Hello, I am trying random forest classification in snap but facing this problem
this error message does not tell much about the reason of the failure.
Can you please describe what you have done so far and how you conducted the classification? Input data, pre-processing, training data…
Hi , I wanna to do the random forest classification with sentienl-2 too.Did you finished your datas processings,and how many bands to classify with sentinel-2？
thanks in advance .
best option is to use all available bands because larger feature space makes the random forest classifier more effective.
fine，and I want to achieve glacier classification.That need to all available bands ethier?
when I did the resameple ,how many should the pixel resolution should be set if i used the all bands to classify which means that different pixel resolutions?like this.
if you want to classify fine structures, you can select 10 meters and bilinear resampling. But if you aim at wider landcover (e.g. forests), 20 or 60 m can be fine as well.
There is no right or wrong, you have to see and check how the resampling affects the image you want to classify.