Detecting Damaged Buildings on Post-Hurricane Satellite Imagery (doi:10.21227/sdad-1e56)
Modified from here by removing duplicates
Source
Data originally taken from here
can be cited with http://dx.doi.org/10.21227/sdad-1e56
the original paper is here: https://arxiv.org/abs/1807.01688
Unfortunately, the source of the satellite imagery is not stated, however, it appears to be Maxar (~0.5 m horizontal grid size)
*Details
Locations are written into the filename
Contents small image tiles, cropped out of larger imagery (128x128x3 pixels)
- 14,022 damaged buildings
- 7,037 undamaged buildings
from the paper: "Essentially, the building coordinates, which can be easily obtained from publicdata (e.g., OpenStreetMap [2]), can be used as the centers of cropping. We use the building coordinates already associated with the damage labels fromTomnod. A window is then cropped from the raw satellite imagery to create a data sample. Tomnod volunteers’ annotation of flooded/damaged buildings istaken as the ground truth for the positive label, ‘Flooded/Damaged building’.At the same coordinates, we crop windows from the imagery captured beforethe hurricane to create negative data samples, labeled ‘Undamaged building’."