Skip to content

TUM - 3D Scanning and Motion Capture class final project: Voxel Carving of 3D Surfaces with AR Markers for Pose Estimation

Notifications You must be signed in to change notification settings

canberkanar/AR_Marker-Voxel-Carving

Repository files navigation

Voxel Carving of 3D Surfaces using AR Markers for Pose Estimation

TUM - 3D Scanning and Motion Capture class final project.

Voxel Carving

"Voxel carving is a technique to reconstruct voxels of a 3D surface from multiple images taken from different angles." [4] [6]

Abstract

In the recent years, Augmented Reality (AR) has gained increasing interest in researches, especially in the field of 3D modelling, and is used for different applicational purposes. In this project, AR will be applied to recognize markers [1] and the results of a 3D reconstruction using voxel carving will be visualized and placed on the markers [5]. Voxel carving is a technique to reconstruct voxels of a 3D surface from multiple images taken from different angles [4] [6]. This is achieved by separating the fore- and background in the images and then shaping the 3D mesh through the object silhouettes from different camera angles, a process called shape-from-silhouette [5].

Prerequisites

  • Image Dataset containing multiple images of the 3D Surface from different points of view
  • Camera Calibration
  • AR Marker Detection
  • Pose Estimation
  • Background Segmentation
  • Voxel Carving
  • Final Rendering

Requirements

  • OpenCV Library for Camera Calibration
  • ArUco Library for Camera Pose Estimation
  • VTK Library for rendering the final reconstruction

Reconstruction Example

References

About

TUM - 3D Scanning and Motion Capture class final project: Voxel Carving of 3D Surfaces with AR Markers for Pose Estimation

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •