🚀 GeoAI Tutorial 6: Building Footprint Regularization Using Deep Learning and GeoAI! 🌍 Learn how to regularize building footprints detected from aerial imagery with three different regularization techniques. Transform building footprints with rugged edges into smooth and rectangular shapes. 🎥 Watch the full tutorial here: 📓 Access the notebook:... 🛠️ Explore the GitHub repository: 📚 Dive into the documentation: 📺 Check out the entire YouTube playlist: #GeoAI #geospatial #AI #Pythonshow more

Is this applicable to low-quality imagery, especially in densely populated areas of Southeast Asia?

Maybe, I have not tested on medium resolution images yet

🚀 Discover how Reinforcement Learning can transform trading strategies! Check out my free Substack for full code and backtest of the completed Q-learning algorithm. Don't miss out! 📈✨

Is there rooftop extraction model?

Seems that your USA buildings example based on the @Esri building footprint extraction model does output nicer geometries, parametrized by rectangularity, orthogonality, angle thresholds. Is it simply a matter of params, or are the underlying methods really different?

Their building regularization algorithm isn't open source. I have to implement it on my own. If you are aware of any other open source algorithms that have better performance, I would be happy to integrate them

Wow that's useful. Amazing work, the docs are great. Going to play with this at the weekend. Is it expensive to run?

You can run it on Colab. It is free.