Land cover classification using RADARSAT data in a mountainous area of southern Argentina

In this paper, a new procedure is proposed for land cover classification in a mountainous area using data derived from a stereo pair of RADARSAT images. Land cover classifications using the RADARSAT tonal and textural information and ancillary terrain data were evaluated. All the derived data were from the same source of the stereo pair of the RADARSAT images. An artificial neural networks (ANN) classifier is applied. The performance of the proposed method was evaluated over a mountainous study area in Southern Argentina. The results showed that the extra information, texture and DEM, extracted from the RADARSAT images can greatly improve the accuracy of classification using ANN. It can be concluded that RADARSAT images and terrain data derived from the RADARSAT images are valuable data sources for land cover mapping, especially in the mountainous areas where optical satellite data and DEM data are not always available.