AIRBORNE
Hyperspectral Remote Sensing
Mineral geologists are relying more on hyperspectral imagery
as on e of their basic tools
By Dr. Bob Agar

Since the first hot-air balloon, man has been using aircraft to view and map the Earth from above. Aerial photography became the prime tool of land surveyors and geologists and for a long time formed the basis upon which mineral exploration data were plotted to create maps. Aerial photography, whether panchromatic or in color, gathers spectral information of the Earth's crust, albeit in analog form, which can assist geologists in producing interpretive maps of geology and geological structure.
    With the first remote sensing satellites launched in the 1970s, Multi-Spectral Scanners (MSS) provided the first large-scale regional digital spectral information, allowing geologists to recognize geological features on a regional or continental scale. However, geologists still used the data as they did air photography, interpreting geology from scaled hard-copy digital data images. Nevertheless, the additional spectral data gathered by Landsat MSS-and later Thematic Mapperª satellites-provided geologists additional spectral information from which a far greater amount of geological detail could be gleaned. With this data, geologists were also able to enhance imagery, highlighting areas of potential exploratory interest by using the spectral properties of iron oxides and common hydrothermal alteration minerals.
    By the early 1980s, users of the satellite data in many different disciplines began noticing the limitations to their work caused by satellite systems designed to provide general-purpose rather than application-specific data. The spectral wavelengths of interest to environmentalists and agriculturists are very different from those of geologists.
    At this time, a new generation of spectral instruments was born as private companies developed airborne spectral scanners with a large number of contiguous recorded wavelength bands specifically located for particular applications. Companies such as Geophysical & Environmental Research Corporation (GER) developed instruments with different band numbers in different parts of the spectrum targeted for specific applications, including geological, environmental, and agricultural. These new data, offering far greater spatial and spectral resolution than hitherto possible, gave exploration geologists the opportunity of doing more than simply producing photogeological interpretive maps or generating exploration targets based upon broad mineral classifications. With airborne multi- and hyperspectral data, geologists could map discrete minerals from aerial platforms.
    Throughout the 1980s, airborne hyperspectral data were available to geologists but met with limited acceptance. The digital data needed to be processed into analog image format and still treated photogeologically. Typically, an in-house expert processed data on large image-processing computer workstations, and geology or mineralogy was determined by interpreting color combinations. Thus, although the data existed to produce mineral maps, in practice the technology was poorly embraced by users in the mineral industry.
    In the early 1990s, however, PCs rapidly expanded computing power, enabling the development of powerful image-processing packages that could produce analog imagery as well as match airborne spectra with reference mineral spectra and quantify and directly map the distribution of important minerals. That such work could now be done by individual geologists, with the same computer used for plotting geochemical or drill log data, brought airborne hyperspectral data to the end-user, resulting in an explosion of interest. Where, in the 1980s there were only two significant developers and operators of airborne hyperspectral instruments, GER and Geoscan, now at least five developers with more than 10 instruments are operating.

What is hyperspectral imaging?
Spectral imaging uses the differential reflectance and absorption properties of natural materials in sunlight to map their distribution. Just as humans see healthy vegetation as green in natural sunlight because the vegetation reflects green light, spectrometers measure the reflected radiation in a large number of wavelengths in order to determine the spectral signature of surface materials. Different materials and minerals have characteristic spectral signatures, some of which may be quite similar or distinct. The ability of an individual spectrometer to recognize specific materials or minerals depends upon its spectral resolution or, in other words, the number of spectral bands it has over a given wavelength range.
    Multispectral scanners are typically those that gather information across several parts of the electromagnetic spectrum, though not necessarily contiguously. For example, the Landsat TM would be considered multispectral as it has 7 bands across the visible, near infra-red, short wave infra-red, and thermal infra-red spectrums. However, its resolution and mineral discrimination power is very low. Hyperspectral instruments are typically multispectral with several contiguous bands in all parts of the spectrum in which they operate. GER's Digital Airborne Imaging Spectrometer for example is hyperspectral, having 63 bands, 27 in the visible, and near infra-red (0.4-1.0 microns), two in the short wave infra red (1.0-1.6 microns), 28 in the short wave infra-red important for mapping clay minerals (2.0-2.5 microns), and 6 in the thermal infra red. The ability to measure reflectance in several contiguous bands across a specific part of the spectrum allows these instruments to produce a spectral curve that can be compared to reference spectra for any number of minerals, thereby allowing the mineral content of a particular piece of ground to be determined.
    The scanners analyze the ground within a narrow field of view that effectively sweeps across the progressing aircraft's path. Each individual data cell, or pixel, contains information for all of the wavelength bands recorded, but may be as small as one square meter for very high-spatial resolution instruments, or as large as 30 meters (m) or more for instruments of low-spatial resolution. As the spatial resolution increases, the total area covered decreases. For most mineral exploration purposes, pixel sizes in the range of 8-12 sq. m allow for good detail and rapid regional coverage in the range of 1,000 to 3,000 sq. km per hour.

When, where, and how should such instruments be used?
All spectrometers rely upon solar energy to record reflectance. Higher amounts of incoming radiation yield greater ranges of energy reflection or absorption. Hence, data should ideally be acquired when the sun is high in the sky. Shadows can be a major difficulty in mountainous terrain. For this reason, acquiring data at a time of year when the sun is high (i.e., summer) and around the middle of the day is normal practice.
    However, summers in some parts of the world are wet or cloudy. Not only do clouds obscure the ground, but wet ground conditions also degrade the data because water absorbs energy in most of the important wavelengths being recorded. Thus, the need for clear skies and dry ground can override sun angle when determining the ideal time to acquire data.
    Heavy vegetation coverage will obscure surface rocks so mineral exploration is best in arid and semi-arid terrains. Vegetated terrains can benefit from this technology, but cover density and ground resolution need to be considered in order to compare its cost-effectiveness to alternative exploration methods.
    The data recorded measure reflectance at surface and provide no information of mineralogy at depth. Thus, even in arid terrains, if the area flown over is covered by transported soils or alluvium, the data will provide little or no information about the mineral prospectivity of the ground beneath the cover. Where soils are in site or outcrop is quite extensive, however, the data will rapidly advance exploration programs by providing extensive coverage, geological mapping, exploration targeting, and mineral mapping. This is particularly the case when operating in remote, rugged, and difficult terrains where ground access is difficult. The airborne survey and subsequent mineral mapping can often save weeks or months of exploration time over field mapping and geochemical techniques.
    The airborne surveys can be optimized by initial use of coarser resolution satellite instruments to outline the broad regional terrain. This might reduce the airborne survey requirements to approximately 1000 sq. km. The survey would provide a geological and mineralogical spectral sample over the entire area every 10m. First pass analysis of the data would map the broad distribution of important alteration minerals and would effectively focus exploration into the most promising areas for ground follow-up, saving time and money in that geochemical sampling is centered only on promising parts of the area that are prospective.
    Where subsequent ground follow-up and geochemistry identifies mineralization, the airborne data can be revisited for that particular prospect to produce very detailed mineral alteration assemblage maps for integration with geochemical data, which collectively can refine the exploration and drill hole targeting process. Airborne data are also often used at this stage in conjunction with ground spectral data collected by field spectrometers that enable field geologists to quickly confirm difficult mineral assemblages without having to resort to time-consuming and costly laboratory analysis.
    Available airborne spectrometers are not located close to remote parts of the world. Often demand for this type of data is highest in parts of the world where the cost of mobilizing such equipment appears to be prohibitive. However, because data acquisition is typically constrained to particular flying seasons for different parts of the world, it is usually possible for an instrument to acquire several separate surveys for different customers in the same region for a shared mobilization cost. Similarly, some operators have flown certain highly active exploration areas speculatively and subsequently sold data "off the shelf" to many different customers. It is likely that as the number of available sensors and operators increases, exploration geologists will have an increasing amount of available data.

The future
The next phase of hyperspectral remote sensing will be moving the platform from aircraft into space. Satellite born hyperspectral sensors are now in development and will greatly add to the options available to exploration geologists. Although operating systems are expected to be launched by 2000, the early instruments have traded high spectral resolution for data storage capacity and coverage. They will have only restricted resolutions, and it will be some time before data are readily available to geologists. Add to this the difficulties of maintaining such sensors in space with extremes of temperature and radiation and the fact that once launched they cannot be adjusted, it is likely that there will be some mission failures before a satisfactorily performing instrument is providing the coverage required to meet exploration needs. In the meantime, airborne sensors will continue to develop and will likely become more sensitive and robust, capable of operation on any aircraft and, as more are built, more available with some ideally suited for broader regional work and others for higher-resolution site-specific surveys.
    Airborne hyperspectral remote sensing has now come of age thanks to the development of user-friendly PC-based software. Mineral exploration geologists are now able to take the data to the mineral mapping stage and focus their exploration cost-effectively in remote and difficult terrains. As more airborne instruments become available and vie for a place in the market, the technology will become increasingly cost-effective. As both airborne and satellite hyperspectral databases expand and archival information becomes available "off the shelf," so its use and value to mineral exploration will increase, and hyperspectral remote sensing will become one of the geologist's basic exploration tools.

About the Author:
Dr. Bob Agar is managing director of Australian Geological & Remote Sensing Services in Perth, Australia. Based in Millbrook, New York, Geophysical & Environmental Research Corporation (GER) is a leading developer of aerial remote sensing equipment for use in a variety of industries, including agriculture, environmental, and mineral exploration.

Back