I would like to use SNAP to perform a supervised classification on Sentinel-1 data as I find it to be very user-friendly and intuitive, however it seems to be missing some functionality (or I just don’t know how to access the functionality!). While the Statistics tool has been great for evaluating my ROIs, I wish I could do more with the Scatter Plot tool. I would like to view multiple classes (ROIs) in the same feature space in order to see if they are separated or overlapping. Is this possible? I would also like to be able to assess the quality of the individual samples (polygons) within each class in case some of my training samples are low quality (i.e. with high standard deviations) so that I can discard them to improve the quality of my training data. Is there any way that I can view the standard deviations of, for example, all 10 samples within my water class at the same time?
Of course, it’s possible to implement supervised classification , in SNAP there are different classifiers, those could be used according to your data, in your case concerning S-1, you have two options , GRD calibrated data, or coherence, the goal of your study identify the use, I got from your demand , you intend to classify water’s surface, many different approaches possible to follow,
Scatter plots, or correlation plot in general give the relation of two parameters, in your case , I suggest the first option is to use statistics , to get the information of your AOI,
It’s possible in this case to create vector polygon , and extract the statistics of the same AOI, in your different images’ date,
There are many ways , bandmath and Statistics,
Also, I suggest to use search of this FORUM, there are a plenty of similar topics.
I think @1sarah means if it is possible to have multiple ROIs in the same scatter plot graph, for example with different colors.
SANP is limited in this case, because the statistics tool only allows to get information for one band only and the scatter plot only allows to refer to one geometry.
@1sarah: You mean if it is possible to create something like this, right?
Source: https://slideplayer.com/slide/7682866/
I don’t have a solution for this, I just wanted to clarify the question.
The opposite is possible: Mark areas in the scatter plot and have them highlighted in the map: Scatter plot regions to image space and viceversa
Thank you very much @ABraun, your interpretation of my question is spot on. It’s a pity that it’s not possible to view multiple ROIs in the same scatter plot, but at least now I can stop spending time trying to carry out a function that doesn’t exist in SNAP. Thank you for the link on highlighting image pixels using the scatter plot, that’s valuable to know.