The predicted soil type is subsequently used by the crop recommendation pipeline to generate nutrient-aware crop recommendations.
Place the file in the root directory of the AgriSense project before running inference.
1AgriSense/
2├── soil_model.h5
3├── app.py
4├── app2.py
5├── predict.py
6├── predict2.py
7└── ...
1from tensorflow.keras.models import load_model
2
3model = load_model("soil_model.h5")
1@inproceedings{agrisense2025,
2 title={A Hyper-Local Two-Stage Machine Learning Framework for Soil Classification and Nutrient-Aware Crop Recommendation},
3 author={Rajadhyaksha, Sahil and Dere, Durgesh and Koli, Preksha and Patil, Shrut and Sankhe, Sanskruti and Mapare, Sheetal},
4 booktitle={2025 18th International Conference on Sensing Technology (ICST)},
5 year={2025}
6}
