A non-local mean algorithm for ship image denoising based on SLIC superpixel segmentation
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
万方数据
针对传统船舶图像去噪算法难以对图像的边缘细节进行保留和分析,以及传统非局部均值去噪算法相似框选择困难等问题,提出基于简单线性迭代聚类(simple linear iterative clustering,SLIC)超像素分割的非局部均值船舶图像去噪算法.通过SLIC算法对图像进行分割处理,界定图像的纹理区域和平滑区域;使用相似框搜索和匹配策略,提升匹配效果,并适当保留更多边缘细节,从而改善图像去噪的效果.实验结果表明,所提出的算法相较于其他传统的船舶图像去噪算法不仅能很好地保留船舶图像的边缘细节特点,而且能在一定程度上提高船舶图像的峰值信噪比,具有良好的去噪效果,可以用于智能航海领域船舶图像的去噪.
Aiming at the problems that it is difficult for the traditional ship image denoising algorithm to retain and analyze edge details of images,and it is difficult for the traditional non-local mean denoising algorithm to select the similar box,a non-local mean ship image denoising algorithm based on SLIC(simple linear iterative clustering)superpixel segmentation is proposed.The SLIC algorithm is used for image segmentation to define the texture region and the smooth region.The search and matching strategy of similar boxes is used to improve the matching effect and appropriately retain more edge details,so as to improve the effect of image denoising.The experimental results show that,compared with other traditional ship image denoising algorithms,the proposed algorithm can not only retain the edge detail characteristics of ship images,but also improve the peak signal-to-noise ratio of ship images to a certain extent.The proposed algorithm is of good denoising effect and can be used for the denoising of ship images in the field of intelligent navigation.
non-local mean denoisingship image denoisingsimple linear iterative clustering(SLIC)superpixel segmentationsimilar box selection
NETL
NSTL
万方数据
