Upload
polito
View
0
Download
0
Embed Size (px)
Citation preview
A DIDACTIC PROJECT FOR LANDSCAPE HERITAGE MAPPING
IN POST-DISASTER MANAGEMENT
Piero Boccardoa,b, Filiberto Chiabrandoa, Anna Facelloa, Loretta Gnavia, Andrea Linguaa, Paolo Maschioa, Fabio Pasqualec and
Antonia Spanòa
aPolitecnico di Torino, Corso Duca degli Abruzzi 24, 10129, Torino, Italy, e.mail: [email protected] ,
[email protected] , [email protected] , [email protected] , [email protected] ,
[email protected] bITHACA (Information Technology for Humanitarian Assistance, Cooperation and Action), Via Pier Carlo Boggio 61,
10138, Torino, Italy, e.mail: [email protected] cSiTI Istituto Superiore sui Sistemi Territoriali per l’Innovazione, Via Pier Carlo Boggio, 61, 10138 Torino, Italy,
e.mail: [email protected]
KEY WORDS: cultural heritage, disaster management, education, risk mapping
ABSTRACT:
In recent years natural and human induced hazards have increasingly turned into disasters with rising frequency and intensity. These
disasters pose threats to prominent cultural and natural heritage sites of the world and, therefore, require the intervention of skilled
technicians and staff with specific scientific and cultural training. It is necessary to contribute towards reducing slow as well as
catastrophic risks in short and long term, through the training of staff for intervention aimed at emergency management and
mitigation of the impact and development of solutions with the collaboration of professionals working in this area.
This paper presents a multidisciplinary project carried out by the cooperation between Politecnico di Torino and ITHACA
(Information Technology for Humanitarian Assistance, Cooperation and Action). The main goal was the training in geospatial data
acquisition and processing for students attending Architecture and Engineering Courses, in order to start up a team of “volunteer
mappers”. Indeed, the project is aimed to document the environmental and built heritage subject to disaster; the purpose is to
improve the capabilities of the actors involved in the activities connected in geospatial data collection, integration and sharing using
the modern geomatic methodology and techniques (terrestrial and aerial LiDAR, close-range and aerial photogrammetry,
topographic and GNSS instruments etc.. All the acquired data and processing results have been implemented in a WebGIS platform
to share information with local authorities and the Civil Protection. The proposed area for testing the training activities is inside the
Cinque Terre National Park (Vernazza), registered in the World Heritage List since 1997. The area was affected by flood on the 25th
of October 2011.
1. INTRODUCTION
Some 75 per cent of the world’s population live in areas that
have been affected at least on one occasion from 1980 to 2000
by earthquake, tropical cyclone, flood or drought. Natural
disaster and their consequences are intimately connected to
human development processes. The loss of human beings and
the destruction of economic and social infrastructure are
expected to worsen as climate change increases the frequency
and magnitude of extreme meteorological events, such as heat,
waves, storms and heavy rains. The community has already
developed a set of instruments to address various aspects of
disaster preparedness, response and recovery (UNDP, 2004).
Therefore, some organizations in the world (UNESCO,
ICOMOS, ICCROM, ecc.) emphasize the importance of the
protection of cultural heritage as a priority of our society. The
heritage conservation field places great importance on the use of
principles in guiding practitioners to appropriate interventions
for heritage properties. Conservation professionals recognize
these principles as being contained within the family of
doctrinal texts loosely linked to the Charter of Venice (1964),
for which ICOMOS is generally recognized as custodian.
ICOMOS has taken responsibility, primarily through the efforts
of its specialized international Scientific Committees, for
extending the basic general principles presented in the Venice
Charter by elaborating complementary texts in related fields
(Massue et al, 2001).
In this context, the “Cinque Terre project” is directed to provide
specific contributions to improve the attention for the natural
and cultural heritage, environmental protection and
enhancement, through the safeguard of the landscape and of the
architectural and archaeological heritage, damaged or under risk
(Spanò and Costamagna, 2010). The purpose is to develop
capacities in the Geomatic field using innovative methodologies
and techniques for achieving advanced metric surveys using
multi sensors and multi scale data (from space, aircraft and
terrain) (Boccardo and Giulio Tonolo, 2008).
The project has involved graduate students attending master
degree in Architecture and Engineering at the Politecnico di
Torino; the students can rely on basic qualification in
topographical and mapping or GIS technologies derived from
core subjects, and the key point for achieving the effectiveness
of the training consists in the organization of all day long
activities on site. The full immersion work, managed in small
groups of students (4 or 5) per each tutor, adding the pre and
post site experimental sessions, allow to obtain a good level of
expertise for the new crisis mappers team.
2. TEST SITE: VERNAZZA (CINQUE TERRE
NATIONAL PARK, WORLD HERITAGE LIST)
The Cinque Terre National Park was selected as test area for
two principal reason: the high value of landscape and
environment, then the severity of the flood event happened on
the 25th October 2011 and the severe damages occurred after
that.
The Cinque Terre area (Monterosso, Vernazza, Riomaggiore,
Corniglia e Manarola) covers approximately 15 km along the
extreme eastern end of the Ligurian coastline, between Levanto
and La Spezia.
The position of the five small towns and the shaping of the
surrounding land-scape, characterised by steep and uneven
terrains, encapsulate the continuous history of human
settlements in this region over the past millennium
(Dongiovanni and Valle, 2007; RSA Parco Nazionale delle
Cinque Terre, 2004). This Park was recognised by UNESCO
on its ‘World Heritage’ list, on the basis of cultural landscape
criteria and its out-standing value, representing the harmonious
interaction between people and nature (World Heritage Report
1997).
Figure 1. The orthophoto of Vernazzola stream (a) pre-event
August 2011 and (b) post event November 2011 (Source:
BLOM-CGR Parma).
The coastal area is typically characterized by high and steep
slopes artificially re-shaped during the centuries by human
activity, through the construction of dry stone walls.
In the Cinque Terre area, slope instability is mainly due to the
presence of land-slide-hill and lithological composition of the
substrate to which has been added, over the last decades, the
gradual loss of maintenance and defence of the territory
operated by human activity (Federici et al. 2001).
The drainage network is poorly developed and consists almost
exclusively of canals, ditches and channels. These are rivers,
with a very limited scope, which, in some cases, may remain
almost dry during the dry season.
2.1 Flood event (25th October 2011)
The surveys have been focused in the area of Vernazza, a
coastal village which reported extensive damages since the
flood event of 25th October 2011 that has been caused by the
overflow of the Vernazzola stream (Figure 1).
The heavy precipitation event, occurred in autumn 2011 over
the Ligurian Sea and the rainfall accumulated on the 25th of
October 2011, over the area of Vernazza, exceeded 500 mm in
12 hours, with peaks above 100 mm/hour, leading to a real
hydrogeological disaster over the zone by the most organized
convective systems.
All along the mountainside and hillside of the area characterized
by an artificially altered substrate (the typical terracing of the
Cinque Terre hillside), the rain-falls have caused an instability
both of land and vegetation.
As a consequence several slides were occurred in the area. The
surface runoff, due to the inclination of the terrain caused a big
amount of detritus resulting from landslides and soil cover and
in few hours the centre of the town (figure 2) was reached and
partially flooded (Ortolani, 2011).
Figure 2 shows the harbour of Vernazza before and after the
flood event.
Figure 2. Vernazza and its harbour before (a) and after (b) the
flood event
The Vernazzola watercourse and the provincial road along the
river, which were seriously damaged by the flood, were the
main objective of ground documentation of the calamities.
Moreover, along the stream of water, several artefacts,
including bridges, stone masonries opposing the slope
pressures, gardens, houses, terraces and other built structures
have been heavily damaged or completely destroyed (Figure
3).Finally, in order to complete the survey, the acquisition of the
main street of the Vernazzola village was realized using the
same methodology.
(a)
(b)
(a)
,
(b)
Figure 3. Examples of damages to buildings.
3. METHODS
3.1 Aerial data acquisition
Since the area is large and complex, the most suitable strategy
for assessing impacts and damages caused by the flood, is the
analysis of images acquired by satellite or aerial platform (pre
and post event), conveniently integrated by ground surveys.
LiDAR data and high-resolution aerial orthophotos, adding to
the basic and thematic multi-scale maps, have been collected for
supporting the project development, in order to plan and locate
the deepening area, to detect the major damages and to check
the accessibility.
The aerial orthophotos and LiDAR flights (post event) were
provided by Blom CGR Parma and Helica for the Friuli
Venezia Giulia Region (Figure 1).
Moreover, the Liguria Region has provided several
geodatabase: the basic maps (Carta Regionale 1:25.000 ed
1994/95; CTR 1:10000 ed 2007; CTR 1:5000 ed. 2007), the
Regional Geological Map 1:25000, the Inventory of landslides
1:10000 (IFFI project), the monitoring network of the slopes
1:10000 (Remover) and post flood erosion/accumulation maps,
performed by the difference of LiDAR data acquired in 2008
and2011. Even this last products have been achieved by the
Friuli Venezia Giulia Region.
3.2 The geodetic network
The geodetic network of the area (Figure 4) has been defined by
the GPS/GNNS method, it was connected to the permanent
ITALPOS stations of BRUGNATO ( 11 km from Vernazza)
and LA SPEZIA ( 15 km from Vernazza). The vertex situated
on the school has was used as local master station. The vertex
coordinates was calculated using Leica Geo Office obtaining a
good accuracy for the application (RMS: 5 mm in 2D, 10 mm in
height). Inside the centre of Vernazza, 3 vertexes were surveyed
using total station with a redundant network schema.
The ellipsoidal heights was converted in geoidal heights using a
local model of geoid (ITALGEO95) furnished from IGM
(Istituto Geografico Militare, the national military geographic
institute).
The network represent the basis for the next maps updates and
for the multi-scale integration that will support the recovery and
rehabilitation projects of the land.
Figure 4. The geodetic network frame.
3.3 Terrestrial laser scanning (TLS) surveys
The scanner Focus 3D (Faro Cam2) effectively fitted to the
needs of damages documentation due to the portable, the handy
characteristics and the simple use. The range of scan distances
is variable from 0.6 to 120 m for reflective surfaces (> 90%),
the error in linear distances is equal to ± 2 mm at 10 m and 25 m
for reflectivity of 10% and 90%; the scan speed is up to 900 000
points per second. In addition, the noise is low and it is possible
to acquire radiometric in-formation thanks to the included
digital camera with the optical axis coaxial to the laser beam.
The Faro instrument was used for surveying the Vernazzola
watercourse (about 2 km from the beginning of the overflow to
the sea estuary), and the area of the village where we planned to
document the state of built heritage (the main street, the square
with the harbour and the castle built on the top of the little hill
dominating the town); finally we surveyed a little new beach
created by the accumulation of debris due to the flood event
(Figure 8).
The scans resolution has been chosen medium-high in order to
be suitable for both natural and man-made land elements (the
scan density was set to obtain 1 point for each 6 mm at
distances of 10 meters). During a first step of elaboration a
regular set of cross sections at the distance of about 10 m have
been briefly extracted, without performing accurate clouds
filtering and colouring procedures.
The availability of automated tools in terrestrial laser data
management software, allow the use of recording techniques
that are endowed with procedures of automatic recognition of
geometric correspondences (best fitting).
However, the detected object features, characterized by very
few geometric elements easily identifiable and the abundance of
repetitive and similar elements (the stones of the river) have
advised the use of procedures strictly controlled by measured
targets. By means of these control points we achieved the
transformation in a single coordinate system based on the
reference system WGS84-ETRF2000.
Other targets, with a spherical shape and unknown position,
have been acquired during points clouds collection with the sole
function of tie points, aimed to the registration of adjacent
scans. The presented workflow together with a good distribution
of tie points and control points enabled rapid registration and
recording of scans, obtaining a contextual georeferencing of
data and an accurate control of the residuals.
Furthermore, two scans were acquired using the RIEGL LMS
420i in order to describe the roof of urban centre (not surveyed
from other TLS scans) and a visible rock face near a landslide to
modelling of fractured rock masses.
Table 1 shows some information about raw data and results of
the performed processing. Figure 5 shows a single coloured
points model, the mesh surfaces and an overview of recorded
clouds representing the entire section of Vernazzola.
Total Scans 50
Used GCPs/CPs 94
Total raw points 1 Billion
Dimension (.xyz file format) 50 Gb
Total points after the post processing (cleaning,
noise reduction, decimation etc.) 260
Million
Dimension (.xyz file format) 13 Gb
Table 1. Data showing large size of LiDAR ground collection.
The large set of cross sections extracted from the recorded point
clouds along the Vernazzola stream (Figure 6) has been used to
develop an hydraulic model, using HEC-RAS software. The
simulations implemented intend to evaluate the stream cross
sections more vulnerable to flood events. The roofs of Vernazza
have been collected using RIEGL scans as shown in figure 7.
Figure 5. A collection of images showing TLS data processing
results. Coloured models, mesh models and (top) overview of
recorded clouds representing the entire section of Vernazzola.
3.4 Report on the new beach valorisation
The heavy precipitation event, occurred in autumn 2011, caused
the surface run-off upstream the Vernazza village; a big amount
of detritus resulting from landslides filled the little creek,
generating a new beach. From the new beach many small and
coloured buildings (Figure 8), standing very closed each other
can be observed; they have been constructed on the top of the
cliff, making a typical scenery of Ligurian villages, that are
usually placed in a surrounding landscape characterised by steep
and uneven terrains. The access to the new beach is possible
through two different tunnels. The first one enable the
Vernazzola river to flow in the sea, while the other underground
passage connects the new beach to the centre of the village
(white ellipse in Figure 8).
In order to realize a complete survey of this new beach, we
opted to use the same acquiring resolutions and equivalent
registration strategies employed in other areas, to achieve an
high resolution model (Figure 9).
Figure 6. The extracted cross sections of Vernazzola stream
Figure 7. The roofs of Vernazza extracted from a RIEGL scan
Figure 8. A view from the sea of the new beach
Figure 9. An example of TLS acquisition
After the flood, this area have been under careful attention
because the uneven coastline of the region offer little place to
accommodate people, and a large stony beach, easily and
rapidly approachable from the train station, is a very interesting
place for the local administration.
After the needful stability controls of the rocky slope and safety
conditions of the area, this place can emerge as a cultural and
tourist resource for the village.
TLS points were used to generate a surface of the beach area
and near rock faces. Figure 10 is a semi-nadiral view to
emphasize the different directions of the tunnels.
The contour lines were extracted by planar sections and the
coastal line was collected using the limit of TLS data (Figure
11).
Section profiles derived from high scale and high resolution
models are not suitable for digital maps, because their geometry
is complex and they present an huge amount of vertexes, due to
triangulated surface. The generalization algorithm PAEK
(Polynomial Approximation with Exponential Kernel;
Bodansky et al.,2007) was applied to obtain a natural and
suitable shape of contours by smoothing lines (Figure 12).
Figure 10. The beach surface
Further more in order to evaluate the seabed in the areas of the
harbour and near the new beach a low cost sounder (GPSmap
421s by Garmin) mounted on a little boat was employed; the
followed trajectory was recorded by a kinematic GNSS
positioning based on geodetic receiver. Figure 13 shows the
obtained result with the used instruments.
Figure 11. Contour lines with new sea border line.
Lastly two diving missions (figure 14) were realized in order to
explore the seabed using GoPro video cameras:
- the first was performed in the normal direction to coastline, in
front of the beach. A sandy and downgrading seafloor (until 18
m), without abundant flora and wildlife was detected.
- the second mission was performed about 500m southern the
beach and allowed the finding of a car (Figure 15) and the trees
rattletrap.
Figure 12. An example of contour lines generalization using
PAEK algorithm
Figure 13. Survey of seabed near coastline using low cost
sounder
4. CONSTRUCTION OF GIS
In the field of natural disasters and emergency management
there is a strong interest in geospatial data sharing, with the aim
of avoiding duplication of efforts and allowing a quicker
operational response. GeoNode (http.geonode.org) is an open
source platform that facilitates the creation, sharing and
collaborative use of geospatial data.
Figure 14. The two diving missions itinerary
Figure 15. The sunk car
The aim is to surpass existing spatial data infrastructure
solutions by integrating social and cartographic tools. The
promotion of data sharing is a main factor. On the one hand data
sharing means that a user is put in condition of being able to
upload data into the system and to give other users the
possibility of exploring these data. Contemporary the user is
provided with the capacity of taking advantage of data shared
by others either through online tools or by downloading them
locally in formats suitable for desktop applications (Heinzelman
and Waters 2010).
In this case, the data collected from available database
collection and from terrestrial surveys will be implemented in a
WebGis. Thus, through a dedicated platform, the end user may
consult and interrogate in a simple and rapid way all the
available information.
5. CONCLUSIONS
The enhancement of documentation processes and the
development of procedures for the planning of emergencies
management has been continuously updated according to the
renewed methods and technical tools.
Regarding the on site training, all the students attending the
apprenticeship are now able to manage independently the
acquisition of topographical and TLS data.
The integration of data obtained by advance ground metric
surveys, with those derived by the analysis of aerial or satellite
images will lead to the achievement of a WebGis, aimed to the
collection and sharing of spatial information suitable for the
updating of the existing maps collection. Even in this phase the
students were involved in maps generation from different
sourced data, assessing metadata editing and evaluating overall
data quality. All the data collected during the activity will be
available for local and regional authorities and for the Civil
Protection through the GeoNode platform, in order to provide
useful information for disaster management and emergency
response.
In perspective, this experience will seek the feasibility of
implementing 3D metric data archives pertaining architectural
heritage, yielded by high detail surveys intending to identify the
architectural features of building complexes, within those
systems currently adopted for high scale 3D urban mapping, and
following their paradigms in geometric and semantic
organization (van Oosterom et al., 2005).
ACKNOWLEDGEMENTS
In order to be able to perform the different steps of project,
including the data acquisition phase, we thank the Liguria
Region for making available all the documentation, basic and
thematic cartography. We also thank the town of Vernazza,
represented by Matteo Spona; the FAROEurope-CAM2,
specifically Alberto Sardo, who made available a second laser
instrument for both rounds of the internship (June and July
2012). We also thank Blom CGR for facilitating the use of low
altitude photogrammetric data and post event Lidar data.
REFERENCES
Boccardo, P., Giulio Tonolo, F., 2008. Natural disaster
management: activities in support of the UN system. In:
Advances in Photogrammetry, Remote sensing and Spatial
information sciences: ISPRS Congress book / Zhilin Li; Jun
Chen; Emmanuel Baltsavias. Taylor & Francis Group, London,
pp. 385-396.
Bodansky, E., Gribov, A., Pilouk, M., 2002. Smoothing and
Compression of Lines Obtained by Raster-to-Vector
Conversion, LNCS 2390, Springer, p. 256-265, 2002.
Bornaz, L., Lingua, A., F. Rinaudo, 2003. Multiple scan
registration in LIDAR close-range applications. In: The
International Archives of the Photogrammetry, Remote Sensing
and Spatial Information Sciences, Vol. XXXIV, Part 5/W12,
pp. 72-77.
Dongiovanni, A., Valle, M., 2007. Il Piano di Gestione del sito
UNESCO “Portovenere, Cinque Terre e le Isole” Subambito
Cinque Terre, SiTI – Istituto Superiore sui Sistemi Territoriali
per l’Innovazione.
Federici, P.R., Baldacci, F., Petresi, A., Serani, A., 2001.
Atlante dei Centri Abitati In-stabili della Liuria, in Provincia
della Spezia- Regione Liguria, Consiglio Nazionale delle
Ricerche, Gruppo Nazionale per la Difesa dalle Catastrofi
Idrogeologiche, Università degli Studi di Pisa, Dipartimento di
Scienze della Terra, Unità Operativa 2.13
Heinzelman, J., Waters, C., 2010. Crowdsourcing Crisis
Information in Disaster- Affected Haiti, (Special Report 252),
Washington, DC, United States Institute of Peace (USIP).
Massue, J. P., Schvoerer M., 2001. Protection of Cultural
Heritage, Handbook of school of civil protection, Module bl-4/c
Ortolani, F., 2011. Alluvione nelle Cinque Terre-Lunigiana:
evento da manuale, difesa da migliorare con l’Allarme
Idrogeologico immediato (www.climatemonitor/?p=20863).
RSA Parco Nazionale delle Cinque Terre, 2004. Relazione sullo
stato dell'ambiente - Parco Nazionale delle Cinque Terre.
www.parconazionale5terre.it/upload/AG21Rsa.pdf
Spanò, A., Costamagna, E., 2010. Geomatic training
experiences for a high vulnerability cultural heritage item. In:
Remote Sensing And Geo-Information For Environmental
Emergencies, Torino, 2-4 Febbraio 2010.
United Nations Development Programme (UNDP), 2004.
Reducing Disaster Risk a Chal-lenge For Development, Global
Report, New York, Printed by John S. Swift Co., USA.
van Oosterom, P., Jansen, E., Stoter, J., 2005. Bridging the
worlds of CAD and GIS, in Large-scale 3D Data Integration:
Challenges and Opportunities, S. Zlatanova and D. Prosperi,
eds., ch. 9, pp. 9-36, Taylor&Francis - CRC Press, Boca Raton
US-FL, 2005.
World Heritage Report, Napoli 1-6 dicembre 1997, Portovenere,
Cinque Terre and the Islands (Palmaria, Tino, Tinetto) n.8267.