%0 Thesis
%9 Skripsi
%A Thoriq Firdaus Arifin, NIM.: 21106050026
%B FAKULTAS SAINS DAN TEKNOLOGI
%D 2025
%F digilib:74489
%I UIN SUNAN KALIJAGA YOGYAKARTA
%K Computer Vision, Internet of Things, Deteksi Objek, Klasifikasi Gambar, Raspberry Pi, Systemd, Pertanian Presisi
%P 191
%T PERANCANGAN SISTEM PENGUSIRAN HAMA BURUNG DAN DETEKSI PENYAKIT PADI BERBASIS IOT DAN KECERDASAN BUATAN
%U https://digilib.uin-suka.ac.id/id/eprint/74489/
%X Declining rice productivity due to bird pests and plant diseases poses a serious challenge to food security, with conventional solutions often limited in range, accuracy, and large-scale analytical capabilities. This research aims to design and build an integrated and autonomous system by combining Internet of Things (IoT) and Computer Vision technologies for real-time agricultural monitoring and mitigation. The research methodology employed the prototyping development method on an edge computing device (Raspberry Pi 5 with a Hailo-8L AI accelerator). Three artificial intelligence models were developed: (1) a YOLOv8s object detection model for bird pest repellent, (2) an EfficientNetV2B2 classification model for leaf disease diagnosis, and (3) an xresnet34 classification model for analyzing field conditions from a top-down perspective. All software services, including REST APIs and the user interface, were orchestrated using systemd to ensure robust and scheduled operations. The results show that the prototype system was successfully implemented and integrated. The bird detection model achieved high performance with a mAP@50 of 97.6%, and the system effectively triggered actuator responses. The system also proved reliable in disease detection and could automatically recover from failures thanks to systemd service management, demonstrating its feasibility as a precision agriculture solution.
%Z Dr. Shofwatul Uyun, S.T., M.Kom.