i.sentinel_2.ndvidiff - GRASS GIS manual

NAME

i.sentinel_2.ndvidiff - Calculates NDVI difference maps from Sentinel-2 data.

KEYWORDS

SYNOPSIS

i.sentinel_2.ndvidiff

i.sentinel_2.ndvidiff --help

i.sentinel_2.ndvidiff [-c] input_dir_first=name input_dir_second=name aoi=name [ndvi_loss_map_vect=name] [ndvi_loss_map_rast=name] [ndvi_diff_map=name] [ndvi_map_first=name] [ndvi_map_second=name] [rgbi_basename=string] [output_dir=string] [ndvi_diff_threshold=float] [relevant_min_ndvi=float] [nprocs=integer] [offset=integer] [cloud_shadow_buffer=integer] [min_size=integer] aggregation_method=string [--overwrite] [--help] [--verbose] [--quiet] [--ui]

Flags:

-c: Run cloud masking (and mosaicking) using t.rast.mosaic/i.sentinel.mask
--overwrite: Allow output files to overwrite existing files
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

input_dir_first=name [required]: Input directory that holds imagery from the first time interval
input_dir_second=name [required]: Input directory that holds imagery from the second time interval
aoi=name [required]: Name of input vector map; Vector map that holds AOI area/s
ndvi_loss_map_vect=name: Output vector map that contains areas of NDVI loss
ndvi_loss_map_rast=name: Output raster map that contains areas of NDVI loss (intermediate result)
ndvi_diff_map=name: Output raster map that contains the NDVI difference map (intermediate result)
ndvi_map_first=name: NDVI raster map of the first time interval (intermediate result)
ndvi_map_second=name: NDVI raster map of the second time interval (intermediate result)
rgbi_basename=string: Basename to save aggregated RGBI-groups
output_dir=string: Output directory to write result files (RGBI/ndvi/loss rasters & vectors)
ndvi_diff_threshold=float: threshold to apply to NDVI difference map. If none is given, the threshold is calculated from the NDVI diff map using thr=Q1-1.5*(Q3-Q1)
relevant_min_ndvi=float: delimit identification of NDVI loss areas to pixels with an NDVI value of at least <relevant_min_ndvi> in the first time interval
nprocs=integer: Number of parallel processes to use; Default: 1
offset=integer: Offset to add to the Sentinel bands to due to specific processing baseline (e.g. -1000)
cloud_shadow_buffer=integer: Buffer in pixels to account for inacurracies in cloud/shadow masks; Default: 5
min_size=integer: Minimum size of detected ndvi loss areas (in map units)
aggregation_method=string [required]: temporal aggregation method used in t.rast.mosaic; Options: average, count, median, mode, minimum, min_raster, maximum, max_raster, stddev, range, sum, variance, diversity, slope, offset, detcoeff, quart1, quart3, perc90, quantile, skewness, kurtosis; Default: minimum

DESCRIPTION
NOTES
EXAMPLE
- Create an NDVI difference map in GRASS only, save intermediate results, skip cloud/shadow detection, use an automatic threshold for the detection of significant NDVi loss
- Export the NDVI difference map and intermediate results to a local directory, use a defined threshold, apply cloud masking and use a cloud/shadow buffer, apply a minimum size
REQUIREMENTS
SEE ALSO
AUTHORS

DESCRIPTION

i.sentinel_2.ndvidiff is a GRASS addon that calculates an NDVI difference map from Sentinel-2 L2A data.

The goal of the addon is to compare the NDVI status from two time intervals for an AOI given by the vector map aoi. For each interval,
Sentinel-2 L2A data is imported from the local file system (using t.sentinel.import) given with` input_dir_first` and `input_dir_second`, cloud-masked (using t.sentinel.mask),
and temporally aggregated to receive one spectral band (using t.rast.mosaic) per time interval.

One NDVI map for each time interval as well as a difference map between the two are calculated. The user may define
a threshold via the ndvi_diff_threshold option, which is used to extract areas of signficant NDVI loss as raster and/or vector maps.

The outputs of the module (RGBI mosaics, NDVI maps, NDVI difference map, NDVI loss areas) can be saved as GRASS raster/vector maps or
exported using the output_dir option.

NOTES

The offset option can be used to take the systematic reflectance offset into account that is added to Sentinel-2
reflectance data since Processing Baseline 4.0.0.

If no ndvi_diff_threshold is given, a threshold is automatically calculated from the NDVI difference map using
thresh = Q1 - 1.5 * (Q3 - Q1), where Q1 and Q3 are the first and third quantile of the NDVi difference map.

For cloud masking, the Sen2Cor cloud mask delivered with L2A data is combined with the output of t.sentinel.mask.

The temporal aggregation method to be passed on to t.rast.mosaic can be defined via the aggregation_method option.
Best results were achieved with minimum aggregation which reduces negative effects due to remaining clouds.

The cloud_shadow_buffer option can be used to define a buffer radius (in raster cells) that is applied to detected
clouds and cloud shadows to compensate for inaccuracies in the cloud/shadow detection.

Cloud/Shadow masking is optional and is activated via the -c flag. If visual inspection shows that all input S-2 scenes
are cloudfree, it may be omitted.

The nprocs option can be used to run the different submodules in parallel.

With the min_size option, small areas of detected NDVI loss may be removed

The rgbi_basename option can be used to define a basename in case the aggregated RGBI maps should be persistent in GRASS

A minimum NDVI value can be passed by the relevant_min_ndvi option. In this case, identification of significant NDVI loss areas is
delimited to pixels where the first aggregated NDVI map has a value of at least relevant_min_ndvi.

EXAMPLE

Create an NDVI difference map in GRASS only, save intermediate results, skip cloud/shadow detection, use an automatic threshold for the detection of significant NDVi loss

i.sentinel_2.ndvidiff input_dir_first=~/S2_2020 input_dir_second=~/S2_2024 aoi=aoi_vector ndvi_loss_map_rast=ndvi_loss_result ndvi_diff_map=ndvi_diff_map_2024_2020 ndvi_map_first=ndvi_s2_2020 ndvi_map_second=ndvi_s2_2024 nprocs=4 offset=-1000 aggregation_method=minimum

Export the NDVI difference map and intermediate results to a local directory, use a defined threshold, apply cloud masking and use a cloud/shadow buffer, apply a minimum size

i.sentinel_2.ndvidiff -c input_dir_first=~/S2_2020 input_dir_second=~/S2_2024 aoi=aoi_vector offset=-1000 nprocs=4 cloud_shadow_buffer=10 aggregation_method=minimum output_dir=~/result_dir ndvi_diff_threshold=-0.25 min_size=200

REQUIREMENTS

- t.sentinel, - t.rast.mosaic - grass-gis-helpers Python module

AUTHORS

Guido Riembauer, mundialis

SOURCE CODE

Available at: i.sentinel_2.ndvidiff source code (history)

Accessed: Thursday Aug 15 12:57:17 2024