A hyperbola-pair based lane detection system for vehicle guidance
—Developing on-board automotive driver assistance systems aiming to alert drivers about driving environments, and possible collision with other vehicles has attracted a lot of attention lately. The ability to perceive, or sense, the surrounding environment is essential to driving and thus to the...
| Main Authors: | , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/9768/ http://irep.iium.edu.my/9768/ http://irep.iium.edu.my/9768/1/A_Hyperbola-Pair%5B1%5D.pdf |
| Summary: | —Developing on-board automotive driver assistance
systems aiming to alert drivers about driving environments,
and possible collision with other vehicles has attracted a lot of
attention lately. The ability to perceive, or sense, the
surrounding environment is essential to driving and thus to the
development of autonomous self-guided vehicles However, the
time requirements of an autonomous vehicle’s vision system are
very demanding, and require careful selection of high speed
algorithms for implementation. This paper presents a recent
vision-based on-road lane detection system. Our focus is on
systems where the camera is mounted on the vehicle rather than
being fixed such as in traffic/driveway monitoring systems. Real
time image sequences are used as inputs, which after being
processed are used for lanes detection. The lanes are detected
using Hough transform and fitted to a hyperbola model. The
proposed lane detection algorithm can be applied on both
painted and unpainted road as well as curved and straight road.
Finally, a critical overview of the methods was discussed, the
assessment of their potential for future deployment were
highlighted. This approach was tested and the experimental
results show that the proposed scheme was robust and fast
enough for real time requirements. Eventually, a critical
overview of the methods were discussed, their potential for
future deployment were assist. |
|---|