- Smooth surfaces with homogeneous appearance and good contrast at the boundaries are the easiest ones.
- If there are buildings or trees near the road and sunshine casts shadows on the road, recognition of the road boundaries may become very difficult; usually, the argument of smooth continuation of road edges gives a good guideline for grouping edge elements. Edge intensity is not a good argument for selection and combination of edge elements.
- Since boundaries of dirt roads are not smooth and crisp but rather jagged, edge detection by large masks with a wide gap between the positive and the negative parts of the mask are favorable.
- Clothoid models are well suited for perceiving these roads also even though they might not have been designed this way.
Video 1990 DissMy.mp4 (initial part)
![A3.3im.001 Road fork](https://dyna-vision.de/wp-content/uploads/2021/05/A3.3im.001-Road-fork-scaled.jpg)
Road fork recognition
At larger distances, using region-based features is more robust for detecting and tracking roads including road forks (see figure):
References
Mysliwetz B. (1990): Parallelrechner-basierte Bildfolgen-Interpretation zur autonomen Fahrzeugsteuerung. Dissertation, UniBwM / LRT. Kurzfassg
Bruedigam C (1994). Intelligente Fahrmanoever sehender autonomer Fahrzeuge in autobahnaehnlicher Umgebung. Dissertation UniBwM / LRT. Kurzfassg
Behringer R (1996). Visuelle Erkennung und Interpretation des Fahrspurverlaufes durch Rechnersehen für ein autonomes Straßenfahrzeug. Dissertation, UniBwM / LRT. Kurzfassg
Lützeler M, Dickmanns ED (1998). Road recognition with MarVEye. Intern. Conf. on Intelligent Vehicles, Stuttgart.
Gregor R, Lützeler M, Pellkofer M, Siedersberger KH, Dickmanns ED (2000). EMS-Vision: A Perceptual System for Autonomous Vehicles. Proc. Int. Symposium on Intelligent Vehicles (IV’2000), Dearborn, (MI) pdf
Luetzeler M (2002). Fahrbahnerkennung zum Manoevrieren auf Wegenetzen mit aktivem Sehen. Dissertation, UniBwM, LRT. Kurzfassg
Siedersberger K-H (2004). Komponenten zur automatischen Fahrzeugführung in sehenden (semi-) autonomen Fahrzeugen. Dissertation, UniBwM, LRT. Kurzfassg
Dickmanns ED (2007). Dynamic Vision for Perception and Control of Motion. Springer-Verlag, (Section 11.3.4.1)