Code for the paper Automatic Alignment of Multi-scale Aerial and Underwater Photogrammetric Point Clouds: a Case Study in the Maldivian Coral Reef, submitted at CVEMP 2023.
Code tested with python 3.10.
Install the requirements with:
pip3 install -r requirements.txt
To extract the features with 3DSmoothNet, please consult the README. inside the 3DSmoothNet folder, and then put the resulting .npz files in the extracted_features folder with the name "<pointcloud_name>_3dsmoothnet".
To run the registration algorithms, adjust the parameters in registration.py and run the script.
To obtain the error metrics run check_gt.py