Oil and gas applications in ArcGIS of highly accurate

www.photosat.ca

1 Gerry Mitchell PhotoSat President

Over 300 stereo satellite elevation mapping projects since 2004

Oil and gas applications in ArcGIS of highly accurate elevation mapping from space

www.photosat.ca

August 28, 2008

WorldView-1 stereo photos

Carbonate anticline, northern Iraq

35 km

August and December WV1 stereo pairs and

processed elevation image

December 5, 2008

WorldView-1 elevation image

www.photosat.ca

Slope direction image WorldView-1 elevation image

Carbonate anticline, northern Iraq

35 km

Stereo WV1 elevation image and slope direction image

www.photosat.ca

Slope direction image WorldView-1 elevation image

Carbonate anticline, northern Iraq

15 km

Stereo WV1 elevation image and slope direction image

www.photosat.ca

Slope direction image WorldView-1 elevation image

Carbonate anticline, northern Iraq

9 km

Stereo WV1 elevation image and slope direction image

www.photosat.ca

Slope direction image WorldView-1 elevation image

Carbonate anticline, northern Iraq

3 km

Stereo WV1 elevation image and slope direction image

www.photosat.ca

Western Libya

Stereo WV1 elevation image and direction of slope image

www.photosat.ca

Himalayas IKONOS and Stereo IKONOS DEM processed in August 2009. The access roads show clearly in the DEM. The south facing slope is 45 degrees.

Steep terrain

www.photosat.ca

Himalayas IKONOS and Stereo IKONOS DEM

45 degree slope

Steep terrain

www.photosat.ca

Himalayas IKONOS draped over the

Stereo IKONOS DEM in ArcGlobe

Steep terrain

www.photosat.ca

IKONOS 1m colour 2004

Geo

Eye S

tere

o S

ate

llit

es

High resolution stereo satellites

www.photosat.ca

IKONOS 1m colour 2004 WorldView-1 50cm greyscale

2008

Geo

Eye S

tere

o S

ate

llit

es

Dig

italG

lob

e S

tere

o S

ate

llit

es

High resolution stereo satellites

www.photosat.ca

IKONOS 1m colour 2004

GeoEye-1 50cm colour 2009

WorldView-1 50cm greyscale 2008

Geo

Eye S

tere

o S

ate

llit

es

Dig

italG

lob

e S

tere

o S

ate

llit

es

High resolution stereo satellites

www.photosat.ca

IKONOS 1m colour 2004

GeoEye-1 50cm colour 2009

WorldView-1 50cm greyscale 2008

WorldView-2 50cm colour 2010

Geo

Eye S

tere

o S

ate

llit

es

Dig

italG

lob

e S

tere

o S

ate

llit

es

High resolution stereo satellites

www.photosat.ca

High resolution stereo satellite photos

Adaptation of seismic processing systems

Graphics Processing Units (GPUs)

Three key technical components enabling geophysical elevation mapping

from space

www.photosat.ca

Graphic Processing Units (GPUs)

GPUs perform numerical processing up to 100 times faster than CPUs. This enables us to do the hundreds of millions of 2D Fourier transforms necessary to automatically produce 1m Digital Surface Models from stereo satellite photos in reasonable times.

512 parallel processors 3 gigabytes RAM 1 teraflop C program compiler

www.photosat.ca

Offshore seismic vessels acquire over a terabyte of data per day, the same quantity of data as the high resolution satellites.

Seismic geophysics tool box

www.photosat.ca

Imaging of stereo satellite data

Using many of the seismic processing tools we calculate and plot the correlation between the stereo satellite photos as vertical profiles of correlation. This produces profiles that we can display and interpret in 3D seismic workstations.

Stereo photo

correlation profiles

Ground Surface

www.photosat.ca

Imaging of stereo satellite data

Cross section of stereo GeoEye-1 satellite photo correlation profiles over a hill in Coahuila Mexico. This cross section is displayed in OpendTect, a free and open source seismic workstation.

Stereo photo

correlation profiles

Ground Surface

www.photosat.ca

Imaging of stereo satellite data

Cross section of stereo GeoEye-1 satellite photo correlation profiles over a hill in Coahuila Mexico.

Positive

correlation

Negative

correlation

www.photosat.ca

Imaging of stereo satellite data

Orthogonal cross sections of stereo GeoEye-1 satellite photo correlation profiles, Coahuila Mexico.

www.photosat.ca

Imaging of stereo satellite data

Orthogonal cross sections and horizontal section of stereo GeoEye-1 satellite photo correlation profiles, Coahuila Mexico.

www.photosat.ca

Imaging of stereo satellite data

Digital surface model displayed in 3D seismic workstation. The colour represent the correlation amplitudes. Bare ground, with strong correlations is shown in dark colours. The trees, with weaker correlations are reds and yellows.

www.photosat.ca

Checking

For each stereo pair of satellite photos there is a unique topographic surface that will exactly match the photos. When the DEM is correct, the ortho of each of the stereo photos should be identical. Alternating the display between the two ortho photos is an effective check for errors in the DEM.

www.photosat.ca

Gerry Mitchell PhotoSat President

Iraq Anticline, PhotoSat DEM, August 2009

www.photosat.ca

Gerry Mitchell PhotoSat President

Iraq Anticline DEM in ArcGlobe

www.photosat.ca

Left and right WV1 ortho, draped over the WV1 DEM, alternating between left and right orthos in Arc Globe.

www.photosat.ca

Gerry Mitchell PhotoSat President

Iraq Anticline, PhotoSat DEM, August 2009

www.photosat.ca

Oil and Gas Seismic: operational application of stereo satellite mapping

www.photosat.ca

Improving seismic safety and seismic quality using stereo satellite elevation mapping for 3D seismic surveying in Ghadames Libya.

Gerry Mitchell PhotoSat

Demonstration Project Location

www.photosat.ca

50cm resolution stereo WorldView-2 satellite photo 220 km2

16 km

www.photosat.ca

1.2m x 2.4m white painted steel sheets were used as ground survey targets.

www.photosat.ca

Three ground survey points were used to reference the stereo satellite elevation mapping.

www.photosat.ca

1m posted elevation model derived from the stereo WorldView-2 satellite photos, 220 km2.

16 km

www.photosat.ca

1m contours created in ArcMap from the stereo WorldView-2 satellite DEM.

www.photosat.ca

Seismic receivers were Geospace Seismic Recorders. These seismic receivers include GPS antennas. They record their locations to within about 2m in X and Y and 3m in elevation.

www.photosat.ca

Seismic sources were vibroseis trucks

www.photosat.ca

Seismic source locations were determined by GPS antennas mounted on the vibroseis trucks. These locations are usually accurate to better than 10cm in X, Y and Z.

GPS Antenna

www.photosat.ca

Shotpoints on 50m grid, except on steep slopes Receiver points on 50m by 450m grid displayed in ArcMap

Shotpoints omitted on steep slopes

www.photosat.ca

Not all shotpoints on steep slopes were omitted. Red = slope > 15% grade. The DEM data was not available for the planning of this seismic survey.

Shotpoints on steep slopes

www.photosat.ca

Shotpoints on 50m grid, except on steep slopes Receiver points on 50m by 450m grid. Image displayed in ArcGlobe.

Shotpoints omitted on steep slopes

www.photosat.ca

70,423 seismic source locations.

www.photosat.ca

Surface Spot Tool from 3D Analyst.

www.photosat.ca

DEM elevations added to survey point attributes.

www.photosat.ca

Differences between the survey elevations and the DEM elevations.

www.photosat.ca

Seismic source locations

Elevation differences between the vibroseis truck GPS and the stereo satellite elevation mapping

www.photosat.ca

Elevation differences between the vibroseis truck GPS and the stereo satellite elevation mapping

www.photosat.ca

Seismic source locations

Elevation differences between the vibroseis truck GPS and the stereo satellite elevation mapping

Standard deviation 30cm Median -35cm

www.photosat.ca

Receiver points were surveyed with a backpack GPS

www.photosat.ca

Receiver points were surveyed with a backpack GPS

www.photosat.ca

8,516 seismic receiver locations.

www.photosat.ca

Seismic receiver locations

Elevation differences between the backpack GPS survey and the stereo satellite elevation mapping

www.photosat.ca

Elevation differences between the backpack GPS survey and the stereo satellite elevation mapping

Standard deviation 33cm Median -14cm

www.photosat.ca

Project Location

Operational stereo satellite elevation mapping project

3,962 km2 mapping area

www.photosat.ca

Stereo satellite mapping eliminates the need for GPS surveying of over 170,000 seismic receiver points in this 3,963 km2 area of western Libya. Geolocated jpegs of the WorldView satellite photo swaths displayed in ArcMap.

109 km

www.photosat.ca

Seismic source points with accurate GPS elevations and ground survey points with some elevation errors for the 3,962 km2 Ghadames stereo satellite elevation mapping. Image displayed in ArcMap.

109 km

53,000 seismic source points

31,983 seismic source points

www.photosat.ca

Elevation differences between the eastern stereo WV2 DEM, two areas of source points and the survey points

44km

31,983 source points

Mean 5cm STD 24cm

35,691 source points

Mean 3cm STD 31cm

www.photosat.ca

Elevation differences between the eastern stereo satellite DEM and the survey points showing the location of 118,000 seismic source points provided after the DEM was generated.

44km

www.photosat.ca

Elevation differences between the 118,000 seismic source points and the stereo satellite elevations; Standard Deviation 28cm. The elevation differences between three of the survey points and the stereo satellite elevations are shown. Image and histogram created in ArcMap.

www.photosat.ca

Differences between the source point GPS elevations and the stereo satellite elevations. The 61cm elevation difference for one of the survey points is also shown. The source point GPS elevations match the stereo satellite elevations about 40cm better than this survey point elevation.

www.photosat.ca

Advanced scouting reduced by ~ 80% Current practice is for field survey crews to scout all vibrator points and access routes in advance to identify inaccessible source-points and to plan for efficient disposition of the vibrator trucks. The satellite imagery with this DEM gives an accurate enough picture so that only a few areas require field visits in advance.

Benefits

www.photosat.ca

Benefits

Quality Control of Source Positioning All source-points are surveyed using the GPS systems mounted on the vibrator trucks. Initialisation time for these systems can be long, with up to 30 minutes before we have full confidence in the elevation accuracy after switch on. Elevation accuracy can also suffer due to poor satellite visibility on occasions. Comparison with the DEM highlights any problem areas. The DEM can be used where the vibrator GPS elevation is obviously in error.

www.photosat.ca

Receiver Positioning

Thousands of recording nodes are deployed for the 3D seismic survey. Non-augmented GPS receivers in each recorder give sufficient x, y accuracy but cannot provide sufficiently accurate elevations in stand-alone mode for seismic processing requirements. The stereo satellite DEM provides the necessary receiver location elevation accuracy for seismic processing and so removes the need for any additional surveying of receiver locations. On the 3,962 km2 mapping area the backpack GPS surveying of approximately 170,000 receiver stations will be eliminated.

Benefits

www.photosat.ca

Stereo Satellite Elevation Mapping Majnoon Oil Field Development, Iraq

www.photosat.ca

Majnoon Oil Field

Shell Press Release January 17, 2010

“Shell, as lead operator, will hold a 45%

share, with partner Petronas holding 30%.

The Iraqi state holds 25% of the

participating interests in all licences.

The consortium targets a production

plateau of 1.8 million barrels of oil per day,

up from a current level of approximately

45,000 barrels of oil per day. Majnoon,

located in southern Iraq, is one of the

world’s largest oil fields.”

www.photosat.ca

1,700 km2 Majnoon mapping area

75km

Stereo Satellite Elevation Mapping Area

www.photosat.ca

WorldView-2 stereo photos April 20, May 9 and June 11, 2010.

Four survey points, GCP 1-4 were provided for ground

reference. Image created in ArcMap. 7

5k

m

WorldView-2 Stereo Pairs

www.photosat.ca

GCP 4

GCP 3 GCP 2

GCP 1

After matching the June 11 and April 20 DEMs and orthos to the May 9 DEM and ortho there was a 10m horizontal mismatch between the April 20 WV2 precision ortho and GCP’s 3 & 4.

www.photosat.ca

GCP 2

GCP 1

1700 km2 , 1m stereo satellite DEM referenced to only two ground survey points. DEM completed in August 2010.

75km

x

x

Copyright of SIEP BV 70 May 2010 UNRESTRICTED

The Photosat processed WV2 DSM has been assessed against 263 surveyed points using a wide-area differential GPS service, using satellite broadcast techniques. Horizontal accuracy is better

than 0.2m.

Unfortunately due to the risks of acquiring survey measurements in Iraq, only 263 points in a constrained area could be captured.

Number of data points analysed 263

Mean 4.321

Sample standard deviation 0.610

Std error of the mean 0.038

Lower 95% limit for mean 4.247

Upper 95% limit for mean 4.396

Mean absolute error 0.311

Sum of Squared Errors 51.895

Mean Squared Error 0.197

RMSE (square root of MSE) 0.444

Accuracy Assessment

www.photosat.ca

Extension to the Majnoon DEM. Four additional ground survey points match to better than 20cm

121km

www.photosat.ca

Stereo Satellite Elevation Profile

Interpolate shape

www.photosat.ca

Stereo Satellite Elevation Profile

Create profile graph

1m

www.photosat.ca

Stereo Satellite Elevation Profile

1,500m

1m