Design And Implementation Of Dsp-based Fingerprint Lock System

With the development of electronic information technology, the traditional identification technology, such as password and certificate, can not satisfy the need of safety. Identification has become much more serious than before. The biometric identification technology can identify someone by their biometric features, such as face, fingerprint and iris, accurately and quickly. The fingerprint recognition technology is developing very fast and used in many fields.The embedded automatic fingerprint lock system is a kind of embedded application on fingerprint recognition technology. Nowadays, some of the fingerprint recognition products in domestic market are based on modules from abroad and do not have information property. Some self-developed products'cost and power consumption are too high to be used in some special environments. To solve these problems, a fixed-point DSP system design is proposed in this paper. It is practical, low power consumption, low cost system and can be used in kinds of environments.The works about this thesis are mainly focused on the following parts:1. A high performance, low power consumption hardware system design is proposed in this thesis. It consists of several modules, which are Processing Module, Fingerprint Sensor Module, Output Control Module, Human-Machine Communicating Module, Power Manager and Power Switchover Module. Especially in the power management module, an automatic power switchover circuit, which can automatically switch the paths of power supply between wall adapter and batteries, is proposed in this paper. This design makes the system design more practical.2. By researches on fingerprint recognition algorithms, fingerprint matching is the key to the whole system and effects on the accuracy and efficiency of the whole system directly. In this thesis, a new minutia matching method based on ridge tracing is proposed. When a minutia has been found, we trace the ridges where the minutia is on and search in a range to check if there are any other minutiae on the same ridges. This new method takes the relationship between minutiae on the same ridge into consideration for the first time and makes it more efficient to find a true reference point. As the experimental result, this method is more accurate and robust than the common minutia matching methods.3. Designed the whole application system on hardware platform by using DSP/BIOS RTOS. The device'drivers and system tasks are programmed to achieve the automatic fingerprint recognition function. The system's efficiency is enhanced by code optimization.In conclusion, our system is designed to be low power consumption, high performance. At the same time, the modularized hardware and software can be used in many different situations.