03-08-2012, 02:28 PM
Nios II Processor-Based Fingerprint Identification System
1Nios II Processor.pdf (Size: 954.96 KB / Downloads: 31)
Design Introduction
With the boom of information technology represented by computers since the 1960s, computer
technology has begun to be used in the fingerprint identification field, bringing new thoughts,
implementation methods, and processing approaches for automated fingerprint identification.
Authorities, institutions, and universities have begun implementing fingerprint analysis and processing
using computers. A computerized system that performs automated fingerprint identification is called an
Automated Fingerprint Identification System (AFIS).
People often need to be identified in society. The common ID authentication methods such as keys,
password, certificates, and IC cards provide identification using objects, which indirectly identify the
object holder. These objects are not very accurate and have significant security risks, including
counterfeited certificates and tokens, and decrypted or stolen passwords. With the development of
image processing and pattern identification technologies, emerging identification technology based on
biometric characteristics has become the focus of research and applications due to its unique reliability,
stability, and convenience. As the earliest and most mature biometric identification technology in the
pattern identification field, fingerprint identification technology integrates sensors, biometric
technology, electronic technology, digital image processing, and pattern identification. Many automatic
fingerprint identification systems are used worldwide, but fingerprint identification technology is not
yet mature. China is behind in fingerprint collection and algorithm study, so the research of fingerprint
identification algorithms and systems will play a significant role in theory and practice.
Function Description
We designed the fingerprint identification system based on Altera’s Nios II processor and FPGAs. This
system can collect real-time fingerprint image signals, extract finger minutiae, and match minutiae in a
database to perform fingerprint identification. The whole system design includes fingerprint image
collection, fingerprint image preprocessing, minutia extraction, minutia matching, and a database.
Fingerprint Image Signal Collection
The fingerprint collector serves as the fingerprint collection module. The module’s fingerprint sensor is
Veridicom’s third-generation product, the FPS200 sensor (with 256 x 300 array numbers and 500-DPI
resolution). The sensor uses Veridicom’s ImageSeek function and high-speed image transmission
technology to obtain quality images of all fingerprint types. The fingerprint collector design performs
the following functions.
Fingerprint Image Preprocessing
During fingerprint image preprocessing, the fingerprint image is enhanced. Accurate fingerprint
identification relies on the identification of the fingerprint ridge texture and minutiae. However, due to
skin condition, collection conditions, devices, the working and living environment of the fingerprinted
person, etc., the raw fingerprint images collected by the fingerprint sensor usually contain noise and
degrade dramatically. Therefore, the raw fingerprint images must be preprocessed after being collected.
Fingerprint image preprocessing procedures include image normalization, orientation and frequency
extraction, filtration, binarization, ridge thinning, etc.
Normalization
Image normalization reduces the diversification degree of the grayscale along the ridges and valleys
without changing the raw image’s structure or texture information. This process gives the image preset
means and variances, and facilitates pattern capture and fingerprint frequency. Nevertheless,
normalization can also enhance some hash in the image background. Equation 2.1 is for normalization
and equation 2.2 is an improved version based on equation 2.1 for the convenience of hardware
implementation.