File size: 1,341 Bytes
c2d4d2f
 
58894b0
 
 
 
 
 
 
 
ebf5c99
 
 
58894b0
 
c2d4d2f
 
ebf5c99
c2d4d2f
338d58f
ebf5c99
338d58f
ebf5c99
58894b0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
---
license: cc-by-3.0
task_categories:
- image-segmentation
language:
- en
tags:
- floods
- earth-observation
- deep-learnig
- Φsat-2
- PhiSat-2
pretty_name: WorldFloods-PhiSat-2
size_categories:
- 10K<n<100K
---

## WorldFloods Φsat-2 dataset

The **WorldFloods Φsat-2 dataset** is an adaptation of the [WorldFloods v2 dataset](https://huggingface.co/datasets/isp-uv-es/WorldFloodsv2) created with the purpose of simulating a Φsat-2-like dataset for training Foundation Models for this mission. 
Φsat-2 images were simulated using the code in the [OrbitalAI Challenge](https://github.com/AI4EO/orbitalAI). For storing the dataset, we followed the [PhiSatNet repo](https://github.com/sirbastiano/PhiSatNet) Data Specification Format.
Since the WorldFloods v2 dataset contains full flood scenes rather than small patches, we created non-overlapping Train and Validation patches of size 512x512 pixels from the simulated images, and only stored patches with less than 20% cloud cover, as we do on the fly to train the flood segmentation models described in the paper [Global flood extent segmentation in optical satellite images](https://www.nature.com/articles/s41598-023-47595-7).
For Test samples, however, we kept the full scenes as this allows a better evaluation of flood detection models. 
The total size of the dataset is 389 GB.