压缩传感理论的研究与实现.docVIP

传感压缩理论的研究与实现 摘　要 当今信息社会需要存储压缩的数据量不断增长，而信息存储的硬件空间增长速度有限，数据压缩图像压缩也就日益重要。根据奈奎斯特采样定理，传统的采样频率不得低于信号最高频率的2倍。但是对于数字图像等信息，传统的采样会导致海量采样，意味着还需要对海量的数据进行后续处理，大大增加了存储和传输的代价。 本文对传感压缩理论进行研究，传感压缩采用较低的采样频率对数字信号进行采样并压缩，将采样和压缩合并为一步而进行，节省了数字信号采样开销、存储开销和传输开销。传感压缩采用非自适应线性投影来保持信号的原始结果，能通过数值最优化问题确定重构原始信号，而且重构信号能保证较高精度。传感压缩不仅可以用在一般数据信号的压缩上，也可以用在图像信号的压缩上。本文利用matlab仿真了传感压缩对图像压缩的存储与重构，将传感压缩算法与JPEG压缩算法做了实验对比，进一步分析传感压缩的应用前景。 在传感压缩理论和多种编码压缩算法的研究基础上，本文对二维图像压缩的新方法进行探索，利用matlab进行实验探索，进一步分析新算法改进方案。文内称此图像压缩的新方法为相似块删除算法。 关键词：传感压缩、图像压缩、采样、重构信号 Todays information society need to store the growing amount of data compression, and information storage growth rate of the hardware space is limited, data compression, image compression will become increasingly important. According to Nyquist sampling theorem, the traditional sampling frequency not less than 2 times the maximum frequency of the signal. But for digital images and other information, the traditional sampling will lead to mass sampling, means that vast amounts of data also need to follow-up treatment, greatly increasing the storage and transmission costs. In this paper, the theory of compressed sensing, sensing a lower sampling frequency of compressed digital signal sampling and compression, and compression into the sample carried a step to save the digital signal sampling overhead, storage overheadand transmission overhead. Non-adaptive compressed sensing to maintain a linear projection of the original signal results determined through numerical optimization problem to reconstruct the original signal and reconstructed signal can guaranteehigh accuracy. Compressed sensing data not only can be used in general signal compression, but also can be used in the image signal compression. In this paper,MATLAB simulation of the sensor storage compression for image compression and reconstruction, the sensor with the JPEG compression algorithm compression algorithm to do the experimental comparison, further analysis of the compressed sensing applications.