基于NSCT变换的遥感图像快速自适应去噪方法  

A Fast Adaptive Denoising Method for Remote Sensing Images Based on NSCT Transform

作　　者：刘洋 张国军 LIU Yang;ZHANG Guo-jun(University of Chinese Academy of Sciences,Beijing 100000,China)

机构地区：[1]中国科学院大学,北京100000

出　　处：《计算机仿真》2022年第1期172-176,共5页Computer Simulation

摘　　要：针对遥感图像受噪声影响导致质量较低的问题,提出了基于NSCT变换的遥感图像快速自适应去噪方法。将Gibbs效应添加至Contourlet变换的图像处理环节,利用非下采样拉普拉斯金字塔与方向滤波器组构建NSCT变换。根据NSCT变换后遥感图像信号能量汇聚至小部分变换系数的特征,利用硬阈值快速去噪方法和NSCT变换系数邻域的一致性特征,推导出遥感图像自适应阈值,将其与硬阈值快速去噪法融合后,实现NSCT变换下快速自适应去噪。仿真结果表明,所提方法去噪优势显著,且用时更短,具有较好的有效性与稳定性。In fact, because noise affects the quality of remote sensing images, this paper reports a fast adaptive denoising method of remote sensing images based on NSCT transform. Gibbs effect was added to the image of Contourlet transform, and non-down sampled Laplacian pyramid and directional filter bank were applied to construct NSCT transform. The signal energy of the remote sensing image converged to a small part of the transformation coefficients after NSCT transformation. The hard threshold fast denoising method and the consistency characteristics of the neighborhood of NSCT transform coefficients were used to deduce the adaptive threshold of the remote sensing image. Concurrently, the threshold was combined with the hard threshold fast denoising method to complete the fast adaptive denoising under NSCT transform. The simulation results show that the method has excellent effectiveness, stability and short time-consuming.

关 键 词：变换 遥感图像 快速 自适应 去噪 矩阵标准差 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]