<![CDATA[Abstrak - Pertumbuhan jumlah kendaraan dan terbatasnya kapasitas area parkir menuntut adanya sistem monitoring yang lebih efisien dan mampu memberikan informasi ketersediaan slot secara real-time. Penelitian ini merancang dan membangun Smart Parking System berbasis Internet of Things (IoT) dan kecerdasan buatan (AI) dengan memanfaatkan modul ESP32-CAM sebagai sensor visual dan model YOLOv8 untuk mendeteksi kendaraan secara otomatis. Sistem diimplementasikan menggunakan arsitektur client–server, di mana ESP32-CAM mengirimkan citra area parkir ke backend Python melalui protokol HTTP POST untuk diproses menggunakan model AI. Hasil deteksi kemudian disimpan pada database MySQL dan ditampilkan melalui dashboard web sebagai monitoring real-time, lengkap dengan indikator status slot parkir serta live streaming kamera. Pengujian dilakukan melalui 50 skenario berbeda, mencakup single-object, multi-object, variasi pencahayaan, dan edge-case. Sistem mencapai akurasi deteksi kendaraan sebesar 90% dan akurasi penentuan status slot sebesar 100%, dengan latensi pemrosesan 0,8–1,2 detik. Hasil tersebut menunjukkan bahwa integrasi IoT dan AI mampu menghasilkan solusi parkir cerdas yang efektif, responsif, serta potensial untuk dikembangkan pada skala implementasi yang lebih luas.Kata kunci : Parkir Cerdas; Internet of Things; ESP32-CAM; YOLOv8; Deteksi Slot Parkir; Pemantauan Real-Time; Penglihatan Komputer; Abstract - The growth in the number of vehicles and limited parking space capacity requires a more efficient monitoring system that can provide real-time information on slot availability. This research designs and builds a Smart Parking System based on the Internet of Things (IoT) and artificial intelligence (AI) by utilizing the ESP32-CAM module as a visual sensor and the YOLOv8 model to automatically detect vehicles. The system is implemented using a client–server architecture, where ESP32-CAM sends images of the parking area to the Python backend via the HTTP POST protocol to be processed using the AI model. The detection results are then stored in a MySQL database and displayed via a web dashboard as real-time monitoring, complete with parking slot status indicators and live camera streaming. Testing was conducted through 50 different scenarios, including single-object, multi-object, lighting variations, and edge cases. The system achieved 90% vehicle detection accuracy and 100% parking slot status determination accuracy, with a processing latency of 0.8–1.2 seconds. These results demonstrate that the integration of IoT and AI can produce an effective, responsive smart parking solution with the potential for development on a wider implementation scale.Keywords: Smart Parking; Internet of Things; ESP32-CAM; YOLOv8; Deteksi Slot Parkir; Monitoring Real-Time; Computer Vision;]]>
