Microsoft Word - 342-2440-1-LE-rev


ACTA IMEKO 
ISSN: 2221‐870X 
September 2016, Volume 5, Number 2, 55‐63 

 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 55 

3D  survey  technologies:  investigations  on  accuracy  and 
usability  in  archaeology.  The  case  study  of  the  new 
“Municipio” underground station in Naples 

Luigi Fregonese
1
,  Francesco Fassi

1
, Cristiana Achille

1
, Andrea Adami

1
, Sebastiano Ackermann

1
, Alessia 

Nobile
1
, Daniela Giampaola

2
, Vittoria Carsana

3
 

1
Polytechnic of Milan, Department ABC, Hesutech laboratory, Campus Mantova, Piazza D’Arco 3 ‐ 46100 – Mantova, Italy 

2
Superintendence Archaeology Campania, Piazza Museo Nazionale, 19 ‐ 80135 ‐ Naples, Italy 

3
Assistant of the Superintendence Archaeology Campania, Via del Marzano, 6 – 80123 – Naples, Italy 

 

 

Section: RESEARCH PAPER  

Keywords: Close Range Photogrammetry; laser scanner; instruments; accuracy; calibration; 3D‐Model; archaeology 

Citation: Luigi Fregonese, Francesco Fassi, Cristiana Achille, Andrea Adami, Sebastiano Ackermann, Alessia Nobile, Daniela Giampaola, Vittoria Carsana, 3D 
survey technologies: investigations on accuracy and usability in archaeology. The case study of the new “Municipio” underground station in Naples, Acta 
IMEKO, vol. 5, no. 2, article 8, September 2016, identifier: IMEKO‐ACTA‐05 (2016)‐02‐08   

Section Editor: Sabrina Grassini, Politecnico di Torino, Italy; Alfonso Santoriello, Università di Salerno, Italy 

Received March 24, 2016; In final form July 22, 2016; Published September 2016 

Copyright: © 2016 IMEKO. This is an open‐access article distributed under the terms of the Creative Commons Attribution 3.0 License, which permits 
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited 

Corresponding author: Luigi Fregonese, e‐mail: luigi.fregonese@polimi.it 

 

1. INTRODUCTION 

Nowadays, three-dimensional models of objects from 
images are a standard in a wide range of applications from 
autonomous robotics to industrial vision and consumer digital 
entertainment. In addition, it has been a topic of intensive 
research since the early days of computer vision and in the field 
of Cultural Heritage [1], but it has eluded a general solution 
regarding the accuracy in photogrammetric systems. 

Image Based Modelling (IBM) is based on multiple 2D 
image measurements to recover 3D object information through 
a     mathematical    model.     This     method    calculates    3D  

 
 

measurements from multiple views with the use of projective 
geometry and a perspective camera model. In addition, it 
guarantees a good portability and implies often low cost sensors 
[2]. 

However, the surveying results must meet specific criteria in 
order to provide the required accuracy for certain applications. 
For that reason, any geometric surveying task, such as the 
photogrammetric one, includes not only the definition of the 
relative positions of points and objects but also the estimation 
of the accuracy of the results [3]. 

With the least squares adjustment method, based on finding 
an approximate solution to overdetermined systems, it is 

ABSTRACT 
Advanced 3D survey technologies, such as Digital Photogrammetry (imaged based) and Laser Scanner, are nowadays widely used in 
Cultural Heritage and Archaeological fields. The present paper describes the investigations realized by the Laboratory Hesutech of the 
Polytechnic of Milan in cooperation with the Superintendence Archaeology Campania in order to examine the potentiality of Image 
Based  Modeling  (IBM)  systems  applied  to  the  archaeological  field  for  advanced  documentation  purposes.  Besides  the  3D  model 
production workflow in an uncommon excavation environment, a special consideration about the reached accuracy will be discussed. 
In the first part of the research, a comparison between photogrammetric camera parameters obtained with IBM systems and the ones 
provided with the calibration certificate by the manufacturer of the camera is performed.  
In the second part of the research, the operational phases of the application of such advanced 3D survey technologies are shown. The 
test field is the archaeological excavation area for the construction of the new “Municipio” underground station in Naples. Due to its 
position in one of the historical area of the city, its construction coexists with the archaeological excavations and it is strictly tied to 
their evolution. In such conditions, the need to reduce as much as possible the time to build the public infrastructure is a very relevant 
feature together with the ability to produce accurate documentation of what is considered archaeologically important. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 56 

possible to obtain reliable information concerning the accuracy 
of the results as well as the accuracy of the observations. 

The building of the new “Municipio” underground station in 
Naples has required extended as well as intensive archaeological 
investigations on almost the entire area of the construction 
yard. The importance of the findings and the evidences 
together with the need to speed up the completion of the 
infrastructure, required to introduce advanced 3D survey 
technologies to satisfy both the requirements. However, the 
high amount of almost every day photogrammetric surveys as 
well as restitutions, has not allowed to follow a rigorous 
workflow regarding the calibration procedure (e.g. with 
dedicated calibration test fields). Anyhow, a self-calibration has 
been always performed by using the same images acquired 
during the survey tasks, obtaining residuals on GCPs of few 
millimetres or 1 to 1.5 cm in the worst cases. In the following 
Section, a comparison between a self-calibration obtained in 
such way and the one certified by the manufacturer will be 
descripted and discussed. In the second part, a few numbers of 
the IBM system applications on different size and complexity 
objects will be shown. 

2. ACCURACY WITH IBM METHOD 

2.1. Method for the determination of the accuracy with IBM 

Nowadays, high-resolution cameras allow to acquire 
excellent quality images in terms of resolution and sharpness 
and can be used to perform precision surveys. The best way to 
evaluate the efficiency and the quality of this instrument as well 
as the whole process is to focus on camera calibration, the 
common issue in photogrammetric applications, especially 
when the precision for dimensional measurements is a not 
negligible variable. 

In order to perform this calibration, the process follows a 
three steps approach: 

1. Image acquisition; 
2. Image alignment with the software Photoscan®, to define 

position of images and camera parameters (internal 
orientation and lens distortions); 

3. Comparison of the calculated values with those reported 
in the metric calibration certificate. 

Compared to the other aberrations, lens distortion is the one 
that mainly affects photogrammetric measurements in terms of 
accuracy, and the images  must be corrected in 
photogrammetry. Lens distortions are of several kinds, but the 
radial one is the most significant. It is the radial displacement of 
a projected image point on the sensor from its theoretically true 
position or, equivalently, a change in the angle between a ray 
and the optical axis. 

In order to define the distortions, the parameters of the 
internal orientation of a camera are calculated by means of the 
self-calibration, in which the distortion and camera parameters 
are included as part of the bundle adjustment solution. 

The digital camera used in this test is the Rollei 6008 (Figure 
1) with fixed digital 72405433 pixel (49.23236.9444 mm) 
resolution back sensor (Phase One) and with a 40 mm optical 
lens. This camera has the following features:  

 it can save Loss-free or RAW-format images, up to 48 bit 
colour depth, 32 MB per image; 

 interchangeable metric lenses PQ (fastest shutter speed: 
1/500 s) between 40 mm and 350 mm, interchangeable 
metric lenses PQS (fastest shutter speed: 1/500 s) 

between 50 mm and 500 mm; 
 metric calibration certificate for each lens; 
 pixel size 6.8 μm. 
The calibration certificate provided by the manufacturer 

includes the following parameters: 

 Ck: 41.195 mm; 
 Xh: 0.161 mm  and  Yh:  +0.460 mm; 
 A1: 3.6010-5 and A2: 2.4410-8 as parameter for the 

radial distortion; 
 R0: 0.0 mm; 
 R: [0, 31] mm. 
The curve of distortion is defined by Rollei with the 

following equation: 

∆ ∙ ∙ ∙ ∙  . (1) 
The radial distortion curve reported in Figure 2 shows a 

 
Figure 1. Rollei 6008AF with Phaseone P45. 

 
Figure 2. Lens radial symmetric distortion values in pixel and μm. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 57 

radial distortion of 374 μm (55 pixels) at 31 mm of radial 
distance. 

2.2. Test for accuracy investigation 

In order to evaluate the overall accuracy of the IBM system 
(in particular Photoscan) a calibration test was realized in a 
closed environment set-up in a square room of about 4 meters 
of length,  with a cross vault on the top (Figure 3). 

In this room, a total amount of 108 coded targets (85 coded 
for PhotoScan and 23 for Leica HDS Cyclone) were displaced 
all around and measured with a Leica Total Station TS30. 

In the first data processing, all acquired targets have been 
used for camera calibration through the steps of image 
alignment and subsequently of optimization. In such a way the 
software Photoscan calculated the inner orientation, with the 
results reported in Table 1. 

These results are congruent with the data provided in the 
original calibration certificate of the camera and the distortion, 
as reported on  also from the distortion curves of Figure 4. 

As concerning the accuracy of alignment of the 
photogrammetric model, the process has calculated its solution 
at high quality and it discovered about 120.000 points (Figure 5) 
with the quality values expressed in Table 2. 

The above described accuracy values clarify that the 
precision of the photogrammetric model (with an order of 0.5 

pixel) well fits the standard deviations obtained for each 
direction X, Y, Z. 

To have a more significant comparison, about 2/3 of all 
GCP were used as check points (CPs) to verify the quality of 
the alignment of the total model: the results of such test are 
summarized in Table 3. 

 
Figure 5. GCP and Tie Points determined in alignment and calibration phase 
of the model of investigation. 

Table 1. Camera Parameters calculated with Photoscan software: exported 
in Australis format. 

Camera 
Parameters 

Value 

Camera   Rollei 6008 P45

H sensor  7240pixel

V sensor  5433pixel

Pixel size  6.8µm 

C  41.150mm

Xp  ‐0.2549mm

Yp  0.3618mm

K1  ‐3.9628710
‐5
 

K2  3.0461610
‐8
 

K3  ‐4.7477510
‐12
 

P1  3.5966310
‐6
 

P2  1.6384410
‐6
 

Figure  4.  Comparison  Lens  radial  symmetric  distortion  in  μm  (above)  and 
pixel (below). 

Figure  3.  The  Polygon  test  in  which  coded  target  for  Photoscan  and  HDS
Laser Scanning were positioned. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 58 

As evident from the comparison, the IBM approach, based 
on photogrammetry and topography, is correct and we can use 
it also in the documentation of Cultural Heritage. Indeed, the 
network geometry adopted for the images acquisition in this 
work is the same as the one adopted during the surveys of 
architectural and archaeological artifacts; in addition, the order 
dimension of the acquired objects, for cases of close range 
survey, are similar. The IBM method is efficient not just for the 
acquisition stage (in terms of instrumental costs and timing for 
data acquisition), but also in terms of achievable accuracy. 

2.3. Test on surface accuracy  

The last part of the test concerns the comparison between 
the point cloud extracted with photogrammetric approach and 
a second one obtained by employing the laser scanner Leica 
HDS 7000 [4].  

The laser scan was made by positioning the scanner in the 
centre of the room and it was registered in the same reference 
system used for the photogrammetric model by aligning it 
through the 23 black/white HDS targets. 

The results of the registration, made with Leica Cyclone, are 
shown in Table 4. 

The evaluation of surface accuracy has been made with the 
software CloudCompare which allows to measure the 
Euclidean distance between two pointclouds. In Figure 6, it is 
clear that the distances between photogrammetric and laser 

scanner pointcloud are included in the interval 
0.0006<x<0.0048 mm. 

These tests confirmed that the photogrammetric approach, 
IBM, is a reliable survey methodology as it guarantees the same 
accuracy of other well-established methods such as laser 
scanner. This method validation allowed us to use both 
methods, IBM and lasers scanner, and to choose the most 
suitable one time by time considering also some other aspects 
such as the possibility to acquire texture, the acquisition time, 
the operational conditions, etc.  

The following Sections contain a few examples of 
applications of these survey methods in real, and complex, 
contexts. 

3. CASE  STUDY:  THE  NEW  UNDERGROUND  STATION 
“MUNICIPIO” IN NAPLES 

The construction of the new underground station 
“Municipio” in Naples, interchange point between two 
underground lines and the touristic harbour, has produced time 
expensive archaeological investigations within the construction 
areas. These investigations were arranged with the Italian 
Cultural and Activities Heritage Ministry, the Municipality of 
Naples and “Metropolitana di Napoli S.p.A.”, concessionaire 
for the construction of the underground, and represent a 
precocious example of “planned archaeology” applied to an 
important public infrastructure: indeed, they have been started a 
few years before the introduction of the “Rescue Archaeology” 
law, issued in 2006. 

Table 2. Residuals on GCPs. 

Parameters  Value 

No. Images  73 

Tie points  117.081

GCP Points  108 

RMS (mm)  0.9 

RMS (pixel)  0.542

σx (mm)  ±0.61

σy (mm)  ±0.65

σz (mm)  ±0.57

Figure 6. Euclidean distances between laser scanner and photogrammetric 
point clouds. 

Table 3. Estimated accuracy of georeferencing images considering both sets 
of points GCPs and CPs. 

Parameters 
Configuration  
with GCPs only 

Configuration  
with GCPs and CPs 

No. Images  73 73

GCPs  108 37

CPs  0 71

RMS (mm)  0.99 0.89

σx (mm)  ±0.63 ±0.61

σy (mm)  ±0.68 ±0.65

σz (mm)  ±0.72 ±0.49

Table 4. Registration Laser Scanner data.

Parameters  Value 

No. Target  23 

RMS (mm)  1.2 

σx (mm)  ±0.95 

σy (mm)  ±0.82 

σz (mm)  ±1.02 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 59 

By following the principles of the “urban archaeology”, the 
investigations have allowed to explore all the human 
settlements originated within the area, from the most ancient to 
the modern ones, denoting an unexampled fact-finding model 
applied to a critical area of the historical centre of Naples, an 
area that couldn’t be differently investigated due to the 
complexity of the urban situation, the depth of the evidences 
and the contextual presence of a phreatic layer. In this second 
part of the research, survey methodologies verified in the 
previous part are applied on the below described evidences and 
adapted to the specific cases will be presented. 

3.1. The archaeological evidences 
The old costal landscape shape in front of the historical 

Neapolitan settlement has been reconstructed thanks to the 
archaeological investigations. However, the area delimited by 
the actual Via Medina, Via Depretis, Piazza Municipio and 
Castel Nuovo was deeply different by the one known as-is 
today, outcome of secular natural and anthropic 
transformations (Figure 7). 

At the beginning, it was the south-west zone of a wide inlet 
delimited, from south-west to north-east respectively, by the 
promontory where now Castel Nuovo is located and by the 
elevation where now the S.Maria di Porto Salvo church is 
located. 

From the end of the 4th century B.C. to the 5th century 
A.D., the above described area matched with the port basin of 
Neapolis and with the neighbouring seaboard as well [5]. The 
laying out of the harbour has dated back to the end of the 4th 
century/beginning of the 3rd century B.C., thanks to the 
founding of tracks related to an extended dredging operation 
(about 3750 m2) on the deepest sea bottoms levels, operations 
that modified the original shape of the inlet. In the same epoch, 
the above slope was regularized by rising up walls for terracing 
purposes made with panelled masonry or tuff blocks technique, 
probably for the protection of the basin. A ramp made with tuff 
blocks was also built, maybe for mooring purposes. 

During the Augustan age, the area was differently 
reorganized, and important harbour and street infrastructures 
were set up. Next to the inside edge of the inlet, a quay made 
with concrete, superimposed over tuff blocks lines and 
delimited by a reticulated work type wall, was built, and the 
natural tuff bank was also worked. The presence of an erosional 
line characterized by the presence of malacofauna (barnacles 

and oysters) proves that the sea level during the roman imperial 
age was about 1.70 to 1.80 m below the actual sea level. In the 
south-east area of the inlet, a much more complex harbour 
infrastructure, still under investigation, has been recently 
discovered: it is composed of concrete structures, built by using 
a wooden formwork, that probably were used as docks or 
breakwaters to protect the entrance of the port basin. The 
regularized coastline area was employed to build a thermal 
structure (1st and 2nd century A.D) along a route, probably the 
via per cryptam from Neapolis to the Phlegraean Fields. 

Among other things, the stratigraphic sequence of 
overlapping sea bottoms that included a high number of 
ceramic and organic evidences (such as waste, ship equipment 
or lost objects), was also excavated. During these investigations, 
six shipwrecks were found: poor rests of two shipwrecks dated 
2nd century B.C. (named “Napoli E” and “Napoli H”), two 
ships dated 1st century A.D. ( “Napoli A” and “Napoli C”), and 
other three boats, maybe sunk due to a storm at the end of the 
2nd century A.D. (“Napoli B”, “Napoli F”, “Naples G”). All 
the boats were built according to the shell-first method, with 
the planks fitted edge to edge and fastened by mortise-and-
tenon joints technique [6]. Shipwrecks A, B and F were sail-
boat types used for commercial purposes, whereas the C and G 
ones had a flat vertical stern, belonging to the horeiae class, 
probably pushed by using both sail and oars with multi-
purposes employment, details known just from iconographic 
sources (a mosaic found in Althuburus, Tunisia) [7]. While the 
A and C boats were entirely removed by constructing a 
fiberglass case around them, the F, G and H boats, found in 
2015, due to their worse condition, have been accurately 
disassembled and each component has been stored within water 
tanks to preserve them. The disassembling process has been set 
up and executed with the cooperation of the ISCR (National 
Institute for Conservation and Restoration) and the naval 
archaeologists G.Boetto (CNRS – Center Camille Jullian) and 
C.Zazzaro (University of Naples “L’Orientale”). 

The archaeological excavations revealed a mutation of the 
inlet morphology at the beginning of the 5th century A.D.: the 
coast line moved toward east and so the port basin. In addition, 
the previously built construction was abandoned or reorganized 
to be used in different manner. 

The urban history of the area completely changed from the 
construction of Castel Nuovo in the 1279 up to the last square 
realized in the second half of the 19th century. 

3.2. Survey method for the archaeology 

One of the issues that affects the Cultural Heritage field is 
the need to storage a huge number of evidences found during 
archaeological excavations. Most of them are usually preserved 
in storehouses: the employment of before mentioned digital 
technologies allows to obtain accurate three dimensional 
models and to share them through the web network, strongly 
changing the way to benefit from such ancient treasures. In 
addition to the possibility to carry out metric and semantic 
information of an object without physically measuring the real 
evidence, such technologies are almost a mandatory step 
especially to document those artefacts that could not be 
preserved, as in the case of the construction of a public 
infrastructure. 

An exclusive use of a traditional two-dimensional drawing 
approach would bring to a loss of important information [8], 
metrical as well as semantic, and to a lack of precision. In 
addition, the total absence of texture information of the 

 
Figure 7. Roman evidences and reconstructed costal landscape shape. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 60 

measured object could make this approach difficult to read if 
not supported by photographs. Because of the three-
dimensionality of architectonic and archaeological structures, 
and also of the importance to the appearance quality of the 
surveyed objects besides the geometrical precision, the best way 
to represent them is to reproduce them by keeping all the three 
dimensions and texture information, in order to have a virtual 
representation extremely similar to the real object. 

Since 2012, the concessionaire for the underground 
construction, together with the “Superintendence Archaeology 
Campania”, has set up a cooperation with the “3D Survey 
Group” of the “Polytechnic of Milan”. 

The main issue that came out in this excavation site was the 
need to work in a complex environment, with working 
machinery close to the areas to be surveyed, logistical 
difficulties, and the necessity to save as much time as possible 
to avoid construction slowdown. These working conditions has 
led to further adapt the modus operandi in particular for the 
acquisition step, but also to better test the adopted survey 
techniques in a stressing situation. Beyond that, the systematic 
usage of Digital Photogrammetry and Laser Scanner (not just 
for isolated cases) has met the need to produce adequate and 
detailed documentation of the found evidences within a huge 
construction area like the one in Piazza Municipio. The 3D 
models obtained from each survey session performed in 
different moments and positions can be successively combined 
together in order to have a global vision of the found evidences, 
making possible also the stratigraphic sequence understanding 
and to reconstruct the ancient scenery and its transformation 
from the Hellenistic era up to the 19th century. The adopting of 
such new techniques sensibly reduced the acquisition time (and 
as a consequence improved the efficiency of the excavation 
process) but guaranteed the accuracy of all acquired data. 

The strong cooperation between surveyors and 
archaeologists required to tune up a process schedule among 
the survey step and the preliminary restitution of the related 
orthophoto, that permit a practical verification by the 
archaeologist before let the excavations go on. In a second 
moment, without the need to halt anymore the excavations, 
phase drawings, architectonic and stratigraphic sections are 
carried out in cooperation with the “Calcagno Architetti 
Associati” company. 

3.3. The new underground station 

The documentation to produce within the “Municipio” 
construction yard with 3D survey technologies is basically 
composed by orthophotos and 3D point cloud models [9], [10]. 
During the last three years, several tests with both survey 
techniques have been conducted on an elevated number of 
evidences that differ in terms of type and dimension, reaching 
the best approach to use depending on the case. 

At the beginning, digital photogrammetry has been mainly 
used to generate orthophotos while laser scanner to obtain 3D 
point clouds. However, the improvements of the actual 
photogrammetric software and the test validation we exposed, 
allow now to obtain accurate dense point clouds in an almost 
fully-automatic way. The obtained results that will follow 
encourages the use of this image-based technique to extract also 
3D point clouds, reserving the use of the laser scanner for 
particular cases and even lower much more the acquisition time 
as well as the halt of the excavations. A series of surveys related 
to particular and different type of evidences and stratigraphy 
will be presented and discussed. In some case, specific 

additional solution were adopted to perform the survey and to 
achieve the requested results in term of precision and resolution 
[11]. 

 A. Dredging tracks 

One of the most interesting and particular discoveries has 
been the finding of the rests of dredging operations on the tuff 
rock stratum. Because these tracks have been found about 7 
meters below the actual sea level, it is supposed they were 
probably realized by employing a dredge mounted on a boat or 
on a buoyant platform [5]. 

These tracks have been localized in different but 
neighbouring excavation areas of the construction yard, and 
they have been excavated in different moments as well. Due to 
these reasons, it was not possible to have a complete physical 
global view of the tracks all together. 

The proposed technologies have been employed in the most 
part of the area in which the dredging tracks have been found. 
Except for the dredging found in 2004 during the line 1 station 
shaft excavations, for which less updated survey methodologies 
were employed, all the remaining areas in which they have been 
discovered, have been surveyed with actual technologies. 

The dredging tracks discovered in the line 6 station shaft 
have been surveyed in 2014. Due to the fact that they were 
partially found below a reinforced concrete slab, built during 
the archaeological excavations two years before, the natural 
light condition was poor, and the laser scanner was preventively 
preferred to carry out the 3D model. The complexity of the site, 
due to the irregularity of the elevation profile of the area, and 
the requested high resolution of the model, required an elevated 
number of scans and an intensive post-processing work as well. 

Thanks to the high quality of the surveys’ results in terms of 
level of detail and degree of realism, and to the potentiality 
allowable by such digital technologies, the archaeologists have 
now the possibility to observe the site overall on a computer, 
from almost infinite points of view, to better recognize the 
directions of the dredging tracks in order to identify the ones 
belonging to a single dredge passage and also a valid support to 
study the possible shapes of the tool used to dredge the sea 
bottom (Figure 8). 

In addition, as the surveys have been topographically 
georeferenced in a global reference system, the combining of all 
the surveys performed - or to be performed - in the 
neighbouring areas, and the possibility to see all the 3D digital 
models together, will permit to have an overall view of the 
whole excavated area. Last, all the a posteriori analysis intended 
to extract in detail measurement information can be performed 
directly on the 3D model and without a specific urgency, 
allowing the contractors the prosecution of the station 
construction and optimizing the infrastructure construction 

 
Figure 8. Dredging tracks found in the line 6 station shaft. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 61 

time as well. 

 B. Sea bottoms stratigraphy 

The use of advanced survey technologies to document the 
stratigraphy sequence of the sea bottoms (Figure 9) entailed to 
plan a process schedule in which a strict interaction between 
surveyors and archaeologists has been necessary.  

A new concept of “measurable” stratum in three-dimensions 
has been adopted, bringing to an important improvement if 
compared to the traditional survey techniques, mainly based on 
2D maps and sparse 3D measured points. 

The “digital assembling” of all the surveyed stratums 
together made possible to reconstruct the entire sequence of 
sediments settled or removed over the centuries, metrically and 
semantically as well, deleting any possible individual 
interpretation [12]. 

Within this archaeological area of the line 6 station shaft, a 
total amount of 16 different sea bottom stratums were 
surveyed: approximatively, a volume of about 5000 m3 was 
removed by following the rules of the archaeologic 
investigations, obtaining a height difference between the first 
and the last stratum surveyed of about 3.60 m (Figure 9 and 
10). 

 C. Thermal Structure 

When a complex ruin has to be excavated, especially if it is 
characterized by great extensions and big dimensions, or in case 
of lack of important structure elements, the understanding of 
the whole structure utility could appear arduous. An 
orthographic vertical view as well as a 3D model of the area can 
give a valid help to understand its shape otherwise not always 
comprehensible just looking at it on the excavation site (Figure 
11). 

The survey of the Roman thermal structure found in the line 
6 station shaft appeared as an extreme difficult task due to the 

complexity of the site and to the restricted working spaces. 
Beyond the survey of each single walls’ façades of the structure, 
a global survey of the entire area, by using a bird’s eye view, was 
necessary. 

Due to the presence of several obstructions that made 
impossible the employment of auxiliary flying units such as 
hanging baskets or UAVs, a flexible and cheap solution has 
been designed and manufactured. In order to take vertical 
photos, the camera was mounted on a horizontally movable 
aluminium frame controlled in one direction by a sort of 
“clothesline” and in a second direction by a movable roof 
previously positioned on two rails to preserve the ruins during 
the excavation progress [8]. 

This stratagem, besides its cheapness, turned up to be a 
good resource to overcome the above described difficulties. 
Thanks to this solution, a true orthophoto (Figure 11) of the 
entire visible parts of the structure was obtained, allowing to 
have a real perception of the building composition. 

The 3D model of the same structure instead (Figure 12), 
obtained by employing a laser scanner, allowed to have a good 
starting point for the prosecution of the excavation activities: 
indeed, a particularity of this structure is the fact that walls 
previously used for different purposes in Hellenistic epoch were 
reused to build it in Roman epoch. 

The surveys performed successively, during the prosecution 
of the excavations and contextually to the disassembling of the 
thermal structures, permitted to document the more ancient 
walls, allowing to digitally reconstruct older situations like a 
backwards travel through the time. 

 
Figure 11. Thermal structure and quay true orthophoto and vector map. 

 
Figure 9. Perspective view of a 3D model of line 6 station shaft excavation
site where the first and last surveyed sea bottoms have been highlighted. 

 
Figure 10. Stratigraphic section carried out from the combined 3D models of
all the surveyed sea bottoms. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 62 

In addition, the possibility to assemble and view together 3D 
models related to surveys performed in different excavation 
moments, even in different years, is an exhaustive tool to 
understand the original uses of the artefacts and the relationship 
among the structure of different epochs. 

The accurate documentation of the much important 
artefacts produced during the excavation progress allowed to 
carry out graphic two-dimensional representation as well, such 
as maps, sections and perspective drawings with wall 
orthophotos façades (Figure 13).  

In addition, stratigraphic sections and 3D multi temporal 
views have been produced in order to show in a clear manner 
the relationship between different epochs structures (Figure 
14). 

 D. Shipwreck “Napoli G” 

At the end of 2014, the partial bottom of a shipwreck, 
including edge-joined hull planks and transversal frames, was 
found (Figure 15). 

Due to the particularity of the archaeological evidence and 
for the location of the finding, there was the necessity to pay 
serious attention to its excavation and to find a quick solution 
to provide exhaustive documentation of the evidence and 
disassemble the shipwreck in as less time as possible. This 
special case required a strictly cooperation between surveyors 
and archaeologists, as the start and the prosecution of the 
disassembling process depended by the preventive and quickly 
production of maps with orthophotos in specific moments, 
useful to the archaeologist to take note of each single 
disassembled part. For these reasons, high resolution 
orthophotos and accurate 3D models were requested.  

The high number of corners for the presence of several 
frames located in their original positon above the bottom of the 
boat, brought to choose the photogrammetry as the best survey 
solution in terms of quality of the final model and time-
consuming. Indeed, with a range-based instrument like the laser 
scanner, a huge number of scans to avoid “holes” on the model 
would have been requested, with a consequent overabundance 
of unnecessary data, an intensive and time expensive editing 
activity and several junction zones between the scans to deal 
with. As additional negative surrounding condition, the boat 
was found partially cut by the bulkheads built before the 
beginning of the archaeological excavation activities. The 
presence of the bulkheads themselves too close to the boat on 
one side would have made impossible the acquisition with a 
laser scanner, while the flexibility of the photogrammetry made 
it possible even if with some difficulty. In order to keep a good 
range of DOF (Depth Of Field) and a suitable degree of 
sharpness in a low light condition, photos were taken with an 
elevated f-number and with the camera set on a tripod. 
Normal/nadiral photos were taken together with a set of tilted 
photos to well reconstruct each particular of the hull and to 
improve the quality of the camera network. A mobile platform 
on which to place the camera was also prepared to move the 
camera itself above the shipwreck. 

 For this evidence, four distinct surveys, which information 
is summarized in Table 5, were performed during the 
disassembling phases, in order to record as much construction 
details as possible. 

Other than being a satisfying documentation, an accurate 3D 
reconstruction of such archaeological found can be a useful 
tool for historians of ancient naval architecture to better 

 
Figure 12. Perspective view of the 3D model of the roman thermal complex
and of the quay. 

 
Figure 13. Architectonic section of the thermal structure together with the
front of the quay. 

 
Figure  14.  3D  multi  temporal  view:  Roman  structures  in  grayscale,
Hellenistic ones in magenta.  Figure 15. Shipwreck “Napoli G”. 



 

ACTA IMEKO | www.imeko.org  September 2016 | Volume 5 | Number 2 | 63 

understand the typology of the vessel, the employed 
construction techniques or to have important elements to 
reconstruct a complete virtual or real model of the vessel 
reducing reconstruction hypothesis. Even a remounting of the 
shipwreck in a museum, as it was found, could be taken into 
consideration as each part was mapped in detail before being 
removed. 

4. CONCLUSIONS 

The results of investigations about the accuracy allow to 
switch between photogrammetry and laser scanner as very really 
efficient method to document Cultural Heritage. The above 
described cases are just a selected group of the most exhaustive 
and interesting evidences found during the excavations 
activities to which advanced survey technologies have been 
applied. Even if these cases could appear as a small part of the 
enormous and continuous survey activity started in 2012 and 
still under way, they fully give an idea of the potentiality of such 
methodologies. The thermal structure in particular is a 
significant example of the possibility to virtually reconstruct the 
several historical stages that came in succession from 
Hellenistic epoch on. 

Beyond the achievable information that can be extracted 
from detailed 3D models, the above shown results demonstrate 
the capability of such technologies to deal with continuous 
excavation activities that take places simultaneously in different 
areas of the construction yard and to perfectly satisfy the 
requirements of all the involved subjects. 

Within the last three years, the improvements introduced in 
the actual data processing photogrammetric software (parallel 
computing, automatic point cloud generation, etc.) brought to 
reorganize the two employed methodologies. According to the 
results of the first part of this paper, much more space has been 
reserved to the photogrammetric approach in the last months, 
sensibly reducing the acquisition time in the perspective of still 
optimize the work. Accurate 3D models in a limited time, 
depending on the complexity of the object, can now be 

obtained with the photogrammetry in a fast manner. In 
addition, this technique demonstrated its capability to be 
employed even in case of outstanding conditions, thanks to its 
flexibility. 

REFERENCES 

[1] L. Quan, Image-Based Modeling, Springer, New York, ISBN 
978-1-4419-6678-0. 

[2] F. Remondino, S. El-Hakim, Image-based 3D modelling: A 
review. The Photogrammetric Record 2006, 2,  pp. 269–291. 

[3] W. Boehler, V. Bordas, A. Marbs, “Investigating laser scanner 
Accuracy”, Proc. of XIXth CIPA Symposium, 30 September–4 
October, 2003,  Antalya, Turkey, pp. 696-702. 

[4] C. Achille, A. Adami, S. Chiarini, S. Cremonesi, F. Fassi, L. 
Fregonese, L. Taffurelli, UAV-Based Photogrammetry and 
Integrated Technologies for Architectural Applications - 
Methodological Strategies for the After-Quake Survey of Vertical 
Structures in Mantua (Italy). Sensors 2015, 15, 15520-15539, 
ISSN 1424-8220, doi:10.3390/s150715520. 

[5] V. Carsana, S. Febbraro, D. Giampaola, C. Guastaferro, G. 
Irollo, M.R. Ruello, “Evoluzione del paesaggio costiero tra 
Parthenope e Neapolis”, in “People/environment relationships 
from the Mesolithic to the Middle Ages, recent geo-
Archaeological findings in Southern Italy”, Méditerranée, 
No.112, 2009, pp. 14-22. 

[6] E. Allevato, E. Russo Ermolli, G. Boetto, G. Di Pasquale, 
“Pollen-wood analysis at the Neapolis harbour site (1st -3rd 
century AD, southern Italy) and its archaeobotanical 
implications”, Journal of Archaeological Science, Vol. 37, 2010, 
pp. 2365-2375. 

[7] D. Giampaola, V. Carsana, G. Boetto, F. Crema, C. Florio, D. 
Panza, B. Pizzo, C. Capretti, G. Galotta, G. Giachi, N. 
Macchioni, M.P. Nugari, M. Bartolini. “La scoperta del porto di 
“Neapolis”: dalla ricostruzione topografica allo scavo e al 
recupero dei relitti”, in Archaeologia maritima mediterranea, An 
International Journal on Underwater Archaeology, No.2, 2005 
(Pisa-Roma), pp. 48-91. 

[8] F. Doglioni, “Stratigrafia e restauro. Tra conoscenza e 
conservazione dell’architettura”, ed. LINT, Trieste, 1997. 

[9] F. Fassi, L. Fregonese, S. Ackermann, V. De Troia, “Comparison 
between Laser Scanning and Automated 3D modelling 
techniques to reconstruct complex and extensive cultural heritage 
areas”, Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., 
Vol. XL-5/W1, 2013, pp.73-80. 

[10] F. Fassi, C. Rossi, A. Mandelli, Emergency survey of remote and 
endangered archaeological sites, Int. Arch. Photogramm. Remote 
Sens. Spatial Inf. Sci., XL-5/W4, pp. 85-91, doi: 
10.5194/isprsarchives-XL-5-W4-85-2015, 2015. 

[11] F. Remondino, Heritage Recording and 3D Modeling with 
Photogrammetry and 3D Scanning. Remote Sensing 2011, 3,  
1104-1138, ISSN 2072-4292, doi:10.3390/rs3061104.  

[12] M. Peripimeno, “Sperimentazione di tecniche 3D laser scanning 
per l’analisi e la conservazione del patrimonio archeologico e 
storico-monumentale. Definizione di procedure e campi di 
utilizzo”, Proc. of National Congress SIFET - Integrazione tra le 
tecniche innovative del rilievo del territorio e dei beni culturali, 
Mondello, 29-31 June 2005. 

 

Table 5. Shipwreck “Napoli G” survey report. 

Survey session  1
st 

2
nd 

3
rd 

4
th 

No. Photos  334  278  285  213 

Images resolution  5616*3744 

Acquisition time 
(mins.) 

90  55  65  65 

Orthophoto 
restitution time 

1 day 

No. points 3D model 
(millions of points) 

117  100  79  74