Content uploaded by Tomás Ramón Herrero-Tejedor
Author content
All content in this area was uploaded by Tomás Ramón Herrero-Tejedor on Jul 10, 2020
Content may be subject to copyright.
Characterization of underground cellars
in the Duero Basin by GNSS, LIDAR and
GPR techniques
Miguel Á. Conejo-Martín1, Tomás R. Herrero-Tejedor1, Enrique Pérez-Martín1, Javier Lapazaran-Izargain2,
Jaime Otero-García2, Juan F. Prieto-Morín3 and Jesús Velasco-Gómez3
1Departamento de Ingeniería Cartográfica, Geodesia y Fotogrametría – Expresión Gráfica; 2Departamento de
Matemática Aplicada; 3Departamento de Ingeniería Topográfica y Cartografía.
Universidad Politécnica de Madrid, Spain.
miguelangel.conejo@upm.es
1 Introduction
The underground cellars that appear in different parts of Spain are part of an agricultural landscape dispersed, sometimes damaged,
others at risk of disappearing. This paper studies the measurement and display of a group of wineries located in Atauta (Soria), in the
Duero River corridor. It is a unique architectural complex, facing rising, built on a smooth hillock as shown in Fig. 1. These constructions
are excavated in the ground. The access to the cave or underground cellar has a shape of a narrow tube or down gallery. Immediately
after, this space gets wider. There, wine is produced and stored [1]. Observation and detection of the underground cellar, both on the
outside and underground, it is essential to make an inventory of the rural patrimony [2]. The geodetection is a noninvasive technique,
adequate to accurately locate buried structures in the ground. Works undertaken include topographic work with the LIDAR techniques
and integration with data obtained by GNSS and GPR.
2 Methods
Surface GPR prospecting and underground LIDAR scanning have been applied
in order to jointly facilitate the determination and location of internal
structures. Through their fitting with the GPR profile, we estimate the Radio
Wave Velocity (RWV) in the ground, required for locating the GPR detections
of the hollow parts and old hidden structures. The radar data were acquired
using a Malå Ramac/GPR ProEx system equipped with unshielded antennae
of 100 and 200 MHz, in order to compare the behaviour of different
frequencies suitable for the area conditions and type of space [3]. Two
profiles were done at each frequency, one along the selected cellar (itinerary
1, Fig 2), and the second transversal to the selected and adjoining cellars
(itinerary 2, Fig 2).
Fig. 1 Underground cellars. Atauta (Soria)
m
Fig. 2 Itineraries (1) and (2), both with GPR
3 Results
Fig. 3 shows the 100 MHz GPR transparent profile obtained from the itinerary 1, superimposed on
the LIDAR profile. The inner profile of the winery detected by LIDAR is represented in blue (hidden
cavities are not detected). The GPR detection is shown in yellow and the road level in white.
We can observe that the dome rests on the road level. Other detectable structures are the chimney
pipe (in red), a wide cavity around it, and a discontinuity over the structure that supports the roof
stairs. The coupling of the results from GPR and LIDAR lets us estimate a value of 130 m µs-1 for the
RWV in the medium (soil and rock), which is appropriate for a reasonably dry limestone.
The radar detection profile does not match with the inner cavity detected by the LIDAR. This is due
to the early GPR reflections in the hollow parts and old hidden structures, in addition to the limited
resolution capability of the 100 MHz GPR, of ca. 50 cm.
A resonance effect appears in the staircase zone produced by multiple reflections between the stair
treads and the ceiling.
Over the cellar roof there may be a layer of fractured rock, which can produce a GPR reflection some
centimetres above the roof.
Once the RWV is characterized using the data of itinerary (1), domes of near cellars have been
detected using 200 MHz GPR itinerary (2). Those cellar domes are at less than 2 m depth (Fig. 4 a
and b)
Fig. 3: GPR transparent profile superimposed on the LIDAR profile
4. Conclusions
The use of Geographic Information Technology allows better geovisualization of the rural heritage, as shown on this article.
Using 200 MHz GPR the penetration depth was scarce, while with 100 MHz we have obtained successfully results. It has detected the different cave-domes, the cave-ceiling and most of the
cave-floors. It is also possible to detect the presence of other structures, as the entrance beam, the chimney or other close entrances.
The joint use of LIDAR and GPR techniques has revealed a faster method than conventional techniques, such as total station or photogrammetry. Also the RWV estimate is faster and more
accurate than using only GPR. The accuracy obtained is centimetric, and GNSS technique makes feasible the combined use of LIDAR and GPR maintaining the accuracy and the survey speed.
The techniques described in this article are suitable to use on other natural cavities, archaeological cavities or multipurpose constructed underground spaces.
This project can help the underground cellars to be declared as Cultural Interest by the Comisión de Patrimonio Cultural de Castilla y León - Junta de Castilla y León (Heritage Department of the
Regional Government of Castilla y León).
References
1. Martín Ocaña, S., Cañas Guerrero, I.: Comparison of analytical and on site temperature results on spanish traditional wine cellars. Applied Thermal Engineering,
vol. 26, nº 7, pp. 700-708. http://dx.doi.org/10.1016/j.applthermaleng.2005.09.004 (2006)
2. Pardo, J. M. F., Guerrero, I. C.: Subterranean wine cellars of central-Spain (Ribera de Duero): An underground built heritage to preserve. Tunnelling and
Underground Space Technology, vol. 21, nº 5, pp. 475-484. http://dx.doi.org/10.1016/j.tust.2005.07.004 (2006)
3. Lorenzo, E.: Prospección geofísica de alta resolución mediante geo-radar. Aplicación a obras civiles. Monografías CEDEX, Madrid. ISBN-8477902569 (1996)
4. Han, J.Y., Perng, N.H., Chen, H.J.: LIDAR Point Cloud Registration by Image Detection Technique. Geoscience and Remote Sensing Letters, IEEE, vol. 10, nº 4
(2013)
5. Pérez-Martín, E., Herrero-Tejedor, T.R., Gómez-Elvira, M.Á., Rojas-Sola, J.I., Conejo-Martin, M.A.: Graphic study and geovisualization of the old windmills of La
Mancha (Spain). Applied Geography, vol 31, nº 3, pp. 941-949. (2011)
Solution
One profile is done joining outside GPR and inside LIDAR measurements. LIDAR gives
real positions, so RWV can be tuned in GPR detecting. This RWV will be also used for
the rest of profiles (LiDAR measurement in blue; in yellow GPR some detections of
cavities and structures over the cellar).
Equipment
•Malå Ramac/GPR ProEx with RTA 100 MHz and
unshielded 200 MHz antennae.
•DGPS Leica 1200
•LiDAR Faro Focus 3D
Problem
Unknown Radio Wave Velocity (RWV).
• Depending on the existing substrate (Table 1) , RWV can take a value between
55 and 175 m/µs.
• Using wrong RWV, values of thickness obtained by GPR can be wrong.
• There is no easy way to measure RWV.
Material
Effective
permittivity
efr
Conductivity
(mS m-1)
Speed
v (m
µs-1
)
Air
1 0 300
Distilled water
80 - 88 0´01 33
Fresh water
80 - 88 0´1 - 10 33
Saltwater (and marine)
80 - 88 4000 10
Snow polar
1´4 - 3 --- 190 – 250
Polar ice
3 – 3´2 0´02 – 0´003 >168
Limestone dry
- wet 4 - 16 10-5 - 25 75 - 150
Shale dry
- wet 5 - 15 1 - 100 77 – 134
Granite Dry
- wet 4 - 15 10-9 - 1 110 - 130
Dry sand
- saturated 3 - 30 10-7 - 1 55 – 174
Dry Limo
- saturated 5 - 30 1 -100 63 – 100
Dry clay
- saturated 4 - 50 0´25 - >1000 60 – 170
Table 1. Typical values for different parameters of
propagation media (modified from Ramak, 2003,
Appendix 3).
The LIDAR used was a Faro Focus 3D unit. It was furnished with a telematic ambiguity unit and a range of 0.6 m - 120 m outdoor with
low ambient light. Point clouds registered by LIDAR and GPR where linked using GNSS techniques [4]. GNSS techniques used in this
study served a double purpose: to serve for geo reference all the survey, and further integrate the observations obtained using GPR
techniques and LIDAR [5].
2
1
010 20 a
b
d
ec
0
5
0
Length (m)
Depth (m)
10 20 30
a b c d e
Fig. 4 (a) : Identification of structures on transversal
itinerary. (b): Radargram showing the different
domes
4a)
Scale 1/100
Scale 1/100
4b)
Depth (m)
Length (m)
