University of Twente Student Theses

Login
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.

Racetrack width and car relative position extraction using an image-based segmentation model

Piibar, O.J. (2025) Racetrack width and car relative position extraction using an image-based segmentation model.

[img] PDF
5MB
Abstract:Computer vision is widely used and researched, however, it is almost never used alone, but assisted by other sensors. This research relies fully on computer vision to extract spatial data from the simulation racing environment Assetto Corsa, where direct access to track dimensions is unavailable. This work addresses the challenge of extracting such data visually to reduce the dependency of reinforcement learning agents on hard-coded track information and improve their ability to generalise to tracks it has never seen. A light-weight image-based segmentation model was developed to detect the track and its road boundary points. The model relies on a histogram-based approach that scans the binary image horizontally to find white pixel value peaks corresponding to road markings. The results are compared to ground truth data using metrics such as R-squared, MAPE, Pearson’s correlation coefficient, and residual analysis. The results: MAPE of 15.91%, R-squared of 0.1164, Pearson’s of 0.917 and a mean residual of 1.76 meters demonstrate that the model performs adequately when extracting spatial data from the the AC sim racing environment relying only on image-based input.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Business & IT BSc (56066)
Link to this item:https://purl.utwente.nl/essays/107493
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page