GAN-based Mango Variety & Disease Detection
This AI/ML research project applies Generative Adversarial Networks (GANs) and computer vision to classify mango varieties and detect diseases from leaf and fruit images. It enables farmers, researchers, and agri-businesses to ensure authenticity, monitor crop health, and reduce yield losses through automated, data-driven insights.
Purpose
Automate mango variety recognition (Chaunsa, Sindhri, Anwar Ratol, etc.) and enable early disease detection (anthracnose, powdery mildew, bacterial black spot). Reduce dependency on human experts in the field, generate synthetic training data to overcome dataset limitations, and support yield improvement and quality assurance.
Who Uses It:
- Farmers & Growers: Capture mango/leaf images via mobile app and receive instant feedback on variety and disease status.
- Agricultural Researchers: Access annotated datasets, GAN-generated synthetic images, and model outputs to study mango genetics and diseases.
- Exporters & Quality Labs: Verify authenticity of mango varieties and certify disease-free shipments.
- AgriTech Companies: Integrate detection APIs into digital agriculture platforms.
Key Capabilities:
- GAN-based Data Augmentation: Creates synthetic mango/leaf images to expand datasets, improving robustness against lighting, seasonal, and noise variations.
- Variety Classification: Identifies mango type with high accuracy from fruit morphology and color features.
- Disease Detection: Detects visual symptoms (spots, fungal growth, discoloration) early; provides severity scores for crop management.
- Mobile/Web Integration: Farmers can upload images directly; offline sync supports rural use cases.
- Analytics & Dashboards: Orchard-to-regional level reports, variety distribution maps, and disease prevalence trends.
Mobile / Web App Highlights:
- Guided image capture to ensure quality.
- Instant AI feedback on variety & disease status.
- Recommendations for treatment or agronomy practices.
- Training mode with healthy vs. diseased comparisons.
Outcomes & Value:
- Improved accuracy in mango classification and disease diagnosis.
- Reduced post-harvest losses through early detection.
- Greater accessibility for farmers and exporters through mobile-first tools.
- Stronger data foundation for agricultural research and AgriTech platforms.