Subband coding methods for seismic data compression

Abstract

Summary form only given. A typical seismic analysis scenario involves collection of data by an array of seismometers, transmission over a channel offering limited data rate, and storage of data for analysis. Seismic data analysis is performed for monitoring earthquakes and for planetary exploration as in the planned study of seismic events on Mars. Seismic data compression systems are required to cope with the transmission of vast amounts of data over constrained channels and must be able to accurately reproduce occasional high energy seismic events. We propose a compression algorithm that includes three stages: a decorrelation stage based on subband coding, a quantization stage that introduces a controlled amount of distortion to allow for high compression ratios, and a lossless entropy coding stage based on a simple but efficient block-adaptive arithmetic coding method. Adaptivity to the non-stationary behavior of the waveform is achieved by partitioning the data into blocks which are encoded separately. The compression ratio of the proposed scheme can be set to meet prescribed fidelity requirements, i.e. the waveform can be reproduced with sufficient fidelity for accurate interpretation and analysis. The distortions incurred by this compression scheme are currently being evaluated by several seismologists. Encoding is done with high efficiency due to the low overhead required to specify the parameters of the arithmetic encoder. Rate-distortion performance results on seismic waveforms are presented for various filter banks and numbers of levels of decomposition.