Detecting Clouds with Machine Learning

New solutions based on existing technology


1. More than 55% of the Earth’s surface is always clouded.
2. Machine Learning is a powerful approach where human intelligence ends.
3. Satellites continuously observe the Earth’s surface with a variety of instruments.

Based on these three facts, we have used our machine learning know-how to repurpose existing satellite data to reliably identify clouds.

Michael Aspetsberger, Industry Focus Leader Aerospace at Cloudflight

Michael Aspetsberger
Industry Focus
Leader Aerospace

Get in touch

What is CI4Clouds?

The Cloud Problem

Satellites continuously provide us with optical data in several wavelengths of the Earth’s surface. The problem is to detect clouds. A misclassification has a strong impact on further algorithms of a processing chain, which assume clear sky conditions. 55% of the Earth surface is always clouded.


The project Computational Intelligence 4 Clouds – CI4Clouds – uses satellite pictures, like from the Korean GOCI satellite instrumentto identify clouds. GOCI, the Geo­stationary Ocean Color Imager, was built for the purpose of ocean surveillance around Korea, Japan, and Eastern China. More than 2 billion people live in this area. Cloudflight’s project partner was the Austrian meteorological office – ZAMG.


We set up a number of machine learning algorithms which compete against each other: Deep Learning (DL) and Convolutional Neuronal Nets (CNN), Random Forests (RF) as well as Support Vector Machines (SVM).

For training high performance hardware is used. The evaluation is against existing cloud masks from GOCI but also European and US satellite cloud products.

RF is the winning algorithm. RF also performs better than the known cloud products. Domain-specific pre- and postprocessing enhances results even further.

GOCI satellite instrument
Testbench: various algorithms compete against each other
Testbench: various Machine Learning algorithms compete against each other (DL = Deep Learning; SVM = Support Vector Machine; CNN = Convolutional Neural Network; RF = Random Forests)
Winner: Random Forests

This project was carried out in a partnership of ZAMG and Cloudflight. CI4Clouds was funded by the Austrian Federal Ministry of Transport, Innovation and Technology (BMVIT, meanwhile BMK) under the program “ICT of the Future” between 2015 and 2016.

  • Have you ever thought that the sky is not the limit?
  • Are you wondering if your problem can be solved with Machine Learning?
  • Do you need high-performance computing expertise to solve your challenges?
Bernhard Niedermayer, Head of Emerging Technologies at Cloudflight

Bernhard Niedermayer

Head of Emerging Technologies

Flip van Eijndhoven

Business Consultant Aerospace
Flip has joined the Cloudflight community in 2020 to help the aerospace industry grow from Amsterdam. He has more than a decade of experience in advising companies on their innovations and technical solutions.

Get in touch

Sign up to our monthly wrap-up and get access to exclusive material and updates about our technologies, research content and events. You can always adjust your settings.