Content uploaded by Jose Ignacio Fiz Fernández
Author content
All content in this area was uploaded by Jose Ignacio Fiz Fernández on Jun 15, 2021
Content may be subject to copyright.
Remote Sens. 2021, 13, 2344. https://doi.org/10.3390/rs13122344 www.mdpi.com/journal/remotesensing
Article
Tests with SAR Images of the PAZ Platform Applied to the
Archaeological Site of Clunia (Burgos, Spain)
Ignacio Fiz 1,2,*, Rosa Cuesta 3, Eva Subias 2 and Pere Manel Martin 2
1 Catalan Institute of Classical Archaeology (ICAC), Rovira i Virgili University (URV), 43001 Tarragona,
Spain
2 Department of History and Art History, Rovira i Virgili University (URV), 43002 Tarragona, Spain;
eva.subias@urv.cat (E.S.); peremanel.martin@urv.cat (P.M.M.)
3 Diputación de Burgos, Universidad de Valladolid, 09003 Burgos, Spain; rcuesta@diputaciondeburgos.es
* Correspondence: joseignacio.fiz@urv.cat
Abstract: This article presents the first results obtained from the use of high-resolution images from
the SAR-X sensor of the PAZ satellite platform. These are in result of the application of various radar
image-treatment techniques, with which we wanted to carry out a non-invasive exploration of areas
of the archaeological site of Clunia (Burgos, Spain). These areas were analyzed and contrasted with
other sources from high-resolution multispectral images (TripleSat), or from digital surface models
obtained from Laser Imaging Detection and Ranging (LiDAR) data from the National Plan for Aerial
Orthophotography (PNOA), and treated with image enhancement functions (Relief Visualization
Tools (RVT)). Moreover, they were compared with multispectral images created from the Infrared
Red Blue (IRRB) data contained in the same LiDAR points.
Keywords: remote sensing; SAR-X; multispectral images; archaeology; GIS
1. Introduction
In this article, we mainly present results of the exploration and analysis of a set of
Synthetic-aperture radar (SAR) images from the Spanish PAZ satellite applied to non-
invasive archaeological prospecting. The works consisted of the generation of images
from SAR data and comparing the results with those received from the exploitation of
other sources, such as multispectral images (National Plan for Aerial Orthophotograph:
PNOA, TripleSat) and LiDAR. In principle, our project—Application of Radar images of the
PAZ Satellite in the Detection of Archaeological Remains (ARQPAZ, Project AO-001-018)—
focused on the archaeological site of La Clunia, (Figure 1), which is located near the town
of Peñalba de Castro (Burgos, Spain), on a hill called El Alto de Castro.
We know from Salustio (Hist., 2, 93), Tito Livio (Periocas, XCII), Plutarco (Sertorio,
9), or Floro (2.10.9) the first references to the pre-Roman Celtiberian city of Clunia, con-
nected to the Sertorian wars. The sources do not make it clear if it was located in the
nearby hills, or at some point at its final location, after it was founded by the Romans. This
foundation would occur sometime between the end of the rule of Augustus, and that of
Tiberius. Clunia would have an important role during the events of 68 AD, receiving from
Galba the epithet of Sulpicia, becoming the capital of Conventus within the province of
Tarraconense. The hill where the site is located has an area of 1.2 km2, of which only 3%
has been excavated since the second half of the 19th century.
Citation: Fiz, I.; Cuesta, R.; Subias, E.
Tests with SAR Images of the PAZ
Platform Applied to the
Archaeological Site of Clunia
(Burgos, Spain). Remote Sens. 2021,
13, 2344. https://doi.org/
10.3390/rs13122344
Academic Editor: Mi Wang
Received: 19 May 2021
Accepted: 11 June 2021
Published: 15 June 2021
Publisher’s Note: MDPI stays neu-
tral with regard to jurisdictional
claims in published maps and institu-
tional affiliations.
Copyright: © 2021 by the authors. Li-
censee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and con-
ditions of the Creative Commons At-
tribution (CC BY) license (http://crea-
tivecommons.org/licenses/by/4.0/).
Remote Sens. 2021, 13, 2344 2 of 22
Figure 1. Location of Clunia. Sources: National Geographic Institute (IGN), National Plan for Aer-
ial Orthophotography (PNOA), Maximum resolution (20 cm). (a) Location of Clunia in Spain (b)
Location of Clunia in the Douro Valley (c) Clunia and the Alto del Castro.
However, the excavated remains correspond to large public spaces such as the thea-
tre, the forum, thermal complexes (Los Arcos), and large private spaces such as the so-
called Casa Taracena. Furthermore, the urban organization of the city does not correspond
to a homogeneous structure, but on the contrary, the provisions of both large public and
private constructions would add up to four possible different orientations (Figure 2).
These peculiar characteristics, together with the great extension of the site and the
economic consequences of the 2008 financial crisis, have led to the need to explore the hill
with non-invasive methods mainly associated with remote sensing. Previous work by our
team focused on exploring the possibilities of analysing infrared images and LiDAR point
clouds obtained by the state through the National Plan for Aerial Orthophotography
Remote Sens. 2021, 13, 2344 3 of 22
(PNOA) and SPOT-5 multispectral images. These were focused on a single area of the
archaeological site and were combined with geophysical prospecting applied in selected
sectors by the SOT (Catalan for “Hole”) —Archaeological Survey Company. For this sur-
vey, electromagnetic technique, using the Grad601 fluxgate gradiometer manufactured by
the Bartington Instruments Ltd. company, was applied in a very specific sector. The geo-
physical survey was carried out on a series of three grids measuring 20 × 20 metres on
each side, each with N-–S orientation; thus, forming a total area of 20 × 60 metres.
This work allowed, among other results, to determine one of the aforementioned four
orientations that would structure the urban organization of the city [1] (pp. 142–143).
Figure 2. The different urban orientations of the city of Clunia: (a) Structure based on the orienta-
tion of the theatre and Taracena House; (b) Orientation based on the orientation of the forum; (c)
Orientation based on the set of Termas de los Arcos; (d) orientation detected in [1]. Source: Hill
shading from LiDAR DEM (5 m/pixel) PNOA.
On the other hand, the first instrumental uses of SAR started being used in 1978, but
it was not until the mid-1980s that the data from SIR A and SIR B were applied in geo-
archaeological studies associated with ancient riverbeds [2,3], ancient cities such as Ubar
(Oman) [4], and the location of Mayan irrigation canals [5,6]. However, these investiga-
tions, due to the spatial resolution of the sensors at the beginning of the new century, were
limited to still-preserved monuments, cultural landscapes, paleo-landscapes, and canal
systems. Since the launch of the TerraSAR-X and COSMO-SkyMed platforms in 2007, it
Remote Sens. 2021, 13, 2344 4 of 22
has been possible to access high-resolution data at the scale of 1 m from Very High Reso-
lution (VHR) SAR sensors, applying them to the detection of archaeological remains.
However, the penetration capacity of these sensors in the X band is very limited, as com-
pared to others that work in the L and C bands [7] (pp. 71–75). It should be added that
PAZ, such as TerraSAR-X, works on the X band.
The effectiveness of the TerraSAR-X image analysis applied to large old buried infra-
structures was shown in the work prepared by Monterroso and Martinez [8]. These au-
thors located sections of the Roman road near the ancient Roman city of Mellaria (Cor-
doba, Spain) under cereal fields and in conditions of high humidity. On the other hand,
they also found remains of medieval roads, paved with gravel and built without elements
that would allow their drainage. This caused the accumulated humidity to be transmitted
to the soils superimposed on said infrastructures. As conclusions, they indicated that SAR
data acquisition in high humidity and rainfall conditions in cultivated fields revealed bet-
ter potentials; in non-cropped landscapes, however, according to the authors, thermal and
RGB photography analyses were better.
The PAZ satellite, equipped with the SAR sensor, was put into orbit in February 2018
by Hisdesat [9]. Specifically, the instruments consist of an X-Band SAR radar which allows
working in four basic image modes—Spotlight, HR (High Resolution) Spotlight, Strip-
Map, and ScanSAR—with various polarizations—single and dual. After the start of its
operation, a call for an opportunity offer was opened by the National Institute of Aero-
space Technology (INTA). This allowed us to access a set of SAR X-Band images to test its
potential as an exploratory tool. We worked specifically with the Spotlight product, which
has a resolution of 1.25 m / pixel.
Furthermore, the importance of the use of various non-invasive techniques for the
detection of archaeological features in the landscape should be noted. There is varied lit-
erature on this topic, where we can see examples such as the use of multiple vegetation
indices applied to hyperspectral images, and combined with geophysical surveys, which
has allowed the highlighting of crop marks in the buried Roman remains of Carnuntum
(Austria), the underground structures of Selinunte in the south of Italy, and the buried
street relics of Pherai (Velestino) in central Greece [10]. Other cases combining non-inva-
sive techniques applied to underwater environments, such as multi-beam echo sounder
(MBES), and sub-bottom profilers (SBPs), can be found in the reconstruction of a medieval
harbour using hydro-acoustics, 3D shallow seismic and underwater photogrammetry in
the Sothern part of the Baltic Sea [11].
Finally, the ARQPAZ project aims to study the behaviour of radar images in three
different environments associated with three archaeological sites, in which we have previ-
ously worked with non-invasive exploratory investigations: Oxirrinco [12], Cosa [13], and
finally—Clunia. This work presents the results of the first evaluation of the possibilities of
the SAR-X images of PAZ in the archaeological site of Clunia.
2. Materials and Methods Used
Our project focused on the following digital sources of information (Table 1): SAR
images of PAZ, LiDAR point clouds of the National Geographic Institute (IGN) with in-
formation from IRRB bands, images of old NIRGB (Near Infrared Green Blue) flights of
the National Plan for Aerial Orthophotography (PNOA, IGN), and TripleSat multispectral
imaging.
The SAR images, with HH polarization, were captured during the months of July to
October 2019 and 2020, and the LiDAR point clouds correspond to 2010 and 2011. On the
other hand, we also wanted to contrast the results with high-resolution multispectral im-
ages. In this case, we chose images captured by the TripleSat 2 satellite, which is part of a
network together with two other satellites with the same characteristics. The TripleSat
satellite, put into orbit in 2015[14], provides a panchromatic band at 0.8 m/pixel resolution,
Remote Sens. 2021, 13, 2344 5 of 22
and four bands at 3.2m/pixel (NIR, Red, Blue, Green). The acquisition of these multispec-
tral images is tried to coincide with the summer and autumn periods, corresponding to
2016 and 2017.
In the case of the PNOA IRGB photographs, after careful analysis, we used in our
work those corresponding to the years 2007, 2009, and 2011. These flights were conducted
during the summer periods, in 2009 in July, and in 2011—in June/July). We cannot confirm
the specific moment of the 2007 image, since the metadata information indicates the cap-
ture between the months of April and July. The images have a resolution expressed as a
ground sample distance (GSD) of 50, 25, and 50 cm correspondingly.
Table 1. Technical information on the data sources used.
Source Dates Resolution Description
PAZ
Spotlight
29 August 2019;
09 September 2019;
20 September 2019;
01 October 2019;
13 July 2020
26 August 2020;
Range asc: 1.25
m/pixel
Polarization: HH
Ascending
Triple SAT
2
18 Jun 2016;
08 October 2017
PAN 0.8 m/pixel
MS 3.2 m/pixel
PAN: 450/650 nm
Blue: 440/510 nm
Green: 510/590 nm
Red: 600/670 nm
NIR: 760/910 nm
LiDAR
IRRB 10 September 2010;
0.25 pixel/m2
0.5 point/m2
First coverage.
Altimetric accuracy cm:
RMSE Z <=20
IR, R, B bands
PNOA
NIRGB
PNOA 2007 (22 April–28 July);
PNOA 2009 (03 July–13 July);
PNOA 2011 (14 Jun–31 July)
GSD: 50 cm
25 cm
50 cm
NIR, G, B bands
We mainly used the SNAP v8.0 (Sentinel Application Platform) program of the Eu-
ropean Spatial Agency (ESA) developed jointly by Brockmann Consult, SkyWatch, and
C-S, for the treatment of both the supplied radar images, as well as the TripleSat optical
images, and also the free software program QGIS v.310.13-A Coruña, an application under
GNU GPLv2, to visualize the results, superimpose them, and contrast them with other
information from other layers. ESA’s SNAP is a free software, distributed under the GNU
license, and developed and supported by the ESA. It is a software that allows the treat-
ment and analysis not only of images from the Sentinel sensor program, but also from
other optical and radar sensors—among which is PAZ—too.
For working with LiDAR point clouds, in order to generate IRGB images and Digital
Surface Models (DSM) at high resolution, we used the free software GIS SAGA v7.9.0
(software for Automated Geoscientific Analyses), which provides LAS (LASer) point
cloud data treatment functions. Moreover, it has a wide variety of filters and interpolation
systems, as well as pansharpened functions applicable to multispectral images—such as
TripleSat from a panchromatic band. The SAGA development was initiated by researchers
from the Department of Physical Geography of Göttingen and the Dept. of Physical Ge-
ography in Hamburg.
Finally, to perform the enhancement of the digital terrain models obtained from the
area, functions of the Relief Visualization Tools program v2.2.1 (RVT)[15] were applied,
which our team had already applied previously in Clunia. The RVT software has been
developed by a research group from the Slovenian Academy of Sciences and Arts, in a
Remote Sens. 2021, 13, 2344 6 of 22
project managed by Žiga Kokalj. At the time, we relied on the studies carried out by Mas-
sini et al. [16] to locate the structures of a medieval fortification in Basilicata (Italy) hidden
under a thick blanket of forest; the Samnitic Hillforts of Civitella (Longano, Italy); the ar-
chaic and Hellenistic urban centre of Satricum (Lazio) analysed using SAGA functions
[17], or the record of archaeological traces in the Roman limes of Dacia in Porolissum (Ro-
mania) carried out by Roman et al [18].
2.1. SAR Sources, Techniques Used
Initially, the first set of applied techniques focused on the individual treatment of the
collected radar scenes. The workflow, described by Meyer [19, and in the work by Mon-
terroso and Martínez[8] (p. 304), is characterized by the application of SNAP radiometric
functions such as calibration, radiometric terrain flattening, and that of geo-referencing as
Range-Doppler Terrain Correction.
On the other hand, one of the issues that affect the quality of SAR images is related
to noise speckle. This effect is inherent to the system and is the result of the interference
between the various echoes produced as they interact on the surfaces of the objects. In
recent years, filters have been developed that make it possible to eliminate or reduce the
distortion caused in images by this speckle. Meyer [19] (pp. 23-24) gives a list of the filters
used, such as the Lee, Enhanced Lee, Frost, Enhanced Frost, and Non Local Means filters,
and lists the publications on these types of filters [20–27].
Generally, the noise speckle was solved using the single product speckle filtering op-
erations of the SNAP program, specifically using one of the filters most mentioned in the
scientific literature, the Lee Sigma filter, testing different sizes of windows. The mottling
is reduced by averaging the values of the neighbouring cells grouped by windows [8] (p.
304).
However, Tapete and Cigna [28](pp. 11–25) have shown that the joint use of multiple
SAR images at different time points in the same study area can reduce speckle by provid-
ing images of greater clarity than treating them individually. This reduction is carried out
through the co-register function, which implies the creation of a stack of images, with one
referent, and the rest—subordinate. The program then places (collocate function) samples
of the subordinate bands on the reference band using their geographic position and an
automated ground control point (GPC) selection. In this group, the multi-temporal
speckle-filtering function is applied, using the Lee Sigma filter. It is from this point that it
is possible to perform statistical averaging operations that facilitate the generation of a
final image that is more precise and sharp as compared to that obtained from a single SAR
image. This method has also been applied by Stewart [29], focusing on various areas of
Lazio (Italy) with COSMO-SkyMed images, and by the same author (Stewart) [30] at the
archaeological site of Qasrawet (North Sinai) using TerraSAR-X images this time. We
have, therefore, followed the aforementioned procedure to generate an image from a stack
of six images obtained in 2019 and 2020 in the periods from July to October. The result
produces clearly significant results, as can be seen in Figure 3, where we compare the
result of this procedure of applying the multi-temporal speckle Lee filter, with that per-
formed with an individual SAR-X radar image filtered with single product speckle Lee,
applied to the area from the theatre and the forum/Casa Taracena. In addition, various
window sizes were tested by applying the Lee filter, with Figure 3 showing those corre-
sponding to 9 × 9 (boxes labelled a and e), and 7 × 7 (labels b and f).
Remote Sens. 2021, 13, 2344 7 of 22
Figure 3. Comparison of a multi-temporal analysis of SAR-X radar images with individual treat-
ment in the areas of the theatre (a, b, c, d), and Casa Taracena/forum (e, f, g, h).
2.2. LiDAR Sources: Techniques Used
First of all, we downloaded the *.las files corresponding to our work area from the
servers of the National Geographic Institute (IGN). We followed the method described by
Massini et al. [16] (pp. 6–7) for the improvement of the DSM, except in the described part
of creating the surface (mesh) from the point cloud. The reason was not to be able to have
the functions described by the authors for the generation of surfaces, because they are
typical of commercial software. For that reason, and to create the DSM with a resolution
of 0.38 m/pixel, the triangulation function of SAGA was used, based on the Delaunay tri-
angulation. The error and noise points had been filtered in advance, and then from the
resulting point cloud were selected those classified 2 to 6 (base, low vegetation, medium
vegetation, high vegetation, and building). Again, to the recommendations of Massini et
al. [16] (pp. 10–12) regarding the use of the Lee filter. However, their same recommenda-
tions for the Lee filter window size could not be adopted, since the SAGA’s Lee-based
multi-direction filter Lee feature does not allow such a selection. However, according to
the implementers, this function that looks for the minimum variance in 16 directions pre-
serves the edges and is useful in eliminating the speckle noise of the SAR images or the
smoothing of the DTMs, preserving the breaks of slopes and narrow valleys [31].
From here, the various forms of visualization derived from the MDS provided by the
RVT program were generated, such as the Sky-View Factor, Openness Positive and Neg-
ative, Simple Local Relief Model (SLRM), etc.
The second type of treatment was to work on the values of the IR, R and B bands
contained in the information of each of the points of the cloud registered in the LAS files
of the PNOA. In the LIDAR flights, carried out by the PNOA from 2009 onwards, a com-
bined flight was applied, allowing to perform simultaneously the capture of the image
and the altimetric information. As of 2015, a simultaneous capture with a four-band image
(RGB and near infrared) was generalized again, with medium-format cameras. This com-
bined process is described by Lorite et al. [32] as the colour assignment to the point cloud
from orthophotographs from the PNOA Image project, or from the flight itself carried out
to obtain the LiDAR, in the event that this flight included a photographic camera. To give
Remote Sens. 2021, 13, 2344 8 of 22
colour to the point cloud, a process is carried out that consists of assigning to each point
of the same the interpolated RGB (Red Green Blue) value of the set of pixels of the ortho-
photography corresponding to that areas that are in the same position on the ground.
Apart from giving RGB colour, which is what the point cloud is distributed with, inter-
nally, false colour (NearInfrared Red Blue) is also assigned to the points, adding the infra-
red value. Therefore, what we conducted was to use the interpolation functions of SAGA,
triangulation, to first generate individual images for each of the bands contained in the
points, and then make a combination of them in false colour. With the resulting images,
the SAGA function of principal component analysis (PCA) was applied, especially due to
the good result obtained by our team when applying them in the published project on the
Egyptian landscape of the Oxirrinquita nomos [12] with Worldview 2 images, and in an-
alytics carried out in the Clunia project with SPOT multispectral images [1]. PCA is a mul-
tivariate statistical method used to reduce the dimensionality of the data resulting in the
reduction of redundant information. In general, the first component explains the maxi-
mum proportion of variance of the original dataset [33]. PCA was estimated in the work
presented by Aqdus et al. [34], as the most effective visualization tool for multi- and hy-
perspectral images.
On the other hand, the multispectral image obtained from the LiDAR PNOA point
cloud was also applied to vegetation index functions, fifteen in total, provided by the
SAGA GIS program. Let us remember that vegetation indices are effective methods to
detect and quantify plants through remote sensing [35]. However, the variation in data
based on habitat and substrate causes the same index not to work the same in dense for-
ests, pastures, drylands, desert, or humid areas. This has led to the creation of about 150
vegetation indices [36]. With reference to archaeology, these indices are useful for the de-
tection of buried structures since they tend to modify the growth and phenology of the
overlying vegetation [37]. Its use in archaeological prospecting has had wide acceptance
and good results both from remote sensing using satellite sensors, and currently—from
UAVs equipped with multispectral cameras.
2.3. TripleSat Sources
The TripleSat images were treated first by adapting their four NIR bands, R, G, B, at
3.2 m/pixel resolution, to the resolution of the panchromatic, 0.8 m/pixel. The QGIS
pansharpening function was used for obtaining a composite image of 4 bands at a resolu-
tion of 0.8 m/pixel. Next, we proceeded to work with the multispectral image applying
the statistical function of the SAGA principal components analysis, and the aforemen-
tioned vegetation indices provided by the SAGA GIS were applied.
3. Results
3.1. Areas Explored in the Results
We observed the results of all the analyses carried out focusing on four areas of the
Clunia deposit (Figure 4). Two of them are within the limits of the site (forum and theatre
/ forum / Cuevas Ciegas area), another corresponds to the borders of the site and one of
the supposed lines out of the city (Cuevas Ciegas), and finally a fourth, not excavated,
located outside the city walls south of the city and corresponding to a funerary route
(Rodeles II).
Remote Sens. 2021, 13, 2344 9 of 22
Figure 4. Areas explored with the results of analyzing the SAR images of PAZ. (a) Clunia forum.
(b) Forum–theatre–Cuevas Ciegas area. (c) Cuevas Ciegas .(d) Rodeles II. Source: Google satellite
QuickMapServices.
There were various motivations for choosing each of the zones. In the first case, the
forum area (Figure 4a), the reason was the observation of various anomalies in the PNOA
images, also identified in the PAZ image. Surprising anomalies, given that in principle the
space where they appear is the forum square, a space that would have to be free of struc-
tures, unless there had been a previous and/or later occupation of the public space.
In the second case, the theatre/forum /Cuevas Ciegas area (Figure 4b), we wanted to
review the explorations carried out in the article published in 2019. Explorations which,
as mentioned, made it possible to determine a new urban orientation that was added to
those previously known and connected with the monumental spaces of the forum, Casa
Taracena, the theatre, or the Termas de los Arcos. In addition, any information that could
be added to the data of a large area not yet excavated would be of great interest.
The third, Cuevas Ciegas (Figure 4c), had also been the result of a study [38–40,1]
both in terms of the access roads to the city from the plane, and in determining the main
networks drainage of city waters towards the plane.
Finally, the fourth exploration zone, Rodeles II (Figure 4d), was chosen as an element
to assess the capacity of contrasting radar images with quite optimal results for the use of
multispectral images in the detection of anomalies in the landscape.
3.2. The Area of the Forum
In the highest resolution RGB images (Figure 5a), no significant changes are appreci-
ated except for a dark area located NE of the forum plaza, a short distance from the line
of the forum portico. It is difficult to see a line with a NE–SW direction, and a transversal
Remote Sens. 2021, 13, 2344 10 of 22
with a NW–SE direction. On the other hand, is the contrast with the image of the 1956
flight (Figure 5b), in which the separation limits of properties with the NE–SW direction,
and the delimiting wall of the hermitage with the NW–SE direction can be seen first. In
this case, a first reasoning may suggest that the traces seen in the first image are actually
remnants of the parcel separation. However, in the photograph of the American flight
1957, there can also be seen two darker zones regarding the rest in the same area, a central
area of the square, and another on its NE side.
Figure 5. (a) Google satellite QuickMapServices. (b) USAF flight 1957.
The possibility of encountering an abnormality in this area increases when we go on
to analyse the IRGB images of the PNOA flights in 2007, and the IRGB image obtained
from LiDAR in 2010. In the first case (Figure 6), we clearly evaluate the lines correspond-
ing to the land division and the separation wall of the hermitage (aligned black and white
triangles), but also observed in the area’s centre in front of the temple, and the lateral NE
of the forum (ellipses dashed line in black) increases in a hue of red due to vegetation
growth. In the second, the IR R B image (Figure 7a) allows to see, in the central part, the
reverse situation in two areas (polygon with a blank dashed line). This indicates a slower
growth of vegetation and a lack of humus area most to this increase, creating two areas in
the centre. The PCA analysis (Figure 7b), indicates that those construction elements (build-
ings, roads, etc.) present a red hue, which is repeated in the two areas mentioned above.
Remote Sens. 2021, 13, 2344 11 of 22
It should be noted that, in this case, both images are apparently present on the W side of
the square, shown as an arrangement of perpendicular lines.
Figure 6. PNOA 2009, 25 cm. Bands NIRGB.
Figure 7. (a) LiDAR IR R G image; (b) Image result of the first component obtained from a princi-
pal component analysis (PCA) applied to the LIDAR IRGB image. Note in the latter how the red
Remote Sens. 2021, 13, 2344 12 of 22
color, associated with altered areas, is observed in the central plaza of the forum and on the
NE/SW side.
The analysis of the 2016 TripleSat image (Figure 8a) shows another abnormality in
the area in front of the NE side of the porch of the forum (ellipse dotted line in black). In
this case, apparently, it gives rise to two quadrangular delimited spaces. The NE–SW
alignments do not coincide with the crop parcel subdivisions from the 1956 flight. The
PCA analysis of the multispectral image shows a similar result in the first component
(Figure 8b).
Figure 8. (a) TripleSat image from 18 June 2016, bands IR, R, G; (b) First component analysis prin-
cipal component applied to image TripleSat of 18 June 2016.
Remote Sens. 2021, 13, 2344 13 of 22
Finally, when we go on to analyze the results of the X-SAR images (Figure 9) of the
PAZ satellite, we find in the central area, in front of the temple, an anomaly with white
tones. This anomaly presents alignments with the NW–SE and NE–SW directions creating
apparent bounded spaces. When we are on the ground, it is not possible to detect reveal-
ing details, such as micro reliefs, that may be associated with the presence of construction
elements, coinciding in principle with what is expected for what would be the forum
square. Perhaps some detail, such as the one indicated in Figure 10, can be appreciated,
but nothing that could suggest that we are on structures or the presence of micro reliefs.
Figure 9. Image resulting from the multi-temporal application of the co-register function with
five SAR-X images of PAZ.
Figure 10. View of the forum from the temple. Arrows in black indicate an apparent alignment in
front of the temple.
Remote Sens. 2021, 13, 2344 14 of 22
We wanted to perform a deeper analysis of the X-SAR image. To do this, we pro-
ceeded to draw eight lines with a NE–SW orientation with the intention of obtaining the
intensity reading profiles (Figure 11). The SNAP program allows the creation of these pro-
files from the tracing of polylines on the image. Observation of the graphs of profiles ap-
plied to the obtained X-SAR image shows a first peak, or followed by a concavity we have
identified as the border of the portico of the forum and the drainpipe. From here, we can
find that the peaks of the curves tend to occur successively in the same position of each of
the samples taken. This could be indicating the presence of longitudinal intrusive ele-
ments (vertical lines in Figure 11). On the other hand, in some of the profiles we evaluated
the presence of profiles, in which the peak is maintained for several meters with few al-
terations (horizontal lines in Figure 11), indicating that we may be facing intrusive longi-
tudinal elements, coinciding approximately with the orientation of the profile line. There-
fore, apparently, it seems that the radar is detecting possible construction elements.
Figure 11. Profiles of eight samples taken from the multi-temporal SAR-X PAZ radar image. The
red lines are indicating a repetition rate of reading the signal, and, therefore, the presence of
aligned elements.
Remote Sens. 2021, 13, 2344 15 of 22
3.3. The Theatre / Forum / Cuevas Ciegas Area
As mentioned, it was in this area where our team [1] had determined the presence of
traces of urban organization with a N/S orientation, in this case applying other techniques
and digital sources, such as SPOT-5 images. In this case, we have contrasted results of the
radar image (Figure 12b) with the application of the SLRM function (Figure 12c) to the
DTM obtained from the LiDAR point cloud. The use of the second type technique pro-
duces better results than in the first case. These can be seen in Figure 12 (a).
Figure 12. (a) Google satellite QuickMapServices; (b) Result of the simple local relief model
(SLRM) function on the DSM obtained from LiDAR PNOA; (c) Image resulting from the multi-
temporal application of the co-register function with five SAR-X images of PAZ.
It is necessary to take into consideration the profound modern high celling division,
observed on the American flight 1956, and which is barely visible today. This greatly con-
ditions and distorts the results in trace detection. It is true that the plot division and sub-
division itself may be in result of a morphological evolution of the urban organization of
Remote Sens. 2021, 13, 2344 16 of 22
Clunia. However, the period of abandonment, sedimentation, agricultural work, and pil-
laging material from its abandonment, in turn may have masked elements as some of the
traces that, if detected in the radar image, have a similar orientation as that of Casa Tara-
cena, and they do not appear in the image of the 1956 flight, nor in the results of applying
the SLR function to the DEM obtained by LiDAR.
3.4. The Cuevas Ciegas Area
In this case, the observation of the treated image of PAZ (Figure 13c) has provided
the identification of two anomalies associated, on the one hand, with one of Clunia’s waste
water systems (Figure 13a), identified in previous works [1] through the application of
GIS functions for the creation of a water network from the DEM. This anomaly is not ob-
served in the rest of the multispectral images IR R B Lidar, PNOA, or TripleSat, except for
the PNOA IRGB image of 2009, and it is only seen in the 1956 flight. This anomaly is not
observed directly on the ground in some change of vegetation (Figure 14).
Figure 13. (a) Google satellite QuickMapServices; (b) USAF flight 1957; (c) Image resulting from
the multi-temporal application of the co-register function with five SAR-X images of PAZ.
Remote Sens. 2021, 13, 2344 17 of 22
The second anomaly that can be displayed in the image PAZ course corresponds to
the section of one of the paths to Clunia studied by Camacho [39-41] and that also is only
observable in flight USAF 1957 (Figure 13b).
Figure 14. Detailed photo with a black-line indication of the scrapping channel.
3.5. The Burial Area, Los Rodeles II
Around the Alto de Castro, there are three large funerary spaces distributed from
south to east, between the middle area of the hillside and the plain of the Arandilla River.
The most obvious sample can be found on a terrace at the foot of the Cuevas Ciegas access
point, in the site called Torreón, where the remains of an imposing structure identified as
a mausoleum are preserved, surely dedicated to a prominent figure [39](p.268). It is, there-
fore, a necropolis on the access road to the city, which is also connected to the presence of
other possible funerary monuments at nearby points on the same hillside. In its vicinity is
the funeral route located in the fertile plain of Arandilla, registered as Rodeles II. In the
latter case, aerial photography made it possible to determine that the site is made up of a
series of structures and enclosures associated with a road that crosses the river [41]. There
is no doubt that these are funeral precincts, such as those mentioned above, distributed
on both sides of the road [40](pp.267–268) [38](pp. 220–222).
The observation of the site of Rodeles II (Figure 15) has allowed us to find PNOA
images from 2011, in which it was possible to identify the structures almost at the same
level of visualization and clarity of the aerial photographs taken in this area. These struc-
tures can also be seen in the infrared image of LiDAR of the PNOA flight. To complete the
works, we carried out PCA analytics and a calculation of the vegetation indices, among
which the Perpendicular Vegetation Index [42] (Figure 15e) has stood out. When com-
pared with the radar image resulting from the multi-temporal treatment, we saw that the
media operation has left evidence of the anomaly of the track, but lost the information
corresponding to the structures associated with the funeral area. For this reason, we car-
ried out an individual treatment of one of the PAZ images, the one from 29 August 2019,
which reveals the presence of the aforementioned structures that perpendicularly cross
the funeral via (Figure 15b).
Remote Sens. 2021, 13, 2344 18 of 22
Figure 15. (a) PNOA 2011 50 cm (Bands B, B, IR); (b) PAZ 29 August; (c) USAF flight 1957; (d) Li-
DAR PNOA IR RG (Bands R R G); (e) Perpendicular Vegetation Index (Walther and Shabaani); (f)
Google satellite QuickMapService..
4. Discussion
According to Stewart et al. [30], remote sensing with SAR has only been fully suc-
cessful in areas covered by sand. On the other hand, Dore et al. [43] maintained that the
characteristics of these types of SAR sensors (independent of the external light source,
penetration areas cloud cover, or soil penetration) have extended the limits of acquisitions
Remote Sens. 2021, 13, 2344 19 of 22
from optical satellites, and the combined use of both types of sensors in archaeological
remote sensing research was advisable. Stewart himself [29] also showed the potential to
also use multispectral reflectance and radar in a combined manner, although in this case
he delved into the use of interferometric coherence. This combination could help in the
detection and interpretation of changes that could affect archaeological sites, such as, for
example, illegal excavations.
However, the use of SAR has had results of interest in non-arid and vegetated areas,
as is the case in Lazio [29,30] with the detection of archaeological forms, combining results
of SAR intensity and interferometry. In both cases, the probability of locating archaeolog-
ical structures has increased.
According to Monterroso and Martinez [8], it is in spaces of cultivated soils where it
seems easier to interpret the SAR-X images compared to bare soils, as Stewart et al.
[44](p.204, 208) in the Portus de Ostia. In addition, continued Monterroso and Martinez,
this interpretation was maintained in cases in which there are periods of intense rain with
soils humidity indexes higher than the annual index.
In the current case, the selection of PAZ images was carried out in meteorological
contexts in months with significant differences between the maximum and minimum high
temperatures, with little rainfall. It should be noted, however, that the best possible pic-
ture has display elements such as the zone Rodeles II, precisely performed after one of
two rainy pulsations occurred in August 2019 (Table 2, T.Max: 25.07; T.Min: 14; T.Med:
19.9 Prec.: 2.4 mm). Let us also bear in mind that the space where this deposit is located
corresponds precisely to cultivated soils, or in cultivation period.
Table 2. Source data from the AEMET meteorological station of Coruña del Conde (Burgos). (Web
source database: https://es.meteosolana.net/estacion/2106B, accessed on 23 April 2021)
Month /
Year
T (Average)
Max
T (Mean)
Min
T. Aver-
age
Prec
(mm)
Images PAZ in-
volved
Aug 2019 29.6 o C 14.4 o C 22 o C 3.6 1
Sep 2019 24 o C 9.9 o C 17 o C 72 2
Oct 2019 18.9 o C 8 o C 13.5 o C 35.8 1
Jul 2020 31.6 o C 14.3 o C 22.9 o C 16.8 1
Aug 2020 28.6 o C 13.5 o C 21.1 o C 15 1
On the other hand, we have seen that the application of the co-register function with
a stack of temporally spaced images provides more clarifying results than the individual
analysis of each of the images, except in one of the areas—the one corresponding to the
Rodeles II.
Our analysis of the SAR images confirms one of the first utilities, and it is the possi-
bility of detecting hydraulic or road infrastructures as we had indicated when analysing
the area of Cuevas Ciegas or Rodeles II, or also, elements that can help to complete one of
the most interesting aspects of the study of Clunia as its urban organization may be. On
the other hand, in Rodeles II, we were able to confirm that better results are obtained on
cultivated soils since these remains were found in a cultivated area in front of spaces of
the site itself that stopped being cultivated at the end of the 20th century.
It remains to be determined whether the anomalies apparently identified in the forum
area actually correspond to building structures. It is interesting to have verified that the
anomalies were detected by analysing sources of information as varied as radar or multi-
spectral images.
5. Conclusions
Above all, throughout this work, we were able to verify the possibilities of using
SAR-X radar images in non-invasive archaeological prospecting. However, we under-
stand up to the time of our research, and as other authors have already indicated, that the
Remote Sens. 2021, 13, 2344 20 of 22
use of this technology must be accompanied by other sources in result of remote sensing,
such as multispectral optical images or thermal images. In that sense, our future interest
lies in adding other analyticals such as polarimetry SAR, as described by Stewart, Lasa-
ponara and Sciavon [45], or to perform flights at low-altitude UAV using thermal and
multispectral cameras, combined with geophysical surveys to check the results. Recent
works are already developing by us in other archaeological areas. As we mentioned at the
beginning of this work, we are experimenting with the PAZ data in other areas with dif-
ferent geographical characteristics, such as the already mentioned Oxirrinco (Egypt) or
Cosa (Italy). The same SAR-X image analysis techniques described in the text were ap-
plied. However, it is still pending a review and comparison with other sources, such as
multispectral images. Our long-term objective is to review and determine in greater detail
the physical and seasonal conditions in which SAR-X data can provide the best results in
the search for archaeological elements.
We see that the application of a multi-temporal treatment on the radar images helps
to enhance the resulting images against individual scenes.
On the other hand, the detection of anomalies in the area of the forum using various
non-invasive analysis systems opens up new urban unknowns associated with the mo-
ments before and after the construction of the forum that will have to be verified in archae-
ological interventions.
Author Contributions: Conceptualization, I.F. and R.C.; methodology, I.F.; software, I.F. and
P.M.M.; validation I.F., R.C. and E.S.; investigation, I.F., R.C. and E.S.; resources, I.F. and R.C.; data
curation, I.F. and P.M.M.; writing—original draft preparation, I.F.; writing—review and editing,
I.F. and P.M.M.; supervision, R.C. and E.S.; project administration, I.F. All authors have read and
agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: The authors would like to thank the INTA-PAZ Science Team for providing
the PAZ data in the framework of the “AO-001-018” project (018 Application of Radar images of the
PAZ Satellite in the Detection of Archaeological Remains, ARQPAZ). Also our thanks to the Rovira i
Virgili University for granting financial support for this open access publication.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Cuesta, R.; Fiz, I.; Subías, E.; Tuset, F.; Iglesia, M.A. Hydraulic and urban management during Roman times based on GIS and
remote sensing analysis (Clunia, Spain). Rev. d'Arqueol. de Ponent 2019, 29, 123–146, doi:10.21001/rap.2019.29.4.
2. McCauley, J.F.; Breed, C.S.; Schaber, G.G.; McHugh, W.P.; Issawi, B.; Haynes, C.V.; Grolier, M.J.; Kilani, A.E. Paleodrainages of
the Eastern Sahara. The Radar Rivers revisited. IEE Trans. Geosci. Remote Sens 1986, 24, 624–647.
3. McCauley, J.F.; Schaber, G.G.; Breed, C.S.; Grolier, M.J.; Haynes, C.; Issawi, B.; Elachi, C.; Blom, R. Subsurface Valleys and
Geoarchaeology of the Eastern Sahara Revealed by Shuttle Radar. Science 1982, 218, 1004–1020.
4. Blom, R.; Crippen, R.; Elachi, C.; Zarins, J.; Clapp, N.; Ledges. G. Space Technology and the discovery of the lost city of Ubar.
In Proceedings of the IEEE Conference: Aerospace Conference, Snowmass, CO, USA, 13 February 1997; Volume 1.
5. Adams, R. Swamps, Canals and the Locations of Ancient Maya Cities. Antiquity 1980, 54, 206–214.
6. Sever, T.L.; Irwin, D.E. Remote sensing investigation of the Ancient Maya in the Peten Rainforest of Northern Guatemala. Anc.
Mesoam. 2003, 14, 113–122.
7. Lasaponara. R., Massini. N. Satellite Synthetic Aperture Radar in Archaeology and Cultural Landscape: An Overview. Archaeol.
Prospect. 2013, 20, 71–78.
8. Monterroso, A.; Martinez, T. COSMO SkyMed X-Band SAR application—Combined with thermal and RGB images—in the
archaeological landscape of Roman Mellaria (Fuente Obejuna-Córdoba, Spain). Archaeol. Prospect. 2018, 25, 301–314.
9. Hisdesat, PAZ Image Product Guide, PAZ-HDS-GUI-001. 2019. Available online: https://www.hisdesat.es/wp-content/up-
loads/2019/10/PAZ-HDS-GUI-001-PAZ-Image-Product-Guide-issue-1.1-.pdf (accessed on 01/04/2021).
10. Cerra, D.; Agapiou, A.; Cavalli, R.M.; Sarris, A. An Objective Assessment of Hyperspectral Indicators for the Detection of Buried
Archaeological Relics. Remote. Sens. 2018, 10, 1–25, doi:10.3390/rs10040500.
Remote Sens. 2021, 13, 2344 21 of 22
11. Pydyn, A.; Popek, M.; Kubacka, M.; Janowski, L. Exploration and reconstruction of a medieval harbour using hydroacoustics,
3-D shallow seismic and underwater photogrammetry: A case study from Puck, southern Baltic Sea. Archaeol. Prospect. 2021, 1,
1-25.
12. Subias, E.; Fiz, I.; Cuesta, R. The Middle Nile Valley: Elements in an approach to the structuring of the landscape from the
Greco-Roman era to the nineteenth century. Quat. Int. 2013, 312, 27–44.
13. Roca, M.; Fiz, I. Reconstrucción, a partir de fotografía aérea, de la topografía de la colonia de Cosa (Ansedonia, Italia) in Roca.
In Proyecto Cosa: Intervenciones Arqueológicas de la Universidad de Barcelona en la Ciudad Romana; Madrid, M., Celis, R., Eds.; Uni-
versidad de Barcelona: Barcelona, Spain, 2013; pp. 69–89. Available online: http://diposit.ub.edu/dspace/bit-
stream/2445/99120/6/Proyecto_Cosa_optimitzat.pdf (accessed on 01/04/2021).
14. Wen, Q.; Hea, J.; Guana, S.; Chena, T.; Hua, Y.; Wua, W.; Liua, F.; Qiaoa, Y.; Kokb, S.; Yeong, S. The TripleSat constellation: A
new geospatial data service model. Geo. Spat. Inf. Sci. 2017, 20, doi:10.1080/10095020.2017.1329266.
15. Kokalj, Z.; Hesse, R. Airborne Laser Scanning Raster Data Visualization: A Guide to Good Practice; Založba ZRC: Ljubljana, Slovakia,
2017.
16. Masini. N.; Gizzi, F.T.; Biscione. M.; Fundone, V.; Sedile, M.; Sileo. M.; Pecci, A.; Lacovara, B.; Lasaponara, R. Medieval Archae-
ology under the Canopy with LiDAR. The (Re)Discovery of a Medieval Fortified Settlement in Southern Italy. Remote Sens. 2018,
10, 1598, doi:10.3390/rs10101598.
17. Garcia Sanchez, J. Archaeological LiDAR in Italy: Enhancing research with publicly accessible data. Antiquity 2018, 92, 1–10.
18. Roman, A.; Ursu. T.; Lăzărescu. V.; Opreanu, C.H. Multi-sensor surveys for the interdisciplinary landscape analysis and ar-
chaeological feature detection at Porolissum in Coriolan. In Landscape Archaeology on the Northern Frontier of the Roman Empire at
Porolissum an Interdisciplinary Research Project; Opreanu, H., Lăzărescu V.A., Eds.; Mega Publishing House: Cluj-Napoca, 2016;
pp. 237–262.
19. Meyer. F. Spaceborne Synthetic Aperture Radar: Principles, Data Access, and Basic Processing Techniques in Ixmucane. In The
Synthetic Aperture Radar (SAR) Handbook: Comprehensive Methodologies for Forest Monitoring and Biomass Estimation; Flores-Ander-
son. A.E., Herndon. K., Bahadur Thapa. R., Cherrington. E., Eds.; SERVIR Global Science Coordination Office: 320 Sparkman
Drive, Huntsville, AL 35805 2019.
20. Bruniquel, J.; Lopes, A. Multi-variate optimal speckle reduction in SAR imagery. Int. J. Remote Sens. 1997, 18, 603–606.
21. Ferretti, A.; Fumagalli, A.; Novali, F.; Prati, C.; Rocca, F.; Rucci, A. A New Algorithm for Processing Interferometric Data-Stacks:
SqueeSAR. IEEE Trans. Geosci. Remote Sens. 2011, 49, 3460–3470.
22. Huang, S.-Q.; Liu, D.-Z.; Gao, G.-Q.; Guo, X.-J. A novel method for speckle noise reduction and ship target detection in SAR
images. Pattern Recognit. 2009, 42, 1533–1542
23. Lee, J.-S.; Grunes, M.; Mango, S. Speckle reduction in multipolarization, multifrequency SAR imagery. IEEE Trans. Geosci. Re-
mote Sens. 1991, 29, 535–544.
24. Lee, J.-S.; Jurkevich, L.; Dewaele, P.; Wambacq, P.; Oosterlinck. A.Speckle fltering of synthetic aperture radar images: A review.
Remote Sens. Rev. 1994, 8, 313–340.
25. Lopez-Martinez, C.; Pottier, E. On the Extension of Multidimensional Speckle Noise Model from Single-look to Multilook SAR
Imagery. IEEE Trans. Geosci. Remote Sens. 2007, 45, 305–320.
26. Novak, L.M.; Burl, M.C. Optimal speckle reduction in polarimetric SAR imagery: Aerospace and Electronic Systems. IEEE Trans.
Geosci. Remote Sens. 1990, 26, 293–305.
27. Sveinsson, J.R.; Benediktsson, J.A. Almost translation invariant wavelet transformations for speckle reduction of SAR images,
IEEE Trans. Geosci. Remote Sens. 2003, 41, 2404–2408.
28. Tapete. D.; Cigna. F. COSMO-SkyMed SAR for Detection and Monitoring of Archaeological and Cultural Heritage Sites. Remote
Sens. 2019, 11, 1–25.
29. Stewart. C. Detection of Archaeological Residues in Vegetated Areas Using Satellite Synthetic Aperture Radar. Remote Sens.
2017, 9, 118, doi:10.3390/rs9020118.
30. Stewart. C.; Oren. E.; Cohen-Sasson. E. Satellite Remote Sensing Analysis of the Qasrawet Archaeological Site in North Sinai,
Remote Sens. 2018, 10, 1090, doi:10.3390/rs10071090.
31. Selige, T.; Böhner, J.; Ringeler. A. Processing of SRTM X- SAR data to correct interferometric elevation models for land surface
process applications. Göttinger Geogr. Abh. 2006, 115, 97–104.
32. Lorite, S.; Ojeda, J.C.; Rodríguez-Cuenca, B.; González Cristóbal, E.; Muñoz, P. Procesado y distribución de nubes de puntos en
el proyecto PNOA-LiDAR. In Nuevas Plataformas y Sensores de Teledetección, Proceedings of the Actas del XVII Congreso de la Asocia-
ción Española de Teledetección, Murcia, Spain, 3–7 October 2017; Ruiz, L.A., Estornell, J., Erena, M., Eds.; Universidad de Valencia:
Valencia, Spania, 2017; pp. 329–332. Available online: https://n9.cl/2y9yf (accessed on 14/06/2021).
33. Zhao, G.; Maclean, A.L. A comparison of canonical discriminant analysis and principal component analysis for spectral trans-
formation. Photogramm. Eng. Remote Sens. 2020, 66, 841–847.
34. Aqdus, S.; Hanson, W.B; Drummond, J. The potential of hyperspectral and multi-spectral imagery to enhance archaeological
cropmark detection: A comparative study. J. Archaeol. Sci. 2012, 39, 1915–1924.
35. Eastman J.R. IDRISI for Windows: Guide to GIS and Image Processing Version 32.20, 2001; Clark University: Worcester (MA, USA),
2001; Volume 2.
Remote Sens. 2021, 13, 2344 22 of 22
36. Verrelst, J.; Koetz, B.; Kneubühler, M.; Schaepman M. Directional sensitivity analysis of vegetation indices from multiangular
CHRIS/PROBA data. In Proceedings of the ISPRS Commission VII Mid-term Symposium Remote Sensing: From Pixels to Pro-
cesses, Enschede, The Netherlands, 8–11 May 2006.
37. Lasaponara, R.; Masini, N.; Scardozzi, G. Immagini satellitari ad alta risoluzione e ricerca archeologica: Applicazioni e casi di
studio con riprese pancromatiche e multispettrali di Quickbird. Archeol. e Calc. 2007, 18, 187–227.
38. Camacho. G. El territorio de Clunia y su evolución entre los siglos I a. C. y X d. C.: Perspectivas arqueológica e histórica. Tesis
Doctoral, Universidad de Barcelona, Barcelona, Spain; Available online:
https://www.tdx.cat/bitstream/handle/10803/666964/GCV_TESIS.pdf?sequence=1, 2018 (accessed on 01/04/2021).
39. Camacho. G. Clunia: Una perspectiva arqueológica. CLIO. Hist. Hist. Teach. 2014, 40, 17–42.
40. Camacho, G. Aproximación descriptiva a las vías y accesos de la Colonia Clunia Sulpicia (Peñalba de Castro-Burgos). Hisp.
Antiqva 2013, 37, 249-270; Available online: https://dialnet.unirioja.es/descarga/articulo/5075608.pdf. (accessed on 01/04/2021)
41. Del Olmo. J. Arqueología aérea en Clunia. Rev. de Arqueol. 2001, 244, 6–9.
42. Walther, D.; Shabaani, S. Large scale monitoring of rangelands vegetation using NOAA/AVHRR LAC data: Application to the
rainy seasons 1989/90 in northern Kenya. In Range Management Handbook of Kenya; 1991; Ministry of Livestock Development,
Nairobi.
43. Dore. N.; Patruno, J.; Pottier. E.; Crespi, M. New Research in Polarimetric SAR Technique for Archaeological Purposes using
ALOS PALSAR Data. Archaeol. Prospect. 2013, 20, 79–87.
44. Stewart, C.; di Iorio, A.; Schiavon, G. Analysis of the utility of Cosmo Skymed strip map to detect buried archaeological features
in the region of Rome. Experimental component of WHERE project. In Towards Horizon 2020: Earth Observation and Social Per-
spectives, Proceedings of the 33rd EARSeL Symposium Matera, Italy, 3–6 June 2013; Lasaponara, R., Masini, N., Biscione, M., Eds.;
European Association of Remote Sensing Laboratories (EARSeL): Matera, Italy, pp. 203–212.
45. Stewart. C.; Lasaponara. R.; Schiavon. G. Multi-frequency, polarimetric SAR analysis for archaeological prospection. Int. J. Appl.
Earth Obs. Geoinf. 2014, 28, 211–219.
