CAJ | 학술논문

由于多尺度小波变换的分块压缩感知算法（MS-BCS-SPL）将每层子带信息进行分块时，使得每层子带中各子块间的采样率相同；但是，当不同的图像子块含有不同的边缘信息时，对这些子块采用相同的采样率会造成资源分配不合理。因此在 MS-BCS-SPL 算法的基础上，利用图像块边缘信息的不同和图像块的方向性，将总的采样率自适应分配给各层子带中的各子块，实现多尺度分块压缩感知的自适应采样。实验结果表明，在不同采样率，尤其较低采样率时，该算法不仅比 MS-BCS-SPL 算法采用了较少的采样数目，节约资源；而且比其可重构较高质量的图像。
유우다척도소파변환적분괴압축감지산법（MS-BCS-SPL）장매층자대신식진행분괴시，사득매층자대중각자괴간적채양솔상동；단시，당불동적도상자괴함유불동적변연신식시，대저사자괴채용상동적채양솔회조성자원분배불합리。인차재 MS-BCS-SPL 산법적기출상，이용도상괴변연신식적불동화도상괴적방향성，장총적채양솔자괄응분배급각층자대중적각자괴，실현다척도분괴압축감지적자괄응채양。실험결과표명，재불동채양솔，우기교저채양솔시，해산법불부비 MS-BCS-SPL 산법채용료교소적채양수목，절약자원；이차비기가중구교고질량적도상。
Same sampling rate of each sub-block in each sub-band is galned when the multi-scale block compressed sens-ing algorithm (MS-BCS-SPL)subdivides sub-band information of each layer into blocks.But unreasonable resource distri-bution would be caused if image sub-blocks have different edge information.In this thesis,on basis of MS-BCS-SPL,multi-scale block compressed sensing adaptive sampling is realized by distributing total sampling rate into all sub-block of each sub-band,which uses the difference of edge information and directivity of image blocks.First the edge information of image blocks is calculated and the direction is estimated,through which the total sampling rate is assigned adaptively into each block of each sub-band by the characteristics of the wavelet transform,realizing adaptive multi-scale block compressed sensing.It is shown by experiments that compared with algorithm of MS-BCS-SPL,by using adaptive multi-scale block compressed sensing algorithm,with different sampling rate especially lower sampling rate,not only less samples is used and resources are saved,also high quality images are reconstructed and reasonable resources distribution is achieved.