X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 22 Nov 2015 08:13 AM by  anon
Correcting color / white balance in R-G-NIR images
 2 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
22 Nov 2015 08:13 AM
    <p>Hello,</p> <p>I have two sets of images of vegetation plots taken on an NDVI camera (standard digital camera modified to record red, green and NIR)</p> <p>One set is correct and produces good NDVI values. This set was taken after the camera's white balance was set to a white card which is visible in the image. The other set does not have white balance set (so it was on auto), although white cards are still visible in the images. This second set is currently not usable and produces unrealistically low NDVI values.</p> <p>I am trying to find ways to correct the second set of images. I can do this by matching the histograms, but this only works when matching a 'broken' image to a correct image of the same vegetation plot.&nbsp;</p> <p>My ultimate aim is to be able to correct images of other plots for which there are no correct photos and no white cards in the images, so histogram matching won't work for this.</p> <p>Surely there must be a way to use the white cards to come up with an average change in the channel gains when white balance is corrected. However when I, for example, take the percentage difference between band 1 on the white card in the broken image and in the the correct image for a given plot, and apply this to the vegetation using band math, this still doesn't produce realistic values.&nbsp;</p> <p>Does anyone have any advice or knowledge of any tools for correcting color balance other than histogram matching?&nbsp;</p> <p>Many thanks in advance.</p> <p><br /> </p>

    Deleted User



    New Member


    Posts:
    New Member


    --
    24 Nov 2015 11:39 AM
    <p>Hi Rachael&nbsp;</p> <p>By white card, do you mean white calibration target that you had placed on the field during your photos?&nbsp;</p> <p>I know in my imagery, I am still having the same issue. The solar irradiance changed through the flight- so the targets used were only relevant for half of the flight lines until the clouds rolled in. I did histogram matching- but you have to make sure the same features are contained between the frames you are matching (i.e. pure vegetation). However, my results were not perfect and I would also like to know another way to do this.....</p> <p>Can anyone else offer some help?&nbsp;</p>

    Deleted User



    New Member


    Posts:
    New Member


    --
    25 Nov 2015 06:08 AM
    <p>Hi Alex,</p> <p>Exactly, the cards were used as calibration targets. Yes that sounds like a similar problem (bit of a relief that it's not just me!)</p> <p>I did find this post about gamma correcting imagery (the answer at the bottom). I haven't figured out how to do it or whether it will help with my problem yet but in case its useful to you:&nbsp;http://gis.stackexchange....ngle-image</p>
    You are not authorized to post a reply.