Scholarlink

<p>Space situational awareness increasingly relies on optical observations to detect and track resident space objects and to estimate spacecraft attitude. Many existing resources are synthetic or restricted, and few provide on orbit, wide field of view imagery with joint labels for space objects and stars. We present a dataset of near-infrared images acquired by the Fast Auroral Imager on the CASSIOPE spacecraft between January and August 2023. The collection comprises 1,378 frames with astrometrically calibrated stars and 4,237 manually verified resident space object instances across 160 transits, accompanied by spacecraft ephemeris, attitude, and image quality metrics. We describe the acquisition conditions, calibration and annotation pipeline, and perform technical validation of pointing stability, astrometric accuracy, annotation reliability, and background characteristics. The dataset supports tasks such as resident space object detection in dense star fields, multi-object tracking under realistic orbital motion, and attitude estimation from star tracker class imagery, and is intended as a shared resource for space situational awareness and navigation studies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An On-Orbit Star Tracker Benchmark for Resident Space Object Detection and Attitude Estimation

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月26日

查看全文
  • Vithurshan Suthakar,
  • Perushan Kunalakantha,
  • Regina S. K. Lee,
  • Gunho Sohn

摘要

Space situational awareness increasingly relies on optical observations to detect and track resident space objects and to estimate spacecraft attitude. Many existing resources are synthetic or restricted, and few provide on orbit, wide field of view imagery with joint labels for space objects and stars. We present a dataset of near-infrared images acquired by the Fast Auroral Imager on the CASSIOPE spacecraft between January and August 2023. The collection comprises 1,378 frames with astrometrically calibrated stars and 4,237 manually verified resident space object instances across 160 transits, accompanied by spacecraft ephemeris, attitude, and image quality metrics. We describe the acquisition conditions, calibration and annotation pipeline, and perform technical validation of pointing stability, astrometric accuracy, annotation reliability, and background characteristics. The dataset supports tasks such as resident space object detection in dense star fields, multi-object tracking under realistic orbital motion, and attitude estimation from star tracker class imagery, and is intended as a shared resource for space situational awareness and navigation studies.