基于压缩传感的重构算法研究

Research of Compressive Sensing Reconstruction Algorithm

传统的奈奎斯特采样定理规定采样率必须是频率带宽两倍,浪费大量采样资源。如果信号可以稀疏表示,那么可以采用压缩传感技术重构原始信号,压缩传感能在采样的同时对数据进行适当压缩,节省系统资源。现存的压缩传感重构算法对图像边缘和纹理的重构效果都不太理想,提出一种基于全变差的图像重构算法,该算法能稳定有效地重构图像的边缘和纹理。

The rate of sampling a signal must be at least twice the bandwidth of its frequency according to the classical Nyquist theorem. It will cost large amount of sampling resource. If the signal has sparse representation, compressive sensing can be used to reconstruct the original signal. CS has the ability to compress signal during the process of sampling. The existed CS algorithms of reconstruction can＇ t recover image＇ s edges and texture ideally. In this paper, a new image reconstruction algorithm based on Total Variation is proposed, which can recover image＇ s edges and texture effectively and stably.