Edge-Enabled Hybrid AI Framework for IoT-Based Crop Management and Irrigation ControlOpen Access

Abstract

The intersection of the Internet of Things (IoT) and Artificial Intelligence (AI) technologies has ushered in a new era of modern agriculture, transforming it into a data-driven, intelligent, and sustainable system. This paper presents an IoT and AI-based Agriculture Framework (IAAF) that integrates heterogeneous sensor networks, drone-based imaging, and hybrid deep learning for real-time precision crop management. The IAAF implemented a CNN-LSTM network for time-dependent sensor data analysis, a transfer-learning-based CNN for multispectral image classification, and a multimodal attention-based fusion model that combines time-dependent and spatial information to facilitate integrative decision-making. Additionally, the framework utilized edge computing to minimize latency and reduce bandwidth consumption and cloud services to manage model retraining and facilitate long-term analysis. Experimental assessments using multi-season field datasets provide evidence that the IAAF framework achieved significantly better performance than baseline models, with a prediction accuracy of 98.78%, a precision of 98.43%, a recall of 98.51%, and an F1 score of 0.986. Accordingly, the IAAF proposes an innovative, intelligent, and resource-efficient model of real-time precision agriculture, promoting a sustainable vision of smart farming.