IMPLEMENTASI METODE PHOTOGRAMMETRY DALAM REKONSTRUKSI MODEL 3D UNTUK OBJEK GAME “BENCANA: NATURE'S TRIALS”

Innaka Dylee, - (2025) IMPLEMENTASI METODE PHOTOGRAMMETRY DALAM REKONSTRUKSI MODEL 3D UNTUK OBJEK GAME “BENCANA: NATURE'S TRIALS”. S1 thesis, Universitas Pendidikan Indonesia.

Abstract

Photogrammetry merupakan teknik rekonstruksi objek tiga dimensi (3D) dari kumpulan gambar dua dimensi (2D) yang diambil dari berbagai sudut pandang. Pada konteks pembuatan aset game edukasi kebencanaan berbasis VR, penggunaan metode photogrammetry masih kerap muncul masalah berupa akurasi dimensi yang belum memadai dan kestabilan rekonstruksi yang rapuh pada pipeline bawaan, terutama pada objek bertekstur rendah sehingga terjadi perbedaan skala, kehilangan detail, dan kualitas visual yang rendah dapat mengganggu interaksi di VR. Penelitian ini bertujuan untuk (1) menemukan konfigurasi photogrammetry yang efektif dalam merekonstruksi model 3D untuk aset game edukatif berbasis Virtual Reality (VR) bertema kebencanaan, serta (2) mengevaluasi tingkat akurasi model menggunakan parameter evaluasi Root Mean Square Error (RMSE) dengan pengukuran dimensi di Blender. Metode penelitian dilakukan dengan mengambil 60 gambar objek menggunakan metode Moving Smartphone Capture (MSC) yang dibagi menjadi lima layer pengambilan sudut. Proses rekonstruksi dilakukan menggunakan Meshroom dengan dua konfigurasi: default pipeline dan customize pipeline. Pada customize pipeline, parameter feature extraction diatur ke describer density dan quality tinggi, serta ditambahkan CCTag sebagai marker pada tahap feature extraction dan SfMTransform untuk kalibrasi skala. Hasil evaluasi menunjukkan bahwa model tanpa CCTag scaling memiliki nilai RMSE sebesar 2,6623 cm, sedangkan model dengan CCTag scaling menghasilkan RMSE sebesar 0.0339 cm, yang berarti terjadi pengurangan error sebesar 98,87% atau peningkatan presisi sekitar 88 kali. Temuan ini menunjukkan bahwa penggunaan customize pipeline dengan CCTag scaling mampu menghasilkan model 3D yang realistis dan presisi, sehingga layak digunakan sebagai aset game edukatif bertema kebencanaan yang membutuhkan ketelitian dimensi tinggi. ------------ Photogrammetry is a technique for reconstructing three-dimensional (3D) objects from sets of two-dimensional (2D) images captured from multiple viewpoints. In the context of creating disaster-education VR game assets, photogrammetry often suffers from insufficient dimensional accuracy and fragile reconstruction stability in out-of-the-box pipelines, particularly on low-texture objects, leading to scale inconsistencies, detail loss, and degraded visual quality that can hinder interaction in VR. This study aims to (1) determine an effective photogrammetry configuration for reconstructing 3D models intended for disaster-themed educational Virtual Reality (VR) games, and (2) evaluate the accuracy of the reconstructed models using the Root Mean Square Error (RMSE) evaluation metric with dimensional measurements conducted in Blender. The research method involved capturing 60 images of the object using the Moving Smartphone Capture (MSC) technique, divided into five layers of different shooting angles. The reconstruction process was performed in Meshroom using two configurations: default pipeline and customize pipeline. In the customize pipeline, the feature extraction parameters were set to high describer density and quality, and CCTag markers were added during the feature extraction and SfMTransform stages for scale calibration. The evaluation results showed that the model without CCTag scaling achieved an RMSE of 2,6623 cm, whereas the model with CCTag scaling reached an RMSE of 0.0339 cm, indicating a 98.87% error reduction or approximately 88 times higher. These findings demonstrate that using a customize pipeline with CCTag scaling can produce realistic and highly precise 3D models, making it suitable for educational disaster-themed VR game assets that require high dimensional accuracy.

	Text S_RPL_2102195_Title.pdf Download (1MB)
	Text S_RPL_2102195_Chapter1.pdf Download (126kB)
	Text S_RPL_2102195_Chapter2.pdf Restricted to Staf Perpustakaan Download (554kB) | Request a copy
	Text S_RPL_2102195_Chapter3.pdf Download (520kB)
	Text S_RPL_2102195_Chapter4.pdf Restricted to Staf Perpustakaan Download (1MB) | Request a copy
	Text S_RPL_2102195_Chapter5.pdf Download (42kB)
	Text S_RPL_2102195_Appendix.pdf Restricted to Staf Perpustakaan Download (2MB) | Request a copy

Item Type:	Thesis (S1)
Additional Information:	https://scholar.google.com/citations?hl=id&authuser=1&user=cEDdpikAAAAJ ID SINTA Dosen Pembimbing: Hendriyana: 6658557 Raditya Muhammad: 6682222
Uncontrolled Keywords:	Photogrammetry, Meshroom, CCTag, RMSE, Virtual Reality
Subjects:	L Education > L Education (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software
Divisions:	UPI Kampus cibiru > S1 Rekayasa Perangkaat Lunak
Depositing User:	Innaka Dylee
Date Deposited:	11 Sep 2025 08:40
Last Modified:	11 Sep 2025 08:40
URI:	http://repository.upi.edu/id/eprint/137517

Actions (login required)

View Item