Synthetic dataset generation for driving scenes in low visibility and light conditions using GTAV
Self-driving cars must be able to operate safely under various environmental conditions, including challenging weather conditions like fog, rain or snow, as well as in different times of day. Collecting reliable real-world training data for these conditions is even more challenging than in favourabl...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/180884 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Self-driving cars must be able to operate safely under various environmental conditions, including challenging weather conditions like fog, rain or snow, as well as in different times of day. Collecting reliable real-world training data for these conditions is even more challenging than in favourable weather conditions, due to the unpredictability of weather and safety conditions.
To tackle these problems, this project aims to build upon the GTAV game engine to create synthetic datasets to supplement the existing inclement weather datasets, as well as to investigate and analyse the viability of using this form of synthetic dataset to train detection models for real life scenes. |
|---|