SPARK 2026 Dataset

The SPARK 2026 dataset contains two streams of data.

Stream 1: Multi-Task Spacecraft Perception

This stream challenges participants to design a single, powerful model capable of performing spacecraft classification, spacecraft detection, and fine-grained segmentation of spacecraft components—regardless of spacecraft type. The focus is on efficiency and performance, encouraging the development of compact, high-performing models suitable for deployment on resource-constrained space platforms.

Stream 2: Event-Based Pose Estimation

Dive into the world of event-based vision with Stream 2, which focuses on pose estimation using the SPADES dataset. Participants will train their models on high-quality synthetic event data and validate their approaches on real event data, addressing one of the most challenging perception problems in space.

To make a request for the data, please use the following contact form.

Stream 1
Stream 2

Related Publication

  • Rathinam, A., Mohamed Ali, M. A., Gaudilliere, V., & Aouada, D. (2024). SPARK 2024: Datasets for Spacecraft Semantic Segmentation and Spacecraft Trajectory Estimation [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10908215
  • Pauly, Leo, Michele Lynn Jamrozik, Miguel Ortiz Del Castillo, Olivia Borgue, Inder Pal Singh, Mohatashem Reyaz Makhdoomi, Olga-Orsalia Christidi-Loumpasefski, et al. “Lessons from a Space Lab — An Image Acquisition Perspective.” arXiv, August 18, 2022. http://arxiv.org/abs/2208.08865.