An integrated system for automated 3D visualization and monitoring of vehicles
- PDF / 3,493,029 Bytes
- 13 Pages / 595.224 x 790.955 pts Page_size
- 69 Downloads / 199 Views
ORIGINAL ARTICLE
An integrated system for automated 3D visualization and monitoring of vehicles Stella Bounareli1 · Ioannis Kleitsiotis1 · Lampros Leontaris1 · Nikolaos Dimitriou1 · Aggeliki Pilalitou2 · Nikolaos Valmantonis2 · Efthymios Pachos2 · Konstantinos Votis1 · Dimitrios Tzovaras1 Received: 21 August 2020 / Accepted: 23 September 2020 / Published online: 18 October 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Recent technological advances in computer vision and the advent of commercial RGB-D sensors have certainly boosted 3D modeling applications. This work presents an integrated system that enables the digitization of big objects, like vehicles, with low-cost RGB-D sensors. The implemented system can be used for visualization and monitoring of vehicles in large fleets that currently require a time-consuming manual inspection process. The main objective is to achieve an efficient consolidation of multiple views of a vehicle inside a moving frame, to acquire color and depth data and generate its 3D representation. The proposed integrated system denoises the acquired depth maps, aligns the produced point clouds captured in different time instances, and builds the 3D-reconstructed mesh. Finally, we apply a texture mapping algorithm to acquire realistic texture details and remove any visible seams. We evaluate all modules of the implemented system by performing several experiments with scanned vehicles. Keywords RGB-D sensors · Calibration · Point cloud registration · 3D reconstruction · Texture mapping
1 Introduction Generating 3D color models of real objects captured with RGB-D sensors is an important task in computer graphics and computer vision with several applications such as 3D visualization [1, 2], inspection [3, 4], and 3D model retrieval [5]. Recently, the advent of low-cost RGB-D sensors has led to the development of many 3D modeling applications [6–9]. Using such sensors, the surface of real objects can be easily captured and digitized. However, the quality of the captured data from low-cost RGB-D sensors sometimes can be insufficient for many applications. For this reason, many researchers have focused on improving the quality of the generated 3D models. 3D modeling systems that are able to produce realistic results can be widely used in many industrial fields. This paper presents a system that deals with
Lampros Leontaris
[email protected] 1
Centre for Research & Technology Hellas, Information Technologies Institute, Thessaloniki, Thermi, 57001, Greece
2
iKnowHow, Kifisias 116 Ave & Davaki 1 Str, Athens, GR, 11526, Greece
the digitization of vehicles for inspection and visualization purposes using commercial sensors. In an indicative overview of 3D modeling systems, authors in [10] deploy an integrated laser scanning system to generate high-resolution point clouds that are fed to a 3D-CNN network for fault diagnosis in microelectronics. In a follow-up work, a deep learning framework is proposed in [11] for simulation by predicting upcoming geometric variations u
Data Loading...
