aalto1 untyped-item.component.html
Camera lidar calibration using transformer
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Sähkötekniikan korkeakoulu |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
ELEC3055
Degree programme
Language
en
Pages
60+7
Series
Abstract
This thesis studies a fundamental problem for the autonomous driving system, vehicle perception, with an emphasis on sensor calibration to improve environmental context awareness, which will eventually contribute to better path planning. In this work, we use the proposed Transformer to demonstrate that, even in cases where extrinsic parameters are unknown, models can substantially benefit by estimating features from raw camera images and point clouds directly without having any prior knowledge of calibrated parameters.
The work attempts to utilize Vision Transformers (ViTs) for extracting visual information from images and also investigates different network designs for point cloud feature representation, such as Stratify models and PointNet++ in the context of LiDAR. The whole idea of the study is to see how one could maximize this by experimenting with both prior information and raw data.