This full-stack web application enables users to upload images of diseased plants to diagnose their illnesses. When logged in, it includes a dashboard with analytics and submission history to provide actionable insights. We used PyTorch to build a Machine Learning classification model on the backend, along with utilizing Google's Gemini Model to provide further insights, augmenting the disease classification and giving solutions to help cure the plants.
The frontend is already hosted on https://crop-guard-peach.vercel.app/. However, the backend needs to be run locally. Instructions to set up local backend:
- Clone git repo into your local folder
- Enter the backend directory
- Create a virtual environment with
python -m venv .
- Install dependencies with
pip install -r requirements.txt
This might take some time. - Start server by running
python main.py
From there, you should be able to use Crop Guard through the above link.
Frontend - Next.js, Tailwind CSS, TypeScript Backend - Python, FastAPI, PyTorch, SQLite
- Improve model + expand to other plant types
- Incorporate DB storage + retrieval of images
- Live cam of model
- Responsive UI for mobile + browser types
- Fix HTTPS proxy security with Vercel - maybe nginx
- Admin analytics