基于SOPC的高性能图像相关器的设计与实现-模式识别与智能系统专业论文.docxVIP

Abstract Image matching is a classical problem in image processing. It has a wide rang of applications in many fields such as computer vision, pattern recognition, medical image processing. Therefore, its very important to research a high performance hardware circuit. The method based on zero-mean normalized cross-correlation matching algorithm is the most commonly and classical image matching method. So this paper designs and realizes a high-performance image correlator based SOPC(System On Programmable Chip). The image correlator is based on the architecture of SOPC system which can give full play advantages to work in the hardware and software together. This paper introduces and analyzes the image matching methods, which analyzes the matching method of the zero-mean normalized cross-correlation matching and its rapid implementation most. Then this paper presentes a hardware implementation overall framework for the the specific requirements of this algorithm based on the characteristics of software and hardware work together. So this paper designes an image correlator based on SOPC system architecture. Then focuses on the designing the key circuit realization method of the image correlator, gives the circuit module and uses ModelSim to simulate the module to validate the correctness of the function. Finally, in this paper, it realizes the image correlator in a specific hardware platform, and then does some optimizations to the key design of the circuit, tests the system and evaluates its performance. As the system is based on the structure of SOPC architecture and the design is full floating-point, the image correlator has the performances of high precision and configurable. Keywords: Image matching, Zero-mean normalized cross-correlation, SOPC, Image correlator  独创性声明 本人声明所呈交的学位论文是我个人在导师的指导下进行的研究工作及取得的 研究成果。据我所知，除文中已标明引用的内容外，本论文不包含任何其他人或集 体已经发表或撰写过的研究成果。对本文的研究做出贡献的个人和集体，均已在文 中以明确方式标明。本人完全意识到本声明的法律结果由本人承担。 学位论文作者签名： 日期： 年 月 日 学位论文版权使用授权书 本学位论文作者完全了解学校有关保留