i.sentinel_1.water - GRASS GIS manual

NAME

i.sentinel_1.water - Automatic thresholding and region growing to detect water areas in S1 GRD scenes.

KEYWORDS

SYNOPSIS

i.sentinel_1.water

i.sentinel_1.water --help

i.sentinel_1.water [-m] input=string output=string hand_rast=string [water_core_thresh=string] [water_boundary_thresh=string] [memory=memory in MB] [iterations=integer] [segment_threshold=string] [hand_max=string] polarization=string [--help] [--verbose] [--quiet] [--ui]

Flags:

-m: Don't run automatic threshold identification. Provide thresholds via water_boundary_thresh and water_core_thresh instead.
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

input=string [required]: S-1 map name; S-1 data must be calibrated and in logarithmic scale (dB). The second last part of the scene name split by '_' must contain the polarisation (VV or VH)
output=string [required]: Output binary water map
hand_rast=string [required]: Name of the HAND index raster; Raster containing Height Above Nearest Drainage (HAND) values
water_core_thresh=string: Manual threshold for core water areas (in dB); All areas below water_core_thresh will be selected as core water areas
water_boundary_thresh=string: Manual threshold for water boundary areas (in dB); All areas below water_boundary_thresh will be selected as water boundary areas
memory=memory in MB: Maximum memory to be used (in MB); Available memory for running i.segment; Default: 300
iterations=integer: Iterations of region growing algorithm (i.segment); Default: 50
segment_threshold=string: Threshold of region merging for region growing algorithm (i.segment); Must be between 0 and 1; Default: 0.1
hand_max=string: Maximum HAND value (meters above drainage network) where water areas are assumed; Default: 15.0
polarization=string [required]: Polarization of dataset; Must be VV or VH; Options: VV, VH

DESCRIPTION
NOTES
REFERENCES
SEE ALSO
AUTHOR

DESCRIPTION

i.sentinel_1.water detects water areas in Sentinel-1 GRD scenes using automatic or manual thresholding and region growing. Water areas are corrected using a Height-Above-Nearest-Drainage (HAND) raster.

NOTES

i.sentinel_1.water allows for automatic or manual thresholding of calibrated Sentinel-1 GRD data in log_dB format. If automatic thresholding is applied, the image is tiles into a regular grid of 500x500 Pixels and the histogram is calculated for each tile. The histogram is smoothed and checked for bimodality by counting the relative maxima. If the histogram is bimodal, the dB-value of the first peak as well as the valley between the peaks (identified using the minimum filter from python's skimage package) are saved. If no bimodal tiles are identified, the process is repeated on tiles of size 400x400 pixels, and then 300x300 Pixels. Global thresholds are then identified by averaging the db-values of first peaks and histogram valleys from all tiles. The average of the first peaks is used as threshold to identify core water areas. The average of the hístogram valleys is used as threshold to identify the maximum possible water extent. Alternatively, both thresholds can be passed by the water_core_thresh and water_boundary_thresh parameters when the -m flag is activated. The core water areas are used as seeds for the region growing algorithm of r.s1.grd, where the Sentinel-1 scene is used as input. The maximum possible water extent is included as boundary map in the region growing process. Only those regions that originate from the seeds are saved from the region growing process and classified as water areas. Finally, the water areas are corrected using a threshold for the provided Height-Above- Nearest-Drainage (HAND) raster.

REFERENCES

Cao, H.; Zhang, H.; Wang, C. and Zhang, B. 2019, "Operational Flood Detection Using Sentinel-1 SAR Data over Large Areas", Water 2019, 11(4),786; https://doi.org/10.3390/w11040786

AUTHOR

Guido Riembauer, mundialis, Germany

SOURCE CODE

Available at: i.sentinel_1.water source code (history)

Accessed: Tuesday Apr 23 08:38:18 2024