海洋测绘2025，Vol.45Issue(5)：11-15,30,6.DOI:10.3969/j.issn.1671-3044.2025.05.003
基于非局部贝叶斯稀疏的散斑噪声降噪
Research on speckle noise denoising based on non-local Bayesian sparse
摘要
Abstract
Sonar images are often severely degraded by multiplicative speckle noise caused by seabed reverberation,which significantly affects image quality and target recognition performance.To address this issue,we propose an image denoising algorithm that integrates non-local similarity-based clustering with Bayesian sparse representation.The proposed method first applies a logarithmic transformation to convert the multiplicative noise into an additive form.Then,an equivalent-view-weighted logarithmic dissimilarity metric is introduced,combined with an improved K-means clustering algorithm,to achieve accurate classification of image patches and enhance intra-class noise consistency.Based on this,a Bayesian Orthogonal Matching Pursuit algorithm is employed for sparse reconstruction,which effectively models the noise characteristics while preserving edge and texture details.Finally,the denoised image is recovered through an inverse transformation.Extensive comparative experiments demonstrate that the proposed method not only significantly suppresses multiplicative speckle noise but also effectively preserves important structural details such as edges and textures.Both quantitative evaluation metrics and subjective visual assessments confirm its clear advantages over existing methods.关键词
声纳图像/乘性散斑噪声/非局部相似性/K-means聚类/稀疏重构Key words
sonar image/multiplicative speckle noise/non-local similarity/K-means clustering/sparse reconstruction分类
天文与地球科学引用本文复制引用
保德彪,邢传玺,黄廷龙,孟轶涵..基于非局部贝叶斯稀疏的散斑噪声降噪[J].海洋测绘,2025,45(5):11-15,30,6.基金项目
国家自然科学基金(61761048) （61761048）
云南省应用基础研究基金(202201AU070047) （202201AU070047）
云南省基础研究专项一般项目(202101AT070132). （202101AT070132）
