I am trying to white balance my acquired image, but instead get a green image, has anyone encountered such a problem?

[jpyung]

Similar to Vision's white balancing utility, my VI allows the user to draw an ROI around a region and ideally reset the red, green, blue gains to make the area white. When I make red, green, and blue values equal one another, instead of white, I get a green region. Has anyone else had this problem or know if I am doing something completely wrong?

Or has anyone tried doing this already and knows of a white balancing algorithm that works?

Thanks in advance.

- josh

[unclebump]

The QuickTime vi's at this site have whiteness control in some of them.

http://labview.epfl.ch/qt/QTVIs.html

[jpyung]

Thanks for the help, but from what I can tell, the vi's don't show how they control the white and if they are changing RGB gains to do it. It seems to be all done in their "code interface node."

Again though, thanks for trying to help.

[chrisger]

hi there,

i bet there's something wrong with the order of the color bytes or their datatypes. compare it with the "RGB to color.vi" from the "Graphics & Sound" -> "Picture Functions" palette. can you please post some code?

[jpyung]

Hi Chris,

I'm using IMAQ ColorImageToArray and I've checked with IMAQ IntegerToColorValue that I'm getting the correct RGB values. I can post my code, however, I am using a QImaging camera which only uses their own VIs to do functions similar to IMAQ functions such as grab setup, grab acquire, initialize, close. The VIs can be downloaded here: http://www.qimaging.com/support/downloads/.

Thanks again for the help.

[Tariah]

Sounds like you are dealing with data from a Bayer array. If that's the case, there are usually twice as many Green elements as there are Red and Blue. The luminance signal, Y, is = 0.25R + 0.5G + 0.25B.  White balance doesn't mean making all of the values equal to each other. It means making a "neutral" portion of the image appear neutral.  This may mean striking an imbalance based upon the color of the light used to form the image.

[DOC-doc]

Right,

I put once my finger in a white balance software and it IS really difficult : that is why the best way to calibrate colors, even on a good digital imager is to shoot on a white point to tune the color levels. There are grey word theory and others... but all have their own limitations.

Sheers

[jpyung]

Right now I can select an area of white to allow for the RGB values to be averaged over the area. Then, I was attempting to do grey world assumption by trying to make the R,G,B values equal to one another. I just assume I'm not understanding the concept and missing something obvious with my code and thinking.

Thanks for the input, Doc-Doc and Tariah.