Abstract
Major developments are now taking place in the integration of remote sensing data with GIS. Techniques such as the use of knowledge based systems and artificial neural networks appear to offer much potential for the extraction of improved geographical information from current remotely-sensed satellite imagery. Whilst the use of such techniques offers much promise in environmental monitoring and management, a number of fundamental problems remain such as choice of appropriate data structures, and procedures for handling error and uncertainty. Over the next ten years the complexity of remotely sensed datasets will grow significantly through use of multi-sensor, hyperspectral, and multi-view angle approaches besides use of time series. Current GIS techniques and technology are not appropriate to handle the much increased dimensionality of such datasets and new developments are needed both in visualization tools, and in spatial and temporal analysis tools.