OpenSAR Insight

SAR imagery

This dataset uses Synthetic Aperture Radar (SAR) images from the European Space Agency's (ESA) Copernicus Sentinel-1 mission, which come from two satellites (Sentinel-1A and Sentinel-1B) in polar orbit that operate in all weather conditions, both day and night. The selected Level 1 SAR products are Ground Range Detection (GRD) in wide interferometric mode (IW) and Single Look Complex (SLC) burst products.

The two available polarisation channels (VV and VH) have been exploited from these data. In this case, the VV polarization is useful for detecting flooding in open water areas. Water acts as a smooth surface, causing specular reflection (similar to a mirror), resulting in a very low backscatter signal. Therefore, flooded areas appear as dark areas in the image. By the same token, VH polarisation is particularly sensitive to double bounce scattering, which occurs when the radar signal first bounces off a vertical structure (such as a tree trunk) and then off a horizontal surface (water). This phenomenon is very characteristic of flooded vegetation, causing these areas to appear bright in the image. VH polarisation is less sensitive to the roughness of the water surface. An important new feature of this dataset is the availability of Level 0 data (L0, RAW data) extracted from the Level 1 data (L1) used.

SAR images location

SAR Data & labeling

The dataset generated for the Flood Detection case study consists of a total of 10 different SAR scenes. Each scene is composed of a time series. Within each time series, there are different SAR products Sentinel-1 Level 1 Ground Range Detected (GRD) and Single Look Complex (SLC). For the selection of the SAR scenes that will make up the dataset corresponding to the Flood Detection use case, the proposed Kuro Siwo database has been used as the main source. This database offers a manually labelled multi-temporal dataset comprising 43 flood events across different regions of the world. Kuro Siwo is built upon SAR Ground Range Detected products, along with a primary SAR Single Look Complex product that has undergone minimal pre-processing. It is specifically designed to support research focused on leveraging both phase and amplitude information, while also providing maximum flexibility for custom pre-processing in downstream applications.

In this use case, we have used the Kuro Siwo database to analyse a total of 10 different scenes, corresponding to different geographical areas. From the total number of proposed scenes, the dataset has been divided into two utilities:

• 6 scenes (2,3,4,5,7,10) will be used to delimit the contours of water bodies.
• 4 scenes (1,6,8,9) to detect temporal changes in the extension of water bodies.

The labelling strategy is based on XML (Extensible Markup Language) format. In the next table are detailed the considered tags, levels, and descendants for all Flood detections files (including metadata, data, and labels).

Level	Tag Name	Descendants	Tag Contents
0	Scene_Info	Scene_ID; Date; Version; CaseStudy	-
1	Scene_ID	N/A	Naming of the SAR scene into the database
1	Date	N/A	YYYYMMDD
1	Version	N/A	Version of the SAR scene
1	CaseStudy	N/A	DarkVesselDetection
0	SARData	SAR_Mission; SAR product; SLCSwath; Time_interval	-
1	SAR_Mission	N/A	S1A-S1B
1	SARProduct	N/A	Name of the SLC product SAR
1	SLCSwath	N/A	Number of swath
1	Time_Interval	Start; Stop	-
2	Start	N/A	YYYYMMDD_HHMMSS
2	Stop	N/A	YYYYMMDD_HHMMSS
0	ProcessingData	Corner_Coord; WindData; StatisticsReport; List_of_ships	-
1	Corner_Coord	SARData_Sample; SARData_Line; Scene_Sample; Scene_Line; Latitude; Longitude	-
2	SARData_Sample	N/A	Number of columns of 512 pixels in the geocoded image
2	SARData_Line	N/A	Number of rows of 512 pixels in the geocoded image
2	Scene_Sample	N/A
2	Scene_Line	N/A
2	Latitude	N/A	Latitudes of corners
2	Longitude	N/A	Longitudes of corners
1	GeoInfo	GRD; AOI	-
2	GRD	N/A	Whether the patch area is within GRD product bounds (1) or not (0). Sometimes GRD products are cropped smaller than the SLC product, resulting in not all SLCs having a valid corresponding GRD image.
2	AOI	N/A	Whether the patch is fully within a Kuro Siwo Area of Interest (0), partially intersecting with an AOI (1), or fully outside all AOIs (2). The original Kuro Siwo dataset does not label full scenes, so this tag was added to ensure users have a clear understanding of whether a patch is fully labelled or only partially.
1	StatisticsReport	SeaArea; Number_of_FloodEvents; Number_of_WaterBodys;	-
2	SeaArea	Scenario
3	Scenario	N/A	Patch Scenario
2	Number_of_FloodEvents	N/A	Number of flood events detected in the patch.
2	Number_of_WaterBodys	N/A	Number of water bodies detected in the patch.
1	List_of_WaterBodys	Body	-
2	Body	Body	Name; Polygon
3	Name	N/A	String with the ID of the water body
4	Polygon	Latitude;Longitude; SARData_Sample; SARData_Line; Scene_Sample; Scene_Line;	-
5	Latitude	N/A	Latitude coordinate of the polygon
5	Longitude	N/A	Longitude coordinate of the polygon
5	SARData_Sample	N/A	Column coordinate of the polygon (patch)
5	SARData_Line	N/A	Row coordinate of the polygon (patch)
5	Scene_Sample	N/A	Column coordinate of the polygon (scene)
5	Scene_Line	N/A	Row coordinate of the polygon (scene)

Data partitions

The data relating to this case study are organised as follows:

Flood use-case	XML labels	Masks	L0 patches	SLC patches	GRD patches
Train set	3481	3481	6962	6962	6962
Val set	423	423	846	846	846
Test set	200	200	400	400	400
Total	4104	4104	8208	8208	8208

Note: XMLs denote the number of unique patches. For this use-case, masks are included. L0, SLC, and GRD are doubled as they include both VV and VH polarisations.