PrintMost remotely sensed data are derived from measurements of electromagnetic radiation. Aerial photographs are analog forms of data that record intensities of electromagnetic radiation within the visible or near-infrared wavelength bands. Digital sensing systems extend the spectral sensitivity of photographic film far beyond the visible band, enabling users to map phenomena that are otherwise invisible. Because many objects exhibit unique spectral response characteristics across a range of wavelengths, multispectral sensing offers the potential to identify features of interest automatically. Recognizing this potential, analysts in many fields have adopted land remote sensing data for such diverse applications as land use and land cover mapping, geological resource exploration, precision farming, archeological investigations, and even validating the computational models used to predict global environmental change. Once the exclusive domain of government agencies, an industry survey suggests that the gross revenue earned by private land remote sensing firms exceeded $2.4 billion (U.S.) in 2001 (ASPRS, 2004).
The fact that remote sensing is first and foremost a surveillance technology deployed by government agencies cannot be overlooked. State-of-the-art spy satellites are rumored to be able to detect objects as small as six inches wide. Meanwhile, GeoEye and other private firms have been licensed to build and launch half-meter panchromatic sensors. As the resolution of remotely sensed imagery increases, and its price decreases, privacy concerns are bound to proliferate. For example, remotely sensed data were pivotal in the case of an Arizona farmer who was fined for growing cotton illegally (Kerber, 1998). Was the farmer right to claim that remote sensing constituted unreasonable search? More serious, perhaps, is the potential impact of the remote sensing industry on defense policy of the United States and other countries. In light of an expected $500 billion investment in commercial satellites (including communications satellites as well as land remote sensing systems) by 2010, some analysts believe that "the military will be called upon to defend American interests in space much as navies were formed to protect sea commerce in the 1700s" (Newman, 1999).
While the ethical implications of remote sensing technologies must not be ignored, neither should their potential to help us to become more knowledgeable, and thus more effective stewards of our home planet. Several challenges must be addressed before remote sensing can fulfill this potential. One is the need to produce cost-effective, high-resolution data suitable for local scale mapping--the scale at which most land use decisions are made. Another is the need to develop more sophisticated image processing algorithms that will enable analysts to extract vector features from raster source data with minimal intervention. Yet another challenge is to develop automated classification techniques to help derive meaningful patterns from the data produced by a new generation of hyperspectral scanners--sensing systems that measure reflected and emitted radiation across 200 or more narrow wavelength bands simultaneously.
