Evidence fusion using the GESTALT-system

A production system is presented that can help coding knowledge about recognizing structured objects in a declarative way. Together with a respective interpreter it forms the GESTALT-system. This system is particularly suited for the fusion of evidence from different sensors giving data from the same scene, for the inclusion of prior knowledge - such as GIS or digital maps - and for the treatment of large numbers of primitive and non-primitive objects. In the last two decades the underlying system structure has been used for diverse applications. There have been several re-implementation endeavors. Three quite different examples of applications are presented.