USGS TerraWeb: Change Detection

Change Detection

Because many aspects of global climatic change reflect changes in the environment, such as aridification and human-induced impacts to ecosystems, change detection and mapping using satellite image data provide a valuable research and monitoring tool. As an example, the automatic change detection results generated using two multitemporal Landsat MSS images of Phoenix, Arizona are shown below. Two radio metrically calibrated MSS images taken 18 years apart were used to generate a digital change image. This image product shows the potential of using remotely sensed images for change detection in arid/semi-arid environments and for studying not only surficial processes in the desert (including subtle changes in vegetation and bedforms), but also to map changes caused by human disturbance. This type of change detection procedure can help document and evaluate changes that have occurred on interannual and/or seasonal basis. In this change image shades of red implies that the surface was more vegetated in 1992 than in 1974 and shades of blue more vegetated in 1974 than in 1992. Notice that changes can be seen in the surrounding desert vegetation, as well as in agricultural and urban areas. The left image was recorded on August 31, 1974 and the middle image on August 27, 1992; the image on the right is the digital 'change image' generated using these two data sets.

Click on the image to see a larger version (525K)

A second change detection example using multitemporal Landsat MSS images is presented to further show the potential use of remotely sensed satellite data to study the desert environment, map eolian erosion, and monitor global change. The images were used to detect vegetation changes in the desert; they cover parts of extreme southwestern Arizona, including the area around Yuma. The first pair of Landsat MSS images detects changes between a wet (April 4) and dry (July 1) season in 1992. Chavez and MacKinnon (1994) found that a radiometrically calibrated visible band is better to detect changes in desert vegetation than a vegetation index (such as the NDVI). Landsat MSS band 5 (red spectral band) for the two dates is shown along with the corresponding digital change image. The second pair of images below are the Landsat MSS band 5 and their corresponding digital change image for the same area; however, these images were recorded on February 24, 1981 (dry year) and February 26, 1984 (wet/El Nino year). In both sets of images the dark tones in the change image represent areas that were more vegetated during the wet season/year and the brighter tones are areas that were more vegetated during the dry season/year (mostly agricultural areas/crops).

Click on these images to see larger versions (518K, 521K)

To Eolian Mapping
To Rainfall Mapping

Back to Eolian and Rainfall Mapping Title Page

U.S. Department of the Interior | U.S. Geological Survey | Jobs
Home | Search | Disclaimers & Privacy Web Rings
Send comments to webman@TerraWeb.wr.usgs.gov
URL:
Last Modified: Thu Oct 31 15:12:07 MST 2002