69

FAQ: Radiometric Calibration of Worldview data

 

  1. Where to locate Gain and Offset values in Worldview metadata ?

Worldview data are not provided with Gain and Offset values. Only Absolute Scale Factor (ASF) and Effective Bandwith (EBW) are provided. Those values are located in the .mdi or .xml files provided with the dataset. ENVI reads those values them from the .xml file

For example: Wordlview-3 data SWIR Band S1

  <ABSCALFACTOR>2.671600000000000e-04</ABSCALFACTOR>

  <EFFECTIVEBANDWIDTH>3.300000000000000e-02</EFFECTIVEBANDWIDTH>

The absolute radiometric calibration values for Gain and Offset are not included in WorldView dataset. They are provided by Digital Globe and are revised regularly.

 

 

  1. Which absolute Gain and Offset values are used in ENVI for Worldview data calibration

In ENVI the absolute radiometric calibration values for Gain and Offset are hard coded in the following file

<INSTALL_DIR>\Harris\ENVI56\resource\filterfuncs\sensor_attributes.json

The file is updated when Digital Globe provides revised data

For example: Wordlview-3 data SWIR Band

- Offset is equal to 0 for all SWIR bands

- Gain values are:

            "S1": 1.030,

            "S2": 1.052,

            "S3": 0.992,

            "S4": 1.014,

            "S5": 1.012,

            "S6": 1.082,

            "S7": 1.056,

            "S8": 1.101

        

 

 

  1. Which version of absolute Gain and Offset values are hardcoded in ENVI 5.6.2

The data hardcoded in the sensor_attributes.json file for ENVI 5.6.2 are compliant with the slide 10 of MAXAR/DigitalGlobe reference file below (page 9 -10)

https://calval.cr.usgs.gov/apps/sites/default/files/jacie/MicheleKuester.pdf

They correspond to the Gain and Offset provided with calibration campaign 2018v0 for VNIR bands and 2019v0 for SWIR bands

 

 

  1. Why Gain values displayed in the ENVI metadata viewer differ from those hardcoded in the sensor_attributes.json file?

ENVI metadata viewer displays Gain and Offset values for Worldview data:

  • The Offset is the one hardcoded in the sensor_attributes.json file
  • However, Gain values as viewed in the ENVI Metadata viewer are different from the one hardcoded in the sensor_attributes.json file.

Gain values included in sensor_attributes.json file are absolute values. They need to be adjusted with a factor corresponding to the ratio ASF/EBW for each dataset

 

For example:

Gain values of WV-3 SWIR bands from sensor_attributes.json are listed below

   G= [1.03,1.052,0.992,1.014,1.012,1.082,1.056,1.101]

ASF values from a given WV-3 data for SWIR bands, extracted from the .xml file associated to this dataset, are listed below

   ASF=[2.671600e-04, 1.728110e-04,1.576480e-04,1.466560e-04,6.666700e-05,6.862700e-05,6.862700e-05,7.254900e-05]

EBW values from a given WV-3 data for SWIR bands, extracted from the .xml file associated to this dataset, are listed below

   EBW=[3.300000e-02,3.970000e-02,3.730000e-02,4.160000e-02,3.890000e-02,4.090000e-02,4.760000e-02,6.790000e-02]

The following computation (ASF/EBW)*G will result in the below output:

   G_ADJ=[0.00833863,0.00457927,0.00419268,0.00357474,0.00173437,0.00181551,0.00152248,0.00117638]

Those values are the Gain values displayed in ENVI metadata viewer for the corresponding SWIR bands of this dataset

 

  1. Why the equation provided by MAXAR/DigitalGlobe for radiance calibration differs from the one in ENVI help

 

ENVI helps shows the below equation to calibrate data in radiance

𝐿 = 𝐺𝐴𝐼𝑁 ∗ 𝐷𝑁 + 𝑂𝐹𝐹𝑆𝐸𝑇

MAXAR/DigitalGlobe (p.5 of the link below) provides a different equation

https://dg-cms-uploads-production.s3.amazonaws.com/uploads/document/file/207/Radiometric_Use_of_WorldView-3_v2.pdf?

𝐿 = 𝐺𝐴𝐼𝑁 ∗ (ASF/EBW) ∗ 𝐷𝑁 + 𝑂𝐹𝐹𝑆𝐸𝑇

 

As mentioned at point 4, Gain values displayed in ENVI metadata viewer are adjusted as follow

G_ADJ=(ASF/EBW)*G

The adjusted gain is then used by ENVI for the radiance calibration in the equation below

𝐿 =  G_ADJ ∗ 𝐷𝑁 + 𝑂𝐹𝐹𝑆𝐸𝑇 = 𝐺 ∗ (ASF/EBW) ∗ 𝐷𝑁 + 𝑂𝐹𝐹𝑆𝐸𝑇

 

So the 2 equations for radiance calibration provided by ENVI and MAXAR/DigitalGlobe are equivalent

 

 

  1. Why the equation provided by MAXAR/DigitalGlobe for reflectance calibration differs from the one in ENVI help

ENVI helps shows the below equation to calibrate data in radiance

𝜌𝜆 = 𝐿𝜆𝑑²𝜋 / 𝐸SUN sin 𝜃

With 𝜃 the sun elevation 

 

MAXAR/DigitalGlobe (p.7 of the link below) provides a different equation

https://dg-cms-uploads-production.s3.amazonaws.com/uploads/document/file/207/Radiometric_Use_of_WorldView-3_v2.pdf?

 

𝜌𝜆 = 𝐿𝜆𝑑²𝜋 / 𝐸𝜆 cos 𝜃𝑆

With 𝜃𝑆  the solar zenith angle

 

In ENVI 𝜃 angle is defined as the sun elevation while in for MAXAR/DigitalGlobe 𝜃𝑆 is the solar zenith angle.

Still from MAXAR/DigitalGlobe reference (p. 11), the solar zenith angle is equal to 90°- the sun elevation angle so

𝜃𝑆 =90- 𝜃= 𝜋 /2 - 𝜃

From trigonometry

sin(𝜃)=cos (𝜋 /2 - 𝜃)

So

cos(𝜃𝑆)=cos(𝜋 /2 - 𝜃)=sin(𝜃)

 

The solar irradiance has a similar definition in both cases:

Eλ =Esun

 

So the 2 equations for radiance calibration provided by ENVI and MAXAR/DigitalGlobe are equivalent

 

 

 

-------------------------------

created by BC on 7/2/2022

reviewed by MM on 7/18/22

 

 

Please login or register to post comments.