PlantUs | Devpost

home page
a simplified, pixel-style simulation of the community garden.
you can add columns and rows to your field
cultivate new crops
Click on plants to access their profiles and real-time health status.
Users upload plant images and test the plant's heath
pretrained AI model diagnoses plant health, detecting diseases, pests, or nutrient deficiencies.
if a diseases, pests, or nutrient deficiencies detected by the Ai model
if the users takes action and the plants heals they get a point they can redeem later

Theme

Sur la terre / Above Ground

Inspiration

Urban living often disconnects people from nature and food production. At the same time, food security and sustainable agriculture are urgent global issues. We wanted to merge technology and community-driven collaboration to reimagine urban farming—making it smarter, more engaging, and scalable. Thus, PlantUs was born. The name combines “Plant” and “Us” to emphasize the power of shared responsibility and community-driven growth.

What it does

PlantUs is a community-driven smart farming platform that brings urban growers together in a shared virtual garden to monitor and optimize plant health. It combines real-world data simulation with an AI-powered plant health detection model to assist users in caring for plants while earning rewards through gamification.

Key Features:

-Interactive Field Dashboard: Displays a simplified, pixel-style version of the shared garden. Click on plants to access their profiles and real-time health status. -AI-Powered Health Detection: Users upload plant images, and a pretrained AI model diagnoses plant health, detecting diseases, pests, or nutrient deficiencies. -notification and alerts:recommended actions to take -Action-Reward System: Users can take recommended actions to solve plant health issues. Successful interventions earn points, which can be redeemed for future rewards (like additional harvest shares).

How we built it

-Frontend and backend: javascript, flask -AI Model: Pretrained image classification model for plant disease detection, optimized for accuracy and real-time feedback.

Challenges we ran into

-AI Integration: Ensuring our pretrained plant health detection model provided accurate and useful -feedback was a significant challenge. -Switching Tech Stacks: Initially built with Node.js, we pivoted to Python for better AI compatibility, which required adapting our backend. -Gamification Design: Creating a balance between fun and functionality while keeping users engaged in plant care was a unique design challenge.

What we learned

What's next for PlantUs

-Connection to real camera for localized info on each crop -Integrate real sensor info for humidity and soil quality check -Integrate weather prediction for recommendations on harvest time -Advanced UI: A game-like, immersive interface for the field where users can watch their avatar tend to plants. -Point Redemption: Use earned points to redeem gardening tools, exclusive plant seeds, and more harvest shares. -Community Money Management: Buy new tools, equipment, and garden upgrades through pooled resources. -User Levels & Avatars: Choose between beginner, intermediate, and advanced gardening modes for a tailored experience. -Global Expansion: Link urban growers across the globe, creating a worldwide knowledge-sharing network for sustainable agriculture. Gardening Knowledge Hub: A forum-like platform where users share tips, experiences, and articles—like Stack Overflow for gardening!

Credits

For Plant Virus Detection Model : link The CNN model architecture consists of CNN Layer, Max Pooling, Flatten a Linear Layers. Using Transfer learning VGG16 Architecture. Using Transfer learning resnet34 Architecture.

For Pest Detection : link

YOLOv8 Nano architecture, which is a lightweight version of the YOLOv8 object detection link

powerpoint template: link