3D Techniques and Tools for Preservation and Visualization of Cultural Heritage

Dimitrios Tzanakis Yannis Psaromiligkos Daune West

Contents
Introduction Research Problem Presentation of the state of the art Case study Conclusions

Introduction

Cultural heritage is priceless Cultural heritage preservation and documentation is essential

Problems
Cultural heritage objects are spread around the world Cultural heritage sites and monuments are always in jeopardy (pollution, rain, sun, wind, fire, earthquakes, war actions)

3D digitization and visualisation of cultural heritage benefits

Restoration study Model study Digital 3D models creation Model data base development of similar objects Virtual museums

Research Problem

What are the main techniques and tools in order to:

model and visualize cultural heritage objects digitally with the use of Information Technology according to the context of cultural heritage documentation and visualisation

Presentation of the state of the art

3D laser scanning techniques

Photogrammetry techniques Visual description of 3D objects

Main 3D laser scanning techniques

Time of flight of a laser pulse This methodology is also known as Light Detection And Ranging (LIDAR). Laser Triangulation

Time of flight of a laser pulse

or Light Detection And Ranging (LIDAR)

A laser pulse is emitted to the object

The distance between the signal sending device and the surface of the object is calculated from the required time between transmission and reception of the laser pulse.

Laser Triangulation

A CCD sensor using the high optical definition of a laser beam, that is projected on the under study object, and via triangulation equations, calculates the position of each point that is illuminated by the laser beam in the 3D space.

laser scanning techniques

(+) More reliable and more usable by non experts in many projects (-) Require many man hours (-) Expensive Suitable for large scale cultural heritage surveying projects => high accuracy products with significant efficiency.

Main photogrammetric techniques

Stereophotogrammetry Convergent or non-stereoscopic photogrammetry Structure from motion

Stereophotogrammetry
Axes

of camera should be parallel in two consequent shooting positions and vertical to the surface of the under study object Photogrammetric triangulation

Convergent photogrammetry
The camera axes converge toward the gravity centre of the object Marked Feature Points (a, b, c)

Structure from motion

Automatic calculation of the orientation from marker-less photos. Un-ordered photo sequences can be used for input

Photogrammetry techniques
(+)

Ideal for cultural heritage objects without complex shapes (+) Ideal when money, location and time constraints exist (+) Much more flexible in the case of multi dimensions projects than laser scanning techniques (-) Experienced operators

Visual description of 3D objects

Point Cloud Each point corresponds to a feature point and embeds 3D coordinates

Visual description of 3D objects

Triangular

mesh is used in most 3D imaging applications Provides much more visual information than the point cloud

Teapot wireframe

rendered teapot

Case study

Photogrammetric techniques selection evaluation:

More appropriate for the case study according to time, money and other constraints Essentially fulfil this case studys scope for visualisation and virtual reality as well as its requirements

Software tools

Arc3D Webservice MeshLab

PixMaker Pro

Equipments

A tripod

Two consumer photo cameras

The Cultural Heritage object

Saint Tryfonas Orthodox Church, Eleonas, Peloponnese

ARC 3D Webservice
The 3D models were created by the ARC 3D Webservice, developed by the VISICS research group of the Katholieke Universiteit Leuven in Belgium The uploading procedure required many attempts and time to achieve usable results Open source

MeshLab
Developed by Visual Computing Lab of the Institute of Information Science and Technologies of the Italian National Research Council CNR-ISTI in Italy 3D triangular mesh processing tool Provides functions for cleaning, editing, simplifying, remeshing, measuring, merging, visualising, converting, etc. Open source

MeshLab

(+) Can import meshes (models) in many formats and export the results in almost all known formats (-) Not so stable, has no undo function and needs great computer processing power (-) Significant training time (-) No complete manual (-) A lot of man hours to process the sub scenes meshes and align and merge them into the final model

Case study
The procedure

ARC 3D is not able to provide 3D data for model creation that encircles an object

In order to create a 360o model, one must divide the 360o scene into individual sub scenes

The upload tool enables the user to be Users connected to the Webservice and after can upload authentication to upload a selected sequence of unordered photos to Arc3D server. and markerless photos.

ARC 3D Webservice
Automatic procedure in the server side Parallel processes running on a cluster of Linux personal computers

The results: camera parameters, range-dense depth maps (i.e. point clouds), texture and quality maps for every photo

3D Webservice calculates the 3D data and camera parameters for each photo

Arc

MeshLab
The 3D models were post processed by MeshLabs Arc3D import interface. Main post processing functions: subsampling, cleaning, noise filtering, hole filling and merging.

MeshLab
The four 3D partial models were imported and cleaned in MeshLab They are all loaded into the same new MeshLab project and while they had originally positioned in different reference spaces

 they were rototranslated in order to be placed into the same single reference space and finally to be aligned and merged.

Snapshots of the VRML model

Three different model versions of reduced resolutions (decimated and compressed) Fast download time for users with slow connections rates

Resolution Full resolution 100% Size (uncompressed) 137,594 Kb

Size (compressed with gzip)

31,735

High resolution
Medium resolution Low resolution

64%
25% 13%

87,431 Kb
34,463 kb 17,977 kb

20,628
8,298 4,405

3D model high resolution (64%)

Case study - Vitrual tour section

PixMaker Pro

Virtual tour Use of Java

www.stvtour.comule.com

Case study conclusions

Photogrammetry techniques:
low cost medium training requirements medium post processing time sufficient accuracy models with high texture quality no need for laboratory conditions fast data acquisition and multi dimension objects compatibility

Case study conclusions

Structure from motion technique suitable for: virtual reality applications and visualisation purposes (although not highly accurate nor stable enough) provision of 3D models from un-ordered and marker-less photos

Final Conclusions
In many methodologies their evolution is proportional to IT evolution In 3D digitization techniques the storage, processing, archiving, management and visualisation of the acquired digital data can only be realized by IT utilization Modern IT technologies (e.g. Web, Virtual Reality) offer easy access to cultural heritage => The contribution of IT in cultural heritage documentation is crucial:

Final Conclusions
Nowadays there is no all-in-one cultural

heritage objects documentation solution The combination of different techniques in order to achieve the matching of their advantages, a hybrid technique, may lead to a complete solution

Final Conclusions
The most crucial limitation of photogrammetric techniques is the software If the development of IT and software continues its current pace, it will be possible in the near future for photogrammetry to overcome its basic limitations providing the panacea in cultural heritage documentation

The future of cultural heritage preservation and documentation belongs to

Photogrammetry !

Questions