【摘要】：隨著現(xiàn)代社會的飛速發(fā)展,人們在生活生產(chǎn)中對于身份安全的需求與日俱增。同時,作為當前生物識別技術中發(fā)展最成熟、最便利、綜合成本最優(yōu)的技術,指紋識別技術在考勤、門禁、移動支付等行業(yè)領域中得到了廣泛應用。而作為指紋識別技術的主要載體,指紋識別系統(tǒng)中的核心器件——指紋識別芯片,由于其能夠單片實現(xiàn)一套完整的指紋識別算法,成為實現(xiàn)一套優(yōu)良的指紋識別系統(tǒng)的關鍵因素,具有極高的科學研究價值。本文在對當前的指紋識別算法以及指紋識別系統(tǒng)的研究現(xiàn)狀與成果進行充分調研的基礎上,結合本學科領域的專業(yè)知識與技能,依托SoC FPGA平臺在軟硬件協(xié)同設計上的優(yōu)勢,探索構造一套安全級別高、價格成本低、識別效率高、具有自主知識權產(chǎn)的指紋識別算法,以及一套能夠支持此算法的、具有自定制指紋識別芯片架構的指紋識別系統(tǒng)。本文將研究設計分為"軟件算法設計"和"硬件平臺構建"這兩個主要部分,進行軟硬件協(xié)同設計:(1)軟件算法設計:首先使用Matlab軟件,進行指紋識別算法的原型設計;其次使用VC++軟件,用C語言對上述算法進行平臺無關性重構建模,構造平臺無關的算法C-Model;然后使用AlteraSoC EDS軟件,針對選定的SoCFPGA平臺,進行算法C-Model的跨平臺移植,并進行算法策略、程序代碼、編譯環(huán)境等層面的軟件算法優(yōu)化;最終通過調試與改良,實現(xiàn)算法的性能、效果最優(yōu)化。(2)硬件平臺構建:首先選定基于Altera Cyclone V SoC FPGA芯片的DE1-SoC開發(fā)套件作為硬件開發(fā)平臺,選定FPC1011F3指紋傳感器作為指紋采集設備;其次使用Altera Quartus Ⅱ軟件以及Qsys工具,在SoC FPGA內部構建一個內嵌ARM Cortex-A9硬核、定制化指紋識別芯片架構的嵌入式微處理器;然后待算法移植到此處理器中,進行軟硬件聯(lián)調與優(yōu)化,并對部分算法步驟進行硬件加速。最終實現(xiàn)功能完整、安全高效的指紋識別系統(tǒng)。本文的指紋識別算法,在實現(xiàn)基本功能的前提下,通過一系列針對性的軟硬件優(yōu)化方案,可達6倍左右的效率提升。同時,本文的指紋識別系統(tǒng),經(jīng)過大量而嚴格的測試,其整體性能最終可達:平均處理時間=722.97毫秒,認假率FAR=0.00%,拒真率FRR=8.2%�？紤]到實際指紋輸入過程中的噪聲干擾,此指標已經(jīng)能夠滿足設計要求,證明本文的算法與系統(tǒng)的高安全性、高識別效率等特性。最后,本文對課題設計過程中遇到的疑難問題與解決方案進行分析與總結,并對課題的后續(xù)研究與深入優(yōu)化進行規(guī)劃與展望。
[Abstract]:With the rapid development of modern society, the demand for identity security is increasing in life production. At the same time, as the most mature, convenient and cost optimal technology in biometrics, fingerprint identification technology has been widely used in the fields of attendance, entrance control, mobile payment and so on. As the main carrier of fingerprint identification technology, fingerprint identification chip is the core device of fingerprint identification system, because it can realize a complete fingerprint identification algorithm. It has become the key factor to realize a set of fine fingerprint identification system, and has the extremely high scientific research value. On the basis of investigating the current research status and achievements of fingerprint identification algorithm and fingerprint identification system, combining the professional knowledge and skills of this subject, this paper relies on the advantages of SoC FPGA platform in hardware and software co-design. This paper explores and constructs a fingerprint identification algorithm with high security level, low price cost, high recognition efficiency and independent knowledge right production, as well as a fingerprint identification system with custom fingerprint identification chip architecture that can support this algorithm. In this paper, the research and design are divided into two main parts: software algorithm design and hardware platform construction. (1) Software algorithm design: first, Matlab software is used to design fingerprint identification algorithm prototype; Secondly, using VC software and C language, the platform-independent reconstruction model is built, and the platform-independent algorithm C-Model is constructed. Then using AlteraSoC EDS software, aiming at the selected SoCFPGA platform, the cross-platform migration of algorithm C-Model is carried out, and the algorithm strategy, program code, compilation environment and other software algorithms are optimized. Finally, through debugging and improvement, the performance and effect of the algorithm are optimized. (2) hardware platform construction: firstly, the DE1-SoC development suite based on Altera Cyclone V SoC FPGA chip is selected as the hardware development platform. The FPC1011F3 fingerprint sensor is selected as the fingerprint acquisition equipment. Secondly, using Altera Quartus 鈪，

本文編號：2320295