Random Forest Classification with Sentinel 2

In this case I think the best solution is, uninstall the SNAP, be sure to delete all the folders related to SNAP after uninstall it for instance, .SNAP .gdal, etcl.
Then redownload and install the latest version, concerning the RAM, I think 4, and 8 GB RAM are not enough, but 16 GB could work fine,

To increasing the memory, please take a look at the following post,

Source of the post


Thank you Falah for your recommendation. I re-installed the SNAP and got some addition classifiers.

The allocated Memory seemed automatically upgraded to 20025 MB (see the below image)



I’m trying to classify my image with Random Forest but nothing hapens and I dont know why.

There is no data in frequency. So, it’s not computing.
What am I doing wrong?
Somebody can help me please?

Uncheck the Evaluate classifier option ,

Training area should be identified under vector data, not under mask, and it should be reproject to WGS84 (unit=degrees), EPSG:4326 , similar project should be done to your S-2 image,

Take a look at this tutorial : CROP MAPPING WITH SENTINEL-2 JULY 2017, SPAIN

Source: https://rus-copernicus.eu/portal/wp-content/uploads/library/education/training/LAND01_CropMapping_Seville.pdf

1 Like

Thank you all. It worked now after to reproject the image. Thank you very much!!!
May I include index images in the RF classification? For example (B1, B2, …B12, NDVI, NVWI, etc)?

Thank you for the tutorial!

Random Forest can handle inputs of different types and ranges, so this is possible.

How can I acces SAR and Optical data fusion for land applications material?

Please take a look at the following post explained by our colleague @ABraun concerning data fusion

Source of the post