Aerial View River Landform Video Segmentation: A Weakly Supervised Context-Aware Temporal Consistency Distillation Approach
Part Of
Proceedings - International Conference on Image Processing, ICIP
Start Page
984
End Page
990
ISSN
15224880
DOI (of the container)
10.1109/ICIP51287.2024.10647662
ISBN
979-835034939-9
Date Issued
2024-10-27
Author(s)
DOI
10.1109/ICIP51287.2024.10647662
Abstract
The study of terrain and landform classification through UAV remote sensing diverges significantly from ground vehicle patrol tasks. Besides grappling with the complexity of data annotation and ensuring temporal consistency, it also confronts the scarcity of relevant data and the limitations imposed by the effective range of many technologies. This research substantiates that, in aerial positioning tasks, both the mean Intersection over Union (mIoU) and temporal consistency (TC) metrics are of paramount importance. It is demonstrated that fully labeled data is not the optimal choice, as selecting only key data lacks the enhancement in TC, leading to failures. Hence, a teacher-student architecture, coupled with key frame selection and key frame updating algorithms, is proposed. This framework successfully performs weakly supervised learning and TC knowledge distillation, overcoming the deficiencies of traditional TC training in aerial tasks. The experimental results reveal that our method utilizing merely $30 \%$ of labeled data, concurrently elevates mIoU and temporal consistency ensuring stable localization of terrain objects. Result demo: https://gitlab.com/prophet.ai.inc/drone-based-riverbed-inspection
Event(s)
31st IEEE International Conference on Image Processing, ICIP 2024
SDGs
Publisher
IEEE
Type
conference paper