Skip to main content

Hi,

I want to detect artificial coverages on agricultural land. Is it possible to do an supervised classification on sentinelhub? You have some help for automatic multiple Scene classification?

The L2A scene classification for example mix some white artificial coverages with clouds. Where can I find the SCL-script?

Where can I find s2cloudless-results from CLP (cloud probabilities) and CLM (cloud masks) on sentinel hub? Are they different to L2A scene classification or they have the same decision tree?

Thanks

Hi,

it is not possible to run supervised classification directly in Sentinel Hub.
You might however find this blog post series relevant, where we demonstrate, start-to-end, on how to do supervised classification using eo-learn and Sentinel Hub, with a Jupyter Notebook sample with all the steps:

360db83263e79cba7951e7ba15aa4495131aa0f1.pngMedium – 30 Dec 20

8bd16f71a837c2d5e59579e399daa3628607f1a4.png

Land Cover Classification with eo-learn: Part 1

Mastering Satellite Image Data in an Open-Source Python Environment

Reading time: 8 min read

 

You should be able to modify this for your use-case pretty easily.

An example of the script for SCL is here:
https://shforum.sinergise.com/t/l2a-scene-classification-for-sentinel-2/51

You can find CLM and CLP cloud masks directly along the Sentinel-2 bands, see:

docs.sentinel-hub.com

d3714e73b38a87afa3c31502a6696052a7395163.png

Sentinel Hub

Sentinel Hub is multi-spectral, temporal satellite imagery service for real-time processing of big remote sensing data.

 

More info here:

docs.sentinel-hub.com

d3714e73b38a87afa3c31502a6696052a7395163.png

Sentinel Hub

Sentinel Hub is multi-spectral, temporal satellite imagery service for real-time processing of big remote sensing data.

 

The masks are not the same as L2A scene classification, as those are done using Sen2Cor algorithm and CLM/CLP use s2cloudless algorithm.


Reply