Salzburg IAMG 2011example is the survey of the Cueva de los Cristales near Naica, Mexico (Canavese...

Mathematical Geosciences at the Crossroads of Theory and Practice

IAMG 2011Salzburg

1

Proposal:

IAMG 2011

Univ. Salzburg (Austria) September 5-9, 2011

R. Marschallinger, F. Zobl

CONFERENCE

September 5–9, 2011University of SalzburgAustria

© S

erge

j See

man

n - F

otol

ia.c

om

Geographic InformationScience Institute

www.oeaw.ac.at/GIScience

Sessions GIN1 - GIN5

Integration of geological base data into the municipal GIS of the town of Straubing (Germany) and time-related modeling

S. Beer, G. Lehrberger, K. Thuro, Germany

Modeling models: understanding of structure of geophysical knowledge by the event bush method

R. Carniel, C. Pshenichny, Z. Kharabrykh,

V. Shterkhun, P. Pascolo, Italy, Russia

Practicing an ontology spectrum for geological data interoperability

X. Ma, E. J. M. Carranza, Ch. Wu,

F. D. van der Meer, The Netherlands, China

Descending to misunderstanding in collaborative geoscientific research projects

C. Pshenichny, P. Diviacco, Russia, Italy

A model for spatial geoscience data and

its implementation

P. Gabriel, J. Gietzel, H. H. Le, H. Schaeben, Germany

General approach to information modeling of earthquake intensity at Oceanic Islands (by the example of Canary Archipelago)

C. Pshenichny, A. Garcia, M. Sanchez Jimenez, V. Anokhin, R. Carniel, V. Voitenko, R. Oritz, M. Llinares Garro, V. Shterkhun, O. Krutkina, V. Shezhko, J.-M. Llinares Marrero, A. Rezyapkin, Russia, Spain, Italy

Multidimensional aspects of GeoBIM data: new standards needed

F. Zobl, K. Chmelina, R. Faber, J. Kooijman, R. Marschallinger, J. Stoter, Austria, The Netherlands

Full-3D surveying of caves: A case study of Märchenhöhle (Austria)

A. Roncat, Y. Dublansky, Ch. Spötl, P. Dorninger, Austria

GIS-based structural analysis on the sphere: an approach for the kinematic interpretation of structural data

I. Görz, P. Hielscher, Germany

1

Full-3D Surveying of Caves: A Case Study of Mär-chenhöhle (Austria)

Andreas RONCAT1, Yuri DUBLYANSKY

2, Christoph SPÖTL2, Peter DORNINGER

1

1 Institute of Photogrammetry and Remote Sensing, Vienna University of Technology, Austria, [email protected], [email protected]

2 Institute of Geology and Paleontology, Leopold Franzens Universität Innsbruck, Austria, [email protected], [email protected]

Peer-reviewed IAMG 2011 publication

doi:10.5242/iamg.2011.0074

Abstract

This study presents the potential of terrestrial laser scanning for full-3D mapping of caves. The technique of laser scanning has become standard for medium and large-scale surveying of topo-graphic surfaces, but has so far only seldom been used in karst studies. On the example of the hypogene cave Märchenhöhle (Austria, cadastre #1742/17) we describe both the theoretical background and the actual work flow from the data capture to the derivation of a digital 3D model of a cave. The model ensures rigorous documentation and offers possibility to comple-ment field observations on cave morphology with subsequent, more-detailed, examination in the office. This makes morphogenetic interpretations more thorough and objective than was possible with conventional cave maps.

Andreas Roncat, Yuri Dublyansky, Christoph Spötl and Peter Dorninger

2

1 Introduction

Mapping of caves is usually done using instruments such as (electronic) distance meters, com-passes and tilt meters. Although computer support is available for data acquisition, processing and creating cave maps, there is still a lot of manual work associated with this task (cf. Plan and Xaver 2010). As a result, the quality of the final map heavily depends on the experience and the cartographic skills of individual surveyors.

In speleogenetic studies (that is, studies aimed at deciphering origin of caves and details of spe-cific processes which led to cave formation) besides gross cave morphology, meso- and micro-morphological features are of great importance (Dublyansky 2012). These smaller-scale features are typically not represented on the standard cave maps. For specialized morphogenetic maps, such small-scale features must be documented and interpreted on-site, and then represented on the final maps qualitatively (mostly, as symbols).

Usage of image-based photogrammetric surveying is limited in cave studies primarily because of the difficulty of achieving homogeneous light conditions.

Laser scanning is independent of light conditions and delivers high resolution 3D point clouds as input for both mapping and quantitative analysis of small-scale features. Its principles and usage in the different scientific fields are presented in Section 2. We further discuss application of the laser scanning technique to mapping of Märchenhöhle (Fairy tale cave), a hypogene cave in Sty-ria, Austria, in Section 3. The laser scanner data recorded in this cave by the authors in a field campaign in July 2010 are presented in Section 4. The subsequent section deals with the derived results, including full-3D models of Märchenhöhle’s main chambers. The article concludes with discussion and outlook in Section 6.

2 Laser Scanning

Laser Scanning is an active 3D measurement technique recording 3D point clouds in regular angular patterns. Instruments are available for both the airborne (known as airborne laser scan-ning, ALS) and the ground-based applications (terrestrial laser scanning, TLS). ALS has become standard for high-resolution topographic data acquisition during the last 15 years (Gomes-Pereira and Janssen 1999; Pfeifer and Mandlburger 2008). For limited areas, TLS has proven to be com-parably accurate and efficient for this task, too (Roncat et al. 2010). Besides the use in topo-graphic mapping, ALS is also in widespread use in other scientific fields, such as forestry (Naes-set 1997; Hollaus et al. 2007), 3D city modeling (Dorninger and Pfeifer 2008), monitoring of infrastructure (Melzer 2007) and hydrology (Mandlburger et al. 2009), just to name a few. In turn, TLS is widely applied for e.g. monitoring tunnels during the construction phase (Fekete et al. 2010; van Gosliga et al. 2006), cultural heritage documentation (Dorninger and Briese 2005) and inspection of industrial and technical facilities (Kopáčik and Wunderlich 2004).

A recent issue of Zeitschrift für Geomorphology summarizes geomorphological applications of (both terrestrial and airborne) laser scanning (Pfeifer et al. 2011), ranging from glacier surface classification and mass balance analysis to studies of fluvial geomorphology. So far, only a li-mited number of case studies reports application of laser scanning to mapping caves. One recent example is the survey of the Cueva de los Cristales near Naica, Mexico (Canavese et al. 2009).

Figure 1 depicts the scanning geometry and the resulting point cloud of a TLS scan (Kraus 2007).

Full-3D Surveying of Caves: A Case Study of Märchenhöhle (Austria)

3

Figure 1: Scan pattern in angular domain (left) and spatial structure of a point cloud recorded in a

single TLS scan (right).

The TLS scanners may employ three operating principles (Beraldin et al. 2010):

Time-of-flight measurement: A short laser pulse is transmitted and the arrival time of its echo is recorded. Multiplication with the speed of light leads to the range of the re-flecting surface. This kind of scanners normally operates in ranges from several meters up to some hundred meters.

Phase measurement: These scanners emit a modulated laser signal, using either conti-nuous wave (CW), amplitude modulation (AM) or frequency modulation (FM). Correla-tion of the return signal with a copy of the emitted signal generated in the instrument leads to the range of the reflecting surface. Phase ambiguity limits the maximum record-able distance to normally approx. 80 − 100 m.

Triangulation-based measurement: The projection of the laser ray on the target surface is captured by a scanner-internal camera. Since the relative position and attitude of the la-ser ray and the camera are known, this allows for the calculation of the three-dimensional coordinates of the projected point by spatial intersection. These scanners operate at dis-tances from some decimeters up to some meters.

The instrument used in this study was a phase-measurement scanner (see Section 4).

3 The Study Site – Märchenhöhle

Märchenhöhle (cadastre # 1742/17) is located at the southern margin of the Northern Calcareous Alps, near Eisenerz (Styria, Austria; see Figure 2 (Wikimedia 2011; BEV 2011)). This 135 m-long cave occurs near the crest of a limestone ridge at approx. 1500 m a.s.l., some 900 m above the valley floor. The cave consists of several near-isometric chambers separated by relatively narrow openings. The lower parts of cave walls are coated with a mammillary calcite crust (cave clouds; see Figure 3). Above this crust the internal surface of the cave exhibits a wealth of pecu-liar smaller-scale corrosional features (e.g., funnel-shaped chambers with flat roofs, cupolas, cup-shaped pockets on inclined walls, etc.).


4

Figure 2: Location of Märchenhöhle.

Figure 3: Inside theMärchenhöhle. Photographs by Harald Auer.

4 Data Sets

The scanning campaign took place on July 22, 2010 using a z+f Imager 5006i terrestrial laser scanner operating according to the phase-measurement principle (Zoller+Fröhlich 2011). This instrument is capable of recording max. 800 million points per scan (40,000 lines x 20,000 points per line, scan mode “Ultra High”) within a scan time of 27 minutes. The maximal range to be recorded is approx. 80 m. The maximum resolution results in a point spacing of 1.6 mm at 10 m distance. Table 1 gives the details to the single scans recorded in the scanning campaign.


5

Table 1: Overview of the 11 recorded scans.

Scan ID Scan Mode Number of recorded points Acquisition time Local Pos. [m]

1 Super High 200,000,000 6 min 44 sec (0.000, 0.000, 0.000)

2 Super High 200,000,000 6 min 44 sec (8.691, 2.736,-1.300)



5 Super High 200,000,000 6 min 44 sec (12.310,14.367,-2.399)

6 High 50,000,000 3 min 22 sec (4.523,12.389,-0.710)

7 High 50,000,000 3 min 22 sec (2.061, 12.059, 0.809)

8 High 50,000,000 3 min 22 sec (9.522, 13.447,-1.438)


10 High 50,000,000 3 min 22 sec (16.678, 15.971,-2.834)


Sum 1,600,000,000 1h 00 min 36 sec

Since the data were acquired from different positions, accurate determination of the relative posi-tions and attitudes of the individual point clouds was necessary. This is known as “relative orien-tation” in photogrammetry and geodesy (Kraus 2007) and as “co-registration” in the field of computer vision (Shapiro and Stockman 2001). In our case, the relative orientation was estab-lished in a two-step approach: the coarse relative orientation was determined using coded targets (as visible in Figure 4) using the software delivered by the scanner’s manufacturer. It was refined using an ICP (iterative closest point (Besl and McKay 1992)) algorithm implemented in the Geomagic Studio software (Geomagic 2011). The first step uses some distinct points in two scans with known point-to-point correspondence whereas the latter makes use of the whole point clouds. However, ICP needs good initial values for the relative orientation to perform adequate-ly.

It has to be mentioned that our study did not require an absolute georeference (i.e. position and attitude of the overall point cloud w.r.t. a superior coordinate system such as national coordinate systems or WGS84). The local coordinate system was defined by taking the first scan position as the origin, defining the local plumb line (z axis) using the tilt sensor of the scanner and taking the scanner’s horizontal 0° line as the y axis. A cartesian right-handed coordinate system was uni-quely determined in this way.


6

Figure 4: Coded targets placed on top of rocks. These targets were used for determining the rela-

tive orientation of the scans. Photograph by Harald Auer.

After the relative orientation, the point cloud underwent subsequent processing steps such as filtering, homogenization and correction of systematic errors (cf. Dorninger et al. 2008; Molnár et al. 2009; Nothegger and Dorninger 2007). These steps were necessary to reduce the amount of data without losing significant details to enable the final 3D modeling of the cave (see Section 5) in reasonable processing time.

5 Results

Co-registration resulted in overall 3D point cloud covering the largest part of the cave. A floor plan of this point cloud was compared to a map of the cave produced by “traditional” surveying technique in 1971. Although absolute georeference was not available for both maps, they show a good coincidence (see Figure 5).

As a next step, horizontal profiles were derived from this point cloud. Each profile consists of the points located at a specific height ±5 cm. Colour coding of the single points according to their local height allows for the detection of vertical structures (Figure 6). The profiles are avail-able as GeoTIFF images and can thus be used for digitization in CAD software packages.

The core result of this study was the derivation of a triangulated 3D model of the whole area of interest which is depicted in Figure 7. Whereas the original point cloud consisted of more than 109 points, the resulting 3D model contains approx. 12 million triangles. Although the point cloud was greatly reduced, sophisticated filtering allowed to preserve also small features such as minor fractures and inclined pocketed walls.


7

Figure 5: Overlay of a map of Märchenhöhle (black and white) and a floor projection of the rec-

orded point cloud (blue). The two map images were manually adjusted, but show nevertheless good coincidence.

Figure 6: Horizontal profiles of the point clouds in the local heights of −1.40 m (left) and −0.60 m (right). Blue corresponds to the local height minus 5 cm, black to the local height and red to the

local height plus 5 cm.


8

Figure 7: Visualization of the triangulated 3D model of Märchenhöhle. Note complex, compartmen-

talized gross morphology of the cave.

Figure 8 shows a visualization of the 3D reconstruction of the cave’s central chamber (viewed from outside). One shortcoming of the current results is the missing intensity and colour infor-mation. Although intensity information was recorded by the scanner for each 3D point, the inten-sity values ought to be radiometrically calibrated before further use. Radiometric calibration will be a future step in the processing of the presented data sets. Photographs providing colour infor-mation were not taken area-wide because of the light conditions.

6 Discussion and Outlook

A hypogene cave in Austria was used as a test-ground for developing measurement techniques and subsequent data processing schemes for surveying caves in full-3D. The outcomes, be it point-based or based on a 3D triangulation, allows for quantitative analysis of the cave’s general shape and also small-scale features. In addition to the geometric data captured with laser scan-ning, intensity information has potential of increasing the information content of the derived models.


9

Figure 8: Visualization of the triangulated 3D model of one of the chambers of Märchenhöhle and

explanatory drawings. Note the complex morphology of the chamber and a wealth of the small-scale features recognizable on the model surface.

Acknowledgements

The authors are thankful to Harald Auer (Steiermärkische Berg- und Naturwacht) and Reinhard Demberger for their great field support in the scanning campaign. Andreas Roncat has been sup-ported by a Karl Neumaier PhD scholarship. Work of Yuri Dublyansky has been supported by the Austrian Science Fund (FWF, grant 213660).


10

References

BERALDIN, J.-A., BLAIS, F. and LOHR, U. (2010). Airborne and Terrestrial Laser Scanning, Chapter Laser Scanning Technology, pp. 1–42. Whittles Publishing.

BESL, P. and MCKAY, N. (1992). A method for registration of 3D shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence 14(2), 239–256.

BEV (2011). AMAP Austria Homepage. http://www.austrianmap.at., accessed: April 2011

CANAVESE, E. P., TEDESCHI, R. and FORTI, P. (2009). The caves of Naica – laser scanning in extreme underground environments. The American Surveyor 6(2), 3–10.

DORNINGER, P. and BRIESE, C. (2005). Advanced geometric modelling of historical rooms. In XX Symp. of CIPA, ICOMOS & ISPRS Committee on Documentation of Cultural Heritage, Turin, Italy.

DORNINGER, P., NOTHEGGER, C., PFEIFER, N. and MÓLNAR, G. (2008). On-the-job detection and correction of systematic cyclic distance measurement of terrestrial laser scanners. Journal of Applied Geodesy 2(4), 191–204.

DORNINGER, P. and PFEIFER, N. (2008). A comprehensive automated 3D approach for build-ing extraction, reconstruction, and regularization from airborne laser scanning point clouds. Sensors 8(11), 7323–7343.

DUBLYANSKY, Y. (2012). Karstification by geothermal waters. In SHRODER, J. J. and FRUMKIN, A (Eds.), Treatise on Geomorphology, Volume 6. San Diego, CA: Academic Press.

FEKETE, S., DIEDERICHS, M. and LATO, M. (2010). Geotechnical and operational applications for 3-dimensional laser scanning in drill and blast tunnels. Tunnelling and Underground Space Technology 25(5), 614–628.

GEOMAGIC (2011). http://www.geomagic.com. Homepage of Raindrop Geomagic, accessed: April 2011.

GOMES-PEREIRA, L. and JANSSEN, L. (1999). Suitability of laser data for DTM generation: A case study in the context of road planning and design. ISPRS Journal of Photogrammetry and Remote Sensing 54, 244–253.

HOLLAUS, M., WAGNER, W., MAIER, B. and SCHADAUER, K. (2007). Airborne laser scanning of forest stem volume in a mountainous environment. Sensors 7(8), 1559–1577.

KOPÁČIK, A. and WUNDERLICH, T.A. (2004). Usage of laser scanning systems at hydro-technical structures. In FIG Working Week, Athens, Greece.

KRAUS, K. (2007). Photogrammetry – Geometry from Images and Laser Scans (2 ed.). de Gruyter.

MANDLBURGER, G., HAUER, C., HÖFLE, B., HABERSACK, H. and PFEIFER, N. (2009). Optimi-sation of LiDAR derived terrain models for river flow modelling. Hydrology and Earth Sys-tem Sciences 13(8), 1453–1466.

MELZER, T. (2007). The mean shift algorithm for clustering point clouds. Journal of Applied Geodesy 1, 159–170.

MOLNÁR, G., PFEIFER, N., RESSL, C., DORNINGER, P. and NOTHEGGER, C. (2009). Range cali-bration of terrestrial laser scanners with piecewise linear functions. Photogrammetrie, Fernerkundung, Geoinformation 2009(1), 563–576.


11

NAESSET, E. (1997). Determination of mean tree height of forest stands using airborne laser scanner data. ISPRS Journal of Photogrammetry and Remote Sensing 52(2), 49–56.

NOTHEGGER, C. and DORNINGER, P. (2007). Automated modeling of surface detail from point clouds of historical objects. In 21st CIPA Symposium, International Archives of the Photo-grammetry, Remote Sensing and Spatial Information Sciences 36 (Part 5/C53), Athens, Greece, pp. 538–543.

PFEIFER, N. and MANDLBURGER, G. (2008). Topographic Laser Ranging and Scanning: Prin-ciples and Processing, Chapter Filtering and DTM Generation. CRC Press.

PFEIFER, N., RONCAT, A., STÖTTER, J. and BECHT, M. (Eds.) (2011). Laser Scanning Applica-tions in Geomorphology, Volume 55 (Supplementary Issue 2) of Zeitschrift für Geomorphol-ogy.

PLAN, L. and XAVER, A. (2010). Geomorphologische Untersuchung und genetische Interpreta-tion der Dachstein-Mammuthöhle (Österreich). Die Höhle 61(1–4), 18–38.

RONCAT, A., DORNINGER, P., MOLNÁR, G., SZÉKELY, B., ZÁMOLYI, A., MELZER, T., PFEIFER, N. and DREXEL, P. (2010). Influences of the Acquisition Geometry of different Lidar Tech-niques in High-Resolution Outlining of microtopographic Landforms. In Fachtagung Com-puterorientierte Geologie - COGeo 2010. doi:10.5242/cogeo.2010.0008

SHAPIRO, L. G. and STOCKMAN, G. C. (2001). Computer Vision. Prentice Hall.

VAN GOSLIGA, R., LINDENBERGH, R. C. and PFEIFER, N. (2006). Deformation analysis of a bored tunnel by means of terrestrial laserscanning. In International Archives of the Photo-grammetry, Remote Sensing and Spatial Information Sciences 36 (Part 5), Dresden, Germany, pp. 167–172.

WIKIMEDIA (2011). Location Map Austria. http://upload.wikimedia.org/wikipedia/commons/6/61/Austria_location_map.svg, accessed: April 2011.

ZOLLER+FRÖHLICH (2011). http://www.zf-laser.com. Homepage of the company Zoller + Fröhlich GmbH, accessed: April 2011.

Salzburg IAMG 2011example is the survey of the Cueva de los Cristales near Naica, Mexico (Canavese...

Documents

Transcript of Salzburg IAMG 2011example is the survey of the Cueva de los Cristales near Naica, Mexico (Canavese...