WP5 FLOV tool webpage


FLOod Vulnerability (FLOV) tools

Jordi Inglada

1. FLOV processing chain

The FLOV processing chain is composed of 3 modules. The first one uses supervised (SVM) or unsupervised (K-Means) classification of images during the flooding in order to produce a flooded area extent map.

The second one uses change detection between pre-event and post-event images in order to produce the flooded area extent map.

The third module uses a Digital Elevation Model (DEM) in order to forecast the flooded area.

2. Software packages used

  • OTB/Monteverdi (CNES)

    • Image visualization

    • Feature extraction

    • Unsupervised change detection

    • Supervised change detection

  • TerraLib/TerraView/TerraHydro (INPE)

    • Feature extraction

    • Unsupervised change detection

    • Flooding simulation

  • Spring (INPE)

    • Image visualization

    • DEM management

3. Processing steps

3.1. Image visualization and compositing

3.2. Feature extraction

3.3. Supervised classification

3.4. Supervised change detection

3.5. Unsupervised classification - TerraView

3.6. Unsupervised classification - Monteverdi

3.7. Vector to vector comparison

4. Additional documentation

The final presentation of the FLOV tools and their results, provided during TOLOMEO Final Meeting, is available here. Moreover, video tutorials of these tools are also available at:

Finally, the so called “protocol documents”, showing how to use the tools in practice, can be obtained http://tolomeofp7.unipv.it/node/74.

5. Where can one find and download FLOV Tools?

6. Complete data set

A test dataset can be downloaded http://tolomeofp7.unipv.it/FLOV/complete.tar.gz. Use the login and password provided after signing the licence agreement.