Inspiration
Farmers often face crop diseases that can significantly affect crop yield and income. Identifying a disease at an early stage can be difficult, especially when expert agricultural guidance is not immediately available.
We were inspired to build CropCare AI as a smart agricultural platform that combines AI-based crop disease detection with practical farming information. Our goal is to make useful agricultural insights more accessible to farmers through a single platform.
What it does
CropCare AI is an AI-powered smart farming platform designed to assist farmers with crop health and agricultural decision-making.
The platform provides:
- 🌱 AI-based Crop Disease Detection — farmers can upload a crop/leaf image and receive a predicted disease.
- 🧪 Soil and Crop Advice — provides information to help farmers make better crop and soil-related decisions.
- 🌾 Mandi Prices — provides crop market price information.
- 🌦️ Weather Information — helps farmers monitor weather conditions relevant to farming.
- 🏛️ Government Schemes — helps farmers discover relevant agricultural schemes.
- 🧫 Seed Laboratory Information — provides information about available seed laboratories.
- 📊 Reports and Notifications — helps farmers and agricultural officers monitor important information.
- 👨🌾 Role-based Access — the platform supports different workflows for Farmers, District Officers, and Administrators.
The main AI component focuses on image-based crop disease detection using a deep-learning model.
How we built it
CropCare AI was developed as a full-stack application with separate frontend, backend, and AI components.
Frontend
We built the frontend using React and Vite, with separate interfaces for Farmers, District Officers, and Administrators. The frontend communicates with the backend through REST APIs.
Backend
The backend follows a microservices architecture using Spring Boot. Different services handle authentication, users, farmer information, disease detection, soil information, seed laboratories, government schemes, reports, notifications, mandi information, and other platform functionality.
We use an API Gateway and service discovery to connect the different backend services while maintaining a consistent API flow for the frontend.
AI Service
For crop disease detection, we use a deep-learning image classification pipeline. Our current baseline model is EfficientNetB0, which was trained locally for the project.
The trained model processes an uploaded crop image and predicts the corresponding disease class. The AI component is integrated with the backend so that predictions can be accessed through the CropCare AI application.
Database and Integration
We use Oracle Database for the main application data and MongoDB for notification-related data. REST APIs connect the frontend, backend services, and AI service into a single application.
Challenges we ran into
One of our major challenges was integrating multiple technologies into one working system.
Some of the challenges we faced included:
- Integrating the React frontend with multiple Spring Boot microservices.
- Implementing authentication and role-based access control.
- Connecting the AI disease-detection service with the backend.
- Handling image uploads correctly between the frontend, API Gateway, backend, and AI service.
- Maintaining consistent data flow between different services.
- Training and evaluating the crop disease detection model.
- Debugging communication issues between microservices and databases.
- Making sure the system remained easy to use while supporting different user roles.
Through these challenges, we learned that building an AI application is not only about training a model; reliable integration between the AI, backend, database, and user interface is equally important.
Accomplishments that we're proud of
We are proud to have developed CropCare AI as a complete working platform rather than only an isolated AI model.
Our key accomplishments include:
- Building a complete React + Vite frontend.
- Developing a Spring Boot microservices backend.
- Implementing authentication and role-based access for Farmers, District Officers, and Administrators.
- Integrating an AI-based crop disease detection pipeline.
- Connecting the AI service with the application through APIs.
- Integrating agricultural services such as mandi prices, weather, government schemes, soil information, and seed laboratories.
- Creating separate workflows for different users.
- Successfully bringing multiple technologies together into one agricultural platform.
What we learned
Building CropCare AI helped us understand how different parts of a real-world software system work together.
We learned about:
- Full-stack application development.
- React frontend development and REST API integration.
- Spring Boot microservices architecture.
- API Gateway and service discovery.
- JWT-based authentication and role-based authorization.
- Database integration with Oracle and MongoDB.
- AI model training and evaluation.
- Image-processing and model-serving workflows.
- Debugging distributed applications.
- The importance of designing AI solutions around a practical real-world problem.
Most importantly, we learned that a useful AI solution requires more than model accuracy. The model, application, data, APIs, and user experience must work together to solve the actual problem.
What's next for CropCare AI
We plan to continue improving CropCare AI by making the disease detection system more robust and expanding its agricultural intelligence.
Our future plans include:
- Improving the crop disease detection model through further experimentation and evaluation.
- Supporting more crops and disease categories.
- Improving prediction reliability under different lighting and image conditions.
- Adding more personalized recommendations based on crop and soil information.
- Expanding weather and market-data integration.
- Providing more useful insights for agricultural officers.
- Improving the platform's scalability and deployment.
- Exploring additional AI models and architectures to improve disease classification performance.
Our long-term vision is to evolve CropCare AI into a comprehensive AI-powered agricultural assistant that helps farmers detect crop problems early and make more informed farming decisions.
Log in or sign up for Devpost to join the conversation.