FPC1 0 11 F based fingerprint recognition system design and implementation

1 fingerprint recognition principle

1684, plant morphology home Grew published its first articles on the fingerprint of a scientific paper.

1809 Bewick his fingerprints as a trademark. 1823 year Anatomy home Purkije will fingerprint is divided into nine categories. 1880, Faulds in the journal nature fingerprint to identify advocates. 1891 Galton made famous Galton classification system. After that, the United Kingdom, United States, Germany and other police departments have used fingerprint verification method for the identification of the main method. Current fingerprint image acquisition principle include: optical technology, semiconductor silicon technology, ultrasonic technology. Which the optical fingerprint capture device has obvious advantages. But because of the requirements for a sufficient amount of optical path, and therefore require sufficiently large dimensions and dryness and overly greasy fingers would also enable the optical effect of fingerprint products. In the late 1990s, developed based on semiconductor Silicon capacitive effect technologies mature. Silicon sensors become a plate capacitor, the finger is another plate, use the fingerprint ridge and Valley line relative to the smooth Silicon sensor capacitance, formation of 8 bit grayscale images. Silicon technology advantage of smaller surface than optical technology better image quality, in 1 cm× 1.5 cm surface get 200-300 lines of resolution (smaller surface also result in cost reduction and can be integrated into smaller units).

2 FPC1011F fingerprint sensor

FPC1011F fingerprint sensor has the following characteristics: (1) FPC1011F chips produced in Sweden, a unique reflective measurements, antistatic up to plus or minus 15 kV, wear-resistant 100 million times, was the domestic financial sector recognized as banks specified parts.

(2) use of fingerprint identification chip PS1802DSP and optimal fingerprint algorithm, fingerprint image is better. (3) processing speed, peak can be reached in 1: 480MIPS, 1 000 mode, time is less than 1 s. (4) power consumption compared to similar products of low, normal working frequency 120 MHz, only 120 mW. (5) module volume 35 mm mm× 26 mm× 1, facilitate the development of various fingerprint products. (6) for dry and wet fingers have Auto-Tuning feature.

3 system hardware design

The entire system as shown in Figure 1.

3.1 fingerprint collection

FPC1011F fingerprint sensor with small capacitor plate, sensor, high sensitivity pixel amplifier so that each pixel is very weak signal FPC1011F can detect, thereby improving image quality.

Use the alternate order and arrange and sensor power plate, alternating plates in the form of two capacitor plate, as well as fingerprint ridges of the Valley and a plate of dielectrics. The constant change of dielectric sensors detect the fingerprint image is generated.

3.2 MCU microprocessor

Because of the fingerprint identification process requires a large quantity of mathematical operations, and procedures also need space for storage, for efficiency, the operational tasks to MCU, and image acquisition part you want as few occupy MCU.

In addition, the use of image acquisition of clearance or image acquisition, from hardware to complete part of simple and complex calculations can share the MCU processing tasks, improve processing of parallelism, meet the requirements of real-time performance. This system uses chipscreen (Mi-croehip) company PIC16C78X series 8-bit MCU chips, mixed in with integrated amplifier, AD, DA, comparator, analog circuit, very suitable for data processing. Through the use of extreme value filtering, smoothing filter, Laplace change, binary, and so on fingerprint image preprocessing and achieved good results. System of acquisition to 8 bits grayscale fingerprint image, each pixel occupies one byte, the image size is 512 x 512 pixel size, storage frame image need 256 k bytes. Due to the image itself to the sheer volume of storage, system need external storage to ensure there is enough storage space, the images stored in USB disk. At the same time, the system can also be accessed via the data cable and is connected to a network, implement remote control function. MCU is the core of fingerprint processing system, responsible for fingerprint for real-time processing.

4 experimental simulation

Use MATLAB7.0 image processing box for simulation, an average of one image data simulation time consuming 0.73 seconds, recognition rate error is less than 1/310 000, fully in line with the actual engineering needs.

Figure 2 is a random one thumb fingerprint image, Figure 3 is a fingerprint library collection corresponds to the image. Green represents the recognition of special enhanced region.

5. concluding remarks

This article presents a new automated fingerprint identification system, using semiconductor sensor FPC1011F and MCU chips of fingerprint recognition system that has the extension interface, you can make the second development, have practical value.