【作者】 王星；

【导师】 马学文；

【作者基本信息】 东北大学 ， 电路与系统， 2013， 硕士

【摘要】 压缩感知理论针对稀疏信号进行采样的同时完成数据压缩,从而节约了大量的计算资源、存储资源和传输资源,这使得其在信号处理领域有着突出的优点和广阔的应用前景。压缩感知理论主要包括信号的稀疏变换、观测矩阵的设计和信号的重构三个方面的内容。这里将重点研究最为重要的信号重构问题。由于压缩感知理论,尤其是信号的重构面临着大数据量的运算,而图像作为二维信号,其涉及的重构算法运算量将更加庞大。在此将FPGA引入图像重构算法的研究,利用其强大的并行运算能力,可以有效地解决大数据量运算的问题。本文是基于压缩感知理论的图像重构系统,在ALTERA公司DE2平台设计完成。系统使用的FPGA是Cyclone Ⅱ系列的EP2C35。整个系统的设计采用了软硬件协同的方法。硬件方面：在系统的总体需求分析的基础上,通过对系统各个硬件模块的设计,完成了对NIOS Ⅱ处理器的定制。系统主要包括图像重构模块和图像显示模块。软件方面：在PC机上用Matlab实现稀疏变换方法和图像重构算法来评价算法性能,通过对图像重构算法效果的对比分析,确定用DWT算法和OMP算法实现图像重构。在Nios Ⅱ IDE环境下采用C语言对算法进行编程并调试,最后对整个系统使用的资源情况进行分析,对图像重构的效果进行评测。通过对硬件模块和软件程序的协同调试,使其达到预期效果。测试结果表明,在DE2开发平台上实现的图像重构系统,能够有效地实现图像重构。在OMP算法复杂的迭代部分采用硬件语言VHDL实现的矩阵乘法器进行运算,极大地提高了图像重构效率。更多还原

【Abstract】 Compressed sensing theory for the sparse signal, which is sampled at the same time to complete data compression, thus saving a lot of computing resources, storage resources and transfer resources, therefore, the compressed sensing theory in signal processing field has prominent advantages and broad application prospects. Compressed sensing theory mainly includes three parts:signal sparse transformation, the design of the observation matrix and the signal reconstruction. This study will focus on the most important problem on signal reconstruction. Since the compressed sensing theory, especially the reconstruction of signal is facing a large amount of data operations, and the image as a two-dimensional signal, which involves the reconstruction algorithm computation will be more substantial. The FPGA will be introduced in the image reconstruction algorithm, using its powerful parallel computing ability can effectively solve the problem of large amounts of data operations.This article is based on the theory of compressed sensing of image reconstruction system, designed and completed on the ALTERA DE2platform. System using a Cyclone Ⅱ series EP2C35FPGA, the whole system was designed with collaborative software and hardware implementations. Hardware aspects:on the basis of the analysis of the system requirements, through the design of each hardware module of the system, complete the custom of NIOS Ⅱ processor. System mainly includes image reconstruction module, image display module. Software aspects:on the PC using MATLAB to simulate sparse transform method and image reconstruction algorithm to evaluate the performance of the algorithm, through the contrastive analysis of image reconstruction algorithm, determine the DWT algorithm and OMP algorithm is used to implement image reconstruction. The NIOS Ⅱ IDE environment using C language programming and debugging of the algorithm, finally analyzing the whole system resources and evaluating the effect of image reconstruction.Through collaborative debugging of hardware modules and software program to reach the image reconstruction system expected effects.The testing results show that the image reconstruction system, realized on the DE2 development platform, can effectively realize the image reconstruction. For the OMP algorithm complex iterative parts, using a matrix multiplier, which is realized by using VHDL hardware language operation, greatly improve the efficiency of image reconstruction.更多还原