AgriSmart.AI

🚧 Progress During the Hackathon During the two-day Hack4Bengal Virtual Hackathon, our team built AgriSmart.Ai, an AI-driven agricultural assistant designed to empower farmers with real-time insights and intelligent recommendations. We began by identifying key problems faced by the agricultural community—lack of accessibility to crop-related data, soil health insights, and early disease detection. With this in mind, we developed a web-based platform equipped with machine learning models, a multilingual chatbot, and real-time forecasting capabilities. We developed and trained a Convolutional Neural Network (CNN) model using the PlantVillage dataset for crop disease detection, optimized for low-resource environments. This model was successfully integrated into the platform, allowing farmers to upload leaf images and get instant disease predictions. Alongside this, we implemented a soil health analysis tool, which recommends suitable crops based on pH, NPK values, and moisture levels. Our multilingual chatbot, built with NLP capabilities, allows farmers to interact in regional languages and receive relevant suggestions on crop selection, fertilizer use, and pest control. We also incorporated a yield prediction module using XGBoost and time-series forecasting, and added weather forecasting features based on external weather APIs. The platform was built using a modern tech stack — React.js, Node.js, Express.js, and Flask, with MongoDB as the primary database and TensorFlow, Scikit-learn, and OpenCV powering the AI components. 🔮 Future Prospects We envision taking AgriSmart.Ai beyond static data by integrating real-time soil sensors. Using microcontrollers and IoT devices, we plan to collect live NPK, pH, and moisture data directly from the soil. This data will be sent to an IoT cloud platform, from where our web app can fetch and analyze it continuously, offering highly accurate and personalized recommendations. This will not only make AgriSmart more intelligent but also turn it into a fully autonomous smart farming assistant. With further advancements, we aim to: - Develop a mobile version of AgriSmart for better accessibility in rural areas. - Add voice assistant features to assist farmers with limited literacy. - Collaborate with agri-tech partners and governmental bodies to deploy AgriSmart at scale. Future Prospect – IoT Integration: - Using ESP32 microcontrollers with soil sensors. - Measuring real-time pH, NPK, and moisture from the field. - Transmitting data via Wi-Fi or LoRa to ThingSpeak/Firebase. - Web app fetches live data for instant analysis. - AI delivers location-specific crop and fertilizer recommendations. - No manual input needed — fully automated smart assistant. - AgriSmart becomes dynamic, adaptive, and continuously learning.