dzetsaka : AI-Powered Remote Sensing Classification
Plugin ID: 995
Next-generation AI for Earth Observation in QGIS with 11 ML algorithms, major runtime optimizations, advanced UX workflows, recipe-driven configuration, and rich report generation.
(142) votes
dzetsaka is a compact AI for Earth Observation plugin for remote sensing classification in QGIS.
Built for fast map production and practical operations:
- 11 production-ready ML classifiers (GMM, RF, SVM, KNN, XGBoost, CatBoost, ET, GBC, LR, NB, MLP)
- Optuna hyperparameter optimization
- SHAP explainability and confidence outputs
- Class-imbalance workflow (SMOTE, class weights, nested CV)
- Guided/compact dashboard for rapid results
2025 trend-aligned focus:
- AI for Earth Observation and Earth Intelligence workflows
- Multimodal EO feature pipelines
- Vision-transformer-ready ecosystem language
- Trustworthy AI: interpretability, uncertainty awareness, and reproducible validation
For documentation and updates: https://github.com/nkarasiak/dzetsaka/
classification
remote sensing
processing
knn
svm
random forest
machine learning
multimodal
earth observation
deep learning
neural network
uncertainty
xgboost
ai for earth observation
open-vocabulary mapping
mlops
vision-language model
reproducibility
bayesian optimization
class imbalance
self-supervised learning
catboost
trustworthy ai
vision transformer
explainability
shap
optuna
smote
nested cross-validation
