So you want to buy a satellite image… What types are available? What resolution can you expect? And how much is this going to cost, anyway? This is a very basic introduction to the types of products available to us today – it’s by no means exhaustive and I’ve tried to keep it simple.
The most important thing to figure out before investing in a product is why exactly you want the imagery. If you’re just interested in identifying buildings or standing walls, there are several free or very cheap options available (like Google Earth). If you’re looking for small-scale changes in vegetation that might reflect archaeological phenomena, then high-resolution, multi-spectral imagery is going to be your best bet.
TYPES OF IMAGERY
There are three basic types of satellite imagery:
1. Panchromatic (essentially black-and-white). The earliest types of satellite imagery were taken by a black-and-white camera mounted to a spacecraft. A great example is the CORONA satellite, launched by the United States National Reconnaissance Office in the 1960s. The CORONA mission used several satellites, and it targeted areas of the world where suspected military action was taking place (especially the Near East). In 1995, the U.S. government declassified thousands of CORONA images, making them one of the cheapest and most accessible sources of historical satellite imagery. Because the photos were taken as stereo pairs, it’s also possible to turn them into digital elevation models (DEMs) (see Casana and Cothren 2008). Today’s CORONA imagery is scanned from black-and-white photographic negatives.
Fun fact: A lot of the CORONA imagery over the Middle East has already been processed and made publicly available by the University of Arkansas Center for Advanced Spatial Technologies. Check it out!
2. Multispectral. First – let’s review “RGB,” which stands for “red, green, blue” and is a standard format for most photos. RGB files have three layers of information, one that corresponds to the reds in the image, another for the greens, and a third for the blues. In other words, RGB images only record visible light. Since the 1980s, satellites have been recording colors beyond the RGB spectrum, meaning they are “multispectral.” GeoEye-1 is an example of four-band (RGBN) multi-spectral imagery. The fourth letter, N, refers to a particular wavelength of near-infrared light. If you purchase a GeoEye-1 image, your file will come with all four bands – red, green, blue, and near-infrared. Landsat has up to 8 bands. If you look at those same images in Google Earth, you can’t see the extra bands, since you don’t have access to the original files.
The non-visible wavelengths are important because they can be used to detect different levels of vegetation growth. The chlorophyll in plants gives off a certain kind of light that can’t be seen by the human eye, but it can be recorded by specialized cameras. These extra bands unlock a whole world of potential analyses that can detect subtle differences in vegetation growth across a landscape. In turn, they can be used to identify subterranean archaeological features (walls, ditches, roads) or ancient waterways. When you purchase a multispectral image, it can be bundled with a higher-resolution, panchromatic image. This allows the user to combine the two in a process called pansharpening, which merges the color bands with the high-resolution black-and-white imagery.
3. Hyperspectral. These images don’t just record a few bands of light – they record sometimes hundreds of very narrow bands. The goal is to cover the continuous spectrum of light, rather than record it in discrete bands. These types of images tend to be used in more specific applications that I won’t go into here.
In summary, RGB color and panchromatic are great for people who are just interested in identifying features on the ground now, tracing changes in the landscape over time, or even detecting seasonal variation that may reflect the presence of subsurface features. Multi- and hyperspectral imagery is best for more advanced applications that rely on subtle changes in vegetation growth.
The next step is to figure out the spatial resolution you need for your project. Spatial resolution can vary pretty widely, depending on when the image was taken and the price-point of the product. It’s usually measured in meters (compared to the resolution of aerial photos or maps, which is usually given as a scale, like 1:5,000). A resolution of 2 meters means that you probably won’t be able to see anything smaller than that. So, if you are looking for walls or terraces, you should consider going with a higher-resolution image.
This info was put together in September 2014 [links updated December 2016]. Prices are only estimates based on online sources, but prices may be lower with certain sales outlets, and academic discounts can range from 20-30%.
- MS = Multispectral resolution
- Pan = Panchromatic resolution
- Scene size = the total coverage of the scene, or the maximum swath width (if just a single number)
|Type||Years||Bands||MS (m)||Pan (m)||Scene size (km)||Cost (per sq. km)|
|CORONA||1960-1972||1||—||2-8||17 x 232||$30 per scene|
|Landsat 4-5 MSS||1982-1999||4||80||—||170 x 185||Free|
|Landsat 4-5 TM||1982-2012||7||30||—||170 x 185||Free|
|Landsat 7 ETM+||1999-||8||30||15||170 x 185||Free|
|Landsat 8||2013-||11||30||15||170 x 185||Free|
|SPOT 1-3||1986-1997||3||20||10||60 x 60||$1,200 per scene|
|SPOT 4||1998-2013||4||20||10||60 x 60||$1,200 per scene|
|SPOT 5||2002-||4||10||2.5 / 5||60 x 60||$2,700 per scene|
|WorldView-1||2007-||1||—||0.46||17.6 x 14||$13|
WHERE TO BUY?
CORONA and Landsat can be purchased or downloaded from the USGS. SPOT and Pleiades can be purchased from the Airbus GeoStore or other online retailers. Other products usually have to be purchased from a third-party retailer, like Satellite Imaging Corporation, Landinfo, or MapMart, just to name a few.
Casana, J., and J. Cothren. 2008. “Stereo Analysis, DEM Extraction and Orthorectification of CORONA Satellite Imagery: Archaeological Applications from the Near East.” Antiquity 82(317):732-49.
Challis, K. et al. 2002-2004. “Corona Remotely-Sensed Imagery in Dryland Archaeology: The Islamic City of al-Raqqa, Syria.” Journal of Field Archaeology 29 (1/2): 139-153.
Morehart, C.T. 2012. “Mapping Ancient Chinampa Landscapes in the Basin of Mexico: A Remote Sensing and GIS Approach.” Journal of Archaeological Science 39:2541-51.
Siart, C., O. Bubenzer, and B. Eitel. 2009. “Combining Digital Elevation Data (SRTM/ASTER), High Resolution Satellite Imagery (Quickbird) and GIS for Geomorphological Mapping: A Multi-Component Case Study on Mediterranean Karst in Central Crete.” Geomorphology 112:106-21.