计算机工程与设计
計算機工程與設計
계산궤공정여설계
COMPUTER ENGINEERING AND DESIGN
2014年
7期
2453-2457
,共5页
彩色图像%分块压缩感知%小波分解%walsh-hadamard变换%联合重构
綵色圖像%分塊壓縮感知%小波分解%walsh-hadamard變換%聯閤重構
채색도상%분괴압축감지%소파분해%walsh-hadamard변환%연합중구
color image%block compressive sensing%wavelet decomposition%walsh-hadamard transform%joint reconstruction
为了解决传统CS方法中测量矩阵储存量较大、运算速度慢的问题，考虑到图像变换域信息的波动性，提出了一种观测算法。将彩色图像进行RGB三色分解，对每一种颜色进行合理分块，分别对每种块进行相应的小波变换或walsh-hadamard变换，最终对3种颜色进行联合重构。仿真结果表明，该算法使得重构图像质量有了明显提高，验证了该算法的有效性。
為瞭解決傳統CS方法中測量矩陣儲存量較大、運算速度慢的問題，攷慮到圖像變換域信息的波動性，提齣瞭一種觀測算法。將綵色圖像進行RGB三色分解，對每一種顏色進行閤理分塊，分彆對每種塊進行相應的小波變換或walsh-hadamard變換，最終對3種顏色進行聯閤重構。倣真結果錶明，該算法使得重構圖像質量有瞭明顯提高，驗證瞭該算法的有效性。
위료해결전통CS방법중측량구진저존량교대、운산속도만적문제，고필도도상변환역신식적파동성，제출료일충관측산법。장채색도상진행RGB삼색분해，대매일충안색진행합리분괴，분별대매충괴진행상응적소파변환혹walsh-hadamard변환，최종대3충안색진행연합중구。방진결과표명，해산법사득중구도상질량유료명현제고，험증료해산법적유효성。
To solve the gigantic data and low operation efficiency problem of traditional CS algorithm,a modified measurement algorithm was developed to sense color image.The color image was decomposed into three channels of R,G and B,every chan-nel image was segmented into tiny blocks properly.The fluctuation of image in transform domain was taken into account,diffe-rent transforms were employed to measure the tiny block images separately.Last,three color images were reconstructed jointly. In the simulation experiment,the proposed algorithm provided superior performance compared with the state-of-the-art compres-sive sensing reconstruction algorithms.
