基于空间自适应变分的条带噪声去除方法

发布时间：2018-01-06 21:31

本文关键词：基于空间自适应变分的条带噪声去除方法　出处：《激光与红外》2016年05期 　论文类型：期刊论文

【摘要】：遥感图像经常被条带噪声污染,导致图像质量下降。为了去除条带噪声,本文提出了一种基于空间自适应变分的噪声抑制方法。首先,对条带噪声建模,并利用信噪比较高的区域估计出模型的增益和偏置;然后,在变分法的框架下,构建能量函数,并引入空间自适应正则因子,根据图像空间信息自适应调整正则参数;最后,采用分裂式Bregman算法优化能量函数,得到去噪图像的最优解。实验表明,本文算法可以将实际遥感图像等效视数由37.26提高到76.48,辐射质量提升因子提高到8.52 d B。本文算法能够有效去除条带噪声,保留图像细节,改善图像质量。
[Abstract]:Remote sensing images are often corrupted by noise band, leads to a decrease in image quality. In order to remove the stripe noise, this paper presents a method of noise suppression based on adaptive variational space. First of all, on the strip noise modeling, model and estimate the gain and offset the relatively high signal-to-noise area using; then, in the framework of change the construction method, energy function, and introduce the space adaptive regularization factor, according to the spatial information of image adaptive regularization parameters; finally, using split Bregman algorithm to optimize the energy function, the optimal solution of the image denoising. Experiment shows that this algorithm can be the actual remote sensing images of equivalent number increased from 37.26 to 76.48. To improve the quality of radiation factor is up to 8.52 D B. this algorithm can effectively remove the stripe noise and keep the image details and improve image quality.

【作者单位】：长春工业大学;
【分类号】：TP751
【正文快照】： 1引言推扫式成像系统经常受到条带噪声的影响[1-2]。条带噪声主要源于不同像元之间的辐射响应非均匀性[3]。条带噪声的存在,影响图像的视觉效果,降低图像的信噪比和后续数据分析的精度[4-5]。因此,研究有效去除条带噪声算法至关重要。去除条带噪声算法主要可以分为以下四类。

【相似文献】