Revolutionizing pest detection in agriculture using artificial intelligence, Internet of Things, machine and deep learning approaches

Abstract

Insect infestations cause major crop losses and often drive excessive pesticide use. Traditional detection methods remain slow, labour-intensive, and subject to human error, limiting timely intervention. Recent advances in Artificial Intelligence and the Internet of Things (IoT) have enabled faster, more accurate, and automated pest detection. Deep learning models such as Convolutional Neural Networks (CNN), Residual Networks, VGG architectures, and lightweight MobileNet variants achieve strong pest recognition performance using high-resolution crop and pest images. Classical machine learning techniques, including Support Vector Machines, Random Forests, and k-Nearest Neighbours, remain valuable when data availability or computational resources are limited. The IoT-based monitoring platforms strengthen these capabilities by integrating distributed sensors, automated imaging systems and edge-processing hardware to provide continuous, real-time field observations. Emerging technologies such as multispectral imaging, UAV-based data collection, and multimodal sensor fusion further enhance detection accuracy in complex outdoor environments. However, several challenges persist, including dataset imbalance, variable lighting, background clutter, and environmental noise, which hinder the reliable transfer of AI systems from controlled laboratory conditions to real agricultural settings. This review synthesises AI techniques, IoT architectures, and edge-intelligence strategies within a unified framework, highlighting their combined potential and identifying future directions for robust, automated pest detection.