Fusion of remote sensing data and geographic information system technology to map buried waste dump sites.

Abstract

Clandestine waste dumps can pose significant problems for planners and developers when their locations are not known. Use of geographic information systems (GIS) and satellite data offer opportunities to reduce cost and improve effectiveness in the search for such burial sites. A negative/positive decision protocol was developed for merging Landsat, Radarsat-1 and JERS-1 data within a GIS environment. A negative decision algorithm was first applied to eliminate areas where clandestine dump sites are not likely to be found: within 100 feet (30.5 meters) of a road right-of-way, on wetland soils, or on areas in continuous agriculture. A positive decision algorithm based on Landsat data was then applied to identify pixels with spectral signatures similar to those of known buried waste sites. The 231 pixels identified included all twelve pixels with dump sites subsequently confirmed through field checks. Overall classification accuracy was 88.4%. Radar data from Radarsat-1 and JERS-1 were merged with the Landsat and other data layers in the GIS. All of the pixels subsequently confirmed to have dump sites had (C band - L band) values smaller than -0.17. Using this as an additional decision role increased the overall accuracy to 99.3%.