This repository accompanies our paper on automated dry stone wall mapping in terraced landscapes from ALS-derived terrain data. It provides the code for the DTM creation, the input data set generation, preprocessing, evaluation, and area-wide inference and mapping.
The deep learning model architecture and training code are available in the separate repository: https://github.com/smfloery/landformAI
In the folder example_use you can find the DTM data set used for model training, validation and testing in the paper. This paper can be directly used with the landformAI repository.
The documentation will be finished by publication of the paper