中文摘要：

压缩感知理论突破了奈奎斯特采样定理的局限，可以极大地减小采集的数据量．压缩感知算法一般假设压缩感知所在的稀疏域中每个位置上信号等概率出现，而事实上，信号在各位置的出现概率差异很大．通过对图像正交变换域内正交系数分布情况的研究，提出了一种基于概率对稀疏域划分的压缩感知方法．该方法可以有效地降低稀疏域的稀疏度，提高恢复图像的质量．与现有算法相比，在采样率相同的情况下，这种算法的重构速度提高了20～60倍，同时不会对图像质量产生负面影响．

英文摘要：

With the development of Compressive Sensing theory in recent years, many new algorithms have been introduced to this field. But still, these algorithms tend to judge the probability of the nonzero signal in each position of the sparse domain as the same, which is in fact not true. In this topic we discuss orthogonal coefficient distribution and divide the whole sparse domain into different parts using probability. With the method called Sparse domain Division using Probability （SDP）, the reconstructed speed would increase 20-60 times without producing any negative effect on image quality at the same sampling rate.