Lunar landing datasets

NAT-DATA-R1: Chang'e 3 navcam images

This dataset only includes the post-processed landing trajectory of Chang'e 3, but does not include the orginal Chang'e 3 images.

The images dataset is processed and produced by “Ground Research and Application System”（GRAS）of China’s Lunar and Planetary Exploration Program, provided by China National Space Administration. It is not re-distributed by ESA, but it can be downloaded at http://moon.bao.ac.cn/ce5web/. We thank the Chang'E-3 payload team for mission operations and China National Space Administration for providing the Chang'E-3 data that made the generation of other datasets possible.

NAT-DATA-L1: Chang'e 3 trajectory replicated in TRON

This dataset contains images acquired in the DLR's TRON facility that replicate the Chang'e 3 descent and landing trajectory on a scaled terrain model.

Number of images: 3658
Data source: DLR's TRON facility

NAT-DATA-L2: Random image pairs acquired in TRON

This dataset is a collection of random image pairs acquired in the TRON facility. It can be used to test dense or sparse optical flow and/or feature tracking algorithms.

Number of images: 7238
Data source: DLR's TRON facility

NAT-DATA-S1: Medium definition synthetic images of Chang'e 3 trajectory

This dataset contains 3655 synthetic images replicating Chang'e 3 trajectory and using a medium definition terrain model.

Number of images: 3655
Data source: SurRender software

NAT-DATA-S2: High definition synthetic images of Chang'e 3 trajectory

This dataset contains 3655 synthetic images replicating Chang'e 3 trajectory and using a high definition terrain model.

Number of images: 3655
Data source: SurRender software

NAT-DATA-S3: High definition synthetic images of Chang'e 3 trajectory with procedural details

This dataset contains 3655 synthetic images replicating Chang'e 3 trajectory and using a high definition terrain model complemented with procedural details.

Number of images: 3655
Data source: SurRender software

NAT-DATA-S5: High definition synthetic images of 3 generic trajectories

This dataset contains synthetic images for three exemplary descent and landing trajectories with different time of flights, and using a high definition terrain model.

Number of images: 6661 (slow), 5591 (med), 3736 (fast)
Data source: SurRender software

How to download

To donwload the datasets, please read and acknowledge our Licence, and then send us your download request by filling this form.

Access to the datasets is only granted for one month upon acceptance of your request. If your access has expired, please send a new request.

References

These datasets and the related documentation were produced as part of ESA Contract No. 4000137956/22/NL/CRS.

If you are using these datasets in your work, consider citing our paper:
Citation: J. Lebreton, I. Ahrns, R. Brochard, C. Haskamp, M. Le Goff, N. Menga, N. Ollagnier, R. Regele, F. Capolupo, M. Casasco, Training Datasets Generation for Machine Learning: Application to Vision-Based Navigation, SPAICE Conference 2024.