Machine Learning for Remote Sensing

A Guide to Earth Observation with Satellite Data, AI and Python

Author

Christoph Molnar

Summary

Silently, thousands of satellites fly through orbit, observing Earth. Satellite imagery provides a non-intrusive way to observe earth, giving insight into land use, population size estimation, and wildfire prediction. But without the technology to analyze the data and make predictions at scale, it’s just a barrage of beautiful pictures. Fortunately, we have machine learning, especially deep learning, to make sense of the data.

At first glance, satellite images are just images. But satellite images are a very special type of image: They can have much more layers than just red, green, and blue channels. All images are related to each other through space and time. Challenges include clouds, evaluating with correlated data, imbalanced labels, and more.

It’s surprisingly hard to find a comprehensive resource on learning how to model satellite and aerial data. Anyways, my favorite way to learn about something is to write a book about it 😉. So, here we are. I’ll publish chapters of the Machine Learning for Remote Sensing Book one by one.

Please let me know if you would be interested in such a book and would like to get updates when new chapters are published:

Who is this project for?

I write this book from my perspective: It’s for someone with a background in data science, machine learning or statistics background who wants to learn how to model satellite data using machine learning. That’s why I think this book will be useful for:

  • People with a basic machine learning background who want go get into remote sensing.
  • Machine learning, AI and data science students.
  • Researchers who apply machine learning to satellite data.
  • Remote sensing experts with at least a basic understanding of machine learning.

What to expect

  • The book is also a way for me to learn, so it will be a bit more raw but hopefully also more authentic.
  • Slowly but surely, I want this book to become the greatest resource on the internet for Machine Learning + Remote Sensing.
  • I’ll add lots of examples using Python code.
  • This website will always be freely accessible.