A small body open-source dataset for image processing algorithms

Zusammenfassung (Abstract) :

The capability to accurately detect surface morphological features such as craters and boulders of multiple sizes and at different scales on the surface of small bodies is of paramount importance for various vision-based applications around small bodies. The development of this capability, however, is hindered by significant challenges: the environmental conditions due to the irregularity of the bodies, properties, and distribution of the features, rapidly changing illumination conditions, and most importantly, the lack of publicly available datasets for training, validation, or testing. In this work, the authors describe the methodology used to generate a general-purpose open-access dataset specifically designed to simplify access to labeled datasets about small bodies and to enable the design and application of advanced data-driven image processing algorithms.

Stichworte :

Dataset; Analog facility; Small bodies; Space exploration; Drone; Segmentation; Navigation

Files:

CEAS-GNC-2024-065.pdf

Links:

https://eurognc.ceas.org/archive/EuroGNC2024/html/CEAS-GNC-2024-065.html