顾及纹理信息的遥感图像NSCT域自适应阈值去噪

发布时间：2018-07-12 20:28

本文选题：NSCT + 自适应阈值　；参考：《遥感技术与应用》2017年03期

【摘要】：非下采样Contourlet变换(NSCT)不仅是一种多分辨率、局域的、多方向的图像变换方法,而且具有平移不变性。本文在其基础上对遥感图像去噪算法进行了研究,基于贝叶斯萎缩阈值法提出了一种自适应阈值选取方法,利用模糊集理论构造了一种自适应模糊阈值函数。分别在原始图像中添加了不同标准差的高斯白噪声和椒盐噪声,采用不同的去噪算法作为对比进行了实验。结果表明:该方法不仅能有效提高峰值信噪比,而且能够保持遥感图像丰富的纹理信息。
[Abstract]:Non-downsampling Contourlet transform (NSCT) is not only a multi-resolution, local, multi-direction image transformation method, but also has translation invariance. In this paper, the de-noising algorithm of remote sensing image is studied. Based on Bayesian shrinkage threshold method, an adaptive threshold selection method is proposed, and an adaptive fuzzy threshold function is constructed by using fuzzy set theory. Gao Si white noise and salt and pepper noise with different standard deviations were added to the original image, and different denoising algorithms were used as contrast experiments. The results show that the proposed method can not only improve the PSNR effectively, but also maintain the rich texture information of remote sensing images.
【作者单位】：南京信息工程大学地理与遥感学院测绘工程系;
【基金】：国家自然科学基金项目“城市边缘区地表组分温度反演模型的构建”(41571350)和“基于GPS/PS-DInSAR综合技术的城市大气水汽时空分布特征研究”(41301400)共同资助
【分类号】：TP751

【相似文献】