How SNAP does classification for multidimensional data

Hi all,
I have some questions regarding how SNAP does classification on more than one image (multidimensional data). I did already classification in SNAP over multidimensional data (for example like 2 or 3 images) by RF and ML classifier but now I plan to do SVM classification in Matlab that’s why I need some helps:
For svm classification method we should follow these steps (I think SNAP does also this):

  1. select training data
  2. train a classifier model
  3. use trained classifier model for classifying a band(for example cohehernce)
    Now I can train a svm classifier model based on these trading bands:
    now based on choosed option we have a classifier model.
    We can apply the classifier model on “image1” or “image2” or “image3” separately.
    But I do not know how about combination of “image1+image2” or “image2+image3” or “image1+image3”. Would you please guide me?


If i understand your question correctly, you try to find out how to combine multiple dataset and feed them into SVM classifier ?

You just need to stack all of your images into a single product. I think in matlab there is a function called stack which does that. If you have 3 images where each image consists of 3 bands, the final product, after stacking will consist of 9 bands (or 9 features)