Autonomous Visual Pest Detection System with ESP32-CAM, Zonal Classification, and Notification via Telegram for Agriculture

Tocto D.C.G.; Cari D.N.A.; Jesús Talavera S. · SSRG International Journal of Electrical and Electronics Engineering (2026) · DOI: 10.14445/23488379/ijeee-v13i1p107

Resumen

This article presents the design and implementation of an autonomous system for visual pest detection using an ESP32-CAM. Electronic modifications are also made to a drone to enable this system, ensuring that it is fully automatic, delimiting the drone's monitoring path, including wireless communication and artificial vision for pest assessment using captured images. The system captures images of the crop at monitoring points defined by the zonal classification of areas where a subsequent harvest will take place, generating a heat map of infestation points. Upon detection of the presence of possible pests, the system will save monitoring images to a microSD card when no network is available and will send captures of the assessment with a critical percentage of the pest when it has a Wi-Fi connection from the ESP32, as well as notifications via TELEGRAM and the location of the critical zonal classification point using a GPS module. This study focuses on corn cultivation, as it is a widely harvested product in areas of Peru such as Arequipa, using low-cost components, providing farmers with real-time alerts to improve response and efficiency by 70% in agronomic decision-making, improving productivity by 21% of total production, and reducing pesticide use. It is adaptable to any environment in the current agricultural industry. It can also be used for other crops such as grapes, avocados, asparagus, rice, etc.

Instituciones: Universidad Nacional de San Agustín de Arequipa