Enhancing Crop Health in Smart Greenhouse Through IoT-Based Data Optimization and Deep Learning Algorithms

Abstract

Integrating convolutional neural networks (CNNs) with the Internet of Things (IoT) is paramount in agriculture, particularly greenhouses. By leveraging IoT capabilities, operators can collect agro-environmental information inside the greenhouse based on installed sensor nodes. This data-driven approach minimizes water, fertilizer, and energy waste. Simul-taneously, CNNs enhance the monitoring systems by facilitating early detection and classification of crop diseases. Our research proposes a comprehensive solution: an online technology platform for intelligent greenhouses based on IoT and CNNs. This platform effectively collects environmental and physical variables and detects diseases in real time using image-based analysis. The results of our study demonstrate that the system architecture is a reliable IoT platform, leading to significant energy savings. Moreover, the disease identification accuracy and classification process achieved an impressive rate of over 98 %, ensuring the system's efficacy in identifying and categorizing diseases. Additionally, the system exhibits a recall rate of over 90 %, indicating its ability to identify and recall crop disease instances accurately.