论文部分内容阅读
常规小波阈值去噪方法未能充分利用地震信号相关性的特点进行去噪,为此在多层小波变换中引入了双变量概率分布模型。基于贝叶斯估计理论,得到了相应的双变量收缩函数;基于层内局域方差估计,得到了一种局域自适应去噪算法。在实验中,将该算法分别应用于实值离散小波变换域和复数小波变换域,并和隐马尔科夫模型的去噪方法进行了比较。图像处理和地震模型测试结果表明,复数小波变换的局域自适应收缩算法去噪效果最好。
Conventional wavelet threshold denoising method can not make full use of the characteristics of seismic signal to denoise. Therefore, a bivariate probability distribution model is introduced in the multi-layer wavelet transform. Based on the Bayesian estimation theory, the corresponding bivariate shrinkage function is obtained. Based on the intra-layer local variance estimation, a local adaptive denoising algorithm is obtained. In the experiment, the algorithm was applied to real discrete wavelet transform domain and complex wavelet transform domain, and compared with the Hidden Markov model denoising method. The results of image processing and seismic model test show that the local adaptive shrinkage algorithm of complex wavelet transform has the best performance of denoising.