Skip to content

Instantly share code, notes, and snippets.

@reachsumit
Last active June 24, 2018 21:54
Show Gist options
  • Save reachsumit/c790f291281661978954ee69bd8450ec to your computer and use it in GitHub Desktop.
Save reachsumit/c790f291281661978954ee69bd8450ec to your computer and use it in GitHub Desktop.
A comparison of AOIs in Spacenet dataset
Area of Interest Satellite Location Panchromatic Resolution Area of Raster (Sq Km) Building Labels (Polygons) Road Labels(LineString) Training Data Test Data
AOI_1 WorldView-2 Rio de Janeiro 50 cm 2544 4082529 N/A 3-band Mosiac = 2.3 GB | 8-band Mosiac = 6.5 GB | Vectors = 0.18 GB | Point of interest = 31 GB N/A
AOI_2 WorldView-3 Vegas 30 cm 216 151367 3685 km Roads: 25 GB of 400mx400m tiles | Buildings: 23 GB of 200mx200m tiles Roads: 8.1 GB of 400mx400m tiles | Buildings: 7.9 GB of 200mx200m tiles
AOI_3 WorldView-3 Paris 30 cm 1030 23816 425 km Roads: 5.6 GB of 400mx400m tiles | Buildings: 5.3 GB of 200mx200m tiles Roads: 1.9 GB of 400mx400m tiles | Buildings: 1.8 GB of 200mx200m tiles
AOI_4 WorldView-3 Shanghai 30 cm 1000 92015 3537 km Roads: 25 GB of 400mx400m tiles | Buildings: 23.4 GB of 200mx200m tiles Roads: 8.1 GB of 400mx400m tiles | Buildings: 7.7 GB of 200mx200m tiles
AOI_5 WorldView-3 Khartoum 30 cm 765 710960 1030 km Roads: 25 GB of 400mx400m tiles | Buildings: 4.7 GB of 200mx200m tiles Roads: 8.1 GB of 400mx400m tiles | Buildings: 1.6 GB of 200mx200m tiles
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment