NDT&E International 39 (2006) 67–75

www.elsevier.com/locate/ndteint

Non-invasive characterisation of ancient foundations in Venice

using the electrical resistivity imaging technique
Nasser Abu-Zeida, Daniele Botteonb, Giovanni Coccoc, Giovanni Santaratoa,*
a
Dipartimento di Scienze della Terra, Via Saragat 1, Blocco B, I-44100 Ferrara, Italy
b
Fassa srl, via Lazzaris 3, I-31027 Spresiano (Treviso), Italy
c
Consultant Engineer-Insula SpA, Venice (Italy), Via Marignana 69, Mogliano Veneto (Treviso), Italy
Received 3 November 2004; revised 8 June 2005; accepted 12 June 2005
Available online 8 August 2005

Abstract
Electrical resistivity is a physical parameter linked to both texture and moisture content of the investigated bodies; therefore, its study
provides reliable information about the internal structure and eventual degradation state of the elements generally used for construction
purposes, like walls, columns, etc. Recent advances in hardware and software allow to acquire resistivity information of the investigated body
in a ‘tomographic’ manner. This technique, called Electrical Resistivity Tomography (ERT), furnishes detailed two (2D) and three (3D)
dimensional models of electrical resistivity distribution of the portion of the subsurface being investigated.
In this work, the application of the ERT technique, to evaluate restoration quality of the building foundations after being consolidated, is
described throughout a test site, located in Venice (Italy). The reconstructed 3D images of the resistivity distribution before and after
consolidation provided details about the internal structural of the foundation itself as well as the volumes which have been occupied by the
mortar after consolidation. Direct field test confirmed the diffusion of the mortar in the foundation and further laboratory analyses, conducted
on a specimen taken at 80 cm, showed that cracks of 40 mm width had been completely filled with mortar.
q 2005 Elsevier Ltd. All rights reserved.

Keywords: Electrical resistivity tomography; Consolidation; Monitoring; Historical foundations

1. Introduction furnish an idea about the degradation status, by comparing

the field result with a previously determined response of a
Vulnerability is defined as a set of conditions and representative set of samples of the materials used in the
processes resulting from physical, social, economical, and construction.
environmental factors, which increase the susceptibility of a NDT can also be advantageously used for quality control
community to the impact of hazards. The ‘Hazard’ concept of restoration works. As in the case studied here, a
is not limited only to human beings but should also be restoration work was rendered necessary on historical
extended to the management and protection of monuments foundations located in Venice, Italy. Foundations of cities
and cultural heritage patrimony as a whole. These hazards like Venice are vulnerable due to two main factors: aging
call for non-invasive assessment for site evaluation. The and deterioration of the foundation materials, being
Non-Destructive Test (NDT) methods offer today architects, weakened by wave action generated by the continuously
geologists and engineers efficient ways to map hidden increasing traffic of vessels in the navigation canals.
characteristics of the investigated bodies, in terms of In order to restore and/or increase the mechanical
geometrical (or spatial) distribution of one or more selected resistance of the foundation walls, engineers use special
physical property. Their mapped distribution, in turn, can types of mortars which are injected into the walls through
pre-drilled holes of about 20 mm in diameter. On one hand,
the quantity of mortar to be injected is difficult to be
* Corresponding author. Tel.: C39 0532 974728, fax: C39 0532 974767. determined a priori due to the scarce information about
E-mail address: g.santarato@unife.it (G. Santarato). voids and fractures distribution. Therefore the decision to
0963-8695/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. terminate injection is left to the site engineers and
doi:10.1016/j.ndteint.2005.06.007 technicians. On the other hand, the availability of an
68 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75

indirect investigation tool to help in assessing the quality of the apparent resistivity, determined at the surface, by the
the consolidation action is surely of interest and is estimated resistivity distribution within the investigated
highly recommended over the direct inspection, although structure through an inversion procedure. This choice is
necessary, which imposes the exposure of portions for certainly more cumbersome than the former, but it is
visual inspection and sample collection for further analyses facilitated by the increasingly efficient 3D inversion
in the laboratory. algorithms used to estimate resistivity distribution. Further-
Nowadays, the same non-invasive surface geophysical more, the computing power available today on desktop
techniques, usually employed in the investigation of the first personal computers allows for getting 3D resistivity
few tens of meters of the subsurface, are being widely distribution images at an acceptable time increase.
applied not only to map the internal geometry and
degradation state of historical buildings, e.g. [1,2], but
also to check for the quality of consolidation action using
the grouting technique, e.g. [3]. 2. The electric resistivity tomography ‘ERT’ technique
Widely used geophysical methods such as Ground
Penetrating Radar (GPR, [4]) and acoustic tomography are The resistivity estimation technique is a non invasive
perhaps the most known non-invasive techniques among geophysical investigation tool based on the application of
architects and engineers to achieve the above-mentioned Ohm’s law. A continuous, or slowly variable electrical
aims, e.g. [5,6]. Nevertheless, the employment of the GPR current is injected into the body to be investigated by means
becomes less efficient in case of building’s materials being of a couple of electrodes, to generate an artificial electric
saturated with salt-water, because the high electrical field E. Its distribution and intensity are functions of the
conductivity dramatically reduces the penetration depth of injected current, position of feeding electrodes and
the electromagnetic (EM) waves. resistivity distribution. As a consequence, an apparent
In the studied section of building’s foundations in resistivity ra is determined by measuring the potential
Venice, the GPR could not be applied, due to the high difference DV1 at a different couple of electrodes:
salinity. On the other hand, the acoustic tomography DV
technique needs the opposite sides of the wall to be ra Z k (1)
I
accessible to allow efficient and cost-effective mapping of
the consolidated volumes. When this is not possible, as in where k is a ‘geometric’ constant that depends only on the
the present case, the use of this non-invasive method reciprocal positions of the four electrodes, and I is the
becomes very cumbersome, since holes have to be drilled on intensity of the injected current.
the accessible side to host acoustic sources and receivers. Placing several electrodes along a profile over the body,
Under these conditions the ‘direct current resistivity modern computer-assisted ‘georesistivity-meters’ will auto-
method’, carried out in tomographic modality, is perhaps the matically span all couples of electrodes, and a correspond-
most suitable technique for the investigation of these ing number of ra values is obtained [11]. Since these values
structures and for tracing pathways occupied by the mortar depend on the true resistivity distribution, this distribution
materials, since the quality of the determined apparent can be estimated by an inversion procedure, based on the
resistivity data is not affected by the highly conductive minimisation of a suitable function, generally the sum of the
environment. Furthermore, a satisfactory reconstruction squared differences between measured and calculated or
of the internal structure can be obtained by making predicted apparent resistivities. To accomplish the latter
measurements on the accessible side only. goal, the space of the investigated body is discretised in a
The use of the geoelectrical resistivity method to monitor two-dimensional (or three-dimensional) grid of cells, where
the degradation state of building walls or columns is not each cell is assigned an initial resistivity value. A finite-
new, as these measurements are being made since, at least, difference or finite-element procedure computes the
one decade, e.g. [7,8]. Resistivity imaging, however, predicted apparent resistivity at the surface.
represents a powerful tool to monitor degradation caused Minimisation of the above-mentioned differences gener-
by moisture capillary resilience and salt re-crystallisation ally leads to an ‘ill-posed’ inversion problem [12]. The
[9]. More recently, apparent resistivity measurements were current method of solution minimises the above difference
also experimented for quality control [10]. while minimising the ‘norm’ or the ‘roughness’ of the
In this paper, we describe the application of the electrical resistivity model (‘Occam’s inversion’, see, e.g. Constable
resistivity Tomography (ERT) technique to evaluate et al. [13]). Minimising the norm while searching for a
restoration quality after consolidation. The obtained results minimum of the sum of the squared differences between
represent a three-dimensional (3D) model of the resistivity measured and calculated apparent resistivity data is perhaps
distribution within the investigated body before and after better known as the ‘damped least squares’ regression
grouting, which is indeed different from the evaluation of
apparent resistivity maps, as was reported by Schueremans 1
DV is the inner product between the two vectors: E (electric field) and
et al. [10]. The difference is related to the substitution of ð !.
ds (distance between the measuring electrodes): DV Z E$ds
 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75 69

Fig. 1. Location of the test site ‘Rio del Malpaga’ in Venice (yellow circle).

(or inversion) algorithm. Occam’s inversion results the foundation inspection, cleaning and sampling for
necessarily in a ‘smooth’ model, i.e. a smooth distribution materials characterization. Before injection, slim holes of
of the estimated true resistivities. However, its main 20 mm in diameter, placed at regular grid of 40!40 cm and
advantage resides in the fact that no initial and structured reaching various depths (generally less than 60–80 cm) are
resistivity model is needed, i.e. the inversion iterative drilled. Hole entrance is being protected by inserting a
process does not necessarily start from an a-priori chosen plastic tube firmly fixed to the hole by a special cement
model. This means that the initial model can be represented reinforced with fibers; the main aim is to protect the wall
by a uniform resistivity distribution in the subsurface or in against weathering action and secondary to prevent the
the interior of the finite body. escape of the mortar out of the same hole.
In Fig. 2, the state of the investigated wall portion
immediately after its exposure is depicted. An advanced
3. Test-site description degradation state can be observed, as testified by the
presence of voids occasionally filled with mud. Ancient
Among the several surveys carried out for consolidation restoration evidences are also observed, which were carried
control, we present as an example the survey performed out using bricks and stones of different dimensions.
along the ‘Rio del Malpaga’ canal, located in Venice, Italy
(Fig. 1), where foundations are certainly several centuries
aged. The restoration of the canal walls was a part of a more 4. Laboratory resistivity measurements
complete project, including also the dredging of the canal
base, recently concluded by ‘Insula S.p.A.’ society, based in The outcome of any non-invasive investigation is a
Venice. The restoration procedure, usually, commences model that illustrates the distribution of the measured
with dewatering the canal using iron barriers to allow for physical property, which in the present case is the electrical

Fig. 2. Picture of the investigated wall portion along ‘rio del Malpaga’ canal (coloured crosses are the electrode positions).
70 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75

Table 1
Summary of the resistivity values obtained from the laboratory tests on different typologies of brick and mortar samples

Sample Sample type dimensions No. of Mean Mean anisotropy

identification W!L!H (cm) tests resistivity (Um) coefficient (l)
M1 Red brick covered with mud and some 2.8!1.9!2.6; 2.6!3.1!2.4 2 64 1.55
encrustations
M2 Grey brick covered with mud and some 2.4!2.4!2.3; 2.3!2.6!2.4 2 20 1.32
encrustations
F1 Mud sample taken from the red brick 1.5!1.5!1.5 1 0.84 1.00
F2 Mud sample taken from the grey brick 1.5!1.5!1.5 1 2.54
C1 Mortar, after 2 weeks maturation 2.3!2.3!2.3 1 82 1.30
C2 Mortar after 2 weeks maturation 2.3!2.3!2.2 1 84 1.20
C1 Mortar, saturated 30‰ NaCl 2.3!2.3!2.3 1 23 1.10
C2 Mortar, saturated 30‰ NaCl 2.3!2.3!2.2 1 21 1.05
CF Mortar-mud mixture after one month 30!30!20 1 5.0–30 1.00
maturation

resistivity. Since this property is a function of texture, therefore, was not further considered as its range is
mineralogy, porosity, water content and salinity of the much less than the overall resistivity range determined for
constitutive materials [9], a calibration operation is thus each material.
needed in order to associate different ranges of resistivity
values to a specific type and/or condition of a material. To With these calibration data in mind, a reliable trans-
this end, several electrical resistivity measurements on formation of the 2D–3D resistivity model into a correspond-
bricks, stones and mud samples were conducted in the ing distribution of the different materials becomes feasible.
laboratory. Moreover, the resistivity of a pure mortar sample
was determined under different conditions of salinity;
similar tests were also conducted on a mixture of mortar 5. Mortar characteristics
and salty mud collected from the canal base. The electrical
resistivity of each sample was measured in the three The mortar product used in this study for foundation
directions, to check for possible electrical anisotropy. Field consolidation, is a trade mark of Fassa S.r.l. (Treviso, Italy),
moisture content was restored by saturating each sample commercially known as the ‘L512’ mortar. The miner-
with a brine solution at 30 g/l of NaCl. The procedure was alogical composition and mechanical properties of mortar
accomplished under vacuum to guarantee the complete samples, collected in the field after consolidation, were
saturation. The obtained results are reported in Table 1. analysed at the producer’s laboratory. Mineralogical
From laboratory measurements the following can be composition, determined using X-ray diffraction and
highlighted: X-ray fluorescence laboratory techniques, is reported in
Table 2, while the mechanical properties are reported in the
1. brick resistivity, saturated with salt water, ranges subsequent Table 3, compared to the standard. Two clean
between 20 and 70 Um (samples M1 and M2); mixtures of mortar were prepared and tested in order to
2. salt mud resistivity is generally around 1 Um (samples determine possible negative influences on the mechanical
F1 and F2); resistance due to the presence of eventual impurities
3. mortar resistivity, measured after two weeks of contained in water or even due to water excess.
maturation, is around 100 Um (samples C1 and C2);
4. mortar resistivity, saturated with salt water, decreases to
20 Um and remains stable (samples C3 and C4); 6. The ERT survey
5. resistivity anisotropy l was found between 1.05 and 1.5 in
the tested samples (M1, M2, C1, C2, C3, C4). Maximum The ERT measurements were carried out along four
anisotropy was observed in the brick samples, while the parallel profiles, located on the exposed portion of the
minimum was observed for the mortar, saturated with foundation’s wall (Fig. 2). Apparent resistivity data were
salt water. Quite obviously the mortar-mud mixture does collected using both ABEM ‘SAS4000/ES464’ and IRIS-
not show any anisotropy. The observed anisotropy, Instruments ‘Syscal Junior’ multi-electrode resistivity
Table 2
Chemical composition of the ‘L512’ mortar

Sample Fe2O3, % MnO, % TiO2, % SiO2, % S, % CaO, % K2O, % P2O5, % Al2O3, % MgO, % Na2O, % CO2, %
L512 0.761 0.018 0.066 6.732 0.455 52.157 0.293 0.020 1.111 0.539 0.048 –
 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75 71

Table 3 7. Results and interpretation

Properties of a field sample of mortar compared to standard specifications
of the L512 product
As it can be easily inferred, 2D inversion of the apparent
Sample L512 Standard L512 Field Test resistivity data along each profile can be misleading, since
(two specimens) the obtained image concerns the 2D section orthogonal to
Dimensions (cm) 4.0!4.0!15.8 4.0!4.0!7.5 the measurement surface, just below the electrode profile. In
Volume (cm3) 252.8 120 the event that the injected mortar flows away off this section,
Mass (g) 444 223
Density (g/cm3) 1.8 1.9
its presence cannot be monitored by the comparison of the
Compr. Str. (kg/cm2) 161 119 resistivity models obtained before and after injection;
Flex. Str. (kg/cm2) 40 n.d. therefore 3D inversion of field data is necessary to get
Elastic Mod (kg/cm2) 125,600 n.d. reliable information.
The 3D inversion was performed using the algorithm
meters. Small nails where used to insure the electric contact developed by Morelli and LaBrecque [14], which is based
with the walls. The distance between profiles was 0.50 m on the ‘Occam’s’ approach. This algorithm implements the
and a total of 24 electrodes, spaced 0.18–0.32 m, were minimisation of a robust functional, to reduce bias due to
employed, resulting in a total length of 5.92 m. possible outliers on the data. The estimated 3D model of the
The apparent resistivity data were collected before and investigated wall section is shown in Fig. 3, where the whole
four weeks after mortar injection using Wenner and volume and the most conductive and most resistive portions
equatorial dipole–dipole arrays. In the former array, the are outlined. In the images, very conductive volumes (less
four electrodes were aligned on the same profile with the than 10 Um) are attributed deep blue-cyano colours, very
potential measuring electrodes placed amongst the current resistive volumes (greater than 150 Um) are attributed red
injection electrodes, at a constant distance. In the latter array, colour and intermediate ones (resistivity ranges between
current and potential electrodes lie on two parallel profiles 100 and 10 Um) yellow-green colours.
where the respective dipoles are orthogonal to the profile These three main resistivity classes can be interpreted,
direction. Depth of investigation is mainly a function of based on the laboratory tests, as follows: observed high
the distance between current and potential electrodes. Joint resistivity volumes (O150 Um, red colour) should be
inversion of apparent resistivity data coming out from both attributed to the presence of voids (highly disconnected
the arrays was then accomplished so as to take into account wall and/or lack of mortar/bricks); intermediate resistivity
eventual resistivity anisotropy, since current is injected in volumes (20–100 Um, light blue to yellow colours) are
orthogonal directions. associated with more preserved wall portions, as can be seen

Fig. 3. (a) 3D images of resistivity distribution in the interior of the investigated wall portion, before consolidation. Very low and very high resistivity volumes
are highlighted in (b) and (c), respectively.
72 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75

Fig. 4. (a) 3D images of the resistivity variations in the interior of the investigated wall portion, after four weeks of consolidation; (b) volumes with resistivity
variations less than K20%; (c) volumes with resistivity variations greater than 20%; (d) volumes with resistivity variations greater than 40%.

in the right lower part of the section (a). Finally the observed increases, from the initial 1–3 Um up even to approximately
low resistivity volumes (!10 Um, deep blue-cyano 80 Um, a value which is in the range of the mortar samples,
colours) can be explained by the presence of salt mud if no mixing occurs. Accordingly, both processes of filling
filling fractures in the bricks as well as voids. Therefore, voids and complete or partial substitution of mud can
large volumes characterised by low resistivity values satisfactorily be monitored, since both processes produce,
indicate the almost complete degradation of the bricks and respectively, a fairly measurable decrease or increase of the
their consequent replacement with the salt mud. resistivity.
The ERT survey was repeated at the end of the hardening Some more comments on the mechanism of the injection
phase and the obtained resistivity model was normalised process could help to better understand the above
with respect to the pre-consolidation one. The results are statements.
shown in Fig. 4. Normalisation was necessary to render the First of all, voids are generally due to partial drying of
legibility of the variations more perceptible. mud after canal’s dewatering. Concerning the mud
A first glance analysis of the resistivity model shows the replacement issue, two hypotheses can be formulated
presence of well-defined regions where resistivity has which are supported by experimental tests. The first one
decreased (blue-cyano colours) and others where it has tries to explain mud replacement by mortar: as the density of
increased (green-reddish colours). A more detailed analysis the mortar is higher than that of the mud, some mud
of these images shows that, in general, regions which were compaction should occur as a consequence of the mortar
characterised by low resistivity before consolidation injection, which in turn contributes to the formation of
became more resistive, while high resistivity regions voids. The second hypothesis tries to explain mud
showed a marked resistivity decrease. Limited exceptions replacement by its mixing with the mortar during the
are observed, although of negligible importance. injection process. Experiments done in the laboratory by the
Bearing in mind the laboratory data, the observed results consultant on different proportions of mud-mortar mixtures,
can be explained, related to the mortar injection which of up to 50% mud and 50% mortar, demonstrated that
either filled the discontinuities (fractures and voids) or mixing is possible. Tests on different samples of this
displaced the filling mud. In the former case, a strong mixture showed that their mechanical properties were close
resistivity decrease and observed within the respective to those of the pure mortar.
volumes, as expected. When partial mixing occurred, this Field observations, during the injection phase, showed
decrease reached values ranging between 10 and 30 Um, in that mud mixed with mortar flowed out from the nearby
agreement with the laboratory test conducted on the mud- holes. This fact supports the latter hypothesis, that is mixing
mortar mixture. In the latter case, resistivity strongly is possible.
 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75 73

Fig. 5. (a) Particular view of the excavated portion of the restored foundation and (b) the corresponding resistivity variation model (after consolidation); the
location of the anomaly is indicated by the arrow.

8. Direct investigation the preliminary site investigation observations, the foun-

dation was judged to be highly deteriorated due to aging and
A hole was excavated, to check directly the consolidation weathering processes caused mainly by the impact of waves
state. The picture reported in Fig. 5 shows the presence of a generated by vessels.
large amount of mortar (the grey colour volumes), which The resistivity tomography (ERT) test demonstrated
corresponds to the increased resistivity volume indicated by its capability in mapping volumes that after consolidation
the arrow (see in Fig. 4, the same image at bottom right). were filled with the ‘L512’ consolidation mortar. This
The mortar specimen, shown in Fig. 6, was collected result was obtained by 3D inversion of apparent
within the test hole depicted in Fig. 5, after 10 days since resistivity data, collected using a layout of four parallel
consolidation phase ended. It was collected at 0.8 m depth, profiles.
while the maximum penetration depth of the pipe used for Direct investigation confirmed the achieved results and
injection was 0.6 m. It was analysed in the laboratory to proved the reliability and suitability of the ERT technique
determine the mortar distribution and penetration conditions for such kind of problems. Sophisticated laboratory analyses
within the sample. To this end, a small portion of the sample as S.E.M. and Energy Dispersion X-ray (EDX) showed that
was examined under the Scanning Electron Microscopy the selected mortar penetrated very thin fractures
(SEM).
(30/40 mm), while still offering a good resistance to salts,
The SEM images are shown in Figs. 7 and 8. In
particular, in Fig. 7a crack width of 40 mm is observed to be
filled with mortar. Further analyses of the sample indicated
almost total absence of discontinuities between the substrate
and the mortar (Fig. 8).
These SEM images confirm that the two main critical
characteristics of a grouting mortar to be used for
consolidation purposes, i.e. (1) ability to penetrate very
thin cracks (of the order of few tens micrometers) and
(2) efficient coating of the internal surfaces after the crucial
hardening phase, are satisfied by the examined ‘L512’
mortar, even in the conditions of salty environment.

9. Discussion and conclusions

A non-invasive test, performed along a portion of the

‘Rio del Malpaga’ canal in Venice, concerning the efficacy
of mortar diffusion within the internal structure of
foundation walls has been accomplished. According to Fig. 6. Picture of the mortar sample collected at 80 cm depth at the test site.
74 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75

Fig. 7. SEM pattern distribution of the L512 mortar sample. Dark zones represent the absence of material (discontinuities or, more generally speaking, voids).

Fig. 8. SEM pattern distribution of the L512 mortar sample and brick material. Blue zones represent the absence of material (discontinuities or more generally
speaking voids).

like chlorides and sulphates, particularly abundant in References

Venice’s environment.
The ERT technique is recommended for such type of [1] McCann DM, Forde MC. Review of NDT methods in the
characterisation over the other, more widely known NDT assessment of concrete and masonry structures. NDT&E Int
2001;34:71–84.
items as GPR and Acoustic Tomography, whenever similar
[2] Cardarelli E, Godio A, Morelli G, Sambuelli L, Santarato G,
severe working conditions are met (i.e. a low resistivity Socco LV. Integrated geophysical surveys to investigate the scarsella
environment and un-accessibility of the rear face of the vault of St John’s Baptistery in Florence. The Leading Edge 2002;
investigated wall). 21(5):467–70.
[3] Berra M, Binda L, Baronio G, Fatticcioni A. 1988: Ultrasonic pulse
transmission: A proposal to evaluate the efficiency of masonry
strengthening by grouting. In Proceedings of the second international
Acknowledgements conference on non-destructive testing, microanalytical methods, and
environment evaluation for study and conservation of works of art.
We acknowledge Insula S.p.A., Venice (Italy), for Rome: Associazione Italiana Prove Non Distruttive and Istituto
Centrale per il Restauro. I/10.1-I/10.19.
supporting this research and for permitting us the
[4] Davis JL, Annan AP. Ground penetrating radar for high-resolution
publication of part of the obtained results. We are also mapping of soil and rock stratigraphy. Geophys Prospect 1989;37.
grateful to Gianfranco Morelli for his help in ERT 3D data [5] Saisi A, Valle S, Zanzi L, Binda L. Radar and sonic as complementary
inversion. and/or alternative tests in the survey of structures. Proceedings of
 N. Abu-Zeid et al. / NDT&E International 39 (2006) 67–75 75

the international millennium congress archi 2000, Bethlehem, [10] Schueremans L, Van Rickstal F, Verderickx K, Van Gemert D.
Palestine. vol. 1 2000 [Section 1b 6]. Evaluation of masonry consolidation by geo-electrical relative
[6] Maierhofer C, Leipold S. Radar investigation of masonry structures. difference resistivity mapping. RILEM Mater Struct 2003;36(1):
NDT&E Int 2001;34:139–47. 46–50.
[7] Ranieri G, Giani GP, Ferrero A. Indagini geofisiche sulle colonne del [11] Barker RD. Offset system of electrical resistivity sounding and its
Duomo di Orvieto. Atti del Seminario di Geofisica per l’Archeologia, use with a multicore cable. Geophys Prospecting 1981;29(1):
Porano, 1988, in Quaderni dell’ITABC, Roma, no 1 [In Italian].
128–43.
[8] Van Gemert D, Janssens H, Van Rickstal F. Evaluation of electrical
[12] Tikhonov AN, Arsenin VY. Solutions of ill-posed problems.
resistivity maps for ancient masonry. RILEM Mater Struct 1996;29:
Washington: Winston and Sons; 1977.
158–63.
[9] Santarato G, Dondi M, Vaccaro C. Valutazione del degrado di cortine [13] Constable Sc, Parker Rl, Constable CG. Occam’s Inversion: a
murarie in cotto mediante indagini di tomografia elettrica integrate practical algorithm for generating smooth models from EM sounding
con dati mineralogico-petrografici. In: Laterizi e terrecotte architetto- data. Geophysics 1987;52(2):289–300.
niche: la conoscenza per la conservazione, University Press Bologna; [14] Morelli G, LaBrecque DJ. Advances in ERT inverse modelling. Eur
2000. p. 79–83 [In Italian]. J Environ Eng Geophys Soc 1996;1(2):171–86.