Match Color on Multiple Data Sources in a Mosaic Dataset

[NathanHeickLACSD]

I am trying to do a proof of concept to create a mosaic dataset that combines a regional aerial imagery dataset with various drone and aerial imagery datasets of facilities and pipeline alignments.

I seem to be able to use the Color Balance Mosaic Dataset tool to match the localized high-resolution datasets to the low-resolution datasets with full coverage, however, the tool creates seamlines where there were none before in the regional aerial imagery.  That imagery was already color-balanced when we received it.  In fact, all of the the different data sources individually may have multiple tiles, but they are color balanced with themselves.

My goal is to match the additional high-resolution imagery to the color of the regional imagery, but I can't find a way to color balance one set of images against another, just the entire mosaic dataset together.

Any ideas image experts?  I appreciate the help.

[CarolineTyra]

Hi Nathan-- Do all of your datasets have the same pixel-depth and the same bands, and have you calculated statistics on the mosaic dataset?

[NathanHeickLACSD]

Two of the sources have 4 bands and one has 3 bands.  I defined the mosaic dataset to have three bands.  All of the images have 8-bit pixel depth.

[PeterBecker]

Try using the GPTool and using 'block Field'. Create an attribute (any field name you want) and give a different value for the two datasets. Then run using this as a block fields. This should ensure same enhancements is defined for all images that are part of the same datasets.

[NathanHeickLACSD]

You know I tried that before and it wasn't working.  Then today, I tried it again with the block fields I had created before and I would swear it was working very well.  Then, I built overviews and I see the seams again.  Either I wasn't paying attention very well earlier today or the overviews somehow messed it up.

[PeterBecker]

Before you create the overviews check the properties of the mosaic dataset (not the layer) to ensure that in 'Defaults' that the 'Apply Color Correction' is turned on. The overview will be generated using this setting. It is possible that it was Off and so not included in the newly generated overviews. If you had previously build overviews, then when building them again, ensure 'Generate missing overview images only' is turned off, else they will not be updated with the new settings. 

[NathanHeickLACSD]

Well, I don't think it is the overviews because I can delete the overviews and I still see the issue when I zoom in.  I went back to just a mosaic dataset with four images in one area and two more in two other locations.  This is our regional imagery which doesn't have any noticeable color shifts between tiles.  When I color balance just the regional imagery, I see the tile boundaries in the middle of the 2 x 2 set of four tiles, but the color difference fades away as you move towards the sides of the square.  I used blocking with a single block.  I wonder if the size has anything to do with it or the big gap between the different sample areas.  I think I didn't notice the issue before with the color difference because our drone imagery sits right on top of the intersection of the four regional imagery tiles.