I'm working for the first time with ArcMap's relatively new image analysis window, and am using the "image difference" function to calculate change between two images which I have performed a supervised classification on.

Each of the two images have only two types of land cover, forest and pasture; however, when I conduct the "image difference" analysis, I receive a raster with a ramp of 1 to -1. My question is how to interpret it? I.e. I'm looking to be able to symbolize where forest remains forest, forest changes to pasture, pasture remains pasture, and pasture changes to forest. I know this should be easy, and is so in ERDAS, but can't figure it out in ArcMap. Thanks in advance for any suggestions, David

  • 1
    What values have you assigned to forest and pasture? If they should be 1 and 2, then I suggest you reclassify the image to convert the 4 and 7 to 1 and 2. – Fezter Nov 18 '13 at 12:18
  • 2
    Also, I don't think the image difference tool in the Image Analysis window is as sophisticated as an image analysis program like ERDAS or ENVI. – Fezter Nov 18 '13 at 12:20
  • @Fezter, good suggestions; yeah, the Image Diff. tool isn't as sophisticated as ERDAS, but unfortunately it's all I've got currently ;) – David Meek Nov 18 '13 at 15:04

Since you are comparing two classified images with a small range of values, instead of using the image analysis window functions (especially given the cryptic result), why not just use spatial analyst functions?

There are a number of ways to do this, but I suggest using Combine: This will output a new grid with a single value for every unique combination of input values, e.g. forest-forest, pasture-forest, forest-pasture, pasture-pasture, etc. The output attribute table will show all original and new values, allowing you to assign new names to the mapping (e.g. deforestation, regrowth, no change, etc.).

  • this was EXACTLY what I wanted to do! Many thanks for the excellent suggestion; just goes to show that the fancy tool is not always the best for a situation. – David Meek Nov 19 '13 at 0:56

Essentially what you are doing is called change detection. When you select the "image difference" function, ensure that you are comparing the same respective bands within each image. I.e. Band 4 vs Band 4, Band 3 vs Band 3 and so on.This will create a raster which you can interpret as follows. Positive numbers represent positive change, (i.e. Areas getting brighter) while negative numbers equal negative changes (i.e. Areas getting darker). I would suggest using band 3 for your comparison. Since vegetation absorbs nearly all red light (it is sometimes called the chlorophyll absorption band) this band can be useful for distinguishing between vegetation and soil and in monitoring vegetation health.

  • thanks @ dpalm, the original raster I classified was 3-bands, and each raster had the same bands (3, 2, 1). So would you suggest re-running the supervised classification on only band 3 and then doing the image difference based on that, or is it possible to specify the image difference to run on only band 3? Thanks! – David Meek Nov 18 '13 at 19:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.