THE WORK ENVIRONMENTS OF earth scientists, engineers, and business planners have turned toward a multisensory future.
In the latter half of the 1990s, the interactive computer workstation, which revolutionized exploration and production in the 1980s, was giving way to various forms of immersive visualization.
Borrowing lavishly from the high-tech arsenal of virtual reality, immersive visualization allowed a worker to seem to move around inside 3D images of the subsurface. And while immersive technology in 1998 concentrated on the visual, workers "immersed" in 3D images were expected eventually to routinely use the sense of touch to study rock textures and motion to examine subsurface characteristics such as dip.
Several oil companies had visualization centers in various stages of development in 1998. Two of them, ARCO Exploration & Production Technology and Texaco Exploration & Production, had built visualization centers and were using them to make business decisions (Figs. 1 and 2).
The ARCO and Texaco centers differed from one another in ways that showed how approaches to visualization vary.
At what it called its Immersive Visualization Environment (IVE) in Plano, Tex., ARCO E&P Technology used a room with five 10-ft sides as a projection cube. The three vertical walls and floor served as projection screens. To workers wearing special eyeware called stereoscopic glasses, images from the four screens seemed to float in the space inside the cube. A worker in the cube thus had the visual sensation of standing within the image. With a technical feature called head-tracking, the image-from the perspective of the worker-changed along with head movement to further enhance the simulation. To the extent the image accurately represented geology, the worker became able to study the subsurface as though standing inside it.
Texaco`s approach to visualization relied less than did ARCO`s on immersion. Its main facility in Houston, called a Visionarium, projected images onto a curved screen 25 ft wide and 8 ft high with a 160° field of view. Users could wear stereoscopic glasses to view images in 3D, but they usually worked from an aspect outside the subsurface representation.
Each approach offered its own advantages and disadvantages. Both of them made use of advanced concepts about human perception and knowledge. These concepts had become crucial in the 1990s as advances in the ability to collect and present geophysical data in three dimensions stimulated development of sophisticated ways to make the quantitative complexity comprehensible to people.
Visualization progress
Visualization has always been near the center of geological and seismic work. Geology employs visual models of subsurface structure and stratigraphy. And seismic work generates images of subsurface features from patterns in sonic reflection and refraction data-starting with a visual representation of data called the trace.
Until the late 1970s, almost all visualization was two-dimensional and occurred on paper records. The basic 2D seismic record makes a vertical, cross-sectional representation of the subsurface. Interpreters interpolate from vertical sections to contour maps, in horizontal perspective, of target structures. In 2D work, perspectives are usually limited to these vertical and horizontal aspects, and vertical sections are confined to locations of seismic lines.
In the 1980s, rapid development of 3D seismic techniques, accompanied by growth in computing power, shifted the focus from paper records to computer screens, on which 3D images more easily could be made to appear three dimensional.
In 3D seismic work, a sonic wave field is sampled over an area rather than along a line. The product of a 3D seismic survey is a volume of data rather than a set of discrete sections related to individual recording lines. And images can be created from any aspect, not just vertical and-through mapping interpolation-horizontal, as in 2D work.
Of course, 3D data can be and usually are presented on flat, 2D paper records and computer screens. Models based on them thus lose a dimension unless the images are somehow adjusted to simulate 3D visual effects. If rendered only in 2D, interpreters must use their imaginations to extrapolate a mental image of 3D reality from the flat image.
The required mental leap creates two large problems. One is that it can require so much concentration two imagine the third dimension that the interpreter can`t focus on geophysical characteristics, which should be the center of attention. The other problem is that several interpreters working on the same geophysical problem often imagine 3D reality in very different ways.
This second problem gained importance in the latter half of the 1990s as companies increasingly assigned interdisciplinary teams to specific exploration and development projects. In the absence of 3D visualization, a single team of professionals working from a common data set bases its discussions and decisions on a number of different mental perceptions of the subsurface.
Through visualization, images on 2D records and screens appear as though they`re in 3D. Interpreters thus don`t need to provide the third dimension with their imaginations, and all interpreters on a team work from a common image.
Visualization solved another problem that developed as 3D seismic techniques advanced. The new techniques and allied computing capabilities greatly increased both subsurface resolution and the types of information that can be extracted from seismic data and used in conjunction with other geophysical information. The advances thus multiplied the number of variables and the level of detail within each variable that an interpreter had to perceive and analyze.
With the interactive computer workstation, interpreters could generate images from a 3D volume of seismic data, combine the data with other types of information, change perspective and scale, and in many other ways manipulate data to achieve the most accurate and understandable visual representation of the information.
The challenge became how to display all that information, in its increasingly fine detail, in such a way that users of the data could keep track of changes and patterns crucial to understanding the subsurface. Advances in visualization technology helped fill the need.
Numbers to images
All visualization techniques essentially turn numbers into images to make them easier than numbers for people to interpret.
A seismic trace is an example. It comes from a string of numbers: measures of sonic-signal strength sampled at specific moments at specific locations and arranged as a function of time. The time is usually two-way travel time-the interval between the sonic impulse and the instant of recording of energy presumed to have been reflected from subsurface horizons.
For a 5 sec record sampled every 4 ms, a single trace is a series of 1,250 numbers. That`s too much for an interpreter to keep in mind, let alone to compare against thousands of other such series of numbers from the same survey in search of patterns.
The basic wiggle trace of seismic work makes use of the keen human ability to identify patterns visually. In the normal trace, the wiggle width is proportional to signal strength (called amplitude), the direction of deflection indicates whether the value is positive or negative, and vertical position along the trace represents time. With a trace, the interpreter takes in the information all at once, free of the need to consider numbers one at a time. And related traces can be arranged side by side-and usually are-so that patterns become evident that would be obscure if the data remained in numerical form.
Modern visualization technology builds on those ideas, taking advantage of the ability of humans to visually discern fine differences in shape, color, and shade and to spot even subtle patterns in complex data. It is a vital part of modern seismic work, in which subtle changes in seismic attributes-such as trace amplitude, phase, or frequency-must be viewed in relation to changes in structural position. Visualization in such a case might provide a 3D model of structure, virtually shaded with artificial lighting to highlight structural relief, and color-code values for the attribute under study.
The user thus views a "picture" of value changes related to structure instead of having to match numbers for the attribute values against other numbers indicating position.
Visualization concepts
According to Geoffrey A. Dorn, research director at ARCO E&P Technology and author of a number of papers on visualization, people use a range of visual "cues" to make sense out of a 3D object. Visualization simulates as many of those cues as possible to make complex 3D objects understandable.
One such cue is projection, or the way in which objects and spatial relationships are portrayed graphically in a 2D plane. In a perspective projection, for example, parallel lines converge at a common point in the distance.
Another cue involves lighting and shading, which help viewers understand shape, incongruities, and texture.
Depth of field, or lack of focus of objects close to or far from an object under study, can indicate relative distance from the viewer and target object.
Related to depth of field, depth cueing depicts distant objects with less intensity than nearer objects.
With obscuration, portions of a 3D object nearest to the viewer block the view of more distant portions. In visualization, the viewer overcomes opacity by controlling the transparency of surfaces at different distances to keep target surfaces in view but retain the 3D context.
Stereopsis is the core principle of the stereoscopic glasses commonly used to create 3D visual effects. It is the ability to compose a 3D image by viewing something from two slightly different directions-as from one person`s two eyes. A stereo viewer rapidly alternates between two images, one for each eye, to simulate 3D effects.
Parallax, another cue, makes close objects seem displaced in relation to far objects when viewed from two different locations.
Similarly, motion takes advantage of human vision`s ability to detect small relative changes in objects under study from one view to another. Visualization techniques often simulate motion through animation.
How it works
Fig. 3 shows several visualization concepts at work in applications very common in seismic work.
The left module in the figure uses projection and perspective to show a 3D data volume on a 2D page. The angle and tilt of the volume create the 3D effect. In the right volume, transparency strips away all data except those for horizons characterized by a selected seismic attribute-strong amplitude. Color-coding highlights a further distinction-whether the strong amplitude is measured on a wave peak or trough.
In a rectilinear opaque volume such as the one in the left module, vertical sides are essentially seismic cross sections, and tops and bottoms are time slices-depictions of amplitude strength and polarity for specific reflection times. But in 3D work, map views-those from a horizontal aspect-often show something other than time. A common map view shows amplitudes across a single horizon rather than across a single time. Such a seismic display is called a horizon slice.
Fig. 4 shows the relationships, from (top to bottom) a horizon interpreted on a vertical cross section in a 3D data volume, to a time slice, to a horizon slice, to an artificially shaded 3D model of the horizon slice.
These are simple images of types routinely generated and manipulated on interactive computer workstations. They obviously provide a much clearer and more-detailed representation of the subsurface than do the cross sections and maps yielded by 2D work.
Whatever their clarity, detail, and three-dimensionalism,however, images on computer screens still fail to make maximum use of human perception.
An important limitation is scale. Even the largest computer screens amount to peepholes in relation to subsurface features. When images are enlarged to approach life scale, so much disappears from the computer screen that there may be no possible context for interpretation. As image scale increases, much disappears from view.
Furthermore, regardless of scale, computer screens and paper records preclude use by interpreters of peripheral vision, a visual cue important to human perception.
Here`s an analogy: Modern photography produces sharp, brilliantly colored pictures. A modern camera loaded with modern film can accurately and in great detail depict a house from a variety of angles and perspectives. Yet no one would buy a house on the basis of a photographic record. Regardless of the quality of such a record, a person cannot learn nearly as much about the house as he or she can by walking through it and experiencing it in life scale through all senses.
Geophysical walk-throughs
Visualization centers attempt to simulate the perceptual equivalent of a geophysical walk-through.
All approaches to visualization at use in 1998 attempted to at least be able to approximate life scale while retaining spatial context and to make use of peripheral vision. ARCO`s IVE did so with an emphasis on immersion; Texaco`s did so with its large, curved screen.
Another approach appeared in the market in 1998, a 3D virtual-reality display unit called VisionDome developed by Alternate Realities Corp. of Durham, N.C., and offered by GeoQuest, a division of Schlumberger in Houston (Fig. 5). A distinctive feature of VisionDome technology was portability: a unit based on it could be moved from place to place.
GeoQuest said VisionDome used an "industry-unique" lens, which had infinite focus, projecting a 180° panoramic view of images. Users didn`t have to wear special eyeware or headsets.
Users and promoters of all the vision centers emphasized the help the centers give to the work of interdisciplinary asset teams. Barton Payne, senior research scientist at Texaco Exploration & Production, said teams using the Visionarium had identified as many as 15 plays in a day, many more than was possible in a day with conventional technology.
Texaco`s center used a feature called Geoprobe, amounting to a 3D cursor of adjustable dimension and shape, which could course through a data volume and produce various types of image in real time.
Where technology headed
Research and development in 1998 promised to take subsurface perception beyond the visual.
Dorn of ARCO Exploration & Production said technology was straining the data-handling capacity of human vision. Audio, he said, might be a next step. Voice commands, for example, might replace the "space mouse" with which users of the IVE gave computer commands and manipulated images. Also, audio tones could represent seismic characteristics, with high tones, for example, representing high values for a selected attribute and low tones representing low values.
A view of the future for immersive technology emerged at an Energy Logistics International conference on 4D seismic technology in Aberdeen, Scotland, during July 1998. H. Roice Nelson Jr. of Continuum Resources, Houston, compared "flat 2D screen displays with the 3D, real-time, interactive, multisensory, dynamic environments" under development in 1998.
"I do know there will be significant differences" between screen displays and immersive environments, he said, "especially when I can:
- Listen to my data, hearing events of interest behind me and turn to find the data, like a fighter pilot turns his head and focuses to see an enemy aircraft he `hears` in his augmented helmet.
- Fly along a seismic horizon and with my vestibular sense of motion feel a subtle change of less than half a degree in the dip of the beds, a change which could represent an invisible fault or lithology change and which I can neither see nor hear.
- Wear haptic devices which allow me to feel geopressure in a seismic volume, to examine temperature changes due to the heat insulation of gas, to grab a virtual proposed well bore with my hand and move the well path to test a bypass pay, and feel force feedback if the driller can neither physically nor economically make the change.
- Drag well bores through a reservoir in shapes obeying optimal well trajectory constraints, through secondary and tertiary targets while minimizing geomechanical stress, all at optimized minimum drilling costs.
- Use computational steering and effective collaboration to optimally match production with a reservoir simulator.
- Pass a significant reservoir development project to another team in Perth, who 8 hr later pass it to a team in London, who 8 hr later deliver it back to me with a detailed object-by-object documentation of progress.
- Use an NT workstation at my house with gigabytes of texture memory, allowing a scalability of computing resources across the home, office, and research environments."
In 1998, Nelson`s view of the future seemed not so very distant.
 |
With stereoscopic glasses and head-tracking, workers in ARCO Exploration & Production Technology`s Immersive Visualization Environment could seem to work inside images generated from geophysical data (Fig. 1).
 |
At Texaco Exploration & Production`s Visionarium, a worker can use a 3D probe to render a surface of interest from a 3D volume of seismic data (Fig. 2).
 |
 |
 |
Inside a VisionDome, geoscientists study a gas-bearing reservoir. The system is portable and can be configured in sizes of 4, 5, and 7 m (Fig. 5).
