Recent advances in crop pest detection, forecasting and early warning: A review

Crop insect pests are critical biotic stressors disrupting plant physiological functions. Their infestations not only cause direct yield losses but also threaten global food security and agricultural economy. Therefore, rapid and accurate pest detection and the ability to predict infestations and issue early warnings have become imperative. Traditional pest management often faces timeliness, accuracy, and scalability limitations due to labor-intensive practices and the dynamic nature of pest behaviors, struggling to achieve precise detection and warnings. Recent advancements in sensing technologies and artificial intelligence have catalyzed transformative intelligent monitoring systems, offering innovative solutions for pest management. This paper reviews recent developments and applications in crop pest detection, prediction and early warning. Sensors and sensing systems, including imaging sensors, acoustic sensors, chemical sensors and wireless sensor network (WSN) are systematically detailed. The methods delve into the characteristics and development of various models, categorizing detection modalities including sound signal-based, electronic nose-based, and computer vision-based methods, while prediction and warning frameworks incorporate traditional time series analytics, machine learning architectures, and deep learning networks. Building upon these advancements, empirical case studies encompassing crop insect detection, prediction, and early warning across heterogeneous agroecological contexts are systematically examined. Despite technological advancements, persistent challenges in data availability, environment changes, and model interpretability are thoroughly discussed. Finally, future research directions of crop insect pest detection, prediction and early warning are prospected. • The research status of pest detection, prediction and early warning is summarized. • Sensors, systems and methods used in pest management are presented. • The applications in various agricultural scenarios are described in detail. • Challenges and future trends are reported.