AgroScan is a deep learning-powered solution that detects crop leaf diseases from images. The model is designed to be deployed via a WhatsApp chatbot using Twilio, making it accessible to farmers and agricultural workers in remote or low-connectivity areas.
AgroScan is currently being developed as part of the Africa Deep Tech Challenge 2025.
- Build an image classifier that detects plant diseases from leaf photos
- Deploy the model through a WhatsApp bot for easy farmer access
- Promote early detection and minimize crop loss in agriculture
- Model: Convolutional Neural Network (CNN)
- Trained on: PlantVillage Dataset (Kaggle)
- Output: Disease classification (e.g., Early Blight, Late Blight, Healthy)
- Evaluation: Accuracy, Confusion Matrix, Precision/Recall
- Python
- TensorFlow / Keras
- OpenCV
- Pandas, NumPy
- Twilio API (WhatsApp integration)
- Hugging Face Spaces / FastAPI for hosting
The model is trained using the PlantVillage Dataset, a labeled image dataset of healthy and diseased crop leaves.
Dataset is not included in this repo due to size.