to have the consistent band numbers and wavelength. Then the 
resampled spectrums and the maximum spectral angle 0.05 
were used to classify. The classification result of SAM was 
shown as Figure 6. Comparison with the result and a known 
distribution map, the extraction effect of oil slick was good. 
  
     
Known 
oil 
Hydro 
carbon 
sample 
  
well 
  
  
  
Figure 6. Hydrocarbon information extracted from Hyperion 
image by SAM 
4. RESULTS 
(1) When the content of crude oil in soil is little, the double 
absorption characteristic in 1748nm and 2330nm for petroleum 
hydrocarbons in soil is very weak, and it is easy to be interfered 
by other soil composition. Then the accuracy of detection using 
the feature for petroleum hydrocarbons in soil was reduced. 
When the crude oil is up to certain amount, it can make a 
hyperspectural remote sensing detection for soil with petroleum 
hydrocarbons using the diagnosis characteristic of spectrum. 
(2) The reflectance of water that was covered with oil slick was 
higher than seawater, while the slope that was indicated the 
changing rate was lower. The broad absorption feature was 
shown in 508nm with the decreasing of the oil slick’s thickness. 
And the absorption characteristic peaks of oil slick reflectance 
were not obvious in the other band regions. 
(3) Based on the three bands in Qaidam Basin’s Hyperion 
image, which was near the absorption characteristic peaks in 
1730nm, Three-Band-Ratio algorithm and Absorption-Depth 
method were used to extract oil-gas hydrocarbon and delineate 
the target area of oil-gas reservoir. 
(4) Based on the identification of characteristic spectrum of 
alteration mineral and Hyperion image in Qaidam Basin, Linear 
Spectral Unmixing (LSU) and Spectral Angle Matching (SAM) 
can be used to determine he mineral composition counterparts 
of endmember. Then the target area of oil-gas reservoir can be 
determined indirectly. 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
   
(5) Combining with the optimal bands in the region of 
visible/near-infrared, Spectral Angle Matching (SAM) was used 
to extract the thin oil slick of microseepage in Liaodong Bay. In 
addition, the spectral angle was determined by experience. 
REFERENCES 
PhilP R P, Crisp P T. Surface geochemical Prospecting methods 
used for oil and gas prospecting[J]. Journal of Geochemical 
Exploration, 1982, 17(1): 1-34. 
Almond, Samuel, 2000. The Remote Sensing Of Oil Slicks 
From Satellite Platforms. 
http://www.geog.ucl.ac.uk/-salmond/essay.html. 
Cloutis E A. Spectral reflectance properties of hydrocarbons, 
remote-sensing implications: Science [J], 1989, 245: 165-168. 
Foudan M.F. Salem. Hyperspectral remote sensing: A new 
approach for oil spill detection and analysis [D]. Doctor 
Dissertation. George Mason University (USA), 2003. 
Heather Freeman. Evaluation of the use of hyperspectral 
imagery for identification of microseeps near Santa Barbara, 
California[ R/OL ].California,Sep.26, 2003. 
Ellis J M, Davis H H, Zamudio J A. Exploring for onshore oil 
seeps with hyperspectral imaging[J], Oil and Gas Journal, 2001, 
10: 49-56. 
Williams A and G.Lawrence, 2002, 'The role of satellite seep 
detection in exploring the South Atlantic's ultra-deep water', in 
Surface Exploration case Histories, etc, AAPG Studies in 
Geology No.48 
Horig B, Kuhn F, Oschutz F, et al. HyMap hyperspectral 
remote sensing to detect hydrocarbons[J]. Int. J. Remote 
Sensing, 2001, 22(8): 1213-1422. 
ACKNOWLEDGEMENTS 
This research is financially supported by National Natural 
Science Foundation of China (Grant No. 40971186). And an 
expression of thank for the help of China Aero Geophysical 
Survey & Remote Sensing Center for Land and Resources and 
PetroChina Exploration and Development Research Institute.