To expand a bit on the excellent advice MrToM gave, I would break up the image into a luminosity component (ie, a layer whose blending mode is luminosity), and then, above that, a color component (ie, a layer above the previous one, whose blending mode is set to color).
This will allow you to work separately on getting the luminosity and the color components in the oddball areas of the map to blend with the luminosity and color components in nearby areas of the "normal" parts of the image. Once you are done with the blending tasks, because of the way you constructed the layers, the luminosity and color components should merge back together to give you a nice, well blended color image.
As MrToM said, you should think seriously about individually blending several relatively small areas instead of trying to blend all of the oddball areas with their surroundings at one time.
With respect to blending the luminosity component, I would look at a histogram of the unaffected areas and use them in a masked levels adjustment layer to set the black and white points of the "oddball" areas of the image. Then, in a separate curves adjustment layer, adjust the curve so the histogram of the adjusted oddball area matches that of the "normal", nearby areas of the map. Making only these three adjustments (ie, black point, white point, and a curve for the mid tones) should get you very, very close to a perfect match for the luminosity component.
An analogous procedure can be used to match the color components of the oddball areas to their surroundings.
HTH,
Tom Mann
(not to be confused with MrToM, LOL)
