CAJ | 학술논문

在分析核范数基础上，提出基于加权鲁棒主成分分析（WRPCA ）的非局部去噪方法。将加权核范数引入鲁棒主成份分析模型，构建加权鲁棒主成份分析模型（WRPCA ），采用增广拉格朗日乘子法对模型进行求解，将WRPCA用于图像去噪。根据图像的自相似性，对噪声图像进行分块，通过块匹配法对图像块进行聚类，获得相似块组矩阵；通过加权鲁棒主成分分析（WRPCA ）算法对相似块组矩阵进行低秩矩阵恢复。实验结果表明，无论对低噪声图像和高噪声图像，该方法去噪效果相比现有的经典算法都有一定提高。WRPCA算法对图像结构保持有很好效果，在保持图像纹理细节方面优于其它去噪算法。
재분석핵범수기출상，제출기우가권로봉주성분분석（WRPCA ）적비국부거조방법。장가권핵범수인입로봉주성빈분석모형，구건가권로봉주성빈분석모형（WRPCA ），채용증엄랍격랑일승자법대모형진행구해，장WRPCA용우도상거조。근거도상적자상사성，대조성도상진행분괴，통과괴필배법대도상괴진행취류，획득상사괴조구진；통과가권로봉주성분분석（WRPCA ）산법대상사괴조구진진행저질구진회복。실험결과표명，무론대저조성도상화고조성도상，해방법거조효과상비현유적경전산법도유일정제고。WRPCA산법대도상결구보지유흔호효과，재보지도상문리세절방면우우기타거조산법。
Based on analyzing nuclear norm ,an approach of non‐local denoising was ,presented by adding weighted nuclear norm to robust principal component analysis ,a model of weighted robust principal component analysis (WRPCA) was obtain ,and image denoising was achieved by using augmented Lagrangian multiplier method to solve the model .According to the self‐similarity of the image ,the noise image was divided into blocks ,and then the image blocks were clustered using block matching method ,similarity matrix blocks were obtained simultaneously .After that ,WRPCA for similarity matrix blocks was used to recovery the low rank matrix .Compared with the existing classic algorithms on both high noise images and low noise images ,the proposed method shows better denoising effect , and it is more robust . Meanwhile , WRPCA has better performance on preserving both image structure and image texture details .