hvent90/lunar-albedo

Lunar Surface Albedo Tiles

High-resolution lunar surface albedo imagery organized as a cube-face tile pyramid, with precise geographic and physical metadata for each tile.

Dataset Description

This dataset contains albedo (reflectance) images of the Moon's surface derived from NASA's Lunar Reconnaissance Orbiter Camera (LROC) Wide Angle Camera (WAC) global mosaic. The images are projected onto a cube-sphere mapping for seamless planetary rendering.

Source Data

• Instrument: LROC Wide Angle Camera (WAC)
• Mission: Lunar Reconnaissance Orbiter (LRO)
• Original Resolution: 100m/pixel global mosaic
• Source: NASA PDS

Dataset Structure

Each sample contains:

• image: 1024x1024 JPEG tile
• Geographic coordinates (center, bounds, corners)
• Physical scale metrics

Metadata Fields

Field	Type	Description
center_lat	float	Tile center latitude (-90° to +90°)
center_lon	float	Tile center longitude (-180° to +180°)
lat_min/max	float	Latitude bounding box
lon_min/max	float	Longitude bounding box
corner_N_lat/lon	float	Exact corner coordinates (N=0-3)
meters_per_pixel_mean	float	Ground sample distance (altitude proxy)
meters_per_pixel_min/max	float	Resolution range across tile
tile_area_km2	float	Physical surface area covered
face	string	Cube face (face0-face5)
tile_x, tile_y	int	Tile position within face
spans_antimeridian	bool	Whether tile crosses ±180° longitude

Cube Face Mapping

Face	Direction	Coverage
face0	+X	Near side (0° lon centered)
face1	-X	Far side (180° lon centered)
face2	+Y	North polar region
face3	-Y	South polar region
face4	+Z	Eastern limb (90° E)
face5	-Z	Western limb (90° W)

Usage

from datasets import load_dataset

ds = load_dataset("hvent90/lunar-albedo")

# Access a sample
sample = ds["train"][0]
image = sample["image"]  # PIL Image
lat, lon = sample["center_lat"], sample["center_lon"]
resolution = sample["meters_per_pixel_mean"]  # meters per pixel

# Filter by resolution (altitude proxy)
high_res = ds["train"].filter(lambda x: x["meters_per_pixel_mean"] < 150)

Intended Uses

• Neural texture synthesis for planetary rendering
• Diffusion model training for lunar surface generation
• Multi-scale/altitude-conditioned image generation
• Geographic ML research

Physical Scale Reference

meters_per_pixel	Equivalent Viewing Altitude
~2400	Orbital (~10,000 km)
~500	High altitude (~2,000 km)
~100	Low altitude (~400 km)

Citation

If using this dataset, please cite the original LROC data:

Robinson, M. S., et al. (2010). Lunar Reconnaissance Orbiter Camera (LROC) Instrument Overview. Space Science Reviews, 150(1-4), 81-124.

License

The source imagery is in the public domain (NASA). This processed dataset is released under CC-BY-4.0.