Communication Dans Un Congrès Année : 2024

Efficient local correlation volume for unsupervised optical flow estimation on small moving objects in large satellite images

Résumé

With the advent of deep learning methods, performance and efficiency of optical flow estimation has significantly increased, especially for supervised models. However, they do not generalize well to more specific data involving small moving objects in large images, such as high-resolution aerial or satellite sequences. In addition, annotation and realistic simulation are difficult for these contents, which calls for unsupervised alternatives. Yet, the latter are still less accurate than their supervised counterparts. In this paper, we introduce an unsupervised local optical flow estimation method adapted to small moving objects in largesize images by involving no downsampling of the feature maps. We adopt a local correlation search and implement it in an original way with a per-shift computation, which minimizes memory consumption and speed up inference computation for large-scale images. We also design a loss function combining similarity, smoothness and sparsity constraints. We demonstrate the performance of our SMOFlow method on real stabilized aerial videos fully representative of future satellite conditions. SMOFlow favorably compares to other methods. Our SMOFlow method is able to accurately capture the motion of small objects in large images, while efficiently reducing memory consumption.
Fichier principal
Vignette du fichier
EARTHVISION2024-7.pdf (8.84 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04907647 , version 1 (23-01-2025)

Licence

Identifiants

Citer

Sarra Khairi, Etienne Meunier, Renaud Fraisse, Patrick Bouthemy. Efficient local correlation volume for unsupervised optical flow estimation on small moving objects in large satellite images. CVPRW 2024 - IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, Jun 2024, Seattle, United States. pp.440-448, ⟨10.1109/CVPRW63382.2024.00049⟩. ⟨hal-04907647⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

More