Copy georeference information between two sets of images

[BrianHovis]

I have two sets of images from historical maps for the same geographical locations.  The black and white images were georeferenced several years ago and for several different years.  There are multiple images for each year. I created a mosaic dataset and shared to ArcGIS Online using a tile package for each year.  From there I made a story map and it was embedded into the Olympia Washington historical society's website.

The color versions are now available and I now want to go through the same process and georeference them.  What is the easiest way to copy the georeference information from the black and white images and paste them to the color versions of the same images? 

• To see an example of one year, here is the link to the 1891 Sanborn web map in ArcGIS Online: https://arcg.is/4byyL
• The website for the historical society and the web page about the map overlay project is here: https://olympiahistory.org/the-sanborn-overlays/ 

#If there is a better place to post this question, please let me know and thanks.

[PeterBecker]

A) If the color maps where the same scans as the original, but just a color version then if you originally georeferenced the bw using the Save (vs Save as) option then there will be an .Aux.xml for each image. This will contain the georeferencing information (control points) and transforms selected. You should be able to just copy the files so they have the appropriate extension for the color images. If you get an error about stats or strange colors, then open the xml file in a text editor and remove the stats section.

B) When you did the original georeferencing did you record the control point files? If so you can use these again for each of the new images. If the new images have a different scale or offset from the originals, consider just editing the values (in excel). This might not be accurate enough but would be an alternative to the TFW defined in C. 

C) If the images are new scans but similar resolution and clipping then instead of using the Aux.xml copy the .tfw or .tfwx (but rename to .tfw) to the new images. If you open then up in Pro the images should fall in approx the same location (do ensure you use the same projection). Now when you use the georeferencing tool with the bw images as the background then the process should be a lot faster as you can just pick points between the images. There is also a auto-georeference feature. Before using it make sure the base map is off and the only layer other layer on is the bw scans.  The system might be able to match between the images and auto-georefernence. This tool is primarily for optical imagery so it might not work with these scanned paper maps, but it is worth the try.

D) If none of the above work, then you will probably need to resort to manually georeferencing, but use the BW images as a base to measure control points of which should be a lot faster.

Once georeferenced you should be able to create a new mosaic dataset and "Import footprints" from the original mosaic dataset, if you used these to clip the images. 

[BrianHovis]

First, thanks Peter for giving me some ideas to work on.   I would like to work through your suggestions.  Option A would be great if I could get it to work. I would like to share with you my trial and make sure I am understanding correctly your suggestion. 

I tested option A method on sheet #2 from the 1891 Sanborn sheets.  I downloaded a small sized .jpeg of that sheet in color and called it "1891Sheet2ColorSmall."  I then found the black and white (B/W) image for that sheet that was correctly georeferenced and renamed it "1891Sheet2BlackWhiteLarge." Using Windows Explorer, I could see the image files with associated files (screen save below.)   There was a world file (.ovr) and no .aux file.

I copied the associated files and changed the file names to be the same as my image file name.  I then went back to ArcGisPro and added the 1891Sheet2ColorSmall file to my map.  Zooming to the file took me to the coast of Africa where lost images go to wait for instruction.  Did I misunderstand the directions?

I thought it might be helpful if you had access to the correctly georeferenced black and white sheet.  I shared it to a map on ArcGIS online at https://arcg.is/1neTuL .  I think I made it so you can download the file if it would be helpful.  I also bookmarked the correct location of the sheet. 

I was able to load the color version too, but AGOL would not let me add an image that was not georeferenced to my map.  It is probably not important, but if you were willing and interested, the color image of the sheet is at: https://www.loc.gov/resource/g4284om.g4284om_g092701891/?sp=2&r=-0.891,-0.039,2.782,1.318,0

 I downloaded the smallest jpeg for the test.  

Anyway, I hope you have some interest in working out the details with me. I know it will help me, and who knows there may be someone else it will help as well.

[BrianHovis]

I found the .aux file for this sheet.  I can't see anything obvious that points to a particular file, so it makes sense that I should be able to associate this geographical information for a black and white copy to the color version both with the same extant.  It does not seem to work with just a change in the file name however.  Is there another way to associate the geographical information with color sheet?

<?xml version="1.0"?>

-<PAMDataset>

<SRS>PROJCS["WGS_1984_Web_Mercator_Auxiliary_Sphere",GEOGCS["GCS_WGS_1984",DATUM["WGS_1984",SPHEROID["WGS_1984",6378137.0,298.257223563]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Mercator_Auxiliary_Sphere"],PARAMETER["False_Easting",0.0],PARAMETER["False_Northing",0.0],PARAMETER["Central_Meridian",0.0],PARAMETER["Standard_Parallel_1",0.0],PARAMETER["Auxiliary_Sphere_Type",0.0],UNIT["Meter",1.0],AUTHORITY["EPSG","3857"]]</SRS>

<Metadata/>

-<Metadata domain="Esri">

<MDI key="PyramidResamplingType">AVERAGE_BIT2GRAYSCALE_MINISWHITE</MDI>

</Metadata>

-<Metadata domain="xml:ESRI" format="xml">

-<GeodataXform xmlns:typens="http://www.esri.com/schemas/ArcGIS/10.1" xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:type="typens:PolynomialXform">

<PolynomialOrder>1</PolynomialOrder>

-<SpatialReference xsi:type="typens:ProjectedCoordinateSystem">

<WKT>PROJCS["WGS_1984_Web_Mercator_Auxiliary_Sphere",GEOGCS["GCS_WGS_1984",DATUM["D_WGS_1984",SPHEROID["WGS_1984",6378137.0,298.257223563]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Mercator_Auxiliary_Sphere"],PARAMETER["False_Easting",0.0],PARAMETER["False_Northing",0.0],PARAMETER["Central_Meridian",0.0],PARAMETER["Standard_Parallel_1",0.0],PARAMETER["Auxiliary_Sphere_Type",0.0],UNIT["Meter",1.0],AUTHORITY["EPSG",3857]]</WKT>

<XOrigin>-22041257.77387803</XOrigin>

<YOrigin>-30241100</YOrigin>

<XYScale>144148035.89861274</XYScale>

<ZOrigin>-100000</ZOrigin>

<ZScale>10000</ZScale>

<MOrigin>-100000</MOrigin>

<MScale>10000</MScale>

<XYTolerance>0.001</XYTolerance>

<ZTolerance>0.001</ZTolerance>

<MTolerance>0.001</MTolerance>

<HighPrecision>true</HighPrecision>

<WKID>102100</WKID>

<LatestWKID>3857</LatestWKID>

</SpatialReference>

-<SourceGCPs xsi:type="typens:ArrayOfDouble">

<Double>7.578015546410608</Double>

<Double>5.9839597622313674</Double>

<Double>1.0146957449615002</Double>

<Double>1.3765682773664594</Double>

<Double>1.0590866280399496</Double>

<Double>5.9965024506227564</Double>

<Double>7.4742407927842578</Double>

<Double>1.3102886595988821</Double>

</SourceGCPs>

-<TargetGCPs xsi:type="typens:ArrayOfDouble">

<Double>-13681350.270692201</Double>

<Double>5949405.9771591695</Double>

<Double>-13681031.958792692</Double>

<Double>5949022.0376892667</Double>

<Double>-13681307.907441538</Double>

<Double>5948991.0001824703</Double>

<Double>-13681076.301056527</Double>

<Double>5949435.3825104004</Double>

</TargetGCPs>

<Name/>

</GeodataXform>

</Metadata>

-<PAMRasterBand band="1">

-<Histograms>

-<HistItem>

<HistMin>-0.5</HistMin>

<HistMax>1.5</HistMax>

<BucketCount>2</BucketCount>

<IncludeOutOfRange>1</IncludeOutOfRange>

<Approximate>0</Approximate>

<HistCounts>4432287|104033349</HistCounts>

</HistItem>

</Histograms>

-<Metadata>

<MDI key="STATISTICS_MINIMUM">0</MDI>

<MDI key="STATISTICS_MAXIMUM">1</MDI>

<MDI key="STATISTICS_MEAN">0.95913648632457</MDI>

<MDI key="STATISTICS_STDDEV">0.19797395608229</MDI>

<MDI key="STATISTICS_SKIPFACTORX">1</MDI>

<MDI key="STATISTICS_SKIPFACTORY">1</MDI>

<MDI key="STATISTICS_EXCLUDEDVALUES"/>

<MDI key="STATISTICS_COVARIANCES">3.91936872868729E-02</MDI>

</Metadata>

</PAMRasterBand>

</PAMDataset>

[PeterBecker]

Brian

I had a quick look at the files. I downloaded the color image (TIF) and georeferenced it based on the layer that you provided earlier. Looking at the extent  it is apparent that these are two separate scans. (See white border from BW around the extent of the color). As a result there will not be a constant offset or scale between the two. The image coordinates in the color image will be approx a scale and offset different to the bw but will change from scan to scan

(Note the AUX.XML file you have above appears to be for a sheet below this one)

I believe you are going to have to manually georeference them. Simplest way is to load the BW image, zoom to its extent. Add the Color image. Go to georeferencing, Fit To Display, and then measure 4 manual control point close to teh corners of each sheet. Then Save and this will will write the new Aux.xml. 

I'm not sure what color mages you are downloading, but for more compact storage I would recommend you use OptimizeRasters and convert to TIF with JPEG compression first.

[BrianHovis · Accepted solution]

Thanks Peter very much.  I appreciate the coaching.