Authors' implementation of the Eurographics 2025 State of the Art "Terrain Descriptors for Landscape Synthesis, Analysis and Simulation"
RESISC45-SigLIP2 is a vision-language encoder model fine-tuned from google/siglip2-base-patch16-224 for multi-label image classification. It is specifically trained to recognize and tag multiple land use and land cover scene categories from the RESISC45 dataset using the SiglipForImageClassification architecture.
SAT-Landforms-Classifier is an image classification vision-language encoder model fine-tuned from google/siglip2-base-patch16-224 for a single-label classification task. It is designed to classify satellite images into different landform categories using the SiglipForImageClassification architecture
Methods that generates imaginary landform contour paths
QGIS Processing Tool for unsupervised landform classification from DEMs using Gaussian Mixture Models (GMM) and Markov Random Fields (MRF).
Geomorphological mapping and Geovisualization
Python scripts supplementing the MGISA landform mapping workflow (preprocessing, analysis, assessment, ensemble) described in Robillard (2025).
To associate your repository with the landform topic, visit your repo's landing page and select "manage topics."