基于WBCT的地震数据感兴趣区域编码算法

提出一种基于小波域Contourlet变换(WBCT)最优截断嵌入码块(EBCOT)的地震数据压缩算法.首先利用WBCT获得地震数据的稀疏表示,根据码块熵最小化原则优化WBCT方向分解,降低码块复杂度,提高编码效率;然后针对不同地质信息对小波各子带的带通特性不同,改进JPEG2000算法感兴趣区域掩模处理方式,对不同的小波子带设置不同权值,进行掩模处理;最后通过EBCOT编码生成压缩码流.实验结果表明,重构的地震数据不仅峰值信噪比(PSNR)得到提高,而且地震数据同向轴局部纹理失真明显减小.

Seismic Data Coding Algorithm with Region of Interest Based on WBCT

We proposed a seismic data compression algorithm based on wavelet based Contourlet transform (WBCT) with embedded block with coding optimized (EBCOT). Firstly, the WBCT was used to obtain the sparse representation for seismic data, and the WBCT direction decomposition was optimized according to the rule of code block entropy minimization, which reduced the complexity of the code blocks and improved the coding efficiency. Secondly, in view of the bandpass features of different geological information in wavelet domain, we improvedthe method of mask treatment of the region of interest (ROI) of JPEG2000 algorithm, set different weights to the different wavelet subbands for mask treatment. Finally, the compressed bit stream was generated by EBCOT encoding. The experimental results show that thereconstructions of seismic data not only improve the PSNR, but also significantly reduce the distortions of local texture of seismic data with the same axis.