification systems on the market at present. U.are.U optical fingerprint collector is designed according to ergonomics, with USB interface, and it is a new intelligent peripheral beside the keyboard on the user's desktop.
1.2 semiconductor fingerprint acquisition technology
Semiconductor sensors appeared in the market only in 1998. These planes containing microcrystals draw fingerprint images by various techniques.
(1) silicon capacitive fingerprint image sensor
This is the most common semiconductor fingerprint sensor, which captures fingerprints through electronic measurement. About 100000 capacitance sensors can be combined on the semiconductor metal array, and the outside is an insulating surface. Each point of the sensor array is a metal electrode, which acts as one pole of the capacitor, and the finger presses the corresponding point on the sensing surface to act as the other pole, and the sensing surface forms a dielectric layer between the two poles. Because the distance between the ridge and the valley of the fingerprint is different from the other pole (the depth of the texture exists), the capacitance value of the silicon surface capacitance array is also different. By measuring and recording the capacitance value of each point, a fingerprint image with gray level can be obtained.
(2) Semiconductor Pressure Sensitive Sensor
The top layer of its surface is elastic pressure-sensitive dielectric material, which is converted into corresponding electrical signals according to the external morphology (concave and convex) of the fingerprint to further generate a fingerprint image with gray scale.
(3) Semiconductor temperature sensor
It can obtain a fingerprint image by sensing the temperature difference between the ridge pressed on the device and the valley far away from the device.
The semiconductor fingerprint sensor adopts automatic control technology (AGC technology), which can automatically adjust the sensitivity of the pixel line of the fingerprint image and the local range of the fingerprint, and can produce high-quality images by combining the feedback information in different environments. For example, you can feel unclear (poor contrast) images, such as dry fingerprints, which can enhance their sensitivity and produce clear (good contrast) images at the moment of snapshot; Because of the ability of local adjustment, areas with unclear image (poor contrast) can also be detected (such as the place where fingers are lightly pressed), and the sensitivity of these pixels can be improved at the moment of snapshot.
Semiconductor fingerprint acquisition equipment can obtain quite accurate fingerprint images with a resolution as high as 600dpi, and fingerprint acquisition does not require a large-area acquisition head like optical acquisition equipment. Because of its small size and low power consumption, semiconductor chips can be integrated into many existing devices, which is incomparable to optical acquisition devices. Nowadays, many fingerprint identification systems are developed with semiconductor acquisition equipment. The main weakness of early semiconductor sensors is that they are easily influenced by static electricity, which makes the sensors sometimes unable to obtain images or even damaged. Salt or other dirt in finger sweat and finger wear will make it difficult for semiconductor sensors to obtain images. In addition, they are not as wear-resistant as glass, which affects their service life. With the continuous development of various technologies, the antistatic performance and durability of the chip have been greatly improved.
Verdicom [5], separated from Lucent, has been devoted to the research and development of semiconductor fingerprint collection technology since 1997. Up to now, a series of CMOS fingerprint sensor products such as FPSll0 and FPS200 have been developed and adopted by some commercial fingerprint identification systems. Its core technology is based on the design of high reliability silicon sensor chip.
FPS200 is a new generation of fingerprint sensor developed by Veridicom Company on the basis of absorbing the advantages of widely used FPSll0 series sensors. The surface of FPS200[6] adopts patented technology of Vefidicom company, which is durable and can prevent various substances from scratching, corroding and wearing the chip. The FPS200 can withstand electrostatic discharge (ESD) exceeding 8KV, so it can be used in harsh environment. This product combines different texture information such as ridges and valleys in fingerprints. Through the image search function of high-reliability silicon sensor chip, no matter whether the finger is dry, wet or rough, one of the best fingerprint images collected by the same finger can be selected and stored in the memory, and the fingerprint resolution can reach 500dpi, which greatly reduces the occurrence of false acceptance and false rejection in the recognition process of sensor chip.
FPS200 is the first fingerprint device with three built-in communication interfaces: USB interface, MCU interface and serial peripheral interface (Sn), which enables FPS200 to connect various types of devices without the support of external interface devices. The overall package size (24 mmx 24 mmxl.4 mm) is only the size of ordinary stamps. Because of its high performance, low power consumption, low price and small size, it can be easily integrated into various Internet devices, such as portable computers, personal digital assistants (PDA) and mobile phones.
1.3 ultrasonic fingerprint image acquisition technology
Ultra-scanning company pioneered the ultrasonic fingerprint image acquisition equipment products. Ultrasonic fingerprint image acquisition technology is considered to be one of the best fingerprint acquisition technologies, but it is rare in fingerprint identification system, with high cost and still in the laboratory stage. The principle of ultrasonic fingerprint identification is: when ultrasonic waves scan the fingerprint surface, follow the reflected signal obtained by the receiving device, and the ultrasonic energy reflected back to the receiver is different because of the different acoustic impedance of the ridge and valley of the fingerprint, and then measure the ultrasonic energy to obtain the gray image of the fingerprint. The dirt and grease accumulated on the skin have little effect on ultrasonic imaging. Therefore, the image obtained in this way is a true reflection of the concave-convex line of the actual fingerprint.
In short, these fingerprint collection technologies have their own advantages and disadvantages. Ultrasonic fingerprint image acquisition technology has not been applied to fingerprint identification system because of its high cost. Generally, the fingerprint collection area of semiconductor sensor is less than 1 square inch, and the fingerprint collection area of optical scanning is equal to or greater than 1 square inch. We can choose which fingerprint collection equipment to use according to actual needs.
Table 1 gives a comparison of the three main technologies.
Table 1
1 optical scanning technology 2 semiconductor sensing technology 3 ultrasonic scanning technology
1 The imaging ability of dry fingers is poor, and the imaging of sweaty and slightly swollen fingers is blurred. Easily affected by dirt and grease on the skin. Dry fingers are good, and wet and rough fingers can also be imaged. Easily affected by dirt and grease on the skin. 3 Very good
The imaging area is large, small and medium.
The resolution is less than 500dpi, the highest is 600dpi, and the highest is 1000dpi.
The equipment is large and small.
Durability is very durable and more durable.
Power consumption is bigger, smaller and bigger.
High cost, low cost and high cost.