How can I match the color of one image to a better reference image?

Asked 12/21/2013

28 views

2 answers

0

I have two photos of the same subject. Image A has incorrect color, while image B has better color but may be lower resolution. I’d like to use B as a reference to correct the color of A rather than simply replacing A.

My idea is to compare the color response of the two images, derive a mapping between them, and then apply that mapping to A as a lookup table or similar correction. I’m especially interested in workflows that could be done in ImageMagick, GIMP, or similar tools.

What is a practical way to match A to B?

Originally by Photography Stack Exchange contributor. Source · Licensed CC BY-SA 4.0

Photography Stack Exchange contributor

12y ago

2 Answers

2

A good way is to work channel by channel. Put your images next to each other or in a split view. Then activate only the red channel. You can then adjust the curve for the red channel until the two grayscale images match up. Do the same for the green and blue channels.

This way you have created a 2D LUT.

Originally by user15615. Source · Licensed CC BY-SA 4.0

user15615

12y ago

0

AI Answer

Generated from our catalog & community — verify before relying on it.

A practical approach is to match the images channel by channel.

View the two images side by side, or in a split view, and isolate one RGB channel at a time. Start with the red channel: display only red so both images appear as grayscale for that channel, then adjust the tone curve of image A’s red channel until its brightness relationships match image B. Repeat the same process for the green and blue channels.

By doing this, you build a per-channel tonal mapping from A to B. In effect, that gives you a simple color lookup transformation you can apply to the full-resolution image A.

This works best when both images show the same subject under similar lighting and exposure. If the lighting differs a lot, a single global mapping may not fully fix the image. But for straightforward color correction, matching RGB curves channel-by-channel is a solid method.

UniqueBot

AI

12y ago

Your Answer