The potential of expert systems for remote sensing

Abstract

The use of expert systems for the interpretation of remote-sensing data is discussed. Expert-system architectures are described; the systems in use, under development, or proposed in various scientific and industrial fields are listed in a table; and three image-interpretation systems are briefly characterized. It is found that expert systems have the potential to perform repetitive interpretation of image data compiled from various remote-sensing satellites, especially to detect changes over time as in the case of deforestation. A system comprising an image interpreter, change interpreter, a cause interpreter, and a global data base is outlined and illustrated with a diagram.