<![CDATA[Bird pest attacks are one of the main causes of declining rice productivity in Deli Serdang Regency, especially during the grain ripening phase. This study develops the smartRicefield innovation, namely an automated bird pest repellent system based on the Internet of Things (IoT) and Artificial Intelligence (AI) using the YOLO11m image detection model. This study begins with collecting bird image datasets in rice fields collected from field image capture or dare sources, then labeled using the YOLO format, and wrapped with augmentation techniques to increase shape diversity. The YOLO11m model consisting of 125 image layers and 20,030,803 parameters with a complexity of 67.6 GFLOPs drilled for the next 100 epochs. The best model in the 86th epoch achieved 100% precision, 83.2% recall, mAP@0.5 of 86.3%, and mAP@0.5–0.95 of 69.3%. The Confusion Matrix Analysis showed good bird detection performance, but a high false positive rate in the background of the trigger image caused false triggers in the object testing. The system was tested in Deli Serdang rice fields with a detection latency of less than 1 second and an expulsion effectiveness of 90% at an effective distance of 10 meters. These results indicate that the integration of AI and IoT in Smart Ricefield is able to provide an effective real-time solution for bird pest mitigation, although improvements are still needed in dataset variations and expulsion mechanisms to increase the system's long-term resilience to all types of bird pests in the rice field environment.]]>
