New Flexible Segmentation Technique in Seismic Data Compression Using Local Cosine Transform

Best-basis searching algorithm based on binary (in general, M-ary) segmentation was constructed by Coifman and Wickerhauser in 1992 [1] (IEEE Trans. on Information Theory, 38, 713-718). However, there are several problems with the binary scheme. First, the binary segmentation is inflexible in grouping signals along the axis. Secondly, the binary-based segmentation method is very sensitive to time/space shifts of the original signal, such that the resulted best-basis will change a great deal if the signal is shifted by some samples. Thirdly, the reconstruction distortion after compression is relatively strong. In this paper, we design a new flexible segmentation algorithm with arbitrary time/space segmentation resolution which addresses the above-mentioned problems caused by the binary segmentation scheme. This new flexible segmentation algorithm is applied to 2-D seismic data compression with two semi-adaptive schemes: Flexible 2-D time-ALCT (Adapted Local Cosine Transform) and Flexible 2-D space-ALCT. From our numerical tests on both synthetic signals and real seismic data using local cosine transform, the advantages of this new flexible segmentation technique over the binary searching scheme can be easily seen, from overcoming the constraint of dyadic segmentations, reducing time/space-shift sensitivity, less reconstruction distortions to superior performance in seismic data compression.