多分辨重叠双正交变换在地震数据压缩中的应用

针对地震数据中信号和噪声的特点，基于多分辨重叠双正交变换(HLBT)，提出了一种新的地震数据压缩方法。HLBT方法的优势是，对信号频谱的非均匀划分可以更加有效地将能量集中在有限的变换域子带上。对多分辨重叠双正交变换的原理进行了描述；给出了地震数据压缩的实现方法，即先对地震数据进行二维重叠变换，再进行量化和系数组织，最后采用自适应算术编码对量化后的数据流进行无损熵编码。利用新的压缩方法对实际地震数据进行了压缩处理，并与小波变换压缩方法进行了比较，结果表明，对于高倍率的压缩，其效果要优于小波压缩方法。同时，对解压信号质量的评价标准进行了初步探讨，提出了形态逼真度的概念。

The application of hierarchical lapped biorthogonal trans- form in seismic data compression

Compression is a relatively new introduced technique for seismic data operations. The main drive behind the use of data compression in seismic data is the very big size of seismic data acquired. Up to now some of the successful algorithms used for seismic data compression were based on wavelet or local cosine transform. Recent works have show that Lapped Orthogonal Transform (LOT) is a very effective technique for compressing seismic data. Difference from the LOT, Hierarchical Lapped Biorthogonal Transform (HLBT) can effectively isolate the energy to a small number of bins in the transform domain by nonuniform division of the signal spectrum. We propose an efficient seismic data compression algorithm. Besides compression, there is a still problem of signal quality evaluation. We present a new index named Morphological Fidelity that can measure the wave shape distortion. Using this algorithm to compress seismic data, and find it provides superior performance to wavelets in low-bits compression.