Depi Ginting, Kurnianingsih, Ramzi Adriman, Kurnianingsih, Sunu Wibirama
Pest disturbances in swiftlet houses reduce edible bird's nest (EBN) production, a valuable commodity in Southeast Asia. Manual pest monitoring is often inefficient and disruptive to the sensitive environments of birds. However, to the best of our knowledge, no study has incorporated computer vision technology and the IoT for smart surveillance in swiftlet farming. To address this gap, we propose a YOLOv10-based detection system to identify small pests in real time under low-light conditions. The proposed system features a Python-based UI, a Region of Interest (ROI) function to improve detection focus, and IoT integration via a Telegram Bot that sends image-based alerts upon detection. The infrared CCTV cameras captured 2,011 images, which were augmented through rotation, resulting in 3,992 images. Six YOLOv10 model variants (n, s, m, b, l, and x) were evaluated. Based on our experimental results, the 'b' variant exhibited the best performance, with an mAP50 of 0.9936 and the lowest latency. Evaluation using a 50-minute video demonstrated accurate and rapid pest identification with cockroaches as the main pests. The experimental study showed the effectiveness of the system in monitoring swiftlet farming while reducing environmental disturbance to birds. © 2025 IEEE.
Universitas Syiah Kuala, Department of Electrical and Computer Engineering, Banda Aceh, Indonesia; Politeknik Negeri Semarang, Department of Electrical and Computer Engineering, Semarang, Indonesia; Universitas Gadjah Mada, Department of Electrical and Information Engineering, Yogyakarta, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock