The Ministry of Transportation of Quebec (MTQ) is preparing the province of Quebec for the arrival of autonomous vehicles. The project has been entrusted to ÉTS and the Innovative Vehicle Institute (IVI), with the aim of mapping Quebec roads using images from a vehiclemounted camera. From these images, objects will be detected using computer vision algorithms. In this way, the MTQ aims to automatically detect potholes, broken streetlights and defective guardrails, in order to improve road safety.
This study focuses on the collection of local data in order to train detection models for road infrastructure. As a result, models based on deep learning and the YOLO architecture were trained. The best result obtained for pothole detection is 0.729 mAP. For the detection of broken lampposts, it is 0.840 mAP. In addition, this project has produced a database containing over 8000 images and 5 detection classes.
| Date | 11 Sept 2023 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Maarouf Saad (Supervisor) & Pier Marc Comtois-Rivet (Co-supervisor) |
|---|
Guillemelle, Q. (Author),
Saad (Supervisor) & Comtois-Rivet (Co-supervisor),
11 Sept 2023Student thesis: Master's thesis › Master in Engineering: Automated Manufacturing Engineering