Adith082 / AgriDoctor

AgriDoctor

All Cloud Deployed/Hosted URLs for this project by us:

Java Spring Boot backend API Url: https://agridoctorbackend-production.up.railway.app
React frontend Url: https://therap-javafest-agridoctor.vercel.app
Fast-API AI/ML backend API Url: https://flamoverse-crop-recommender-from-weather-npk.hf.space

Status: Full Project is Cloud Deployed by Us.

Presentation Video

Presentation Video: Youtube Video Link

AI and ML backend

Related codes and notebooks Folder:   AI_API_Docker_codes
Hosted doc link: AI API DOC
AI/ML backend API Url: https://flamoverse-crop-recommender-from-weather-npk.hf.space

Technologies

Build tool : Docker

Framework: Fast-API

ASGI Server: Uvicorn[standard]

AI libraries: torch, tensorflow, scikit-learn, pillow, torchvision

Other Libraries: pydantic, python-multipart, starlette, numpy, requests

Crop Recommendation

Dataset used (kaggle):   Crop Recommendation Models Traning Dataset
Training notebook (Containing all details):   Crop_Recommendation_Model.ipynb
Total number of unique predictions: 22 unique different plants as follows: ['rice' 'maize' 'chickpea' 'kidneybeans' 'pigeonpeas' 'mothbeans' 'mungbean' 'blackgram' 'lentil' 'pomegranate' 'banana' 'mango' 'grapes' 'watermelon' 'muskmelon' 'apple' 'orange' 'papaya' 'coconut' 'cotton' 'jute' 'coffee']

Models trained: Random Forest, Naive Bayes, SVM Tree, Decision Tree.
Best Performing Model: Random Forest (Accuracy: 99.777%)

Crop Disease Prediction

Trained after merging following Datasets (Kaggle):

1. New Bangladeshi Crop Disease
2. New Plant Diseases Dataset

Model Architechture: Based on ResNet (CNN)
Model Performance(Accuracy): 98.5%
Total plants: 16
Total Diseases: 35
Total Classes: 45 as follows: ['Apple___Apple_scab', 'Grape___Black_rot', 'Rice___Brown_Spot', 'Orange___Haunglongbing_(Citrus_greening)', 'Tomato___Bacterial_spot', 'Potato___Late_blight', 'Apple___Black_rot', 'Corn_(maize)__Northern_Leaf_Blight', 'Potato___Early_blight', 'Corn(maize)_Common_rust', 'Peach___healthy', 'Grape___healthy', 'Cherry(including_sour)__healthy', 'Cherry(including_sour)__Powdery_mildew', 'Strawberry___Leaf_scorch', 'Tomato___Tomato_mosaic_virus', 'Rice___Neck_Blast', 'Corn(maize)_healthy', 'Potato___healthy', 'Tomato___Leaf_Mold', 'Wheat___Yellow_Rust', 'Tomato___Early_blight', 'Wheat___Healthy', 'Grape___Leaf_blight(Isariopsis_Leaf_Spot)', 'Tomato___Target_Spot', 'Tomato___Spider_mites Two-spotted_spider_mite', 'Apple___healthy', 'Squash___Powdery_mildew', 'Pepper,bell___healthy', 'Tomato___healthy', 'Tomato___Late_blight', 'Soybean___healthy', 'Peach___Bacterial_spot', 'Rice___Healthy', 'Blueberry___healthy', 'Grape___Esca(Black_Measles)', 'Corn(maize)___Cercospora_leaf_spot Gray_leaf_spot', 'Apple___Cedar_apple_rust', 'Pepper,_bell___Bacterial_spot', 'Wheat___Brown_Rust', 'Strawberry___healthy', 'Tomato___Tomato_Yellow_Leaf_Curl_Virus', 'Tomato___Septoria_leaf_spot', 'Raspberry___healthy', 'Rice___Leaf_Blast']

Training notebook (Containing all details): AI_API_Docker_codes/crop_disease_detection/plant-disease-classification-resnet-data-merged.ipynb

React Frontend

Code Folder: frontend
React frontend cloud deployed Url: https://therap-javafest-agridoctor.vercel.app

Technologies

Build tool : create-react-app

Framework: React

Template Engine: JSX

Styling: CSS

Prebuilt Design Framework: react-bootstrap

HTTPs Request Library: Axios

Other libraries: react-toastify, react-router-dom

Picture Credits

Image and Icon Credits: Flaticon, Freepik