For a long time now, the remote sensing industry has been seen to have much promise in the acceptance and integration of satellite remote sensing imagery within the general GIS community. There have been many forecasts and prognostications that with the launch of newer and better satellite sensors and platforms, the "critical mass" of acceptance was just around the corner.Some segments of the GIS marketplace such as natural resources, local government and utilities have made effective use of satellite remote sensing in a cost and time effective manner.Business geographics on the other hand has made little use of remote sensing with the exception of limited and aerial photography, while very little use has been made of satellite remote sensing to any great extent.
Within the business geographics community, trying to acquire and make use of satellite remote sensing imagery has been a somewhat frustrating experience for the average desktop GIS user. There are a variety of reasons for this including:
- Size of remote sensing data sets - individual satellite images or air photo-derived ortho-images often approach many hundreds of megabytes in size. Displaying these datasets in GIS is often slow as a result. Image compression software tools are becoming more readily available and thereby helping to alleviate this problem.
- Many satellites, many resolutions - someone has to be an experienced remote sensing specialist to understand and make sense of the many sources and corresponding optimal scales for each sensor and camera. In addition, users are often confused about where to go to get the data.
- Matching remote sensing data to GIS data sets - things often just don't line-up (ie.- georeferencing problems) that require sophisticated and specialized software, not to mention time and effort to rectify.
- High cost of data for the perceived value in a GIS context - to many desktop GIS users, remote sensing is just a "pretty picture" and the perceived high cost doesn't justify using remote sensing imagery for simply that reason.
In some ways, satellite remote sensing has been written-off by many who looked at imagery from the 1980's through the mid 1990's as being too coarse or "blocky" to provide any meaningful use other than for regional land use/land cover assessments. Probably the single biggest recent development in the remote sensing community as it pertains to the business geographics sector, which might change this view, is the launch and availability of data from new high-resolution satellites. The IKONOS (Space Imaging) and QuickBird (Digital Globe) satellites which were launched in the fall of 1999 and fall of 2001 respectively, now offer one meter or better resolution on images that provide air photo detail from space. With these new satellites the "great divide" between air photography and satellite remote sensing is now spanned where the convenience of satellite frequent and repeat coverage can now compete with custom acquired air photos.
This article is intended to provide the remote sensing novice and skeptic with a fresh perspective on what some of the newer developments in satellite remote sensing can do for business geographics. Some background on the potential areas for remote sensing applications is presented along with an exploration of how remote sensing can provide a unique view that typical geodemographic data does not provide. The future of satellite remote sensing is here now - are there, are you making use of the data?
What is Remote Sensing Used For?
To the business geographics community (and the general public as a whole), remote sensing scenes or images are certainly very visually attractive. However, in a GIS context what exactly can you do with them? There are three primary ways in which remote sensing imagery has been used and these include:
(1) Generating Maps and/or Updating Existing Maps
This is one of the more traditional uses for aerial photography and remote sensing imagery. In areas in which mapping is not available due to coverage or scale limitations or simply through the data being out-of-date, remote sensing imagery is used as a source for generating and/or updating maps and associated features.
(2) "Image Mapping"
Image mapping describes the use of an image as a stand-alone map product unto itself. This usually requires the use of high quality printers or film recorders to generate a hard-copy product of good visual quality.
(3) Backdrops within GIS
This involves taking digital remote sensing images into GIS software to use as a backdrop for other geographic data themes. In this way, remote sensing imagery is useful for generating composite images and "on-screen" digitizing and general data verification.
Practical Uses of Imagery in Business Geographics
Throughout the years, remote sensing imagery derived from the workhorse Landsat Thematic Mapper (TM) satellite series has been extremely useful for large geographic/regional applications. The TM sensor continues to produce images which coverers 180 by 180 kilometers for a full-scene at a resolution of 30 meters. The recent Landsat 7 satellite has a 15-meter resolution panchromatic channel that produced some improvements in the features that could be detected. The two image products can be "fused" together to make a single composite image that gets the colors from the 30 meter TM image and the detail of the 15 meter image to provide enhanced visual detail.
Some of the "large-scale" remote sensing applications in business geographics include:
- Site selection/optimization - locating new structures and transportation routes
- Network planning - analyzing existing facilities and infrastructure data
- Visual orientation - "image mapping" which presents a more visually familiar view of the earth
- Land-use/Land Cover classifications - provides summary information on land use and land cover along with the ability to derive statistics for the land cover categories within the area of choice
In the later case, the telecommunications industry has seen a use for large-scale remote sensing imagery that has been classified for particular land-use categories of urban and vegetated areas referred to as "clutter classifications". These interpretations have seen use in modeling cellular telephone signal propagation, which in turn helps engineers to locate cell towers.
High-resolution satellites are now providing imagery that yields much finer detail of the earth's surface. The resolutions that are now commercially available include panchromatic imagery from the IKONOS 1 meter and 61 centimeter QuickBird satellites. These satellites essentially produce images that approach air photo quality in terms of discernable detail.
Some of the potential application areas for this type of high-resolution imagery in business might include:
- Insurance claims/disaster monitoring - insurance companies having imagery that helps to quantify the geographic damage extents after a natural disaster such as floods, tornadoes, etc.
- Site appraisal and location analysis - commercial and even residential real estate companies can quickly acquire images that portray very current conditions on the ground to assist in site evaluation, development and visual assessments
- Utility and telecommunication planning - areas of new housing development can be located and reviewed to assist in the network development and expansion planning process
Many other applications will undoubtedly emerge as more usage of this type of high-resolution imagery happens with the business geographics community.
"Drilling-Down" Into Remote Sensing Imagery
Have you ever spent some time lately working with remote sensing imagery in GIS? All of the satellite choices and associated resolutions mentioned previously won't be convincing in terms of how it makes a difference in day-to-day business geographics without actually seeing the visual results. In this light, the following figures have been produced.
Figure 1 shows how the relatively large-scale Landsat TM imagery can produce an interesting backdrop to GIS when compared with larger geo-demographic boundaries. The City of Calgary, Alberta is a rather large urban concentration with over 860,000 people within the city proper. The Census Metropolitan Area (CMA), which is depicted by the purple boundary in Figure 1, contains a much larger area, including the City of Calgary itself as well as other larger communities such as Cochrane and Airdrie, along with a number of other smaller communities.
Figure 1 - Landsat 7 TM Image Along With the Calgary Census Metropolitan Area (CMA) Boundary Superimposed
The image also reveals that within the CMA is a rather diverse range of land covers that are not revealed through a typical GIS or geo-demographic view of the world.
Landsat TM Imagery is not without its limitations and if one enlarged the image in this figure to it's full resolution, urban detail and in particular building details are not revealed due to its relatively coarse resolution. At best, Landsat TM imagery can used to update neighborhoods in which there is new growth or development. In some ways this type of user experience is in fact likely what has shaped GIS user perception of what satellite remote sensing products are like in a general sense.
In contrast to this, high-resolution satellite imagery allows for a whole new level of geographic scale and urban detail to be realized. To provide a practical and relevant context for business geographics, the typical GIS demographic and street network view of the world is compared and contrasted with the enhanced view that high-resolution satellite imagery can provide.
Figure 2 below illustrates a very typical GIS map view of a neighborhood in the City of Calgary, which features a typical street network file and census boundaries. In this image, two Enumeration Area (EA) census units are presented along with associated street networks. These EA's - numbers 48008367 and 48008368 - are found in a relatively upscale portion of the city.Table 1 below presents some detail on the geo-demographic characteristics as drawn from the 1996 census.
Table 1 - EA Population Characteristics (1996 Census)
In a GIS and geo-demographic sense, the street network configuration does infer some information as to pattern of the housing and seems to imply that the housing density is relatively low, but it does not reveal any indication of the likely land cover or land use. Aside from obtaining parcel boundary data from the city itself, there is no other direct way to gain much more detail about the neighborhood characteristics.
Figure 3 presents an "image map" view of the exact same EA's as in Figure 2, made up of an IKONOS 1 meter and 4 meter "fused" or combined panchromatic and color combined image with the EA boundaries superimposed. In this instance, the IKONOS image provides so much more detail than can be obtained from the simple GIS portrayal. Green spaces are brought out in significant detail along with some institutional land use that is not discernable without highly detailed municipal data. What is clearly highlighted is the non-homogeneous nature of the distribution of households within the EA. Should the user be interested, the image can be enlarged even further to reveal details such as the buildings themselves along with the cars on the road. Even this type of imagery has limits and only the outlines of these features are clear, and the viewer is unable to tell the type of car that is seen. Quite unlike the resolutions obtained from classified US government spy satellites, the ability to read newspaper headlines is not yet possible in the commercial realm.
There is no question that things have gotten much simpler in regards to the use of aerial photography and satellite imagery. New high-resolution satellite platforms can provide a whole new level of accessibility to imagery. The companies that own and operate these satellites provide significant information on image quality and characteristics on the corporate web sites, along with many image examples and educational material to make the use and access to imagery a much more pleasant experience for the novice user. Desktop GIS software has been enhanced to handle basic geo-referenced image formats such as GeoTIF as well as compressed image formats such as MrSID.
High-resolution imagery is now witnessing a number of practical uses within the business geographics community. Some of the early adopters so far have included the insurance and real estate sectors. It is important to note that this type of imagery is a complement to the Landsat TM, SPOT and other large-scale satellite platforms. These types of imagery are still very actively used in the resource industries and as mentioned in the telecommunications example, still used in clutter and other types of land-use classification.
High-resolution satellite remote sensing data can now provide additional detail to supplement the GIS view of the world that was once the exclusive domain of aerial photography. With data like that, satellite remote sensing is indeed worth another look in 2002.
The author would like to thank Space Imaging for allowing the use of the IKONOS imagery for this paper. Demographic and street network data are provided courtesy of DMTI Spatial.
Ikonos - Space Imaging corporate web site (www.spaceimaging.com)
Landsat - NASA web site (http://geo.arc.nasa.gov/sge/landsat/landsat.html)
QuickBird - Digital Globe corporate web site (www.digitalglobe.com)
SPOT Image - SPOT Image Corporation web site (www.spot.com)
Van Wyngaarden, Robert, 1998, "Satellite Data and Remote Sensing Applications for Business Geographics" In: GIS '98/RT '98: Pathways to Knowledge Integration Conference Proceedings, GIS World Inc.pp.25-28.
Figure 2 - Typical GIS View of Census and Street Network Data.
Figure 3 - GIS "Image Map" of Ikonos Imagery and EA Boundary Data.
Robert van Wyngaarden
Golder Associates Ltd.- GeoGraphic Information Services (GGIS)