Page 322 - Acquisition and Processing of Marine Seismic Data

CHAPTER

6

Deconvolution

OUTLINE

6.1 Convolutional Model 316 6.4 Predictive Deconvolution 335
6.2 Assumptions for Deconvolution 319 6.5 Determination of Deconvolution
6.2.1 Assumption 1: A Simple Earth Parameters 337
Model 320 6.5.1 Autocorrelation Time Gate 339
6.2.2 Assumption 2: Stationary Wavelet 320 6.5.2 Deconvolution Design Gate 340
6.2.3 Assumption 3: Noise Component 320 6.5.3 Operator Length 341
6.2.4 Assumption 4: Source Waveform 320 6.5.4 Prediction Lag 341
6.2.5 Assumption 5: Wavelet Causality 320 6.6 Poststack Deconvolution 349
6.2.6 Assumption 6: Random
Reflectivity 321 6.7 Maximum Entropy (Burg)
Deconvolution 349
6.3 Spiking Deconvolution 324
6.3.1 Deconvolution With Inverse 6.8 Shaping Filters 350
Filter 325
6.9 Surface Consistent Deconvolution 356
6.3.2 Inverse Filtering With Least
Squares 328 6.10 QC in Deconvolution 357
6.3.3 Optimum Wiener Filters 330
6.3.4 Prewhitening 334

By its simplest definition, deconvolution is the the source signal from the seismic trace. In seis-
inverse of the convolution process. In the convo- mic data processing, it is widely used for improv-
lutional model theory explained in Section 6.1, ing the temporal resolution of the seismic data,
the seismic trace is formed by a simple convolu- since it is always preferred to work with seismic
tion of the source signal and the earth’s reflectiv- data having a broad amplitude spectrum, includ-
ity series in depth. Deconvolution tries to remove ing frequencies both low and as high as possible,

Acquisition and Processing of Marine Seismic Data 313 # 2018 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/B978-0-12-811490-2.00006-2