Skip to content

nitin-dominic/AgriPathogenDatabase

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 

Repository files navigation

Crop Disease Foundation Models: Technical Articles, Codes, and Resources

A single, well-documented repository that collects papers, datasets, code, pretrained models, simulation/digital-twin environments, baseline notebooks, evaluation scripts, and demo UIs — all specifically for crop disease detection, monitoring, and management (text + vision + multimodal + decision-making). Curated resources, code, and benchmarks for foundation-model-enabled crop disease detection, monitoring, and management.

What this repo contains

  • papers/ — annotated list of recent studies (LLMs, VLMs, multimodal) focused on crop diseases.
  • datasets/ — links + preprocessing scripts for PlantVillage, in-field datasets, hyperspectral disease sets.
  • models/ — baseline models (classification, detection, VLMs) and pointers to weights.
  • sim/ — digital twin environments and sim-to-real experiment configs.
  • experiments/ — RL/adaptive-learning recipes and evaluation scripts.
  • notebooks/ — walkthroughs: detection → explanation → prescription report → Q&A.

⚠️ Disclaimer: This repository is a work in progress. Content, papers, and code are continuously being added and updated.


Table of Contents


Introduction

This repository curates recent studies on crop disease detection, monitoring, and management, with a focus on:

  • Vision-Language Models (VLMs), Large Language Models (LLMs), and multimodal approaches
  • Early disease detection and targeted interventions
  • Human-in-the-loop collaboration for validation and adaptive learning
  • Simulation and digital twin environments for model testing and sim-to-real transfer

2024

LLM-focused Studies 2024

Approach / Models Model Type Open-access source Use Cases Journal Reference
Used LLM with Agricultural Knowledge Graphs (KGs), Graph Neural Networks (GNNs) LLM Not specified Plant disease diagnosis, reasoning over symptoms, linking textual disease corpora with structured knowledge MDPI Agriculture Zhao2024
GPT-4 (OpenAI API) for automated literature synthesis on pest controllers LLM Proprietary (OpenAI) Automating systematic reviews, reducing expert workload Methods in Ecology and Evolution Scheepens2024
GPT-based models for sensor + text queries LLM Proprietary (OpenAI) Query-based plant health monitoring, e.g., explaining yellowing leaves using sensors + LLM reasoning Int. J. Computer Applications in Technology Ahir2024
Q&A systems using GPT-4 + knowledge graphs LLM --- Plant disease diagnosis, pest identification, Q&A for sustainable management Resources, Conservation and Recycling Yang2024
GlyReShot (Chinese agricultural NER, few-shot + GROM module) LLM --- Recognizing entities (disease, crop, pest, drug) in Chinese agricultural text Heliyon Liu2024
RAG chatbot with hybrid DeiT + VGG16 VLM Not explicitly available Medicinal plant identification + bilingual insights using images + RAG Telematics and Informatics Paneru2024
Agricultural Knowledge Graph (AGKG) + LLMs LLM Not specified Entity retrieval and Q&A via domain-specific AGKG Displays Wang2024

VLM-focused Studies 2024

Models Type Open-access source Use Cases Journal Reference
DINOv2 Vision model Hugging Face Self-supervised feature extraction, clustering of disease symptoms ScienceDirect Bai2024
BLIP / BLIP-2 Multi-modal model Hugging Face Image captioning and visual reasoning for disease explanation - Liang2024
LLaVA Multi-modal model Hugging Face Multi-modal reasoning for plant disease recognition - Xu2025
SAM VLM GitHub Wheat disease diagnosis through reasoning ScienceDirect Zhang2024
ViT + GPT-2 VLM OpenAI / Hugging Face Align plant disease phenotypes with trait descriptions Plant Phenomics Zhao2024
Inception-v4 + LSTM VLM - Align crop disease images with question embeddings Plant Phenomics Zhao2024
PDC-VLD Multi-modal (vision + text) - Tomato leaf disease detection with unseen class generalization Plant Phenomics Li2024
FHTW-Net Vision-language model GitHub Retrieve matching text from a query image (and vice versa) for rice leaf disease descriptions Plant Phenomics Zhou2024
ILCD Multi-modal visual question answering GitHub Addressed complex questions about crop disease stages and attributes Plant Phenomics Zhao2024
PhenoTrait (GPT-4 / GPT-4o) Multi-modal (image-to-text) PlanText Generates plant disease descriptions from images Plant Phenomics Zhao2024
PepperNet Multi-modal VLM - Detecting pepper diseases and pests using natural language descriptions Nature Scientific Reports Liu2024
Qwen-VL Pre-trained VLM Google Drive Generate text descriptions for disease images as prompt for classifiers MDPI Sensors Zhou2024
Segment Anything Model (SAM) Image segmentation SAM-Meta AI Identifies and segments diseased regions in leaves IEEE Access Moupojou2024
Visual Answer Model (VQA) Multi-modal VQA - Answer questions about fruit tree diseases using images + Q&A knowledge Frontiers in Plant Science Lan2023

2023

LLM-focused Studies 2023

Approach / Models Model Type Open-access source Use Cases Journal Reference
GPT-3.5 for agricultural extension services LLM Proprietary (OpenAI) Farmer advisory chatbots, pest/disease diagnosis, local language support Nature Food Tzachor2023
ChatAgri (ChatGPT-based agricultural text classification) LLM GitHub link Cross-lingual agricultural news classification, few-shot and prompt-based learning Neurocomputing Zhao2023

VLM-focused Studies 2023

Models Type Open-access source Use Cases Journal Reference
ITLMLP Vision-language pre-training - Few-shot cucumber disease recognition using image, text, and label information Computers and Electronics in Agriculture Cao2023
YOLO + GPT Multi-modal model GitHub / OpenAI Generate agricultural diagnostic reports with deep logical reasoning Computers and Electronics in Agriculture Qing2023
ITF-WPI Cross-modal feature fusion GitHub Wolfberry pest identification using image + text Computers and Electronics in Agriculture Dai2023
Neuro-symbolic AI Deep learning + knowledge graph GitHub Improves prediction accuracy and explains results for farmers Expert Systems with Applications Chhetri2023
ShuffleNetV2 + TextCNN Multi-modal model - Extract textual features and semantic relationships from descriptive text Nature Scientific Reports Qiu2023
MMFGT Multi-modal transformer - Few-shot pest recognition combining image + text information MDPI Electronics Zhang2023
ODP-Transformer Multi-modal image-to-text + classification - Two-stage pest classification + caption generation Computers and Electronics in Agriculture Wang2023

2022

LLM-focused Studies 2022

Approach / Models Model Type Open-access source Use Cases Journal Reference
AgriBERT (BERT-based, knowledge-infused with FoodOn/Wikidata) LLM --- Semantic matching of food descriptions, cuisine classification, agricultural NLP tasks IJCAI 2022 Rezayi2022

2021

VLM-focused Studies 2021

Models Type Open-access source Use Cases Journal Reference
ITK-Net (Image-Text-Knowledge Network) Multi-modal - Identify invasive diseases in tomato/cucumber using image + text + domain knowledge Computers and Electronics in Agriculture Zhou2021

2018

LLM-focused Studies 2018

Approach / Models Model Type Open-access source Use Cases Journal Reference
Original GPT pre-training paper (OpenAI) LLM Not open-source initially Laid foundation for later agricultural LLM applications OpenAI Radford2018

Datasets

  • PlantVillage: Lab images of diseased crops
  • In-field disease datasets: Multi-season field imagery
  • Hyperspectral datasets: Spectral bands for early detection

Models & Benchmarks

  • Classification: EfficientNetV2, ResNet
  • Detection/Segmentation: Faster R-CNN, YOLOX, U-Net
  • Multimodal / VLMs: CLIP, BLIP2, vision-language pipelines
  • RL / adaptive learning setups for targeted interventions

Simulation & Digital Twin Studies

  • Virtual plant canopies with disease progression
  • Targeted spraying simulation with reward optimization
  • Sim-to-real transfer experiments with domain randomization

Human-in-the-Loop Approaches

  • Shared autonomy for disease management
  • Validation of uncertain cases by humans
  • Annotation loops for continual model improvement

Experiments / Workflows

  • Prescription report generation from visual and textual data
  • Q&A and extension service platforms
  • Grad-CAM explainability for disease detection

References

  • Add full citations here (BibTeX or numbered list)
    s-by-year)

Quick start

  1. Clone: git clone https://github.com/<you>/Crop-Disease-Foundation-Models
  2. Read datasets/readme.md for dataset setup.
  3. Launch demo: python demos/gradio_demo.py

How to Contribute

If you have suggestions, come across missed papers, or find useful resources, we welcome your contributions via pull requests.

Guidelines:

  1. Use the following Markdown format when adding a new paper:

  2. For preprints with multiple versions, use the earliest submitted year.

  3. Display papers in descending chronological order (latest first).

  4. Indicate clearly whether the study is LLM-focused or VLM-focused.

  5. For code or datasets, provide a direct link and a short description.

Example:

Zhao, J., Li, H., and Wang, P. GlyReShot: Few-shot Chinese Agricultural NER. Heliyon 2024. [PDF]; [GitHub].

Cite

If you use this repo, please cite: @misc{...}

Contributions welcome — see CONTRIBUTING.md.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published