Dickmanns and Schick 1991: Real-time machine vision is at the verge of becoming an
economically feasible development step in technology since very recent times.
The development of digital microprocessors both with respect to single unit
performance per unit cost, volume and power consumption and with respect to
parallelization is reaching a state where high frequency image processing
becomes affordable. At the same time, the development of algorithms for dynamic
scene understanding has advanced to a state where tasks like taxiway automation
may become solvable.
The 4D-approach to real-time vision as developed at UniBwM is discussed for this application. Results achieved in the framework of the EUREKA project PROMETHEUS and other activities are transferred to the airport environment. In this application area, imaging sensors may be CCD-TV cameras, infrared cameras or imaging radar devices, or combinations thereof for all-weather conditions.
Both central surveillance and onboard environmental sensing may be realized using different versions of the same technology. Results achieved in road vehicle guidance and surveillance are discussed.