基于BEMD和QWT的遥感图像去噪算法

Remote sensing image denoising algorithm based on two-dimensional empirical mode decomposition and quaternion wavelet transform

遥感图像在传输过程中容易受噪声影响,需要对其进行去噪处理。在传统遥感图像去噪方法的基础上,设计了基于二维经验模式分解和四元数小波变换的遥感图像去噪算法。首先,对遥感图像进行二维经验模式分解,得到相应的内蕴模式函数和残差分量;然后,用四元数小波变换方法仅对第1个内蕴模式函数进行变换得到新的第1个内蕴模式函数,并进行硬阈值去噪,通过四元数小波逆变换得到第1个内蕴模式函数;最后,通过二维经验模式分解重构得到最终去噪遥感图像。实验表明,所提出的方法有效可行,可以突出遥感图像的细节信息。

Because remote sensing images are easily affected by noise during transmission,it is necessary to denoise remote sensing images.Based on the traditional remote sensing image denoising method,this paper designs a remote sensing image denoising algorithm based on two-dimensional empirical mode decomposition and quaternion wavelet transform.First,the remote sensing image is decomposed by two-dimensional empirical mode,and the corresponding intrinsic mode function and residual component are obtained.Then use the quaternion wavelet transform method to transform only the first intrinsic mode function to obtain the new first intrinsic mode function,and perform hard threshold denoising,and obtain the first intrinsic mode function through the quaternion wavelet inverse transform mode function.Finally,the final denoised remote sensing image is obtained through two-dimensional empirical mode decomposition and reconstruction.The experiment results show that the method is effective and feasible,and can highlight the details of remote sensing images.