【作者】 赵娟；

【导师】 王典洪；

【作者基本信息】 中国地质大学 ， 通信与信息系统， 2005， 硕士

【摘要】 在当今数字化信息化时代，信息安全保障的实际需求日益增强，与此相关的生物特征识别技术也逐渐受到了人们的高度重视。指纹识别是生物识别技术的一种，它以其唯一性和不变性等独特的优势成为目前应用最广泛的身份验证与识别技术。它涉及到模式识别、图像处理、数字信号处理、人工智能和计算机等领域，是一门综合性技术，有着重要的理论意义和实用价值。传统的指纹识别系统大多是基于PC平台的，而当今蓬勃发展的民用市场对自动指纹识别系统提出了更高的要求，自动指纹识别技术正向着小型化和嵌入式的方向发展。 结合上述背景，本文研究并总结了指纹识别系统的基本原理和关键技术，详细介绍了指纹图像预处理、特征提取和特征匹配各部分对应的算法。通过对目前各种算法的对比，主要针对DSP嵌入式系统的性能特点和处理能力，结合指纹图像的自身特点，选用了比较适合于DSP硬件系统的指纹识别算法，并在指纹图像的细化、伪特征点的删除和特征匹配方面进行了改进、简化或是创新，使整套算法不但能够更好地完成各自的工作，而且运算更简单，还能有效地融合在一起，为其在DSP系统中的应用奠定了基础。 本文的指纹识别算法以实践为基础，指出了图像的预处理、特征提取和特征匹配过程中的一些特定参数和方法。 在编程运行验证了上述算法实现的可行性后，本文设计了一种以DSP为核心的指纹识别系统，并对其外围电路模块进行了详细地介绍。主控芯片采用的是TI公司的TMS320VC5410芯片，该芯片具有良好的性能价格比，且对大量的加乘运算具有优异的表现，这对指纹图像处理算法至关重要。同时TI为该系列处理器提供的开发环境CCS给开发者提供了丰富的调试手段，而且，它支持C语言和汇编语言的混合编程，这有助于本文识别算法在硬件中的实现。另外，为了满足系统在资源和速度上的要求，在系统的硬件体系结构和存储器资源管理上都有合理的安排和设计，既充分发挥系统各部分的最高性能，又兼顾到成本因素。试验证明，无论是从运行时间还是处理的结果来看，整个系统在实时性和可靠性上都达到了令人满意的效果，基本满足了此课题最初的设计要求。更多还原

【Abstract】 In current society, the demand on the security of information is increasing. Biometrics, which refers to identifying an individual based on his or her characteristics, has the capability to reliably distinguish between an authorized person and an imposter. Among all biometrics, fingerprint-based identification is one of the widest used identity and identification technologies for its uniqueness and invariability. It relates to a lot of fields, including pattern recognition, digital image processing, digital signal processing, artificial intelligence and so on. It has great worth, and has not only theoretical significance but also practical value.Now most of the automatic fingerprint identification systems are realized on the PC-based platform. But the newly developed market for civil application has demand more for system’s portability and convenience. So the trend of miniaturizing and embedding the system for automatic fingerprint identification become more and more distinct.In these backgrounds, I investigate and sum up keystones and the key of technology in fingerprint identification. I introduce some algorithms about image preprocessing, feature extraction and matching in detail. Through comparing the algorithms at present, I selecte the fingerprint recognition algorithms relatively suitable for DSP hardware system, with combining their own characteristic of the fingerprint images, which are mainly against the performance characteristic of DSP embedded system and handling capacity. In the procedure of fingerprint images thinning, false features elimination and fingerprint-matching, I have the corresponding algorithms improved, simplified and innovated. So, the whole set of algorithms not only can finish their own work better, but also operation is simpler. They can also merge together effectively. All these, have established the foundation for its application in DSP system.The textual fingerprint identification algorithms depend on practice again and again. Some particular parameters and methods about fingerprint preprocessing, feature extraction and matching are pointed out in the paper.After running the correlative program to verify the feasibility, I design a kind of fingerprint identification system on DSP, and introduce its peripheral circuit modules in detail. What the top management chip is adopted is TMS320VC5410 chip with high proprieties of TI Company. TMS320VC5410 has excellent behavior to a large amount of adding and multiplication operations, which are essential to pattern process algorithms of fingerprint. At the same time, CCS that TI offered to this serial processors provide the abundant debugging means for developer, and, it supports C language and the assembler language to mix programming, this contributes to the fingerprint identification algorithms realized in the hardware system. In addition, for更多还原