Remote sensing for petroleum exploration,Part 1: Overview of imaging systemsFLOYD F. S ABINS, Remote Sensing Enterprises, Fullerton, Californiahis paper, the first of two thatsummarize remote sensing as ap-plied to petroleum exploration, pro-vides an overview of the science andthe computer techniques used toprocess the data. The second (to bepublished in TLE’s May issue) willdescribe successful oil-explorationprojects that employed remote-sens-ing technology. The examples arefrom projects by colleagues and my-self at Chevron, from which I retiredin 1992.Remote sensing has several ap-plications to geophysical explorationof onshore regions. Images are inter-preted to produce geologic mapsthat show structural trends and po-tential prospects which are, in turn,used to plan an efficient seismic pro-gram. Furthermore, if an area lacks areliable base map, one can be readi-ly generated from remote-sensingdata, as can maps showing accessand trafficability which can also im-prove the efficiency of field opera-tions. (When Lee Lawyer, a formerSEG President, was chief geophysi-cist of Chevron Overseas Petroleum,he insisted that any onshore seismicsurvey be preceded by a remote-sensing interpretation.) Landsat im-ages have also been employed inshallow offshore areas to identify un-charted reefs and other hazards toseismic surveys.Remote sensing is defined as thescience of acquiring, processing, and in-terpreting images from satellites and air-craft that record the interaction betweenmatter and electromagnetic energy. Re-mote-sensing images of the earth areacquired in three wavelength inter-vals, or regions, of the electromag-netic spectrum. The visible regionranges from 0.4 to 0.7 µm and is di-vided into the blue, green, and redbands. The infrared (IR) regionranges from 0.7 to 30 µm and is di-vided into the reflected IR and ther-mal IR portions. The reflected IR por-tion ranges from 0.7 to 3.0 µm; theenergy is predominantly reflectedsunlight at wavelengths longer thanvisible light. The Landsat and SPOTsatellite systems acquire valuable im-ages in the visible and reflected IR re-gions. The thermal IR portion rangesfrom 3.0 to 15.0 µm; the energy is ra-diant, or heat, energy. Thermal IRimages have considerable potentialfor exploration in arid and semiaridterrains; however, the method hasbeen underutilized in recent yearslargely owing to the lack of suitableimages.Images in the visible, reflected IRand thermal IR are acquired in thepassive mode by systems that simplyrecord the available energy that isreflected or radiated from the earth.The microwave region ranges from0.1 to 30 cm; images for explorationare primarily acquired in the activemode called radar. This paper focus-es on three systems: Landsat The-matic Mapper and SPOT panchro-matic images in the visible andreflected IR regions, and aircraftradar images.Landsat Thematic Mapper images.Landsat is an unmanned satellitethat orbits the earth in a sun-syn-chronous pattern at an altitude of705 km. The two second-generationsatellites carry the thematic mapper(TM) system. TM is a multispectralsystem which records seven separateAPRIL 1998 THE LEADING EDGE 467TCoordinated by M. Ray Thomasson and Lee LawyerGEOLOGIC COLUMNFigure 1. Reflectance spectra ofvegetation and sedimentary rocks,showing spectral ranges of Land-sat TM and SPOT systems. Allfigures in Part 1 are from Sabins(1997).Figure 2. Landsat TM bands. (a) = 2 (green), (b) = 4, and (c) = 7 (both reflected IR). (d) = geologic features nearThermopolis, Wyoming, as interpreted from Figure 3.images, or bands, for each scene.Figure 1 shows wavelength ranges ofthe three visible bands (1, 2, 3) andthree reflected IR bands (4, 5, 7).Band 6 records thermal IR energybut is rarely used for exploration.The reflectance spectra of commonsedimentary rocks and vegetation inFigure 1 provide insights for select-ing the optimum TM bands for in-terpretation. The spectra of the dif-ferent rocks are very similar in thevisible bands but have major dis-tinctions in the reflected IR bands;therefore, TM bands 4, 5, and 7 areespecially useful for mapping differ-ent rock types.TM images are acquired by across-track scanner with an oscillat-ing mirror that sweeps across the ter-rain normal to the satellite groundtrack. A spectrometer separates thereflected sunlight into the six spectralbands. The image data are teleme-tered to ground receiving stations viathe tracking and data relay satellites(TDRSS). A TM image covers 170 185 km2of terrain with a spatial res-olution of 30 m. The digital data arestored in a raster format. An individ-ual band consists of 5667 scan lines,each of which contains 6167 pictureelements (pixels) for a total of almost35 million pixels. A pixel represents a30 30 m ground-resolution cell andrecords the intensity of reflected en-ergy on an eight-bit scale, rangingfrom 0 (minimum reflectance) to 255(maximum reflectance).Any three bands may be mergedin any combination of blue, green,and red to produce a color compos-ite image. There are 120 possiblecolor combinations, but theory andexperience show that a small numberof combinations are suitable for mostregions and applications. Figure 2shows TM bands 2, 4, 7 for a smallsubarea that includes the town ofThermopolis in the Bighorn Basin ofWyoming. These bands are general-ly optimum for arid and semiarid ter-rain, such as central Wyoming. Fig-ure 3 is a color composite of bands 2,4, 7 merged in blue, green, and red.Figure 2d is a map showing major ge-ologic features interpreted from thissmall-scale color image. Four anti-clines are seen in the image. The Lit-tle Sand Draw and Gebo anticlinesare oil fields that were discoveredlong before the launch of Landsat.Images of existing oil fields are valu-able examples for interpreting im-ages of frontier regions. Detailedmaps at scales as large as 1:50 000 are468 THE LEADING EDGE APRIL 1998Figure 4. Terrain returns and image signatures for a pulse of radar energy. Figure 3. Color composite image of TM bands 2, 4, and 7 combined inblue, green, and red. Thermopolis, Wyoming subarea.interpreted from larger-scale versionsof TM images. Part 2 of this articlewill describe how TM images con-tributed to oil discoveries in the Cen-tral Arabian Arch.Landsat was launched by NASAand is operated by Space ImagingEOSAT. Images, in digital or hard-copy format, are sold by Space Imag-ing EOSAT, 4300 Forbes Boulevard,Lanham, Maryland 20706 (http://origin.eosat.com/) and by the USGSEROS Data Center, Sioux Falls,
View Full Document
Unlocking...