Domain-adaptive lane detection for transitioning weather conditions

Lane detection is a critical task handled by autonomous vehicles to perceive their environment accurately. Ensuring robust and reliable lane detection capabilities is essential for the safety of the vehicle's operations. However, adverse weather conditions like rain can significantly impede...

Full description

Saved in:
Bibliographic Details
Main Author: Gibson, Gideon Thomas
Other Authors: Soong Boon Hee
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
Online Access:https://hdl.handle.net/10356/177111
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Lane detection is a critical task handled by autonomous vehicles to perceive their environment accurately. Ensuring robust and reliable lane detection capabilities is essential for the safety of the vehicle's operations. However, adverse weather conditions like rain can significantly impede the performance of lane detectors. They introduce rain artifacts in images, which can distort the lane features, making them more difficult to distinguish. Rain also causes road surfaces to be more reflective to light, distorting much of the markings on the road, further causing a greater hindrance to lane detection. When it rains, the road surface also becomes more slippery, making correct and accurate lane detection even more crucial in autonomous vehicles for their safety. Therefore, it is crucial to adapt lane detectors to perform reliably in both clear and rainy conditions.