I have been using the unsupervised classification (ISODATA clustering) on some vegetation index layers recently. I am using eCognition Server to run the ruleset on subsets of a large area, so the VI layers on all subsets were processed in the same way (externally to eCognition). However, the algorithm fails on a third of my subsets. Manually amending the algorithm parameters does not seem to change a thing and I always end up with a single cluster at the end. I know for a fact that there is a variety in my data and I should definitely have more than one cluster. Any idea what's causing it to fail for some subsets while others are fine?
Was this article helpful?