This repository was built to streamline the comparison of different DSM pipelines.
The Dockerfile can be used to install several pipelines, such as Nasa's ASP, CNES' CARS, Centre Borelli's s2p-hd, IGN's Micmac.
The functions in comparison.py can be used to compute statistics for the DSM reconstructions of each method. The code used to compare DSMs to ground truths is dicretly inspired from the library demcompare which we thank.
You need to install the demcompare library from this repository:
git clone https://github.com/centreborelli/dem-evaluation.git
cd dem-evaluation
pip install -e ./demcompare
If you use this software please cite the following paper:
s2p-hd: Gpu-Accelerated Binocular Stereo Pipeline for Large-Scale Same-Date Stereo, Tristan Amadei, Enric Meinhardt-Llopis, Carlo de Franchis, Jeremy Anger, Thibaud Ehret, Gabriele Facciolo. CVPR EarthVision 2025.