Calibration routine

Processing steps below.

SLC

Sentinel-1 SLC split
Thermal noise removal
Calibration
Deburst
Merge - optional when multiple subswaths selected
Range doppler terrain correction

GRD

Thermal noise removal
Calibration
Range doppler terrain correction

Requirements

~13 GB RAM for full scene (3 subswaths)
~1.3 GB GPU memory for full scene (depending on DEM tiles)

Arguments

alus-cal

ALUs - Calibration
Version 1.6.0

Arguments:

  -h [ --help ]              Print help
  -i [ --input ] arg         Input SAFE dataset (zipped or unpacked)
  -o [ --output ] arg        Output folder or filename
  -w [ --wif ]               Write intermediate results (will be saved in the 
                             same folder as final outcome). NOTE - this may 
                             decrease performance. Default OFF.
  --sw arg                   Subswath for which the calibration will be 
                             performed, one of the following - IW1;IW2;IW3. Not
                             used for GRD.
  -p [ --polarisation ] arg  Polarisation for which the calibration will be 
                             performed - VV;VH
  --bi1 arg                  First burst index - starting at '1', leave 
                             unspecified for whole subswath and GRD
  --bi2 arg                  Last burst index - starting at '1', leave 
                             unspecified for whole subswath and GRD
  -a [ --aoi ] arg           Area Of Interest WKT polygon. Or shapefile (.shp) 
                             consisting similar geometry.Overrules first and 
                             last burst indexes and subswath selections. Not 
                             effective for GRD.
  -t [ --type ] arg          Type of calibration to be performed, one of the 
                             following - sigma;beta;gamma;dn
  --dem arg                  DEM file(s). SRTM3 and Copernicus DEM 30m COG are 
                             currently supported.

  --ll arg (=verbose)        Log level, one of the following - 
                             verbose|debug|info|warning|error
  --gpu_mem arg (=100)       Percentage of how much GPU memory can be used for 
                             processing


https://github.com/cgi-estonia-space/ALUs

Performance

Please see the latest and always updated comprehensive evaluation in the Wiki

Following contents below might be deprecated

Reference laptop computer details:
CPU: Intel i7 10750h
RAM: 32GB
GPU: NVIDIA GeForce GTX 1660 Ti 6GB
SSD (NVMe): SAMSUNG MZALQ512HALU-000L2

Compared against SNAP 8 GPT on Ubuntu 20.04.
Configuration:
-Xmx22381m
-Xms2048m
-XX:+AggressiveOpts
-Xverify:none
-Dnetbeans.mainclass=org.esa.snap.main.Main
-Dsun.java2d.noddraw=true
-Dsun.awt.nopixfmt=true
-Dsun.java2d.dpiaware=false
Tile size 512
Number of threads 12

Input:

S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE

Processing options:

Subswath IW2 (all bursts)
Polarisation VV
calibration type sigma

For more raw computing comparison, the IW2 subswath GeoTIFF file has been stored on ram disk - this way I/O influences results less. Also outputs are stored on ram disk.

SNAP using SRTM3

./gpt calibration_routine_snap_graph.xml 
-Pinput=S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE/manifest.safe
-Poutput=/mnt/ramdisk/snap_cal_iw2.tif -Psw=IW2 -Ppol=VV

~42.5 seconds

SRTM3

./alus-cal -i S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE -o /mnt/ramdisk --sw IW2 -p VV 
-t sigma --dem srtm_37_02.tif --dem srtm_37_03.tif --dem srtm_38_02.tif --dem srtm_38_03.tif --ll info

Output:

[2023-02-22 16:10:55.524778] [0x00007fd0ad7df000] [info]    Processing parameters:
Input product - ../S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE
Subswath - IW2
Polarisation - VV
Calibration type - sigma
First burst index - 0
Last burst index - 0
AOI - 
Write intermediate files - NO

[2023-02-22 16:10:55.699153] [0x00007fd0ad7df000] [info]    Sentinel 1 split done - 174ms
[2023-02-22 16:10:55.699182] [0x00007fd0ad7df000] [info]    Using 'NVIDIA GeForce GTX 1660 Ti' device nr 0 for calculations
[2023-02-22 16:10:55.709495] [0x00007fd0ad7df000] [info]    Target bands count: 1
[2023-02-22 16:10:55.709515] [0x00007fd0ad7df000] [info]    Processing band i_IW2_VV_VV
[2023-02-22 16:10:58.266871] [0x00007fd0ad7df000] [info]    Thermal noise removal done - 2567ms
[2023-02-22 16:10:58.289810] [0x00007fd0ad7df000] [info]    target bands count 1
[2023-02-22 16:10:58.289835] [0x00007fd0ad7df000] [info]    Processing band Sigma0_IW2_VV_VV
[2023-02-22 16:10:59.024776] [0x00007fd0ad7df000] [info]    Sentinel1 calibration done - 757ms
[2023-02-22 16:10:59.137422] [0x00007fd0ad7b5000] [info]    Loading SRTM3 tile ID 3702 to GPU
[2023-02-22 16:10:59.219146] [0x00007fd0ad7b5000] [info]    Loading SRTM3 tile ID 3703 to GPU
[2023-02-22 16:10:59.286741] [0x00007fd0ad7b5000] [info]    Loading SRTM3 tile ID 3802 to GPU
[2023-02-22 16:10:59.365022] [0x00007fd0ad7b5000] [info]    Loading SRTM3 tile ID 3803 to GPU
[2023-02-22 16:11:00.287164] [0x00007fd0ad7df000] [info]    TOPSAR Deburst done - 1186ms
[2023-02-22 16:11:00.287185] [0x00007fd0ad7df000] [info]    Merge done - 0ms
[2023-02-22 16:11:02.285407] [0x00007fd0ad7df000] [info]    Terrain correction done - 1998ms
[2023-02-22 16:11:02.539286] [0x00007fd0ad7df000] [info]    Algorithm completed, output file @ S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_cal_deb_tc_srtm3.tif
[2023-02-22 16:11:02.539304] [0x00007fd0ad7df000] [info]    Another test
[2023-02-22 16:11:02.701453] [0x00007fd0ad7df000] [info]    Unloading SRTM3 tiles from GPU
~/alus/alus-github/cmake-build-release/alus_package/alus-cal -i  -t sigma -p   10.14s user 4.80s system 204% cpu 7.319 total

~7.4 seconds

COPDEM 30m

time ~/alus/alus-github/cmake-build-release/alus_package/alus-cal -i ../S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE -t sigma -p VV --sw IW2 --dem Copernicus_DSM_COG_10_N49_00_E004_00_DEM.tif --dem Copernicus_DSM_COG_10_N49_00_E005_00_DEM.tif --dem Copernicus_DSM_COG_10_N50_00_E004_00_DEM.tif --dem Copernicus_DSM_COG_10_N50_00_E005_00_DEM.tif --dem Copernicus_DSM_COG_10_N50_00_E006_00_DEM.tif --dem Copernicus_DSM_COG_10_N51_00_E004_00_DEM.tif --dem Copernicus_DSM_COG_10_N51_00_E005_00_DEM.tif -o S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_cal_deb_tc_copdem.tif --ll info

Output:

[2023-02-23 11:24:53.843607] [0x00007f32c100b000] [info]    Processing parameters:
Input product - ../S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE
Subswath - IW2
Polarisation - VV
Calibration type - sigma
First burst index - 0
Last burst index - 0
AOI - 
Write intermediate files - NO

[2023-02-23 11:24:54.030635] [0x00007f32c100b000] [info]    Sentinel 1 split done - 186ms
[2023-02-23 11:24:54.030664] [0x00007f32c100b000] [info]    Using 'NVIDIA GeForce GTX 1660 Ti' device nr 0 for calculations
[2023-02-23 11:24:54.040753] [0x00007f32c100b000] [info]    Target bands count: 1
[2023-02-23 11:24:54.040771] [0x00007f32c100b000] [info]    Processing band i_IW2_VV_VV
[2023-02-23 11:24:56.622863] [0x00007f32c100b000] [info]    Thermal noise removal done - 2592ms
[2023-02-23 11:24:56.645418] [0x00007f32c100b000] [info]    target bands count 1
[2023-02-23 11:24:56.645439] [0x00007f32c100b000] [info]    Processing band Sigma0_IW2_VV_VV
[2023-02-23 11:24:57.373840] [0x00007f32c100b000] [info]    Sentinel1 calibration done - 750ms
[2023-02-23 11:24:57.486067] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 184041 loaded to GPU
[2023-02-23 11:24:57.512886] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 185041 loaded to GPU
[2023-02-23 11:24:57.532775] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 184040 loaded to GPU
[2023-02-23 11:24:57.553562] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 185040 loaded to GPU
[2023-02-23 11:24:57.573801] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 186040 loaded to GPU
[2023-02-23 11:24:57.592789] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 184039 loaded to GPU
[2023-02-23 11:24:57.611409] [0x00007f32b98d7000] [info]    COPDEM COG 30m tile ID 185039 loaded to GPU
[2023-02-23 11:24:58.635122] [0x00007f32c100b000] [info]    TOPSAR Deburst done - 1175ms
[2023-02-23 11:24:58.635144] [0x00007f32c100b000] [info]    Merge done - 0ms
[2023-02-23 11:25:00.550532] [0x00007f32c100b000] [info]    Terrain correction done - 1915ms
[2023-02-23 11:25:00.803102] [0x00007f32c100b000] [info]    Algorithm completed, output file @ S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_cal_deb_tc_copdem.tif
[2023-02-23 11:25:00.803121] [0x00007f32c100b000] [info]    Another test
[2023-02-23 11:25:00.951947] [0x00007f32c100b000] [info]    Unloading COPDEM COG 30m tiles
~/alus/alus-github/cmake-build-release/alus_package/alus-cal -i  -t sigma -p   11.87s user 4.84s system 229% cpu 7.288 total

~7.2 seconds

Comparison

Resulting diff file statistics when using SRTM3:

gdal_calc.py -A S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_tnr_Cal_deb_TC_snap.tif 
-B S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_tnr_Cal_IW2_deb_tc.tif --calc="A-B" 
--outfile=/tmp/snap_man_min_alus_cal_diff.tif --NoDataValue=0

Band 1 Block=14224x1 Type=Float32, ColorInterp=Gray
    Computed Min/Max=-0.004,0.004
  Minimum=-0.004, Maximum=0.004, Mean=0.000, StdDev=0.000
  NoData Value=0
  Metadata:
    STATISTICS_MAXIMUM=0.00408935546875
    STATISTICS_MEAN=1.6726067428608e-09
    STATISTICS_MINIMUM=-0.0042724609375
    STATISTICS_STDDEV=4.1099733990715e-06
    STATISTICS_VALID_PERCENT=4.028

Legend

Difference when using SRTM3 DEM

Difference when using Copernicus DEM 30m COG

There are known bugs for both of the tools when using Copernicus DEM documented in - https://forum.step.esa.int/t/copernicus-dem-complications-when-coregistering-s1/38659/2

gdal_calc.py -A snap/S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_Cal_deb_TC_copdem_no_data.tif -B S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_cal_deb_tc_copdem.tif --calc="A-B" --outfile=S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7_split_IW2_tnr_cal_deb_tc_copdem_snap_min_alus_diff.tif --NoDataValue=0

Resulting diff file statistics:

Band 1 Block=14224x1 Type=Float32, ColorInterp=Gray
  Minimum=-0.000, Maximum=0.000, Mean=0.000, StdDev=0.000
  NoData Value=0
  Metadata:
    STATISTICS_MAXIMUM=4.5299530029297e-05
    STATISTICS_MEAN=4.0933211844668e-10
    STATISTICS_MINIMUM=-2.6583671569824e-05
    STATISTICS_STDDEV=1.6678710960146e-07
    STATISTICS_VALID_PERCENT=0.2109

The match is nearly identical for Copernicus DEM, since its resolution is higher and less floating point arithmetics introduced rounding errors between interpolated values of the pixels. The relative difference picture is almost completely dark. The 'nodata' values are a box with white color, inside, the difference can be seen (black == equal).

Here is a zoomed in picture where pixels are different on byte level (maximum difference 4.5299530029297e-05) are marked with white, highlighted with red.

Sources

Input

S1A_IW_SLC__1SDV_20210703T055050_20210703T055117_038609_048E45_35F7.SAFE

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

Calibration routine

Requirements

Arguments

alus-cal

Performance

Comparison

Sources

Files

README.md

Latest commit

History

README.md

File metadata and controls

Calibration routine

Requirements

Arguments

alus-cal

Performance

Comparison

Sources