The complexity and poor accessibility of tropical rain forests render satellite images potentially very useful for geological and vegetation studies in these areas. We analysed a Landsat-MSS image from Peruvian Amazonia to find out which methods of digital procesing give the most useful results for detecting and delimiting different vegetation types and geological formations. The best results were obtained with enhanced colour composites, especially when histogram equalization was applied: the vegetation types that were known from the area were clearly visible in the image products, and also previously unknown regional patterns were found. Digital classifications proved far less satisfactory. Apart from producing gross misclassifications, they failed to provide spatially continuous patterns, and hence the products were less visually clear than the unclassified image. Some ground cover types could not be distinguished from each other by their spectral chracteristics alone, and therefore adequate field knowledge of the geology, geomorphology and vegetation of the study area proved indispensable in interpreting the satellite image.
