Photogrammetry in Cultural Heritage: Experiments with a Cloud Based Service and an iPad

Analysis of Basilica Notre Dame

Figure 1.  Surface deviation analysis of Notre Dame Basilica in Old Montreal.  Software used in the project: Autodesk Recap 360 and BuildIT

 

Introduction

Can we document cultural heritage monuments with affordable tablets / smartphones and cloud-based photogrammetry services such as Autodesk Recap 360?  And how can the performance of these systems reach the accuracy and resolution levels of established devices?  What is the procedure for obtaining a 3D model of a monument by photos?  This blog post will try to present some of  the potential answers to these questions.

Notre Dame Basilica Scan and Analysis

Figure 2.  Notre Dame Basilica in Old Montreal: 3D Model from Autodesk Recap 360 (Left) and Surface Analysis result from BuildIT (Right)

As a preliminary experiment, we took photos of the Notre Dame Basilica in Old-Montreal.  Next, the photos were converted to 3D .stl models through Autodesk Recap 360.  The points were then analyzed with BuildIT software, to observe the effects of deviations and statistical distribution (Figure 2).

 

Cultural Preservation through Digital Models

Historical monuments play a crucial role in the preservation and propagation of a community’s culture and education; they define our cultural identity on many levels.  They must, therefore, be well documented and made available to the general public.  The Venice Charter states that in all works of preservation or excavation, there should always be precise documentation in the form of analytical and critical reports, illustrated with drawings and photographs.

Cyark Mount Rushmore

Figure 3.  Digital 3D Model of Mount Rushmore from Cyark project.

There are several projects around the world for the digital preservation of notable monuments, such as Cyark (Figure 3) , Reark and the Virtual World Heritage Laboratory at Indiana University.  Projects of this kind can prove to be very useful in case the heritage sites get damaged or destroyed.

 

Data Acquisition Techniques for Cultural Heritage Projects

Data Acquisition Techniques

Figure 4.  Categories of data acquisition techniques for cultural heritage projects

There are several categories of data acquisition for cultural heritage projects depending on the size of the object we are interested in (Figure 4):

Regional – Large scale projects: These mainly concern topographical surveys of large archeological or historical areas.  As an example, the images acquired by high-resolution satellites such as Ikonos (with a resolution of 1 m) can be used to map the detailed land changes in a city [Xiao]

Local – Medium scale projects: This kind of project mainly includes historical monument documentation.  The use of terrestrial laser scanners for the documentation of underground water cisterns in Istanbul [Temizer] can be given as an example of this category.

Object – Artefact – Small scale projects: The scanning of small historical artefacts.  An example of this is the use of handheld scanners to digitize small Egyptian statues.  The models were subsequently 3D printed and exhibited in place of the originals.

Of course, many projects will involve some combination of all of these categories.  The choice of the most appropriate technology depends on the object or area under study, the experience of the technical personnel, the available time and  the financial budget.

 

General Comparison of Laser Scanning and Photogrammetry Devices

Terrestrial Laser Scanner vs Photogrammetry

Figure 5.  Comparison of terrestrial laser scanning to photogrammetry

As an example of a terrestrial laser scanner, we shall present a Phase Shift laser scanner from Faro industry, the Faro Focus 3D.  The Phase Shift laser scanner belongs to the family of Time of Flight laser scanners (ToF).  This device calculates the spatial position of the points that define the surface of the scanned feature by evaluating the time needed for a pulse of laser to reach the object and be reflected back to the scanner.  It can capture up to a million points per second and has a defined accuracy of ±2 mm.

Photogrammetry is the stitching of photos to obtain 3D models based on common features between the photos.  We will use Autodesk Recap 360 for photogrammetric purposes in this study.  Autodesk Recap is an online service that can create 3D models from photos you submit.  Autodesk Recap 360 is a subscription-based application; for fees please refer to the related Autodesk website.

As shown in Figure 5, the laser scanners can give a great number of points with direct access to the data during measurement.  Photogrammetry devices offer a comparatively lesser number of points and we can only access the data once the all of the photos have been taken.  Undoubtedly, the primary motivation for photogrammetry would be the price of the devices, where we can use off-the-shelf digital cameras, tablets or smartphones.

 

Measurement Comparison of Laser Scanners and Photogrammetry Devices from the Literature

Haut Landau

Figure 6.  Courtyard of Haut Landau castle, associated point cloud and deviation analysis of points on extracted planes.  [Grussenmeyer]

The study of Grussenmeyer on Alsace Haut Landau castle compared terrestrial laser scanning with photogrammetric techniques.  They showed that the two techniques can be applied equivalently at least for flat and regularly shaped objects (Figure 6).  They also concluded that there was not a single recommendable technique and that all surveying techniques have limitations as well as complementary advantages.  In their example laser scanners could not capture the very low ditches or very high donjons, so photogrammetry had to be used.

Similarly, other researchers [Haddad, Bayram] compared measurements from terrestrial laser scanners and photogrammetry.  They found comparable results, again on flat and regular surfaces.

In most other studies [Al-Kheder, Baz, Rizzi], photogrammetric and laser scanner techniques were used in combination.  In these studies, photogrammetry has been mostly used as a complementary technique for either more detailed documentation (such as oblique angles of building facades) or texturing.

 

Workflow for Photogrammetry with Autodesk Recap 360

This experiment represents a simple example of the workflow for photogrammetry through Autodesk Recap 360.  As we wanted to push the envelope of the service and compare it with more professional conditions later on, we didn’t bother to have ideal conditions for accuracy and resolution: We used an iPad camera (instead of a DSLR) on a cloudy autumn day, with variable atmospheric conditions.

Our workflow can be summarized as follows (Figure 7)

Data Acquisition Procedure

Figure 7.  Photogrammetry workflow with related software websites: [http://www.apple.com/ca/ipad/] [https://recap360.autodesk.com/] [http://www.builditsoftware.com/]

The details of image capture and cleaning were explained in a previous post.

 

Surface Deviation Analysis with BuildIT Software

 Surface deviation analysis of Notre Dame Basilica

Figure 8.  Surface deviation analysis of Notre Dame Basilica in Old Montreal.  Flat surfaces selected for analysis.

We imported the point cloud into BuildIT.  Then, points in selected flat zones were best-fitted to planes of the church wall.  Afterwards, we carried out a surface deviation analysis in BuildIT.  To adjust the dimensions to nominal values, an approximate scaling was applied using the nominal dimensions of the basilica.

After the analysis, point distribution with respect to the best fit plane was analyzed (Figure 8).  The distribution was found to be different than a normal distribution.  This non-normality has also been investigated by other articles on mobile mapping systems [Toschi].  A potential source for this non-normality can simply be the filtering by Autodesk for ideal registration or mesh purposes, rather than a device related dispersion.

surface deviation analysis of Notre Dame Basilica with BuildIT

Figure 9.  Surface deviation analysis of Notre Dame Basilica in Old Montreal.  Most outliers are detected near the edges of the flat surfaces.

The analysis also shows most outliers are near the edges of the best fit planes for flat surfaces (Figure 9).  This effect mainly derives from the noise coming from the neighboring surfaces and shows the importance of outlier elimination in a preliminary analysis.

The number of points provided by Autodesk Recap is limited.  We only had 300K points, which is very small when compared with laser scanner data that would be obtained from such a large monument (Though the number of points would increase if we had used a DSLR camera instead of an iPad and took more photos).  We don’t have control of raw data since points are provided by the Autodesk application.  Further processing (such as point manipulation, filtering, outlier removal) is relatively restricted.

 

Conclusions

In this post, we investigated the possibilities of photogrammetry techniques with simple devices, an iPad camera, and an online photogrammetry service, Autodesk Recap 360.  The results were promising, but much remains to be done.

Given that we are at the preliminary stages of our research and want to explore the full capability of the product, in the discussed experiments we used photos taken in non-ideal conditions.  In future studies, we plan to experiment with calibrated DSLRs, terrestrial laser scanners and to perform a benchmarking study with different kinds of large structures.

A preliminary literature review shows that photogrammetric techniques can be a suitable alternative to terrestrial laser scanners, at least for flat and regularly shaped surfaces.  But more importantly, they can also be used in combination with laser scanning.  Their integrated use would allow an improvement to the resolution of 3D models, increased accuracy of objects and the addition of color characteristics.

Santa Maria in Trastevere, ceiling and apse
by Matthew Brennan
on Sketchfab

 

References:

Al-Kheder, S., Al-Shawabkeh, Y., Haala, N., 2009.  Developing a documentation system for desert palaces in Jordan using 3D laser scanning and digital photogrammetry.  Journal of Archaeological Science 36 (2), pp.  537-546.  doi: 10.1016/j.jas.2008.10.009  http://www.sciencedirect.com/science/article/pii/S0305440308002513

Baz, I., Kersten, Th., Büyüksalih, G., Jacobsen, K.  , 2008.  Documentation of Istanbul Historic Peninsula by Static and Mobile Terrestrial Laser Scanning.  International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol.  XXXVII, ISPRS Congress, Beijing, July, 3-11.  http://www.isprs.org/proceedings/XXXVII/congress/5_pdf/172.pdf

Rizzi, A., Baratti, G., Jimenez, B., Girardi, S., Remondino, F.: 3D recording for 2D delivering – The employment of 3D models for studies and analyses.  Int.  Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol.  38(5/W16), o n CD-ROM.  4th Int.  Workshop “3D-ARCH 2011: “Virtual Reconstruction and Visualization of Complex Architectures”, 2-5 March 2011, Trento, Italy.  http://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XXXVIII-5-W16/301/2011/isprsarchives-XXXVIII-5-W16-301-2011.pdf

Bayram B., Nemli G., Comparison Of Laser Scanning And Photogrammetry And Their Use For Digital Recording Of Cultural Monument Case Study: Byzantine Land Walls-Istanbul, ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume II-5/W3, 2015 25th International CIPA Symposium 2015, 31 August – 04 September 2015, Taipei, Taiwan, http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-5-W3/17/2015/isprsannals-II-5-W3-17-2015.pdf

Grussenmeyer, P., Landes, T., Voegle, T., Rongle, K., 2008.  Comparison methods of terrestrial laser scanning, photogrammetry and tacheometry data for recording o f cultural heritage buildings.  Inter.  Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol.  37(B5), pp.  213-218.  http://www.isprs.org/proceedings/xxxvii/congress/5_pdf/38.pdf

Haddad, N., Akasheh, T.  2005.  Documentation of archaeological sites and monuments: ancient theatres in Jerash.  In: Proceedings of CIPA XXth International Symposium, September 26–October 1, Torino, Italy, 350–355.  http://cipa.icomos.org/fileadmin/template/doc/TURIN/350.pdf

T. Temizer , G. Nemli , E. Ekizce , A. Ekizce , S. Demir , B. Bayram , Askin F.H A.  V. Cobanoglu , H. F. Yilmaz, 3D DOCUMENTATION OF A HISTORICAL MONUMENT USING TERRESTRIAL LASER SCANNING CASE STUDY: BYZANTINE WATER CISTERN, ISTANBUL, International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-5/W2, 2013 XXIV International CIPA Symposium, 2 – 6 September 2013, Strasbourg, France, p.623-628.  http://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XL-5-W2/623/2013/isprsarchives-XL-5-W2-623-2013.pdf

Toschi, I., Rodríguez-Gonzálvez, P., Remondino, F., Minto, S., Orlandini, S.  & Fuller, A.  (2015), Accuracy evaluation of a mobile mapping system with advanced statistical methods.  The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XL-5/W4, 245-253.  http://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XL-5-W4/245/2015/isprsarchives-XL-5-W4-245-2015.pdf

Yinghui Xiao; Qingming Zhan, “A review of remote sensing applications in urban planning and management in China,” in Urban Remote Sensing Event, 2009 Joint , vol., no., pp.1-5, 20-22 May 2009 doi: 10.1109/URS.2009.5137653,  http://www.researchgate.net/publication/224545459_A_review_of_remote_sensing_applications_in_urban_planning_and_management_in_China

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply