A Web-Based Agriculture Recommendation System using Deep Learning for Crops, Fertilizers, and Pesticides

Abstract

Agriculture plays a pivotal role in the economies of developing countries like India, making a substantial contribution to their gross domestic product (GDP). However, farmers often struggle with selecting appropriate crops, fertilizers, and pesticides, frequently neglecting site-specific factors like soil type, water requirements, temperature, and regional crop profitability. These oversights can lead to reduced crop quality, yield, and profitability. To address this challenge, a recommendation system is proposed. It employs an ensemble model with a majority voting technique for crop selection based on site-specific parameters, aiming for high accuracy and efficiency. This system utilizes NPK (Nitrogen, Phosphorus, and Potassium) values, temperature, relative humidity, and pH data for real-time crop prediction. Furthermore, it includes modules for fertilizer and pesticide recommendations. For pest identification, a convolutional neural network (CNN) model is employed, allowing users to upload images of pests, with the system recommending corresponding pesticides. Developed as a web-based platform, this system aims to provide farmers with accessible tools for informed decision-making to improve crop quality, yield, and profitability.

Topics & Concepts