X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 04 Jun 2018 09:05 PM by  Tharindu Abeysinghe
Atmospheric correction with Worldview 3 image
 2 Replies
Sort:
You are not authorized to post a reply.
Author Messages

Tharindu Abeysinghe



New Member


Posts:6
New Member


--
03 Jun 2018 10:13 AM
    I have a worldview 3 image (Level 1B product with 17 bands, 1 Pan, 8 MS, and 8 SWIR). I need to apply an atmospheric correction to the image. After I apply radiometric calibration with ENVI 5.4, SWIR bands result in no data pixels in some areas covered with water. This happens with both FLAASH settings and default settings in radiometric calibration. As I understand, only option I have with ENVI is I can apply QUAC for the original image. Then apply (B1 le 0)*0+(B1 ge 10000)*1+(B1 gt 0 and B1 lt 10000)*float (b1)/10000 equation to scale the data to 0 to 1 range.
    Can you help me to find a better method to use radiometric calibration and FLAASH or QUAC to correct my image to BOA level?

    Thank you in advance..!

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    04 Jun 2018 10:25 AM
    NoData should only be assigned to pixels that are outside of the image boundary. In areas of water, the returned radiance values can be very low due to low reflectance over water in SWIR bands from any sensor. If the pixels actually show NoData, you can remove that setting by going to View Metadata->Edit Metadata and remove the Data Ignore Value. This should reveal the actual pixel value of the pixels in the water when in units of radiance (with or without FLAASH Settings).

    It is correct that if you choose to use the image DN and not radiance values, then you will not be able to use FLAASH which expects data to be in radiance units. When using both FLAASH or Quac, the output reflectance is scaled by 10000. To convert the image to percent reflectance that ranges between 0-1.0 (approximately), you can use the equation:
    float(b1)\10000
    then use Map variable to input file and select the scaled reflectance image as input. This way each band will be divided by 10000 to return percent reflectance.

    Tharindu Abeysinghe



    New Member


    Posts:6
    New Member


    --
    04 Jun 2018 09:05 PM
    Thank you Mari. This is so helpful..!
    You are not authorized to post a reply.