Inland Water Mapping using Sentinel-1 and Open Street Map data

This past training was on

Friday 22nd of October 2021

This RUS Copernicus training session focuses on the processing of Sentinel-1 products with the SNAP machine learning capabilities and Open Street Map as auxiliary data to detect and map inland water areas over areas with man-made structures (a difficult case in terms of classification).

This is a hands-on training session where participants will access from their own laptop a Virtual Machine to exploit the open source software and toolboxes available in the RUS environment (mainly SNAP and QGIS here) to process Sentinel-1 data.

Please note that the RUS Copernicus Virtual Machines used for this training course can only be provided to citizens/residents of the Copernicus programme member countries (EU plus Iceland and Norway). 

Background

With their night and day all weather imaging capabilities, SAR Imagers onboard the Sentinel-1 missions (A and B) offer the input material to produce High Resolution water masks over wide areas.

In this training session, we will introduce the main theoretical concepts for water mapping from C-band SAR imagery. This short introduction will be followed by a hands-on session to illustrate how the Sentinel-1 GRD products can be exploited with the SNAP toolbox to obtain a high resolution water mask in raster (DIM format) and then vector masks (KML and ESRI-Shapefile). Open Street Map layers will be selected and imported into SNAP to support the machine learning classification stage. The consolidation of the mask (merging polygons and some clean-up) will be done with QGIS.

Participants will learn how to access the Sentinel-1 data, be familiar with the RUS environment and discover very interesting functionalities of the freely accessible SNAP toolbox.

Equipment

Attendees must be in possession of their own computer with a microphone to follow the online training. They will be provided with a pre-installed VM, accessible through an internet browser on their personal laptop. The VM will contain all the software and data necessary to follow the exercise during the course and will also be accessible to users during the following week.

Cost & Participation

Attendance to the training course is free of charge.

Audience & Selection

This training session is designed for a beginner audience, wishing to discover Sentinel-1 data download and processing. Space is limited to 25 participants on a first-come first-served basis.

Trainers

Radar Signal Processing Engineer

Pierre Fabry, Along-Track

Share this training:

Training:

Inland Water Mapping using Sentinel-1 and Open Street Map data

Training date & time

Friday 22nd of October 2021

13:00 – 17:30 CEST

Location:

Due to the uncertainties related to the coronavirus situation, this hands-on session will be held fully online.

Contact