一种基于轮廓波变换的高效率去噪算法研究

An Efficient Denoising Algorithm Based on Contourlet Transform

为提高地震属性剖面的信噪比和空间分辨率,提出了一种基于轮廓波变换的高效率去噪算法。该方法利用地震剖面能量分布的空间特征,结合9-7正交滤波器组和方向滤波器组对原始采集的地震剖面进行多尺度分解,根据有效地震信号和噪声信号分解后的差异性设计了一种自适应的阈值滤波器,通过拉普拉斯金字塔重构地震数据实现轮廓波变换去噪。试验结果与小波变换算法的去噪结果对比论证表明,基于轮廓波变换的高效率去噪算法能够有效压制噪声,提高地震信号的信噪比,处理后效果较小波变换有明显提高。

In order to improve the signal to noise ratio and spatial resolution of seismic attribute profile,a highly efficient algorithm based on contourlet transform is put forward.The method made full use of seismic energy and spatial distribution of performance characteristics,and combined the 9-7orthogonal filter banks with the directional filter banks,multi scale decomposition is carried out for the original acquisition of seismic profile,then an adaptive threshold filter is designed according to the useful seismic signal and noise signal decomposition differences.The denoising of contourlet is implemented by reconstructing the seismic data through Laplace Pyramid.The result using contourlet transform compared with that of denoising shows that the highly efficient algorithm based on contourlet transform could suppress the noise effectively,improve the signal-to-noise ratio of seismic signal.Compared with the wavelet transform,the treatment effect is improved obviously.