化工自动化及仪表
化工自動化及儀錶
화공자동화급의표
Control and Instruments in Chemical Industry
2015年
11期
1206-1210,1267
,共6页
张岩%任伟建%唐国维
張巖%任偉建%唐國維
장암%임위건%당국유
图像重构算法%压缩感知技术%Curvelet变换%稀疏表示
圖像重構算法%壓縮感知技術%Curvelet變換%稀疏錶示
도상중구산법%압축감지기술%Curvelet변환%희소표시
image reconstruction algorithm%compressed sensing technology%Curvelet transformation%sparse representation
提出一种基于Curvelet稀疏表示的图像分块压缩感知重构算法，分析了基于Curvelet稀疏表示的图像各尺度之间系数的相关性，结合分块压缩感知技术设计多尺度相关的双变量收缩阈值迭代重构方法，解决了Curvelet变换高频仍存在大量细节信息，而且各尺度之间的Curvelet变换系数相关性较强等问题。实验证明：该算法能更好地保持重构图像的细节信息，提高峰值信噪比。
提齣一種基于Curvelet稀疏錶示的圖像分塊壓縮感知重構算法，分析瞭基于Curvelet稀疏錶示的圖像各呎度之間繫數的相關性，結閤分塊壓縮感知技術設計多呎度相關的雙變量收縮閾值迭代重構方法，解決瞭Curvelet變換高頻仍存在大量細節信息，而且各呎度之間的Curvelet變換繫數相關性較彊等問題。實驗證明：該算法能更好地保持重構圖像的細節信息，提高峰值信譟比。
제출일충기우Curvelet희소표시적도상분괴압축감지중구산법，분석료기우Curvelet희소표시적도상각척도지간계수적상관성，결합분괴압축감지기술설계다척도상관적쌍변량수축역치질대중구방법，해결료Curvelet변환고빈잉존재대량세절신식，이차각척도지간적Curvelet변환계수상관성교강등문제。실험증명：해산법능경호지보지중구도상적세절신식，제고봉치신조비。
The Curvelet sparse representation-based algorithm for image compressed sensing reconstruction was proposed and the multi-scale correlation among different scales was analyzed.Basing on the block compressed sensing technology, the bivariate shrinkage threshold iterative reconstruction method designed for the multi-scale correlation can keep details in Curvelet high-frequency transformation and have strong correlation of Cur-velet transformation coefficients between each scale considered.Experimental results show that this algorithm can hold the details of reconstructed images and improve peak signal to noise ratio.
