The 4D-approach to dynamic machine vision

In this survey paper covering a decade of experience with recursive estimation methods for visual dynamic scene understanding, the common methodical background of the 4D approach exploiting spatio-temporal models of objects and processes observed is discussed. Applications range from `balancing an inverted pendulum´ over `vehicle docking´, `road vehicle guidance, `on-board autonomous aircraft landing approaches´, `landmark navigation for land vehicles´ to `grasping of an object freely floating in a satellite orbit with large delay times. Simultaneous recognition of rough 3D shape and motion parameters for moving road vehicles including occlusions has been achieved in well structured Autobahn scenes in real time (5 Hz). Recognition and tracking of moving humans in well structured environments has been advanced to a stage that real-time performance seems possible with the next generation of micro-processors just around the corner. After a brief review of the general method the aspects of exploiting object-oriented, model-based predictions for very efficient intelligent image feature extraction are discussed for some of the tasks considered