>  Docs Center  >  Using ENVI  >  Calibrating Images Tutorial
ENVI

Calibrating Images Tutorial

Calibrating Images Tutorial

In this tutorial, you will calibrate a QuickBird Level-1 image to spectral radiance and reflectance while learning about the various metadata fields that ENVI uses to perform calibration.

This tutorial uses ENVI 5.3. You can use versions 5.1 or later, but the steps may vary slightly.

Files Used in This Tutorial


The tutorial data files are available in a single ZIP file from our website. Extract this file to a local directory. Go to the folder named rigorous_ortho\005606990010_01_P008_MUL.

Files

Description

005606990010_01_P008_MUL\05JUL*.TIL
QuickBird Level-1 multispectral imagery for Phoenix, AZ from 11 July 2005

QuickBird files are courtesy of DigitalGlobe.

Background


Calibrating imagery is a common pre-processing step for remote sensing analysts who need to extract data and create scientific products from images. Calibration attempts to compensate for radiometric errors from sensor defects, variations in scan angle, and system noise to produce an image that represents true spectral radiance at the sensor.

ENVI's Radiometric Calibration tool provides options to calibrate imagery to radiance, reflectance, or brightness temperatures. See Radiometric Calibration for more information on how each option is computed.

The available calibration options depend on what metadata is included with the imagery. Most vendors distribute a metadata file or ephemeris data along with the image data.

Note: It is important to select the correct metadata file (using the File > Open menu option) when opening data from various satellite sensors so that ENVI reads the required calibration parameters. Refer to the following table:

Sensor

Calibration Options

Metadata File to Open

Radiance

Reflectance

Brightness
Temperature

ALOS AVNIR-2 and PRISM Level-1B2 data

 

HDR*.txt

AlSat-2A

 

 

.dim

DMC DIMAP

 

.dim
EO-1 ALI

 

Use the File > Open As > Optical Sensors > EO-1 > HDF4 menu option and select a *_HDF.L1G file. A *_MTL.L1G file must be in the same directory.

EO-1 Hyperion Level-1R

 

Use the File > Open As > Optical Sensors > EO-1 > HDF menu option and select an .L1R file. Calibration metadata is hard-coded into the application and not read from any metadata files.
Gaofen-1

 

Use the File > Open As > Optical Sensors > CRESDA > GF-1 menu option and select an .xml file.

GeoEye-1

 

.til

IKONOS

 

metadata.txt

KOMPSAT-3

 

 

*_aux.xml

Landsat TM, ETM+, and Landsat-8 OLI/TIRS data

*_MTL.txt, *WO.txt, *.met

OrbView-3

 

.tif

Pleiades-HR 1A, 1B Primary and Ortho (single or mosaic)

 

 

DIM*.xml

Pleiades-HR DIMAP V2 (JPEG2000, TIFF)

 

DIM*.xml

QuickBird

 

.til

RapidEye Level-1B and -3A (TIFF, NITF)

 

*_metadata.xml

A NITF/NSIF license is required to open NITF files.

ResourceSat-2

 

 

.h5

Sentinel-2

 

*.xml

Sentinel-3

 

 

*.xml

SPOT DIMAP

 

DIM*.xml

SSOT DIMAP

 

 

METADATA.DIM

TripleSat/DMC3

 

*.xml, *.dim

UrtheCast Theia

 

 

*.xml

WorldView

 

.til
Ziyuan-1-02C

 

Use the File > Open As > Optical Sensors > CRESDA > ZY-1-02C menu option. Select from the following files: *.orientation.xml opens the MUX-PAN data product with metadata, *-MUX.xml opens multispectral data with metadata, *-PAN.xml opens panchromatic data with metadata, and *.xml opens the HRC data product (images with two parallel cameras) with metadata.
Ziyuan-3A

 

Use the File > Open As > Optical Sensors > CRESDA > ZY-3 menu option. Select from the following files: *.orientation.xml opens the TLC data product (images with nadir, forward, backward view) with metadata, *.xml opens multispectral data with metadata, *-NAD.xml opens TLC nadir-view data with metadata.

To open QuickBird or WorldView data, select the image file. ENVI will read the necessary metadata from the accompanying *.IMD file.

Open a QuickBird Image and View Its Metadata


  1. Start ENVI.
  2. From the menu bar, select File > Open. A file selection dialog appears.
  3. Navigate to the folder where you saved the tutorial data and select the file 05JUL11182931-M1BS-005606990010_01_P008.til. Click Open.
  4. In the Layer Manager, right-click on the filename and select View Metadata.
  5. Click the Spectral category on the left side of the Metadata Viewer. This shows several metadata fields related to calibration. ENVI needs gain and offset values in units of W/(m2 * µm * sr) to calibrate imagery to radiance. You can see these values under the Gains and Offsets columns. The gains and offsets are already in the correct units in this image. If they are not in the correct units, you can use the Scale Factor field (discussed in the calibration steps below) to scale the calibrated image to the correct units.
  6. Select the Image Parameters category. You can see the values for Sun Azimuth and Sun Elevation that were derived from the QuickBird metadata.
  7. Select the Time category. The Acquisition Time for this scene is listed in Coordinated Universal Time (UTC). The Sun Azimuth, Sun Elevation, and Acquisition Time are used in combination with the various fields under the Spectral category when calibrating the image to reflectance.
  8. Close the Metadata Viewer.

Calibrate the Image to Radiance


  1. From the Toolbox, select Radiometric Correction > Radiometric Calibration. The File Selection dialog appears, with the QuickBird file already selected.
  2. Click OK. The Radiometric Calibration dialog appears.
  3. Leave the default options as-is. You will create a floating-point radiance image (BSQ interleave) in units of W/(m2 * µm * sr). Keeping the Scale Factor at 1.00 ensures the units will remain the same as the original gain and offset values.
  4. The Appy FLAASH Settings button is for users who will subsequently perform atmospheric correction using the FLAASH® tool. You can skip this step for the tutorial, but here is some background information if you plan to use FLAASH in the future:
  5. The use of FLAASH requires a separate Atmospheric Correction Module: QUAC® and FLAASH® license. FLAASH requires input imagery to meet the following criteria:

    • The image must be calibrated to radiance in units of µW/(cm2 * nm* sr).
    • The input image can be floating-point, long integer (4-byte signed), or integer (2-byte signed or unsigned).
    • The image can be in band-interleaved-by-line (BIL) or band-interleaved-by-pixel (BIP) format.

    Clicking the FLAASH Settings button will create a radiance image in BIL, floating-point format. It will apply a scale factor of 0.1 to the radiance image to get it in units of µW/(cm2 * nm* sr). Clicking this button prevents you from having to separately convert the interleave of the radiance image and figuring out the appropriate scale factor for use with FLAASH.

    When you start FLAASH, select the radiance image that you just created with the Radiometric Calibration tool. When the Radiance Scale Factors dialog appears, leave the default value of 1 for the Single Scale Factor field.

  6. Click the Browse button next to Output Filename, and save the radiance image as qb_radiance.dat in a directory of your choice.
  7. Ensure that the Display Result check box is selected.
  8. Click OK. When processing is complete, the calibrated radiance image is displayed.
  9. To visually compare the original and calibrated images, toggle the qb_radiance.dat layer off and on in the Layer Manager.
  10. With both layers selected in the Layer Manager, click the Cursor Value icon in the main toolbar.
  11. The On demand updates button in the Cursor Value dialog is enabled by default. Click it to turn off the red probe.
  12. Move the cursor around the radiance image, and look for the "Data" values that are reported for the radiance (floating-point) and original (integer) image. The following figure shows an example where the calibrated image is displayed in true color:
    • Band 3 is assigned to the red channel
    • Band 2 is assigned to the green channel
    • Band 1 is assigned to the blue channel
  13. For the current pixel location in this screen capture, the calibrated image has a radiance value of 97.46099 W/(m2 * µm * sr) in the red band, while the original image has a raw DN value of 546 in the red band.

  14. Close the Cursor Value dialog.

Another way to quickly verify the radiance values is to display a spectral profile:

  1. Uncheck the original QuickBird image (05JUL*) in the Layer Manager so that only the calibrated image is displayed.
  2. Click the Spectral Profile button in the main toolbar.
  3. Click anywhere inside the image to display a plot of radiance values for the selected pixel location. You can use a spectral profile to help identify features of interest:

    The following example shows a pixel that represents soil. The radiance values peak in the red wavelength region (~ 660 nm).

    The next example shows a pixel that represents water. The radiance values peak in the blue wavelength region (~ 485 nm).

    The next example shows a pixel that represents vegetation. The radiance values peak in the near-infrared wavelength region (~ 900 nm).

  1. When you are finished, close the Spectral Profile dialog.
  2. Right-click on each layer name in the Layer Manager, and select Remove.

Calibrate the Image to Reflectance


Next, you will calibrate the QuickBird image to top-of-atmosphere reflectance. This image has all the metadata needed to calibrate to reflectance:

  • Gains
  • Offsets
  • Solar irradiance
  • Solar elevation
  • Acquisition time

Follow these steps:

  1. From the Toolbox, select Radiometric Correction > Radiometric Calibration.
  2. In the File Selection dialog, select the original QuickBird image (05JUL*), then click OK.
  3. In the Radiometric Calibration dialog, change the Calibration Type to Reflectance.
  4. Leave the other options at their default values.
  5. Click the Browse button next to Output Filename, and save the reflectance image as qb_reflectance.dat in a directory of your choice.
  6. Ensure that the Display Result check box is selected.

  7. Click OK. When processing is complete, the reflectance image is displayed.
  8. Click the Cursor Value icon in the main toolbar.
  9. Look at the "Data" values for each band in the Cursor Value dialog and verify that the values are less than 1.0. Reflectance values theoretically range from 0 to 1, but since no surface is a true reflector of solar radiation, the brightest features typically have a maximum value of 0.7 to 0.8.
  10. Close the Cursor Value dialog.
  11. Click the Spectral Profile button in the main toolbar.

  12. Click anywhere inside the image to display a plot of reflectance values for the current pixel location. The following figure shows an example of a pixel that represents water. Reflectance values range from 0.04 to 0.09 across all four bands, with the lowest value (0.04) in the near-infrared wavelength region:
  13. When you are finished, exit ENVI.

For more information on the topics covered in this tutorial, see , Radiometric Calibration or click the help button in the Radiometric Calibration tool.

QUAC and FLAASH are registered trademarks of Spectral Sciences, Inc.



© 2019 Harris Geospatial Solutions, Inc. |  Legal
My Account    |    Store    |    Contact Us