The results of an image classification may vary A LOT depending on the variables you use from the beginning. Number of classes, minimum class size, sample interval, atmospheric corrections, conversions, decisions in category assignation...
I'm interested to know how do you decide these variables. For instance, it is said than the minimum class size should be approximately ten times the number of bands.
But what about the number of classes?
Atmospheric corrections, do you really need them if you just run an unsupervised classification? I mean, the differences would be qualitatively the same..
Conversions, if you standardize the values of all bands to have the same ranges is it supposed to work better?
Etc.
In my experience there are not any "fixed" rules and it is more a test-error process. But I want to learn with your insights, and maybe we can all learn something.
It would be also interesting to share any "tricks" or tools you think people may overlook or not know at all and explain why do you think are useful. As an example, I discovered a while back the ArcGIS Dendrogram tool. The result looks like this:
It tells you how the categories in your signature file relate to each other, you can easily see which ones are more similar. It can be really helpful some times when you have doubts in a certain category.
أكثر...
I'm interested to know how do you decide these variables. For instance, it is said than the minimum class size should be approximately ten times the number of bands.
But what about the number of classes?
Atmospheric corrections, do you really need them if you just run an unsupervised classification? I mean, the differences would be qualitatively the same..
Conversions, if you standardize the values of all bands to have the same ranges is it supposed to work better?
Etc.
In my experience there are not any "fixed" rules and it is more a test-error process. But I want to learn with your insights, and maybe we can all learn something.
It would be also interesting to share any "tricks" or tools you think people may overlook or not know at all and explain why do you think are useful. As an example, I discovered a while back the ArcGIS Dendrogram tool. The result looks like this:
It tells you how the categories in your signature file relate to each other, you can easily see which ones are more similar. It can be really helpful some times when you have doubts in a certain category.
أكثر...