graphics and real-time feature extraction
In the late 1970’s, a Picture System 2 of E&S
allowed generating perspective views of road scenes by line
drawings like the one shown in the figure. The requested high
image rate was one of the basic ingredients to the success of the
4-D approach to real-time machine vision since it forced everybody
involved to think according to this dynamic constraint.
The other advantage was that quick changes from
one application area to another were possible with this setup.
This allowed test environments of different road types to be
available on a mouse click. The figure shows some real-time
results of road recognition [Meissner
Generated Images (CGI): real-time
feature extraction for crossroad detection 1990
The initial steps in developing this capability
have been done in the HIL-simulation loop since this
maneuver is rather complex needing bifocal vision and gaze
control: Beside road and crossroad detection
it also requires estimating
distance to the point of intersection
of both roads as well as
intersection angle between
experiments is essential since
timing for correct triggering and performance of the elements of
the turn-off maneuver is important; small errors in maneuver
execution may lead to dangerous situations in the real world.
This effort took several man-years until
sufficient fitness had been achieved for trying the transition to
real-world experiments with the test vehicle VaMoRs. Of course,
turning off to the right or to the left
are rather different maneuvers needing special attention.
in changing the geometry of intersections
(road widths with numbers of lanes, intersection angles) helps
increasing efficiency in
testing the capability achieved.
Above all, the main
advantage of CGI is that the scene to be recognized by vision is
already available in the computer for comparison and that
scene parameters can be varied at will.
Video 20 Turn-off
(1982): Steuerung dynamischer Systeme aufgrund bildhafter
Informationen. Dissertation, UniBwM / LRT.
Zapp A, Otto KD (1984). Ein Simulationskreis zur Entwicklung einer
automatischen Fahrzeugführung mit bildhaften und inertialen
Signalen. In Breitenecker, et al. (ed): Simulationstechnik,
Informatik-Fachberichte 85, Springer, pp 554-558
Zapp A (1985): Guiding Land Vehicles Along Roadways by Computer
Vision. Proc. Congres Automatique 1985, AFCET, Toulouse, pp
(1988): Automatische Straßenfahrzeugführung durch Rechnersehen,
Dissertation, UniBwM / LRT.
Müller N (1995). Navigation with an Autonomous Vehicle on
Arbitrary Roads. In Linkwitz et al. (eds): High Precision
Navigation. Dümmler Verlag, Bonn, pp. 415-42
(1996). Autonomes Manövrieren und Navigieren mit einem sehenden
Straßenfahrzeug, Dissertation, UniBwM, LRT