Discover the SciOpen Platform and Achieve Your Research Goals with Ease.
Search articles, authors, keywords, DOI, etc.
In the field of close-proximity space operations, RGB-D cameras have been used widely. However, publicly available spacecraft datasets for pose estimation are primarily based solely on image data. There is no publicly available dataset based on RGB-D cameras that can be directly used. Therefore, this paper introduces a novel RGB-D camera based spacecraft (RDCS) dataset. Due to high acquisition costs and confidentiality issues, it is difficult to obtain real spacecraft data and their associated pose information. To address this, the space environment and camera trajectories are constructed in Blender to capture pixel-aligned RGB images and depth maps of spacecraft from various observation angles. In addition, the RDCS dataset also includes pose annotations, bounding box labels, and corresponding mask images. The process focuses on addressing the challenges of acquiring the RGB image and the depth map of the spacecraft at the same keyframe while ensuring their pixel alignment, as well as determining the corresponding pose annotations. Finally, a spacecraft pose estimation network based on the pixel-level fused feature is applied to verify the availability of the RDCS dataset.
Distributed under a Creative Commons Attribution License (CC BY 4.0).
Comments on this article