Krishi-Sahayak: AI-Based Farming Assistant for Enhanced Agricultural Productivity and Sustainable Development

Abstract—This paper presents Krishi-Sahayak, an AI-powered farming assistant application designed to revolutionize agricul- tural practices in India. The system generates weekly tasks for farmers based on field size, location, soil analysis, and time-series analysis of crop prices in nearby markets. The application inte- grates multiple advanced technologies to provide comprehensive solutions, including personalized crop recommendations based on soil analysis, disease detection through leaf image classification using Convolutional Neural Networks, weather-driven dynamic farming calendars, and multilingual support for enhanced acces- sibility. A unique feature of the system is its interactive platform where farmers can submit and upvote agricultural concerns, facilitating communication with policymakers. The application addresses critical challenges in agriculture by leveraging tech- nologies such as MobileNetV2 for disease detection, K-Nearest Neighbors (KNN) for crop recommendations, and real-time APIs for weather and market data. Our evaluation demonstrates significant improvements in crop yield optimization and disease detection accuracy, indicating the system’s potential for wider adoption and impact in the agricultural sector. Index Terms—Agriculture, Artificial Intelligence, Machine Learning, Crop Disease Detection, Smart Farming, Agricultural Technology, CNN, KNN, MobileNetV2