I use S1 data ,In Random Forest, There is an error: “bound is must be positive”
How should this problem be solved? Looking forward to your reply!
what training data has been used?
what is the projection which was selected during Range Doppler Terrain Correction?
traning data is vector data
was the data imported before or after TC?
traning data imported after TC,My training data is a square plot.
can you please try to select WGS84 (lat/lon) during TC and then import the training data again? It was reported that the RF classifier struggles with UTM coordinates in some cases.
The situation you said is ok, but I want to do S1 and S2 collocation, S2 Projection is ,so,I want to set S1 the same as S2.
then it would be easiert to reproject S2 to WGS84 instead, so you can collocate both within this coordinate reference system.
But, I set S2 Projection is ，then,imported vector data,In Random Forest,The result is ok.
I set S1 Projection is ，then,imported vector data,In Random Forest,There is an error: “bound is must be positive”.
Why S2 is ok and S1 is not??
maybe the extent of the products is different?
Is the same research area，Okay, I will think of other ways.
Thanks for your reply！
Hello. I have a question about this. I have a mosaic of 3 image, ok. Then I try to do RF, but during the process snap give me the ‘extractor band’ a file very large and the processing of RF is charging but this doesn’t have advance…please. what am I doiung wrong
I 've made a reprojectation of my image and I’ve had all the cares during the preprocessing of data, but I don’t have results. Only I see Computer raster data… and 0% charging that’s all
Well, at first in snap I have posted a question about if I can allow the permission to Firewall Windows Defender for install snap.exe that’s what I have had in the moment of my installation. I denied that, but It’d posible if this are making the error now?
I will close this thread because this question is duplicated. Discussion should go on here: