INTEGRASI MODEL DETEKSI OBJEK YOLOV3 DAN IOT DALAM PERANCANGAN ALAT PENGUSIR HAMA BURUNG

Indonesia, as one of the largest agrarian countries in the world, faces many challenges in the agricultural sector, one of which is the problem of bird pests that often damage rice crops. Birds such as sparrows, house sparrows, and weaverbirds can reduce crop yields by up to 80%, which is highly detrimental to farmers. This issue needs to be addressed urgently, especially with the development of technology that can be used as an alternative solution. This research aims to develop a bird repellent tool that utilizes Artificial Intelligence (AI) and the Internet of Things (IoT), integrated with each other. This tool uses the YOLOv3-640 object detection model, which has been trained to detect birds with an accuracy performance calculated using the mean Average Precision (mAP), yielding a value of 76.28% after 8000 training iterations. Additionally, the tool is designed to operate automatically, where the system can detect birds, repel them, and store detection data in a monitoring system accessible by farmers through a dashboard. This dashboard provides information on daily, weekly, monthly, and annual detection counts, as well as the history of detection images. This research not only offers a solution to the existing problem but also opens opportunities for further development in integrating AI and IoT technologies in the agricultural sector.
Keywords: Object Detection, Artificial Intelligence (AI), Internet of Things (IoT)