Detecting and segmenting destructive anomalies in farmland from satellite images, improving time, efficiency, and crop yield.

基于YOLOv11+Flask+Vue+SQLite3的杂草检测系统

Solution IA dédiée à la surveillance des cultures tropicales, permettant la détection automatique de la mosaïque du manioc et des dégâts causés par la chenille légionnaire d'automne sur le maïs grâce à la vision par ordinateur et à YOLOv11. — https://huggingface.co/kjd-dktech/agbledo01

PyTorch MLP that forecasts crop yield (kg/ha) one year ahead for 165 countries & 102 crop types using multi-source climate, soil, and land-cover data. R²=0.9452 | Pearson r=0.9681 | 52K+ training samples.

🍅 AI-powered tomato classification system using ResNet-50 and color analysis to sort tomatoes into ripe, unripe, and damaged categories. Includes video frame extraction, batch processing, and pre-trained model with 95%+ accuracy.

Open-source AI agriculture platform — TensorFlow crop yield prediction, OpenCV plant disease detection, JWT-authenticated Flask REST API, SQLite crop marketplace, and a React 18 + Vite dashboard. Self-hosted, fully documented, and live in under 2 minutes.

A multi-task deep learning pipeline leveraging a pretrained MobileNetV2 backbone to jointly classify disease type, progression stage, and days since infection from leaf imagery. Outputs feed an urgency scoring function for actionable treatment recommendations..

Hybrid deep learning pipeline: MobileNetV2 embeddings + GLCM/HSV features → XGBoost classifier. 92.36% accuracy across 12 sugarcane disease classes on 12,000 images.

Folder with code related to object detection in the CCTV cameras placed in the agricultural field and also down streaming for agricultural use-case

Deep learning feasibility study for automated clove quality classification using CNN architectures on a novel Zanzibar dataset. AI for East Africa Conference (AI4EAC) 2026, Kigali, Rwanda.

Multi-LLM, multi-tier prompt engineering benchmark for CNN-generated training pipelines on coffee leaf disease recognition (RoCoLe dataset)

Streamlit ML app predicting optimal crop types from N/P/K ratios, soil pH, temperature and rainfall using RandomForest + SHAP explainability. Batch CSV inference supported.

Reproducible benchmark for Black Soldier Fly larvae detection and image-level counting with classical computer vision and lightweight YOLO models.

Plant health AI platform — leaf disease classification (38 classes) from photos, soil/climate care recommendations, growth stage tracking, and LLM-powered plant Q&A.

Official benchmark dataset and code for clove quality grading using classical texture features and fine-tuned deep models. CVPR 2026, Vision for Agriculture (V4A) Workshop.

Curated agricultural and plant AI models relevant to grass AI-genomics.

End-to-end crop prediction ML deployment — joblib-serialised RandomForest/XGBoost served via FastAPI or Streamlit, with Docker containerisation and health-check endpoint.

Plant AI internship project — leaf disease classification (38 classes) from photos, soil/climate care recommendations, LLM-powered plant Q&A chatbot, and growth stage tracking via Streamlit.

Hardware-aware classical ML for silkworm Grasserie disease detection on PYNQ-Z2 — MYZ307E term project, ITU 2026