摘要
压缩感知理论突破了奈奎斯特采样频率的限制,利用该理论研究和实现了二维图像的压缩采样和重建。该方案利用小波变换实现图像稀疏化,利用标准伪随机数均匀分布和二维中心傅里叶变换生成随机测量矩阵,并对小波变换后的高频子带进行加权采样,用改进的基追踪算法实现二维图像压缩感知重建。仿真实验结果表明,该方案重建图像的客观评价PSNR效果较好。
Theory of compressed sensing releases us from the restriction defined by Nyquist sampling frequency.This paper focuses on research and realization of a new compressing and sampling scheme for two dimensional image data based on this theory.Sparseness processing is realized eby using wavelet transform in this scheme,and the random measurement matrix is generated using standard pseudo-random number matrix of uniformly distributed sequences and two dimension central Fourier transform,also weighted sampling is implemented for wavelet transformed high frequency sub-bands.Compressed Sensing reconstruction of two dimensional image is solved using improved basis pursuit algorithm.Simulation experiment shows that better PSNR of objective evaluation of the reconstructed image is achieved.
出处
《电子设计工程》
2011年第11期163-166,共4页
Electronic Design Engineering
基金
国家自然科学基金项目(61072111
60672156)
吉林省科技厅项目(20100503)
关键词
压缩感知
基追踪算法
重建
小波变换
Compressed Sensing
basis pursuit algorithm
reconstruction
wavelet transform
