Datasets:

Modalities:
Image
Text
Formats:
text
ArXiv:
Libraries:
Datasets
License:
Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringlengths
16
20
M-33-20-D-c-4-2_0
M-33-20-D-c-4-2_1
M-33-20-D-c-4-2_10
M-33-20-D-c-4-2_100
M-33-20-D-c-4-2_102
M-33-20-D-c-4-2_104
M-33-20-D-c-4-2_106
M-33-20-D-c-4-2_108
M-33-20-D-c-4-2_109
M-33-20-D-c-4-2_11
M-33-20-D-c-4-2_111
M-33-20-D-c-4-2_112
M-33-20-D-c-4-2_114
M-33-20-D-c-4-2_115
M-33-20-D-c-4-2_118
M-33-20-D-c-4-2_119
M-33-20-D-c-4-2_12
M-33-20-D-c-4-2_121
M-33-20-D-c-4-2_122
M-33-20-D-c-4-2_123
M-33-20-D-c-4-2_13
M-33-20-D-c-4-2_130
M-33-20-D-c-4-2_131
M-33-20-D-c-4-2_132
M-33-20-D-c-4-2_133
M-33-20-D-c-4-2_137
M-33-20-D-c-4-2_139
M-33-20-D-c-4-2_14
M-33-20-D-c-4-2_140
M-33-20-D-c-4-2_141
M-33-20-D-c-4-2_142
M-33-20-D-c-4-2_144
M-33-20-D-c-4-2_146
M-33-20-D-c-4-2_148
M-33-20-D-c-4-2_149
M-33-20-D-c-4-2_15
M-33-20-D-c-4-2_151
M-33-20-D-c-4-2_152
M-33-20-D-c-4-2_154
M-33-20-D-c-4-2_155
M-33-20-D-c-4-2_156
M-33-20-D-c-4-2_157
M-33-20-D-c-4-2_159
M-33-20-D-c-4-2_16
M-33-20-D-c-4-2_160
M-33-20-D-c-4-2_162
M-33-20-D-c-4-2_163
M-33-20-D-c-4-2_164
M-33-20-D-c-4-2_166
M-33-20-D-c-4-2_167
M-33-20-D-c-4-2_168
M-33-20-D-c-4-2_169
M-33-20-D-c-4-2_170
M-33-20-D-c-4-2_171
M-33-20-D-c-4-2_172
M-33-20-D-c-4-2_173
M-33-20-D-c-4-2_174
M-33-20-D-c-4-2_175
M-33-20-D-c-4-2_176
M-33-20-D-c-4-2_177
M-33-20-D-c-4-2_18
M-33-20-D-c-4-2_180
M-33-20-D-c-4-2_181
M-33-20-D-c-4-2_182
M-33-20-D-c-4-2_184
M-33-20-D-c-4-2_185
M-33-20-D-c-4-2_187
M-33-20-D-c-4-2_188
M-33-20-D-c-4-2_189
M-33-20-D-c-4-2_190
M-33-20-D-c-4-2_191
M-33-20-D-c-4-2_192
M-33-20-D-c-4-2_194
M-33-20-D-c-4-2_196
M-33-20-D-c-4-2_198
M-33-20-D-c-4-2_199
M-33-20-D-c-4-2_20
M-33-20-D-c-4-2_200
M-33-20-D-c-4-2_201
M-33-20-D-c-4-2_203
M-33-20-D-c-4-2_204
M-33-20-D-c-4-2_206
M-33-20-D-c-4-2_208
M-33-20-D-c-4-2_209
M-33-20-D-c-4-2_211
M-33-20-D-c-4-2_212
M-33-20-D-c-4-2_213
M-33-20-D-c-4-2_214
M-33-20-D-c-4-2_216
M-33-20-D-c-4-2_218
M-33-20-D-c-4-2_219
M-33-20-D-c-4-2_220
M-33-20-D-c-4-2_221
M-33-20-D-c-4-2_223
M-33-20-D-c-4-2_224
M-33-20-D-c-4-2_225
M-33-20-D-c-4-2_23
M-33-20-D-c-4-2_230
M-33-20-D-c-4-2_234
M-33-20-D-c-4-2_235
End of preview. Expand in Data Studio

The Dataset

The LandCover.ai (Land Cover from Aerial Imagery) dataset is a dataset for automatic mapping of buildings, woodlands, water and roads from aerial images.

Dataset features

  • land cover from Poland, Central Europe (1)
  • three spectral bands - RGB
  • 33 orthophotos with 25 cm per pixel resolution (~9000x9500 px)
  • 8 orthophotos with 50 cm per pixel resolution (~4200x4700 px)
  • total area of 216.27 km2

Dataset format

  • rasters are three-channel GeoTiffs with EPSG:2180 spatial reference system
  • masks are single-channel GeoTiffs with EPSG:2180 spatial reference system

(1): Image source: Head Office of Geodesy and Cartography, Poland

Reproduce and compare

We provide split.py to split images into 512x512 pieces, and following files: train.txt, val.txt and test.txt containing lists of pieces used for training, validation and testing respectively.

Versions

Version 1

  • classes: building (1), woodland (2), water(3), road(4)
  • areas: 1.85 km2 of buildings, 72.02 km2 of woodlands, 13.15 km2 of water, 3.5 km2 of roads

Paper

Citation

To cite our work, please use the following:

@InProceedings{Boguszewski_2021_CVPR,
    author = {Boguszewski, Adrian and Batorski, Dominik and Ziemba-Jankowska, Natalia and Dziedzic, Tomasz and Zambrzycka, Anna},
    title = {LandCover.ai: Dataset for Automatic Mapping of Buildings, Woodlands, Water and Roads from Aerial Imagery},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month = {June},
    year = {2021},
    pages = {1102-1110}
}

License

This dataset is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Contact

If you encounter any problem or have any feedback, please contact [email protected]

Downloads last month
215