Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- image-classification
|
| 5 |
+
- image-segmentation
|
| 6 |
+
tags:
|
| 7 |
+
- lunar
|
| 8 |
+
- moon
|
| 9 |
+
- remote-sensing
|
| 10 |
+
- multimodal
|
| 11 |
+
- foundation-model
|
| 12 |
+
- planetary-science
|
| 13 |
+
pretty_name: Moonstone
|
| 14 |
+
size_categories:
|
| 15 |
+
- 10K<n<100K
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
+
# Moonstone: A Multimodal Foundation Model Benchmark for Lunar Remote Sensing
|
| 19 |
+
|
| 20 |
+
28-channel, 128 pixels-per-degree (~237 m/pixel) global multimodal lunar dataset assembled from
|
| 21 |
+
seven instrument families across five missions (LRO WAC/LOLA/Diviner/Mini-RF, Chandrayaan-1 M3,
|
| 22 |
+
GRAIL, Lunar Prospector GRS, Clementine). All channels are aligned to a common equirectangular
|
| 23 |
+
grid (46,080 x 23,040 px, lunar sphere a=b=1,737,400 m) and organized into 7 physical modality
|
| 24 |
+
groups (surface, thermal, spectral_M3, gravity, radar, hapke, composition).
|
| 25 |
+
|
| 26 |
+
## Contents
|
| 27 |
+
|
| 28 |
+
| Path | Description |
|
| 29 |
+
|------|-------------|
|
| 30 |
+
| `aligned/` | 28 source-of-truth GeoTIFFs at 128 ppd (+ M3 geometry, crater_mask) |
|
| 31 |
+
| `mmap/` | Pre-normalized (z-scored) memory-mapped float32 arrays for pretraining (unlimited random crops) + NaN masks + channel_index.json |
|
| 32 |
+
| `lunar_patches_v4.h5` | 16,200 patches (180x90 grid) x 28 x 256 x 256, with geology/age/mare metadata and fixed 70/15/15 split, for downstream evaluation |
|
| 33 |
+
| `channel_stats.json` | Per-channel (mean, std) normalization statistics (200 random 256x256 windows) |
|
| 34 |
+
|
| 35 |
+
## Channels (28)
|
| 36 |
+
|
| 37 |
+
surface: wac_morphology, elevation, slope, roughness · thermal: diviner_tbol_midnight,
|
| 38 |
+
diviner_temp_night, rock_abundance, christiansen_feature · spectral_M3: m3_{750,950,1000,1250,1580,2000,2817,2857} ·
|
| 39 |
+
gravity: grail_{freeair,bouguer,uncertainty} · radar: minirf_{cpr,s1} (log1p) ·
|
| 40 |
+
hapke: wac_hapke_{415,566,604,689}nm · composition: clementine_uvvis_750nm, lpgrs_{tio2,feo}
|
| 41 |
+
|
| 42 |
+
## Benchmark tasks
|
| 43 |
+
Geology (49-class), Age (5-class), Composition (FeO/TiO2 regression), Cross-modal thermal
|
| 44 |
+
prediction, Mare/highlands segmentation, Crater (>10 km) segmentation.
|
| 45 |
+
|
| 46 |
+
## Provenance
|
| 47 |
+
All data derived from public NASA PDS / USGS / ODE archives. Built via the 15-step pipeline in
|
| 48 |
+
the Moonstone code repository (steps 01-15 + fix_minirf). Normalization: z-score, NaN->0 after norm.
|