BP Turns To Image Processing for Visualization of Seismic Data
By Jacki Bannon and Paul Batey

BP Exploration Ltd. of London has dramatically expanded the ways in which its exploration staff processes and visualizes data by switching from an out-dated, in-house image processing software to a user friendly commercial product.
      Among the more significant advantages of the transition to state-of-the-art image processing software is the ability to process and visualize a variety of geophysical data using algorithms and programs previously applied only to satellite imagery. The second major benefit is that many personnel within BP's worldwide exploration community are now able to perform complex, multi-data processing that once was the domain of a select few image processing technicians.
      For more than six years, BP had conducted most satellite image processing on its own VAX-based software called IGAS. In 1991, the company upgraded to ER Mapper image processing software developed by Earth Resource Mapping of Perth, Australia. This Unix-based package has allowed BP exploration scientists to integrate several raster and vector data types stored in multiple databases maintained by BP.
      One of the most useful applications of the software has been displaying gridded seismic data and combining it with well information and surface vector data to detect correlations that may pinpoint potential petroleum exploration targets. Using the IGAS system, this process was restricted to remote sensing specialists.
      BP stores its gridded seismic data in a common petroleum exploration package called Z-Map Plus, a product of Landmark Graphics. The ER Mapper software has a built-in command to import Z-Map raster seismic data into the image processing system so the seismic data can be enhanced and displayed similar to the ways satellite images are processed.
      BP has used this feature to import and process seismic horizons in Intensity-Hue-Saturation and pseudo-color displays. This process allows exploration scientists to view the seismic data as a subsurface topography which can be interpreted in much the same way as surface topography.
      These IHS transformations are often used with multispectral satellite imagery such as Landsat or SPOT to maximize the color contrast, thereby visually highlighting very subtle differences in raster values. BP finds this technique can also enhance very slight differences in seismic values and lineations, thereby resulting in a more accurate subsurface interpretation.
      The ER Mapper software is equipped with a Dynamic Link which allows data in other formats to be imported in real-time into the image processing system. For instance, BP keeps extensive well and seismic line information in an ORACLE database. Vector data such as road networks, pipelines, coastlines and infrastructure are contained in ARC/INFO. These data can be imported without changing formats and integrated directly with the image processing data.
      This technique has been used in the North Sea to detect oil seeps. Multi-temporal Landsat, SPOT or ERS-1 imagery is examined on the image processing system for recurrence of oil slicks on the ocean surface. When slicks are found, vector data such as well locations and shipping lanes are imported from ARC/INFO and laid on top of the satellite imagery to ensure that the slicks are not man made.
      Once their natural origin has been substantiated, the satellite images are integrated in the image processing system with the seismic horizon data. BP geologists examine the datasets to determine if the location of the oil slicks can be correlated to features such as faults, which may link from the subsurface to the seabed.

Planning Exploration
BP's management staff has been finding multiple data processing functions with the image processing system that are saving time and money in the planning of field exploration and logistics. The best example of this occurred in the recently opened exploration basins in China and the Former Soviet Union.
      Many of the blocks recently opened to outside exploration companies have already been drilled by national organizations. Unfortunately, the well data is often not available, which puts the new exploration company in the potentially expensive position of unwittingly exploring in an area where areas have already been drilled.
      BP has acquired high-resolution SPOT and Russian MK-4 photographic data to find locations of existing infrastructure in China's Tarin Basin. This information, as well as outcrop data, geologic maps, and existing seismic is stored in the GIS until it must be imported into the image processing system. There the geologists will compare topography found in the satellite image with correlative vector data so they can determine what areas should be targeted for further exploration.
      A number of factors are examined when deciding how to proceed with the exploration project. Any existing well data, outcrop information and geologic structure will be correlated with the image processing system to determine which are likely oil plays. The imagery is also examined to lay out the seismic survey in difficult terrain to yield the most efficiency with least expenditure.

Conventional Exploration Methods
In addition to the complex, multiple-data processing projects, BP also uses ER Mapper for the routine types of projects that are the backbone of modern petroleum exploration. In the Perdenales Delta of Venezuela, BP sedimentologists are using satellite imagery to prove the geologist's favorite adage, "The present is the key to the past."
      As is true with many exploration companies, BP conducts a great deal of petroleum exploration in depositional environments created in river deltas.
      BP has collected Landsat data for many deltaic environments around the world to unravel the complex patterns and processes that are involved in forming the structures associated with many hydrocarbon traps in their area of interest. BP has acquired several images of the Perdenales Delta (Figure 1) and processed the Landsat TM data to highlight sediment suspended in the river. The sediment flow and deposition is being examined to better understand the mechanics involved. The objective is to be able to more accurately predict which subsurface depositional environments are likely to trap hydrocarbons.
      Although much of this interpretation occurs on the image processing display, BP also outputs the data to large scale color plots. With only a few hours of processing, the scientists are able to print the scaled images on Versatec, Calcomp or photographic quality color plotters for visual, hard-copy examination.
      In a similar application, BP has processed a Landsat scene of the Apsheron headland in Azerbaijan to create a natural color image. In this region, mud volcanoes have been found to be associated with the subduction zones that also relate to hydrocarbon accumulations. In the image [Figure 3], mud volcanoes are characterized by a light blue center and a radial pattern coming out from the cone.

Back