Application of ARM kernel at75c220 in fingerprint identification system

In recent years, the pattern recognition technology has become more and more widely used in access control, security and finance. A typical fingerprint recognition system is composed of a fingerprint sensor and a DSP processor. Fingerprint sensor acquisition fingerprint image, DSP processor real-time implementation of fingerprint recognition algorithm. At the same time, the usual fingerprint identification System also has strong communication ability, in addition to have RS485 communication, high-grade fingerprint Identification System also requires an Ethernet interface and Internet connectivity capabilities. Atmel's new smart Internet chip at75c220 integrates the arm core, DSP coprocessor, and Ethernet Mac interface, combining at75c220 and fingerprint sensors to form a high-performance fingerprint identification system.

1 at75c220 main features

AT75C220 is a high performance arm processor located in smart Internet applications with a ARM7TDMI microprocessor core with OADDSP core, powerful peripherals such as interrupt controller and peripheral data controller, and its DMA channel enhances the real-time performance of the microcontroller. The main features are:

• Integrated ARM7TDMI ARM Thumb processor Core

• A 16b fixed-point DSP core
• Dual Ethernet 10/100mb/s Mac interface
• 88kB of Full fast RAM (DMA supported mode)
• Multimedia Digital signal codec (codec interface)
• Integrated SDRAM Controller

  In this way, at75c220 is ideal for devices that require image processing capabilities and have Internet communication capabilities, such as high-performance fingerprint identification devices and VoIP applications. A single chip integrates multiple functions to reduce costs and improve system reliability. At the same time, the complete development of the chip software support (including Linux operating system and DSP algorithm program support, etc.) also helps reduce time to market.

  2 chip structure

  The at75c220 consists of a ARM7TDMI-based RISC processor that runs at speeds of up to mi/s, a DSP processor running at a speed of 60mi/s, and a dual Ethernet 10/100mb/s Mac interface. As shown in its internal body 1.

  The at75c220 contains 2 different bus ASB and APB,ASB for connectionArm Core, DSP core module, and the APB bus connects serial communication interface, general I/O port, and so on, between the two buses with Amba Bus bridge connection.

  ARMTTDMI microprocessor is a high-performance embedded CPU, low power consumption, fast operation, including thumb decoder, can support 16b instructions, while supporting debugging and fast multiplication and other functions, and through the module cross-bus and other functional modules for data exchange.
  ARMTTDMI (THUMB) is one of the most authorized users of the company, combining the ARM7 instruction set with the THUMB extension to reduce memory capacity and system costs. He also uses embedded ice debugging techniques to simplify system design and improve performance with a DSP-enhanced extension. Another reason to choose arm -based products is that ARM's processor core is supported by a large number of manufacturers to facilitate product development and upgrading.

  Compared with other chips with integrated DSP cores, AT75C220 's DSP subsystem resources are quite rich, and 2 shows that, in addition to a 60mi/s,16b fixed-point DSP core, there are different uses of Ram:2kbxl6x-ram,2kbxl6y-ram, 16KBXL6 Universal Data ram,24 kBxl6 downloadable program RAM, a 256x16 dual-port RAM mailbox for MCU-DSP communication. The complete codec interface allows a large number of industrial sound, multimedia or data connections.  

• 
  The chip-integrated Ethernet interface includes 2 separate 10/100mb/s Ethernet MAC interfaces and a 32kB packet buffer. The dual-port Mac design allows the device to be embedded in an existing Ethernet without the need for additional cables, and due to the local bus connection between the 2 Mac interfaces, a large number of packets can be transmitted from one Mac to another without delay, without the need for software involvement. The packet buffer is operated by the Mac interface through DMA, which is used to temporarily save the received and sent packets, which in most cases can satisfy large traffic volumes.
  
  3 fingerprint recognition algorithm

  Fingerprint automatic recognition as a more difficult pattern Recognition Research Branch, its research difficulties mainly focus on how to filter and enhance the various noisy images collected, how to extract the global and local characteristics of fingerprints and how to do in the image can not be fully positioned and the image can be distorted in the case of feature matching. Generally can be divided into the following steps, 3, the fingerprint identification algorithm flow mainly includes gray-scale filtering, binary, binary filtering, refinement, refinement of the steps to remove noise, and finally to feature extraction and matching.

  
  In this system, the gray filter of fingerprint recognition algorithm is based on the directional filter of Block square graph, then binary and thinning the fingerprint image, further denoising in the process of binary value and refinement, eliminating the possibility of the characteristic point which leads to the error. The feature points of fingerprint, such as endpoint, fork point and so on, were obtained by template matching method. The matching algorithm based on the center point is used for fingerprint matching. by verifying on PC, the algorithm can achieve better results.

  4 Fingerprint identification System

  Fingerprint Identification System is generally composed of fingerprint sensor, DSP processor and peripheral circuit, in order to increase the processing power of the system, the system often includes additional processor management such as communication, human-computer interface, fingerprint database and so on. The system structure 4, the capacitive fingerprint sensor FPSll0 acquisition fingerprint image and by the at75c220 internal DSP coprocessor for fingerprint identification, the completion of the image preprocessing, extraction of fingerprint signatures, and fingerprint database with the fingerprint features to match, give the discriminant results. The arm core runs on the Linux operating system, realizes the TCP/IP protocol stack, because the fingerprint recognizer's network system to the real-time performance and the throughput rate request is not high, but the request reliability and the confidentiality is relatively high, so the Transport layer Protocol chooses the TCP protocol, and transmits the data to be encrypted. The ARM processor also includes a human-computer interface and control outputs, which are used to control the electronic door lock. In addition, the ARM software includes the access control system, the attendance system, some of the proprietary modules, such as access Rights Management, attendance database management.

  
  5 Porting of the algorithm

  The fingerprint recognition algorithm is implemented on the PC at the beginning, and it needs to be ported to the at75c220 DSP subsystem. If the direct use of C language transplant, it is difficult to play the actual processing capacity of DSP, execution efficiency is very low. After the key algorithm module is implemented in assembly language, the fingerprint acquisition, image preprocessing and matching can be completed in 1s, and the algorithm efficiency and recognition rate are equivalent to that of PC, which achieves the practical application requirements.

  6 conclusion

  As the hardware adopts integrated multi-function at75c220 composing system, the system hardware design is simple, which makes the whole system meet the requirements of miniaturization, networking and intellectualization. The practice proves that the equipment processing ability with this kind of structure has been improved, which represents the development direction of the future access control system.

  The integration of hardware functions makes the main design work focus on the software design, how to further improve the recognition rate, the perfect distributed database system, the data confidentiality transmission is the main point of the future design.

Application of ARM kernel at75c220 in fingerprint identification system