IMPLEMENTASI INTEGRASI YOLOV11 DAN INTERSECTION-BASED METHOD PADA SMART PARKING SYSTEM UNTUK ESTIMASI KARAKTERISTIK PARKIR

Muhammad Kamal Robbani, - and Yudi Wibisono, - and Eddy Prasetyo Nugroho, - (2025) IMPLEMENTASI INTEGRASI YOLOV11 DAN INTERSECTION-BASED METHOD PADA SMART PARKING SYSTEM UNTUK ESTIMASI KARAKTERISTIK PARKIR. S1 thesis, Universitas Pendidikan Indonesia.

Abstract

Permasalahan keterbatasan lahan parkir di tengah pertumbuhan kendaraan menjadi tantangan serius di kota besar. Penelitian ini mengembangkan sistem smart parking berbasis computer vision untuk mengestimasi karakteristik parkir secara otomatis. Model YOLOv11 diimplementasikan untuk mendeteksi mobil pada rekaman CCTV, lalu hasil bounding box diterapkan pada pendekatan intersection-based method untuk mengklasifikasikan status masing-masing spot parkir. Akurasi klasifikasi ditingkatkan dengan menerapkan object tracking menggunakan BoT-SORT, perhitungan jarak centroid, serta manajemen transisi status parkir. Sistem ini mengekstraksi enam karakteristik parkir, yaitu parking accumulation, parking index, parking volume, parking duration, parking turnover, dan dynamic parking capacity yang selanjutnya disimpan ke database PostgreSQL. Hasil evaluasi menunjukkan mAP@50-95 sebesar 87,3% setelah fine-tuning, serta akurasi estimasi karakteristik parkir dengan R² 0,989 (terang) dan 0,91 (gelap). Selain itu, tactical dashboard bernama ParkSight berbasis Power BI dikembangkan untuk memvisualisasikan hasil, sehingga dapat membantu manajer parkir memahami pola parkir dan mendukung keputusan berbasis data secara efisien. The issue of limited parking space amid rapid vehicle growth has become a serious challenge in major cities. This study develops a smart parking system based on computer vision to automatically estimate parking characteristics. The YOLOv11 model is implemented to detect cars from CCTV footage, and the resulting bounding boxes are applied using an intersection-based method to classify the status of each parking spot. Classification accuracy is enhanced through the application of BoT-SORT object tracking, centroid distance calculation, and transition status management. The system extracts six parking characteristics—parking accumulation, parking index, parking volume, parking duration, parking turnover, and dynamic parking capacity—which are then stored in a PostgreSQL database. Evaluation results show an mAP@50-95 of 87.3% after fine-tuning, and the accuracy of parking characteristic estimation achieved R² values of 0.989 (daylight) and 0.91 (low-light). Furthermore, a tactical dashboard named ParkSight was developed using Power BI to visualize the results, enabling parking managers to better understand parking patterns and support data-driven decision-making more efficiently.

	Text S_KOM_2102313_Title.pdf Download (717kB)
	Text S_KOM_2102313_Chapter1.pdf Download (2MB)
	Text S_KOM_2102313_Chapter2.pdf Restricted to Staf Perpustakaan Download (16MB)
	Text S_KOM_2102313_Chapter3.pdf Download (245kB)
	Text S_KOM_2102313_Chapter4.pdf Restricted to Staf Perpustakaan Download (21MB)
	Text S_KOM_2102313_Chapter5.pdf Download (161kB)

Item Type:	Thesis (S1)
Additional Information:	https://scholar.google.com/citations?hl=en&user=bupqrnUAAAAJ ID SINTA Dosen Pembimbing: Yudi Wibisono: 260167 Eddy Prasetyo Nugroho: 5990993
Uncontrolled Keywords:	Computer Vision, Object Detection, YOLOv11, Manajemen Parkir, Tactical Dashboard. Computer Vision, Object Detection, YOLOv11, Parking Management, Tactical Dashboard.
Subjects:	Q Science > QA Mathematics T Technology > TA Engineering (General). Civil engineering (General) T Technology > TE Highway engineering. Roads and pavements
Divisions:	Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Ilmu Komputer
Depositing User:	Muhammad Kamal Robbani
Date Deposited:	08 Sep 2025 10:49
Last Modified:	08 Sep 2025 10:49
URI:	http://repository.upi.edu/id/eprint/138074

Actions (login required)

View Item