IEEE Signal Processing - March 2018 - 84

on human visual attention models, mimicking the process of a
human interpreter perceiving and analyzing a large volume of a
seismic data set. We hope that this article will make the challenges found in seismic interpretation more accessible to the signal
processing community. The start-up tools and data sets provided
here will enable the interested readers to get a jump-start in this
domain, especially if they do not have a background in geophysics. Furthermore, analyzing the seismic volume from the HVS
point of view can unveil excellent opportunities that may not
exist otherwise. Finally, solving challenges that are rooted in the
unique characteristics of seismic data will help advance the corresponding signal processing and machine-learning theories and
algorithms as well.

Subsurface structures and data sets
Before proceeding to the review of common interpretation tasks,
in this section, we provide a brief introduction to several subsurface structures of interest, followed by a list of data sets that
are commonly used for developing and testing the algorithms.
Throughout this article, we will use oil and gas exploration as
the example to illustrate subsurface understanding through seismic interpretation for a number of reasons. Compared to other
applications, data is more abundantly available with oil and gas
exploration. Also, there is rich literature on this subject with various contributions that can provide a benchmark. Furthermore,
the depth of the imaged subsurface is deeper in the application of
consideration, and that adds to the complexity of the structure to
be interpreted. Therefore, the structures and data sets introduced
here are typically related to hydrocarbon reservoir identification
and characterization, although some of them can also be used

for other applications such as earthquake monitoring and environmental engineering, which is a common practice in the community. Nevertheless, all reviewed and proposed algorithms are
applicable in all applications that image the subsurface to locate
certain structures.

Common subsurface structures
Subsurface structures are complex because of the massive
geologic evolution and deformation over millions of years. A
migrated seismic volume, therefore, can contain multitudes of
geologic structures such as horizons, unconformities, faults, salt
domes, channels, and gas chimneys. Horizons, represented as
seismic reflections, are the dominant geologic structures apparent in a seismic volume while the other structures can often be
recognized as discontinuities or edges of seismic reflections.
We will focus the discussion in this article on the interpretation
of faults, salt domes, channels, and gas chimneys. Examples are
shown in Figure 2. All of these structures are of great geological implications for hydrocarbon migration and accumulation
as discussed next.

Faults
A fault is defined as a lineament or planar surface across which
apparent relative displacement occurs in the rocks' layers. The
movement of impermeable rocks and sediments along the fault
surface creates membranes that hinder the migration of hydrocarbons from source rocks and create structural hydrocarbon
traps. Because of the lateral changes in texture across a fault,
interpreting such structures can be treated as an edge detection
problem that is common in digital image processing. However, different from edges in natural
images, faults do not always display
explicit edges, and the visual appearance is typically noisy.

Salt domes
A salt dome is defined as a dome-shaped
structure formed by the evaporation of a
large mass of salt in sedimentary rocks.
Salt domes are impermeable structures
that prevent the migration of hydrocarbons and provide entrapment for oil and
gas reservoirs. Chaotic reflections are
often observed in a salt dome in a form
of a distinct texture; thereby interpreting such structures can be treated as a
texture segmentation problem in digital
image processing. Nevertheless, in salt
domes, the boundaries are not explicit
due to the underlying physics as well as
noise and low-resolution data.

Fault
Salt Dome

Structure of
Interest
Channel
Gas Chimney

Channels
A channel is defined as a remnant of
an inactive river or a stream channel
that has been either filled or buried by

Figure 2. Examples of four types of subsurface structures essential for hydrocarbon exploration.
84

IEEE Signal Processing Magazine

March 2018

Table of Contents for the Digital Edition of IEEE Signal Processing - March 2018