Perception for autonomous driving systems under low-light conditions | |
| Author | Bhatti, Muhammad Omer Farooq |
| Call Number | AIT Thesis no.DSAI-23-04 |
| Subject(s) | Automated vehicles--Data processing Automated vehicles--Control Machine learning |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Data Science and Artificial Intelligence |
| Publisher | Asian Institute of Technology |
| Abstract | A great deal of progress has been made recently on autonomous driving systems (ADS). However, despite this progress, we are as of yet unable to achieve fully autonomous driv ing capability without human supervision. The problem areas occur in the perception system of the ADS, where unexpected conditions can lead to a failure of the perception module, causing accidents. Such accidents undermine public trust in the technology and set back the adoption of fully autonomous vehicles. Many of the errors in percep tion are caused by variations on adverse lighting conditions. In this thesis, I explore the particular issue of low-light situations, in which, due to low exposure, normal methods for high-level vision tasks may not work as well. This thesis presents a comparative study of various state-of-the-art image enhancement models for the purpose of facilitat ing high-level vision tasks. The results enable us to evaluate the suitability of particular models for the required task under low-light conditions. In addition, joint-training of En lightenGAN with YOLOv5 is performed. The resulting model enables us to learn image features which are more robust under varying illumination conditions as well as produce state-of-the-art results (79.5% mAP) for object detection under low-light conditions. |
| Year | 2023 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Data Science and Artificial Intelligence (DSAI) |
| Chairperson(s) | Dailey, Matthew N. |
| Examination Committee(s) | Mongkol Ekpanyapong;Chaklam Silpasuwanchai |
| Scholarship Donor(s) | AIT Scholarships |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2023 |