General description of the dataset

The Master Flood Database consists of flood images picked from different publicly available datasets. The origin of the images is specified in the “sources.csv” file. The train- val split was made using 3:1 ratio. For images taken from the “Flood Area Segmentation” dataset and the “Water Dataset” (url links are given in “sources.csv”), we normalized the binary masks in the {0,1} values, so we provide our version of masks. For images taken from the “Roadway Flooding Image Dataset”, we use the annotations path from the original dataset, since the masks are already normalized in the wanted range.

Moreover, Test videos contain video frames from real flooding scenarios in Greece and Italy, to test the trained model in real world data. We extracted frames from these videos. Given the test video, its frames are named with our convention “frame.jpg”, where is the frame number as extracted from the video. We provide the segmentation annotations for these frames, with the naming convention “frameIds.png”. The Greek video is too big, so we selected and annotated 567 frames from it. As for the Italian video, we selected and annotated all 1406 frames.

Dataset Structure

Inside each folder train-val-test, there is a csv file that contains the real image path (on the source dataset), the annotation path (on our database) and the source of each image.

To advance research in the relevant field, AUTH made the dataset publicly available for research purposes via the AIIA lab. All requests for access to/use of the dataset must be submitted in writing by researchers. In order to access the Flood Master Database created/assembled by Aristotle University of Thessaloniki, please complete and sign this license agreement. Subsequently, email it to Prof. Ioannis Pitas (using “TEMA – Master Flood Database availability” as e-mail subject) so as to receive FTP credentials for downloading.