Learnable Earth Parser: Discovering 3D Prototypes in Aerial Scans
This provides a practical tool for aerial surveying and mapping by enabling unsupervised analysis of 3D scenes, though it is incremental as it builds on existing unsupervised techniques.
The paper tackles the problem of parsing large 3D aerial scans without annotations by proposing an unsupervised method that decomposes point clouds into learned prototypical shapes, resulting in improved decomposition accuracy over state-of-the-art methods while maintaining visual interpretability.
We propose an unsupervised method for parsing large 3D scans of real-world scenes with easily-interpretable shapes. This work aims to provide a practical tool for analyzing 3D scenes in the context of aerial surveying and mapping, without the need for user annotations. Our approach is based on a probabilistic reconstruction model that decomposes an input 3D point cloud into a small set of learned prototypical 3D shapes. The resulting reconstruction is visually interpretable and can be used to perform unsupervised instance and low-shot semantic segmentation of complex scenes. We demonstrate the usefulness of our model on a novel dataset of seven large aerial LiDAR scans from diverse real-world scenarios. Our approach outperforms state-of-the-art unsupervised methods in terms of decomposition accuracy while remaining visually interpretable. Our code and dataset are available at https://romainloiseau.fr/learnable-earth-parser/