Contento

Journal of Archaeological Science 107 (2019) 10–22
Contents lists available at ScienceDirect
Journal of Archaeological Science

journal homepage: www.elsevier.com/locate/jas
Theory and practice for an object-based approach in archaeological remote T

sensing
Luigi Magnini, Cinzia Bettineschi∗
Department of Cultural Heritage: Archaeology and History of Art, Cinema and Music – University of Padova, Italy
A R T I C LE I N FO A B S T R A C T
Keywords: Object-based image analysis (OBIA) is rapidly emerging as a valuable method for integrating the data processing
OBIA techniques and GIS approaches classically employed in archaeology. OBIA is intended to replicate human per-
Archaeology ception by using a protocol of (semi)automated image segmentation and classification. However, the lack of a
Landscape evolution theoretical background adapted to the specificities of the archaeological discipline is still preventing researchers
Knowledge integration
from finding a shared language and a common protocol of investigation necessary to allow the comparability of
Image interpretation
Diachronic semantic models
the results.
Archaeo-objects This article discusses a series of crucial theoretical issues linked to the incompleteness and the equi-/multi-
finality of the archaeological record and introduces the core concept of Diachronic Semantic Models (DhSM) as a
means to integrate the long-term evolution of the archaeological landscape in the conceptual, digital and real-
world frameworks of the object-based approach.
We also present an assessment of the limits and potential of this method, built from a set of case studies from
published and unpublished research. Finally, we propose a general workflow of an Archaeological Object-Based
Image Analysis (ArchaeOBIA) project, designed for stimulating the development of an operational routine for
object-based applications in archaeology.
1. Introduction interpretation of the remotely sensed imagery, even in the archae-

ological field. This approach has a long history in the domains of en-
In recent years, archeological research has seen an increasing vironmental, material and biomedical sciences (Heidrich et al., 2013;
number of remote sensing (RS) applications with the use of new sensors Caie et al., 2016; Feuchtinger et al., 2016; Hawkins et al., 2016), but its
and data types, such as multi/hyper-spectral imagery (Traviglia, 2011; potential is yet to be fully exploited for systematic research on cultural
Lasaponara, Masini, 2012; Doneus et al., 2014; Agapiou et al., 2014; De heritage.
Guio, 2015; Moriarty et al., 2019), radar (Wiseman, El-Baz, 2007; As previously pointed out, the number of papers dealing with ob-
Lasaponara, Masini, 2013; Chen et al., 2016; Tapete, Cigna, 2017; ject, pattern and scenery recognition (OPSR) of archaeological contexts
Burigana, Magnini, 2018) and LiDAR data (Bewley et al., 2005; is still very limited and includes applications of template matching,
Devereux et al., 2005; Doneus et al., 2008; Challis et al., 2011; Opitz, machine learning, convolutional neural network (CNN), custom algo-
Cowley, 2013), that have joined the classic aerial photographs. How- rithms and object-based methods (see Traviglia, Torsello, 2017 and
ever, these innovations had only little impact on the traditional pho- Davis, 2018 for a general overview on the topic). In this paper, we will
tointerpretation, which remains essentially a work for the human op- focus our discussion on object-based image analysis (OBIA or GeOBIA,
erator via visual inspection (Brophy, Cowley, 2005; Cowley, 2015; with a geographic connotation) which was described as “an evolving
Crutchley, 2015; Wilgocka et al., 2016; Quintus et al., 2017). paradigm with specific tools, software, methods, rules, and language
The reduction of the instrumental costs and the exponential increase (that) is increasingly being used in studies which need to conceptualize
in the volume of datasets of the last few years prompts for an overall and formalize knowledge representing location based reality” (Blashke
revision of the methods traditionally used in archaeology (Bennett et al., 2014).
et al., 2014). In this context, the automation or semi-automation of In order to promote the interoperability of the rule-sets, it is ne-
image analysis seems to offer an opportunity to speed up and grant cessary to make explicit what it is implicit in the classification and
better reproducibility for the classification and the subsequent interpretation process. For this purpose, we propose a theoretical
∗
Corresponding author.
E-mail addresses: [email protected] (L. Magnini), [email protected] (C. Bettineschi).
https://doi.org/10.1016/j.jas.2019.04.005
Received 3 January 2019; Received in revised form 20 April 2019; Accepted 26 April 2019
Available online 06 May 2019
0305-4403/ © 2019 Elsevier Ltd. All rights reserved.
L. Magnini and C. Bettineschi Journal of Archaeological Science 107 (2019) 10–22
framework aimed at formalizing expert archaeological knowledge using influence of the shape parameter during the segmentation can offer
ontologies (i.e. formal, explicit specifications of a shared con- better results in the task of landform classification (Eisank et al., 2011).
ceptualization, according to Gruber, 1993). Moreover, we introduce the As previously noted, image-objects produced via image segmenta-
concept of Diachronic Semantic Models (DhSM), developed to better tion possess a series of intrinsic (object features) and extrinsic (class-
explain the long-term evolution of the landscape in machine-readable related features) descriptors which can be used to direct the classifi-
language (§ 4.1). cation process. In fact, the classification phase is intended to distinguish
As object-based applications make their way into archaeological image-objects into meaningful classes based on their attributes and
practice, it becomes increasingly important to find a shared language relationships and according to the specific aims of the research
and a common protocol of investigation, ideally passing from opera- (Castilla, Hay, 2008). The selection of the classification parameters and
tional practice to operational routine. In this paper, we suggest a gen- their threshold values can be both derived from selected training areas
eral workflow for OBIA applications in archaeology built from a wide and a set of derived statistics (standard nearest neighbor classification,
range of published and unpublished case-studies to ease the compar- NN) or directly evaluated by the operators according to their expert
ability of data. Finally, we argue that there is a growing urgency to find knowledge (rule-based classification). It was demonstrated that rule-
a common way for publishing rule-sets and rule-set libraries to be semi- based classification outperforms pixel-based and object-based NN
automatically or automatically implemented for archaeological in- classifications in the accuracy of the results (Gibril et al., 2017). It
vestigations. This topic is of general interest for the OBIA community, should, however, be noted that the expert knowledge is subjective and
but it should be stressed with a particular emphasis in view of further cannot be used as such for the creation of exportable rule-sets, if not in
increasing the role of object-based applications in archaeology, as most very limited cases (Andres et al., 2012) as it will be better discussed
of the operators have a humanistic background and, consequently, later.
longer learning times in the development of customized rule-sets. The most common software for object-based image analysis is
eCognition Developer, owned by Trimble Inc.; however, in recent years
2. Overview of the method the interest for Free and Open-Source Software (FOSS) for GeOBIA is
progressively increasing as testified by numerous papers employing
2.1. The OBIA approach combined solutions using Orfeo ToolBox, R, GRASS GISS, QGIS or
Doker (Van De Kerchove et al., 2014; Böck et al., 2016; Grippa et al.,
RS imagery are composed of pixels (or voxels, in a 3D coordinate 2016; Knoth, Nüst, 2017) but also by the development of specific tools
system), whose dimension is a function of the sensor used and of the such as GeoDMA (Körting et al., 2013) and InterIMAGE (Costa et al.,
parameters employed for the acquisition (Gonzalez et al., 2008). While 2010).
pixel-based classifications rely only on the information contained in
each single pixel (Lillesand et al., 2004), the basic entity of OBIA is 2.2. OBIA in archaeology: a general outline of the literature
represented by image-objects (sometimes also called image-segments)
(Hay et al., 2001; Blaschke et al., 2004). An image-object is “a discrete Automation is still a controversial issue in archaeological photo-
region of a digital image that is internally coherent and different from interpretation (Hanson, 2008, 2010) as clearly highlighted by the quote
its surroundings” (Castilla, Hay, 2008); each image-object is char- “Why does there even need to be an automated process for satellite
acterized by a set of additional spectral, textural, morphometric and archaeology?” (Parcak, 2009). The answer is strictly entangled with the
relational parameters that can be used to fine-tune the results of the potential of OPSR to speed-up the examination of large amounts of data
image classification if compared to the per-pixel approach (Baatz et al., and to grant a better reproducibility to the task of image analysis,
2008; Gamanya et al., 2009; Wuest, Zang, 2009; Blaschke et al., 2014). contributing at the same time to the management and protection of the
Moreover, image-objects can represent geographic objects (sensu archaeological record (see, among others: Magnini et al., 2017; Davis,
Castilla and Hay, 2008) more accurately than single pixels and, thus, 2018; Lasaponara, Masini, 2018).
they offer an improved basis for classification. Despite the challenges and misunderstandings between the tradi-
Image analysis using object-based methods can be essentially di- tional archaeologists and the RS experts, the scientific literature on the
vided in two main sequential steps: segmentation and classification. subject is rapidly growing (Traviglia et al., 2016). Among the papers
Segmentation is used to partition the image into homogeneous regions, dealing with (semi)automated OPSR in archaeology, however, only a
called primitive-objects. However, primitive image-objects are not al- minority employ object-based image analysis (Lambers, Traviglia,
ways meaningful, i.e. they do not always readily correspond to the real- 2016; Davis, 2018).
world entities (Castilla, Hay, 2008). This is the reason why numerous The earliest employ of OBIA in archaeology dates back to 2007,
cycles of segmentation and classification are sometimes employed for when Jahjah et al. (2007) applied this approach for change detection
further refining. analysis around the area of the ancient Babylon (Iraq). In the same year,
The typology of segmentation algorithms is wide (Pal, Pal, 1993) De Laet et al. (2007) tested a procedure for the extraction of archae-
and in constant increase, starting from the simplest (such as the ological structures from multispectral satellite images comparing a set
chessboard and the quad-tree algorithms) to the most complex proce- of different automated methods and visual inspection. The outcomes of
dures (as the multi-threshold or the multiresolution segmentation). The this last case study were discouraging and contributed to a contraction
multiresolution segmentation (MRS) is generally recognized as one of in the archaeological applications of OBIA stricto sensu. However,
the best performing solutions both in the fields of biomedical sciences starting from the same period, there was a progressive growth of OPSR
and RS. MRS is a bottom-up technique designed to emulate human applications (such as Bescoby, 2006; Menze et al., 2006; De Boer, 2007;
perception. Basically, it maximizes intra-object homogeneity while Trier et al., 2009; Menze, Ur, 2012; Trier, Pilø, 2012; Schuetter et al.,
maximizing inter-object heterogeneity (Baatz, Shape, 2000). The algo- 2013; Caspari et al., 2014; Schneider et al., 2015; Sanger, 2015;
rithm starts from generator-pixels called seeds and groups adjacent Toumazet et al., 2017; Guyot et al., 2018; Trier et al., 2018; Matos-
pixels in numerous subsequent steps. When the computed heterogeneity Machado et al., 2019), which are well worth mentioning here as they
surpasses the threshold value defined by the scale parameter, the pro- provided the foundation for a combined use of OBIA and OPSR, which
cess interrupts and image-objects are generated (Benz et al., 2004). is one of the most recent trends in automated detection (Davis et al.,
Simplifying, the higher is the scale parameter, the bigger will be the 2019).
resulting image-objects (Drăguţ, Blaschke, 2006). The homogeneity It was only in 2012–2013 that object-based image analysis stricto
criterion can also be manually adjusted by operating on the shape and sensu was given a second chance, this time for automating the deli-
compactness of desired values. It has been stated that omitting the neation and classification of landforms starting from Digital Elevation
11
Models (DEMs) (Verhagen, Drăguţ, 2012, 2013). The promising results the everyday field practice. The recent birth of eCognition Essentials
of this project triggered a new season of experimentation. In 2013, the (Trimble Inc.) has tried to overcome this problem and to offer an in-
method was used with success in the identification and classification of tuitive workspace for object-based classifications, especially with NN.
mountain pools on aerial photographs for ethnoarchaeological purposes However, the Developer version still retains the most complete collec-
(De Guio et al., 2013). In the same period, it was employed for the tion of tools and algorithms which are crucial for rule-based classifi-
textural characterization of ancient marbles from petrographic micro- cation of archaeological RS data.
graphs (Hofmann et al., 2013). The following year, OBIA was tested for Segmentation is a second point of interest. This phase is in fact the
analyzing magnetic anomalies deriving from geophysical surveys most criticized step of an OBIA project, as highly dependent on the
(Pregesbauer et al., 2014) and it was cited as an engaging novelty for personal choices of the operator. Different supervised and unsupervised
archaeological RS in a theoretical paper by Sevara and Pregesbauer methods have been proposed over the years to automate the selection of
(2014). An article on semi-automatic photointerpretation of aerial the scale parameter, as recently synthetized in the review paper by
images combined with near infra-red (NIR) data was published in 2015 Zhang and Du (2016). However, the problem is yet to be completely
(De Guio et al., 2015). From 2016 to 2018, there was a steep increase in solved, as segmentation is highly dependent on the data source, the aim
the quality and quantity of the publications, pointing to rapid ad- of the study, the geographical context, the surrounding contrast and the
vancements in the near future. Sevara et al. (2016) compared pixel- internal heterogeneity (Zhang et al., 2018). In general terms, it was
based and object-based image analysis in two test areas starting from stated that over-segmentation is usually preferable than under-seg-
different LiDAR visualizations. Moreover, Freeland et al. (2016) re- mentation (Witharana and Civco, 2014); in fact, any specialized soft-
ported the results of a study in the Kingdom of Tonga, where they ware offers the chance to refine the image-objects to better fit with the
compared the results of the GeOBIA approach with an inverted pit- real-world objects during all the steps of the workflow. This is parti-
filling algorithm called iMound. OBIA was also experimented for the cularly obvious for archaeological case studies, where an apparent over-
first time at artifact-level for classifying a set of prehistoric stone tools segmentation is often necessary for granting the necessary detail to
based on morphometric characteristics (Lamotte, Masson, 2016). A traces that may appear secondary in the investigated context, such as in
custom procedure was tested by Cerrillo-Cuenca (2017) for the detec- the case of evanescent crop-marks in cultivated and uncultivated fields,
tion of megalithic barrows on LiDAR data. Furthermore, it proved within a network of agricultural channels and rural houses (De Guio
helpful in predicting the location of control places suited for human et al., 2015).
occupation in mountainous environment (Burigana et al., 2017) and for The incompleteness of the residual record is a major problem ar-
mapping and monitoring of the vanishing heritage connected to the chaeology. This is even more evident in the field of RS, because this
First World War (WWI) (Magnini et al., 2017). New approaches were approach mostly relays on mediation factors (vegetation/soil/snow/
employed on Maya sites for the classification of the land cover con- shadow-marks, micro-morphology, etc.) which can sometimes limit the
nected with different types of archaeological structures (Inomata et al., interpretative potential of the work, if disconnected from ground-truth
2017); in the USA, the method was employed for the identification of assessment strategies. In fact, the archaeological remains are always
ethnoarchaeological charcoal hearths (Witharana et al., 2018) and, affected by post-depositional processes which progressively alter their
again, for mound and shell-ring detection (Davis et al., 2018, 2019). original characteristics, with different degrees of impact (Schiffer,
What emerges from this general chrono-history of OBIA applications 1972; Wood, Johnson, 1978; Nash, Petraglia, 1987; Leonardi, Balista,
in archaeology is that the method is generally used to identify small, 1992; Harris, 1997). These processes can often lead to multifinality, i.e.
round archaeological structures: most often mounds, but also barrows, a high variability in the physical outcomes of the same category of
charcoal pits and shell-craters. The source datasets for landscape-level evidences (von Bertalanffy, 2003; Forbes, 2017). On the other hand,
analyses are usually constituted by LiDAR-derived DEMs and their vi- different archaeological entities can end up sharing similar character-
sualizations. However, aerial and satellite imaging (both orthophotos istics at a certain point of their morphogenetic path, so much as to be
and multispectral data) were also used, sometimes in a multi-layer not easily distinguishable from a remote point of view. This problem
process. Moreover, it should be noted that there is a wide variability in can be framed in the general context of equifinality, which accounts for
the scale of the analyses, from the regional to the microscopic level the possibility of a convergent or similar behavior in open natural and
(Hofmann et al., 2013; Heidrich et al., 2013; Gerisch et al., 2018). anthropic systems of high complexity (Fig. 1) (von Bertalanffy, 2003;
Graham, Weingart, 2015; Forbes, 2017).
3. Results and discussion Moreover, the archaeological traces on RS imagery are often asso-
ciated to a complex palimpsest of historical and contemporary infra-
3.1. Practical issues structures of both natural and anthropic origin. Their presence affects
the spatial continuity and/or the integrity of the ancient remains and
Despite the strengths and opportunities offered by OBIA, a series of can significantly hamper a proper semi-automatic recognition.
drawbacks and weakness must be taken into account when considering As previously noted, human perception can differ widely from a
the slow emergence of this method in archaeological RS. The criticism digital classification based on numeric values. For example, despite the
of conservative archaeologists has significantly slowed down in the last straightforward nature of mountain pools in Alpine environment, ap-
few years as the technological advances in the field of computer-aided plying the object-feature “roundness” is often not enough to correctly
OPSR have demonstrated their significant contribution to the research identify them all (De Guio et al., 2013). This kind of logical con-
(Bennett et al., 2014); however, a series of practical issues is still pre- sequentiality has value in the conceptual domain, but can rarely be
venting from an extensive diffusion of the method beyond the restricted applied to real-world case studies. The presence of outliers (i.e. the twin
circle of RS experts. pools) requires a revision of the original mental model and the use of
The basic obstacle should probably be associated to the software different morphometric, textural, chromatic or relational parameters
solutions (both proprietary and FOSS) for object-based image analysis, able to incorporate the widest variability of the considered archae-
which are far from being user-friendly. Back in the early days of ological entities.
GeOBIA, Hay and Castilla (2008) noted that “under the guise of ‘flex- Further limits sometimes highlighted in the literature are the re-
ibility’, some commercial object-based software provides overly com- producibility of the method and the exportability of the rule-sets
plicated options, resulting in time-consuming analyst ‘tweaking’”. If this (Sevara et al., 2016; Freeland et al., 2016). These questions are strictly
was true for computer scientists less than a decade ago, it is even more entangled with the willingness of sharing the descriptors and para-
true for most of today's archaeologists that have just managed to reach a meters used in the classification process and are primarily connected to
systematic integration of Geographic Information Systems (GIS) into the operators rather than with the method per se. Nevertheless, both are
12
Fig. 1. The multifinality (top) and equifinality (bottom) problem from an archaeological RS perspective.
also rooted in the dilemma posed by the so-called semantic gap, which is conceptualizations can be also regarded in terms of mental models, that
the lack of coincidence between the information extracted from the are meant to simplify complex real-world situations and are often dis-
visual data and the interpretation given by one or more experts on a cipline-specific (Bishr, 1998). The same author describes semantics and
given dataset (Smeulders et al., 2000). In this sense, the association of semantic modeling as “the relationship among the computer re-
semantics based on formal and material ontologies (Husserl, Moran, presentations and the corresponding real world feature within a certain
2001) to the archaeological domain seems one of the most promising context”. This process involves three main domains of knowledge: 1)
answers, as it will be better discussed in the following paragraphs. the real-world domain, made up of concrete entities (e.g. an existing
village, cemetery, road); 2) the conceptual domain, constituted by ab-
stract ideas of real-world entities based on expert knowledge; 3) the
3.2. Towards a theoretical framework
digital domain, associated to the virtual representation of real-world
entities in computer-generated data.
The application of OBIA in archaeology is essentially based on ex-
However, the application of this approach to the archeological re-
pert knowledge. This means that the archaeologist retains an essential
search is not straightforward as the chronological depth introduces a
role in the whole process of image classification and data interpreta-
new variable in the system. Consider the case of an ancient settlement.
tion. Hence, automation should be regarded as an aid rather than a
In the real-world domain, it was subject to specific transformations
substitute of traditional visual inspection. At the same time, “the pro-
(during use, obliteration, destruction, re-use …) that altered its original
cess transforms object-based image analysis into a type of computer-
characteristics irreversibly. When trying to formalize a general on-
aided photointerpretation in which two experts analyzing the same data
tology, it is necessary to consider its history through time: how could
will obtain two different results because of their different experiences”,
have it looked like in the past? How could it be now, after the trans-
as it was already pointed out in the field of Earth Observation (EO)
formations induced by the post-depositional processes in that specific
(Arvor et al., 2013).
context? And finally, how does it effectively appear, from a RS per-
With this in mind, it is now necessary to propose an explicit theo-
spective?
retical framework to provide a common ground for further develop-
In extreme synthesis, landscape archaeology is the study of the re-
ments of the method in the archaeological practice and to foster object-
lationship between humans and nature in a diachronic perspective. The
based applications in archaeology towards higher levels of heuristic
combined transformations in the landscape promoted by these two
awareness. In turn, this will help to minimize the human bias in-
actors trigger the formation and the evolution of the archaeological
troduced during the classification, contribute to the performance of
record that can be studied through RS imagery. As shown in the scheme
(semi)automatic image interpretation and ease the interoperability of
of Fig. 1, this uneven and often unpredictable succession of anthropic
the data.
and natural actions can cause both a high rate of morphogenetic
In this chapter, we discuss the use of ontologies as a mean to for-
variability within the same class of archaeological evidences (multi-
malize archaeological knowledge. We also stress the role of the ar-
finality) and similar outcomes deriving from the nonlinear development
chaeological landscape, which is a four-dimensional entity derived from
of different entities (equifinality). This palimpsest results in varying
long-lasting human/nature interactions. Moreover, we define the no-
degrees of shift between the conceptualized idea of a specific archae-
tion of archaeological objects in the context of OBIA and we use them to
ological evidence and its physical, real-world appearance.
explain the complexity of the archaeological palimpsest. By doing so,
For this reason, semantic models should be applied with a dia-
we also suggest a possible method to address the problem of classifying
chronic perspective. Fig. 2 summarizes the biunivocal connections be-
multi-/equi-final entities.
tween the digital, the conceptual, and the real-world domains in the
study of the archaeological record at landscape-level. At the same time,
3.2.1. Diachronic Semantic Models (DhSM) it introduces the core concept of Diachronic Semantic Models (DhSM),
The use of ontologies and semantic modeling has been proposed on which integrates the idea of evolution/transformation in the for-
various occasions as an approach for linking the conceptualized idea of malization of ontologies derived from expert knowledge. In other
the geographic entities and their digital representations (among others: words, having a clear picture of how an ancient context should have
Dehn et al., 2001; Eisank et al., 2011; Arvor et al., 2013; Ghazouani looked like is not enough. One needs to be aware of the possible
et al., 2018). As already noted, ontologies were defined as a formal, modifications that occurred during the millennia to develop an efficient
explicit specification of a shared conceptualization (Gruber, 1993) ea- conceptualized model that can be translated in machine language and
sily transferrable in machine-readable language. Shared
13
Fig. 2. Graphical representation of the biunivocal connections linking the real-world, the conceptual and the digital domains for modeling the evolution of the
archaeological landscape through time.
used to maximize the results of a semi-automatic image analysis. prevents the possibility to fully model or predict them. As all archae-
Further consequence is that classification does not necessarily co- ological objects share a certain degree of unpredictability, our defini-
incide with interpretation. While classification can be neutral (para- tions of simple and complex archaeo-objects will also relay on the
metrically speaking), interpretation is subjective because it relies both concept of heterogeneity, according to what was proposed by
on the information extracted from the actual data and from additional (Weinberg, 1975) who developed a classification based on small-
knowledge sources which can be a priori and contextual (Ghazouani number, middle-number and large-number systems.
et al., 2018). Interpreting is giving meaning; in the case of OBIA, it Essentially, simple archaeo-objects are made up of one or a few
consists in assigning a semantic label to a group of image-objects that homogeneous parts with similar behaviors and can be defined with a
share a set of meaningful parameters. According to this view, devel- limited set of descriptors and parameters. Despite their unpredict-
oping ontologies based on DhSM is an essential point within the OBIA ability, it is however possible to model them with a decent level of
framework because it helps in formalizing explicit knowledge models accuracy within a given context (Fig. 3). It seems no coincidence that
which can be shared among the archaeological RS community and thus most object-based archaeological case studies have dealt with the in-
offer a common ground in the image interpretation and decision- vestigation of simple archaeo-objects such as mounds (among others:
making domain. Kramer, 2015; Freeland et al., 2016; Davis et al., 2018), shell-craters
(Magnini et al., 2017) and charcoal hearths (Witharana et al., 2018).
Complex archaeo-objects comprise a low to high number of het-
3.2.2. Archaeological objects: simplifying complexity erogeneous parts, require a vast amount of information to be described,
In computer vision, ‘objects’ are autonomous portions of the real- are not fully predictable and can be modeled as a whole in a given
world defined by specific properties which express an enduring identity context only with a low degree of accuracy. In other words, despite
in contrast with the characteristics of the surroundings (Smith, 2001). being generally perceived by the human observer as a single entity,
To avoid any misunderstanding, it should be stressed that objects are complex archaeo-objects are composed by elements belonging to one or
real-word entities, while image-objects in OBIA are self-coherent seg- more hierarchically interconnected sub-classes. This is clearly ex-
ments of a digital image (Castilla, Hay, 2008). emplified by the remains of an ancient settlement (Fig. 4a) or by the
According to our proposal, archaeological objects (in short, archaeo- fortification published in Sevara et al. (2016).
objects) constitute a sub-set of geographic objects as defined by Castilla It is interesting to note that even archaeological objects that might
(2003). Specifically, archaeo-objects are scale-dependent, bounded be conceptualized as simple, in specific environmental (season, level of
geographical areas with archaeological significance that can be iden- disturbance, etc.) and representational (sensor typology, scale, etc.)
tified as the residual referent of the original archaeological record at a conditions are more consistent with the category of complex archaeo-
specific stage of its morphogenetic path. Despite the focus on the cur- objects. Consider for instance the case of the paleo-channel and the
rent situation, archaeo-objects are intrinsically dynamic and their ancient road in Fig. 4b, where the presence of modern structures affects
transformations through time can be broadly modeled using DhSM. the spatial continuity and alters the integrity of the crop/soil-marks to
Demarcating the boundaries is the result of a cognitive process and such an extent that the traces can be no more classified with a limited
can be viewed as a conceptual activity; the borders of an archaeo-object set of descriptors and parameters (e.g. high ratio between length and
may be gradual or abrupt, and eventually a single archaeo-object may width, uniformity in color) (De Guio et al., 2015).
be physically split up by the presence of modern disturbance. However, For this reason, it is crucial to work on multiple layers with a variety
as Castilla and Hay (2008) have demonstrated, this constitutes the of datasets and/or to use OBIA in conjunction with other OPSR ap-
peculiarity of certain geo-objects with respect to the conventional proaches for accurately identifying objects at landscape scales (e.g., De
geographic entities that are ontologically autonomous. Guio et al., 2015; Cerrillo-Cuenca, 2017; Davis et al., 2018).
It is possible to distinguish two main categories of archaeological The formal distinction between simple and complex archaeo-objects
objects: simple and complex. Complex systems theory distinguishes is important as it can help to address the problem of semi-automatic,
between simple, complicated and complex systems (Forbes, 2017). digital classification of multifinal entities or of complex archaeo-ob-
Simple systems are composed by few elements, are fully predictable and jects, which are a long-lasting problem in applying an object-based
can be easily modeled. Complicated and complex systems, instead, approach to archaeological RS.
possess a higher number of parts and require more information to be As already noted, complex archaeo-objects can be seen as a hier-
described. Yet, while complicated systems are fully predictable, com- archically interrelated group of simple archaeo-objects. Hence, the next
plex systems are characterized by a certain degree of entropy that
14
Fig. 3. Classification of simple archeo-objects: an example of WWI shell craters in Alpine environment (68).
step in the analysis of complex archaeo-objects is to split up the overall first classified according to their post-depositional history and their
DhSM into single, hierarchically interconnected elements, to evaluate appearance on the LiDAR visualizations in the three sub-classes: re-
their presence/absence in the scene and then to reconstruct the com- stored, filled and residual. Secondly, the three classes were merged to
plex object through relational parameters between the individual sub- obtain a general representation of the class “trench”.
classes. The opportunity to create specific rule-sets for each element The expression Archaeological Object-based Image Analysis or
and to modulate them in a global rule-set through a hierarchical pro- ArchaeOBIA was first introduced in its French form (ArchéOBIA) by
cedure opens new perspectives for the use of this method also in a Lamotte and Masson (2016), who intended to promote “une métho-
multiplicity of palimpsestic archaeological contexts. dologie d'extraction automatique d'informations quantitatives sur des
Fig. 5 shows a preliminary classification of the multifinal class collections archéologiques quelle que soit la période” (“a methodology
“trench” related to the WWI trenching system around the Austro- for the automatic extraction of quantitative information from archae-
Hungarian fort in Luserna (province of Trento, Italy). The trenches were ological collections regardless of the period”). The authors limited the
Fig. 4. Classification of complex archeo-objects. (a1) Multispectral WorldView 2 satellite image of the ancient village of Tebtynis (Egypt), with classification of the
excavation progress (a2) for time series analysis; (b1) aerial image of Ponte Moro (Po Plain, Italy) with crop/soil-marks related to a Bronze Age infrastructural system
interrupted by a modern hydric and road network, (b2) with classification of a paleo-channel (blue) and an ancient crossroad (yellow). (For interpretation of the
references to color in this figure legend, the reader is referred to the Web version of this article.)
15
Figure 6. ArchaeOBIA levels of application. (a) regional-level: classification of

“control places” in Alpine environment (Isarco Valley, Bolzano Province, Italy)
starting from LiDAR data; (b) local-level: classification of the perimetral bank of
the Bronze Age Terramare settlement of Castello del Tartaro (Po Plain, Italy);
(c) item-level: classification of the recalcified osteological tissue in a Bronze Age
Fig. 5. Complex archeo-objects: the case of the WWI trenching system around cranium of the Olmo di Nogara necropolis (Po plain, Italy) after medical sur-
Fort Lusern (province of Trento, Italy). (a) Classification according to the post- gery; (d) classification of the glassy matrix and the different types of crystalline
depositional history in 3 different classes: restored trenches (yellow), filled inclusions in a SEM-BSE image of a Ptolemaic glass sample from Egypt.
trenches (blue) and residual trenches (green); (b) merging of the previous
classes in the multifinal class “trenches”. (For interpretation of the references to a series of numerical values which can be used for further statistical
color in this figure legend, the reader is referred to the Web version of this
processing. The final phase is the systematic validation of the results by
article.)
means of ground/aerial and geophysical surveys, excavation trenches,
remote cross-validation or literature research to minimize the bias de-
field of applications of object-based methods in archaeology to the rived from the equivocality of the archaeological record. According to
morphometric computation of material remains at item-level, in terms our view, assessment needs to be fully integrated in the workflow and
of artifacts and ecofacts (Lucas, 2012). However, as we hope to point not treated as a separate, optional component (as previously suggested
out throughout this chapter, the potential of the OBIA approach impacts in e.g. Ainsworth et al., 2013; Bennett et al., 2014; Freeland et al., 2016;
various fields of the archaeological research, from EO at regional and Magnini et al., 2017; Davis, 2018). At a following stage, verified rule-
local scale, to material studies at item and microscopic levels (Fig. 6). sets can be exported and applied in different (but similar) contexts, with
eventual refining to cope with differences in scale, contrast, resolution
3.3. ArchaeOBIA in action etc. of the new case study.
For this reason, we propose a conceptual revision of the term. In

particular, ArchaeOBIA will be employed throughout the text when 3.3.1. A matter of scale
referring to the use of object-based methods in archaeology within the The following paragraph will provide a scalar overview of OBIA in
theoretical framework of DhSM and archaeo-objects. The workflow archaeology, with the aim of exemplifying the variety of the possible
concept of an ArchaeOBIA project is summarized in Fig. 7. The en- applications and the potential of the method in answering specific ar-
visaged workflow is constituted by 5 main interconnected steps, com- chaeological questions. The discussion will distinguish four scale-levels,
prising: data input, segmentation, classification, data output and vali- the first two related to EO and the others to material studies: 1) the
dation, which can eventually lead to refining or to a direct re- regional-level considers extensions bigger than fifteen square kilo-
application of the rule-set. The first step includes the know-why re- meters; 2) the local-level ranges between fifteen square kilometers and
presented by the mental model of the operator formalized as DhSM and metric scale; 3) the item-level goes from metric to millimetric scale; 4)
the raster/vector files that constitute the basic data for the analysis. the sub-item, microscopic level considers magnifications in the order of
Subsequently, segmentation and classification are performed in itera- millimeters to nanometers (Fig. 6). The thresholds used are arbitrary
tive cycles to produce meaningful image-objects; the more suitable but based on the average values derived from a selection of OPSR/OBIA
descriptors and parameters for the specific case study are evaluated case studies published in the archaeological literature (see Table 1).
accordingly. The results consist of new raster/vector files, a rule-set and The earliest approaches of regional-level OBIA in archaeology can
16
Fig. 7. Chart representing the general workflow of an ArchaeOBIA project.
be traced back in the works of (Veraghen, Drăguţ, 2012, 2013), who group further explored the potential of object-based image analysis for
used the method to create geomorphological maps from LiDAR data for archaeological predictive modeling, by testing a protocol for locating
assisting the archaeological predictive modeling. The results are of high “control places” in two test areas in Alpine environments (Northern
significance from a methodological point of view, as they show the Italy). The rule-set was developed on the Western Asiago Plateau
reliability and speed of OBIA in automatically classifying landforms at (Vicenza province, Veneto) (Burigana et al., 2017) and subsequently re-
regional scale. Starting from those pioneering experiences, our research applied automatically in the Isarco Valley (South Tirol). Firstly, we
Table 1
A selection of 35 archaeological papers using OBIA (white background) and other OPSR approaches (gray background) divided per scale range. If more than one case
study is present in the same publication, the various areas analyzed are separated by semicolons (;). Whenever the dimensional information is not explicitly defined in
the text, we estimated the values using the images available (signaled in the table with *).
Article Year Dimension of Area/s Scale Notes
Hoffmann et al. (2013) 2013 micrographs/not reported Microscopic

Aprile et al. (2014) 2014 *10–12 mm2 Microscopic Area of each thin section (14)
Bettineschi (2018) 2018 10 mm2 Microscopic Mean area of 64 SEM-BSE images
Hein et al. (2018) 2018 6 mm2 Microscopic Area of each thin section (59)
Lamotte, Masson (2016) 2016 *46 cm2 Item Mean area of 52 artefacts
Magnini (2017) 2017 8.87 cm2; 86.91 cm2 Item
De Laet et al. (2007) 2007 *4–5 km2 Local
Jahjah et al. (2007) 2007 7.74 km2 Local
Schneider et al., 2015 2014 total not reported/*8–9 km2 (validation) Local
De Guio et al. (2015) 2015 0.78 km2; 1 km2 Local
Kramer (2015) 2015 *3.5–4 km2; *1 km2 Local
Sevara et al. (2016) 2016 0.26 km2; 0.9 km2 Local
Magnini et al. (2017) 2017 4 km2 Local
Toumazet et al. (2017) 2017 2 km2 Local
Lasaponara, Masini (2018) 2018 *0.24 km2; *0.75 km2 Local
Witharana et al. (2018) 2018 10 km2; 10 km2 Local
Davis et al. (2019) 2019 *10 km2; *15 km2; 3 km2 Local
De Boer (2007) 2007 12 km2; 15 km2; 52 km2 Local/Regional
Caspari et al. (2014) 2014 not reported Local/Regional?
Bescoby (2006) 2006 42 km2 Regional
Menze et al. (2006) 2006 38400 km2 Regional
Trier et al. (2009) 2009 not reported/no scale bar Regional?
Menze, Ur (2012) 2012 23000 km2 Regional
Trier, Pilø (2012) 2012 29.3 km2; 400 km2 Regional
Veraghen, Drăguţ (2012) 2012 192 km2 Regional
Schuetter et al. (2013) 2013 69 km2 Regional
Veraghen, Drăguţ (2013) 2013 192 km2 Regional
Freeland et al. (2016) 2016 259 km2 Regional
Burigana et al. (2017) 2017 52 km2 Regional
Inomata et al. (2017) 2017 441 km2 Regional
Magnini (2017) 2017 110 km2 Regional
Davis et al. (2018) 2018 2481 km2 Regional
Guyot et al. (2018) 2018 246.7 km2 Regional
Trier et al. (2018) 2018 0.92 km2 (training); 400 km2 Regional
Matos-Machado et al. (2019) 2019 100 km2; 2.9 km2; 4.2 km2 Regional
17
considered the physiographic, climatic and morphological character- expectancy of the human subjects after medical surgery.1 The possibi-
istics of the selected areas. After a preliminary evaluation of the local lity to reapplying the original rule-set on various crania optimizes time
landscape, it soon became apparent that numerous factors classically and workload, providing an objective and repeatable methodology for
used for predictive modeling, such as the availability of water, timber, the measurement and therefore the comparison of the different wounds.
pastures and quarries, were essentially ubiquitous in the area (Burigana The accuracy was evaluated with reference to a repeated set of manual
et al., 2017). For this reason, we applied three DEM processing tech- trials. The results obtained are essentially comparable with the manual
niques: slope, local dominance and solar radiation. Classification was extraction of the measures; yet, there is a substantial advantage in terms
implemented with a rule-set that selected only image-objects with high of time necessary to perform the task. Tentative tests using pixel-based
local dominance and solar radiation, but low slope, which were con- classification, instead, offered very little advice as they were unable to
sidered more suitable parameters for human occupation and territorial detect the textural changes that distinguish the original bone from the
control (Fig. 6a). recalcified tissue.
The project returned five areas on the Asiago Plateau that were Again, OBIA has been tested for the analysis of archaeometric data
ground controlled, confirming an anthropic exploitation covering a only in very limited occasions, as well as OPRS in general (e.g. Aprile
time span from protohistory to WWI. Subsequently, the same model et al., 2014; Hein et al., 2018). Looking at OBIA applications, Hoffmann
was applied to the Isarco Valley to verify the replicability of the et al. (2013) proposed a method for the extraction of mineral grains in
method. The procedure resulted in 36 potential “control places”, the microscopic images of marbles thin sections. The data were used as a
majority of which (24) found good correspondence with the archae- basis for the morphological and textural measurements intended for the
ological sites known in the area. The remaining 12 archaeo-objects identification of the provenance of the material. Recently, Bettineschi
identified by the predictive model were remotely controlled analyzing (2018) employed OBIA for analyzing SEM-BSE (Scanning Electron Mi-
orthophotos and, where available, a LiDAR-derived DTM with 0.5 m croscope – Backscattered Electrons) images selected to represent the
resolution. The interpretation of the data confirmed a possible an- most significant color classes of the investigated set of opaque archae-
thropic exploitation covering a time span from protohistory (Bronze/ ological glasses2 (Fig. 6d). After the classification, the data were
Iron Age) to the XIX century for half of them (Magnini, 2017). The quantitatively analyzed to obtain indications on the number of the
outcomes of the analysis proved the feasibility of the approach that can coloring and opacifying inclusions, their maximum and average di-
be exported and applied to similar mountainous landscapes for site mensions and on the ratio between the total volume of the glassy matrix
predictivity analysis. and the volume of the crystalline inclusions, porosity excluded (as
A significant portion of the archaeological efforts in the use of OBIA proposed in a per-pixel approach by Artioli et al., 2008). The accuracy
were performed at local scale (Fig. 6b, Table 1). The aim of the works is in terms of phase identification and classification was evaluated based
generally to implement an automated approach for identifying and/or on a combination of chemical analysis and visual inspection on a subset
mapping archaeological sites or specific types of ancient (infra)struc- of the 64 analyzed micrographs and it returned a preliminary value of
tures in a given area (De Laet et al., 2007; Jahjah et al., 2007; Kramer, over 90%. This objective and reproducible method provided a quanti-
2015; De Guio et al., 2015; Witharana et al., 2018; Davis et al., 2019). tative and qualitative estimation of the textural characteristics of the
In certain cases, mapping is not only devoted to quantifying the ar- different glasses, offering new hints on the production technologies
chaeological record, but also performed for monitoring its evolution used and on the standardization of the manufacturing processes.
through time in view of improving the management of the cultural It should be stressed that, up to now, the application of OBIA to
landscape (Sevara et al., 2016; Magnini et al., 2017). material studies is mostly devoted to quantification, rather than iden-
The average rate of true positives in landscape-level OPRS and OBIA tification. In fact, there are multiple benefits in employing OBIA as an
applications, calculated from the review published by Trier (et al., automated method of quantification at item and sub-item level. While
2018), ranges in the order of 81%; a similar accuracy (84%) was also manual measurements are generally subject to random errors of vari-
reached in other publications, considering both omission and commis- able magnitude, automatic procedures generate precise and repeatable
sion errors (Magnini et al., 2017). The study by Trier (et al., 2018) data. Assuming the existence of systematic errors (of omission/com-
covers a selection of papers published in the period from 2012 to 2018 mission), the resulting shift from the real value will be constant for all
and one can perceive a progressive increase in the performance of the measures, thus granting better comparability of the data. Moreover,
classification, up to a pick-rate of 98% true positives and 1% false ne- OBIA can simultaneously consider morphometry, color, texture and
gatives touched in 2018. In other words, the value of OPSR and OBIA relational parameters, helping in refining the final classification, if
for archaeology is no more in question, as demonstrated by the abun- compared to the traditional per-pixel approaches.
dant bibliography and the high number of successful case studies. What
is still missing at this point is the development of dedicated, sound and
4. Conclusions and perspectives
shared theoretical bases, that can surpass the boundaries of pure
computer sciences for adapting to the specificities of archaeological
OBIA is a growing trend in archaeological RS. In this paper, we
research. With this paper, we hope to offer a further contribution to-
offered a synthesis of its basic principles and discussed the most re-
wards this goal.
levant papers dealing with archaeological case studies. Moreover, we
Looking at the published literature on OBIA, there are only two
showed the practical problems which still preclude a wider diffusion of
papers dealing with item-level applications (Lamotte, Masson, 2016;
the method among the archaeological RS community and the field
Masson, Lamotte, 2018). In their research, the authors were able to
operators. It was argued that some of these issues can be overcome by
characterize a group of prehistoric handaxes in terms of colorimetric,
introducing a theoretical framework able to formalize expert knowl-
textural and morphometric indices. The results were used as a statistical
edge. The incompleteness and dynamic nature of the archaeological
guide to compare the different objects and to propose typological cor-
record led us to propose the use of ontologies based on Diachronic
relations with chronological significance. In fact, manual measurements
Semantic Models (DhSM). This formal approach can provide a possible
are characterized by limited repeatability due to the difficulty of
key for the description of the evolution of the archaeological record
identifying the correct orientation of the object and of choosing the
through time. Moreover, it can prove particularly useful for the clas-
reference points. By combining OBIA and 3D models, it is possible to
sification of multi-/equi-final entities which constitute a long-lasting
obtain repeatable and exportable metric data. Fig. 6c presents the ap-
plication of LiDAR visualizations and OBIA on the laser scanner data of
Late Bronze Age (XIV-XII century BC) cranial samples for the quanti- 1
The publication is currently in preparation.
fication of the recalcified osteological tissue linked to the life 2
A dedicated paper is currently in preparation.
18
Fig. 8. Integration of Dhiacronic Semantic Models (DhSM) into the general workflow of an ArchaeOBIA project. Semantic model (SM) 1 represents an approximation
of how the ancient context was expected to be in ancient times; SM2 includes the multi-/equi-final appearance it may have today; and SM3 considers its possible
representation on different types of RS imagery.
problem in archaeological photointerpretation. Additionally, we in- and LiDAR) (Bosco et al., 2015; Stek, 2016; De Reu et al., 2016;
troduced a formal definition of archaeo-objects, which are the focus of Campana, 2017; Colombatti et al., 2017) it is just a matter of time
all RS applications in archaeology, especially in the field of computer- before OBIA will start to be employed in the post-processing of data
aided methods for automated and semi-automated image analysis. from UAV surveys.
This overall theoretical framework represents the backbone of the Future research is expected to further expand the role of OBIA in
ArchaeOBIA concept. In general terms, ArchaeOBIA can be defined as archaeology. Although this is only the beginning of a long-term process,
the application of object-based image analysis to archaeological re- we anticipate that the proposed ArchaeOBIA approach will be able to
search, irrespectively of the scale of investigation. Furthermore, this enhance the interoperability of the rule-sets and disclose new possibi-
approach is designed to systematically integrate OBIA and result as- lities towards an explicit method for extracting archaeological in-
sessment, to achieve an appropriate balance between processing speed formation from RS imagery. Building from this preliminary discussion,
and reliability of results. A scheme on the integration of DhSM and we hope that the archaeologists will become more and more aware of
ArchaeOBIA is proposed in Fig. 8. their fundamental value in the (semi)automatic recognition of archaeo-
The paper also presented a series of case studies at regional, local, objects. As their knowledge and mental models will be systematically
item and microscopic scale to highlight the versatility of the method. integrated in DhSM, translated in machine-readable language and used
Judging from those data, ArchaeOBIA shows promising growth op- as a reference for OPSR, the overall accuracy of the results will hope-
portunities with regards to the fields of application and the type of fully reach new standards.
sensors whose data might be processed in the near future. Besides, it
proved as a reliable and reproducible method to deal with the com- Author contributions
plexity of the archaeological record. In fact, ArchaeOBIA offers an ef-
ficient and robust protocol to help (semi)automatic photointerpretation Conceptualization, L.M. and C.B.; methodology, L.M. and C.B.;
and data analysis, capable of simultaneously operating on multiple software, L.M.; validation, C.B. and L.M.; investigation, L.M.; data
layers for the classification of archaeo-objects. It also grants the op- curation, L.M.; writing—original draft preparation, C.B.; writing—re-
portunity to speed up the process of image analysis and object review and editing, L.M.; visualization, C.B.; project administration, L.M.
cognition when working at landscape level or with huge amount of data
thanks to the exportability of the rule-sets. Finally, it is a powerful tool Funding
for protecting the archaeological record as a multi-temporal, multi-
level, (semi)automatic monitoring system. In detail, object-based image This research was carried out within the framework of the post-
analysis is an instrument of exceptional potential for the time-series doctoral grant “Remote diagnostic for monitoring endangered archae-
diagnostics of the degradation processes both at landscape and item ological landscapes” funded by the project Horus 2.0 (Department of
level. In this sense, it is also an effective means to identify and prevent Cultural Heritage, University of Padova).
illegal excavations and looting actions in the perspective of an ‘Applied
Archaeology’ (Downum, Price, 1999). Furthermore, the method can be Conflicts of interest
used to monitor processes of destruction (voluntary or accidental) of
cultural heritage areas impacted by military conflicts or natural ha- The authors declare no conflict of interest.
zards, that are essentially inaccessible for direct interventions on the
field (as seen in Lasaponara, Masini, 2018 using automatic feature ex-
Acknowledgments
traction). Another possible direction is the application of ArchaeOBIA
to Unmanned Aerial Vehicles (UAV)-derived data. After the success and
Numerous colleagues have contributed to the debate on the issues
capillary diffusion in the archaeological practice of aerial platforms
presented in this paper, but major thanks are due to Armando De Guio
equipped with a variety of sensors (from optical to multi/hyper-spectral
for the constant inspiration and scientific support. The authors would
19
also like to thank the anonymous referees whose critical reading helped clinical reporting. Oncotarget 7. https://doi.org/10.18632/oncotarget.10053.
in improving the accepted draft of the manuscript. Campana, S., 2017. Drones in archaeology. State-of-the-art and future perspectives:
drones in archaeology. Archaeol. Prospect. 24, 275–296. https://doi:10.1002/arp.
1569.
References Caspari, G., Balz, T., Liu, Gang, Wang, Xinyuan, Liao, Mingsheng, 2014. Application of
Hough Forests for the detection of grave mounds in high-resolution satellite imagery.
In: 2014 IEEE Geoscience and Remote Sensing Symposium. Presented at the IGARSS
Agapiou, A., Alexakis, D., Sarris, A., Hadjimitsis, D., 2014. Evaluating the potentials of
2014 - 2014 IEEE International Geoscience and Remote Sensing Symposium. IEEE,
sentinel-2 for archaeological perspective. Rem. Sens. 6, 2176–2194. https://doi:10.
Quebec City, QC, pp. 906–909. https://doi.org/10.1109/IGARSS.2014.6946572.
3390/rs6032176.
Castilla, G., 2003. Object-oriented Analysis of Remote Sensing Images for Land Cover
Ainsworth, S., Oswald, A., Went, D., 2013. Remotely acquired, not remotely sensed: using
Mapping: Conceptual Foundation and a Segmentation Method to Derive a Baseline
lidar as a field survey tool. In: Opitz, R.S., Cowley, D.C. (Eds.), Interpreting
Partition for Classification. PhD Thesis. Polytechnic University of Madrid.
Archaeological Topography. Oxbow Books, Oxford, pp. 206–222.
Castilla, G., Hay, G.J., 2008. Image objects and geographic objects. In: Blaschke, T., Lang,
Andres, S., Arvor, D., Pierkot, C., 2012. Towards an ontological approach for classifying
S., Hay, G.J. (Eds.), Object-based Image Analysis: Spatial Concepts for Knowledge-
remote sensing images. In: 2012 Eighth International Conference on Signal Image
Driven Remote Sensing Applications, Lecture Notes in Geoinformation and
Technology and Internet Based Systems. IEEE, Naples, pp. 825–832.
Cartography. Springer, Berlin, 9783540770572, pp. 91–110.
Aprile, A., Castellano, G., Eramo, G., 2014. Combining image analysis and modular neural
Cerrillo-Cuenca, E., 2017. An approach to the automatic surveying of prehistoric barrows
networks for classification of mineral inclusions and pores in archaeological pot-
through LiDAR. Quat. Int. 435, 135–145. https://doi:10.1016/j.quaint.2015.12.099.
sherds. J. Archaeol. Sci. 50, 262–272. https://doi.org/10.1016/j.jas.2014.07.017.
Challis, K., Forlin, P., Kincey, M., 2011. A generic toolkit for the visualization of ar-
Artioli, G., Angelini, I., Polla, A., 2008. Crystals and phase transitions in protohistoric
chaeological features on airborne LiDAR elevation data: visualizing archaeological
glass materials. Phase Transitions 81, 233–252. https://doi:10.1080/
features in airborne LiDAR. Archaeol. Prospect. 18, 279–289. https://doi:10.1002/
01411590701514409.
arp.421.
Arvor, D., Durieux, L., Andrés, S., Laporte, M.-A., 2013. Advances in Geographic Object-
Chen, F., Masini, N., Yang, R., Milillo, P., Feng, D., Lasaponara, R., 2016. A space view of
Based Image Analysis with ontologies: a review of main contributions and limitations
radar archaeological marks: first applications of COSMO-skymed X-band data. Rem.
from a remote sensing perspective. ISPRS J. Photogrammetry Remote Sens. 82,
Sens. 7 (1), 24–50. https://doi.org/10.3390/rs8030208.
125–137. https://doi:10.1016/j.isprsjprs.2013.05.003.
Colombatti, G., Aboudan, A., Bettanini, C., Magnini, L., Bettineschi, C., Deotto, G.,
Baatz, M., Schäpe, M., 2000. Multiresolution segmentation — an optimization approach
Toninello, L., Debei, S., De Guio, A., Zanovello, P., Menegazzi, A., 2017. Horus — a
for high quality multi-scale image segmentation. In: Strobl, J., Blaschke, T.,
drone project for visual and IR imaging. In: 2017 IEEE International Workshop on
Griesebner, G., AGIT-Symposium (Eds.), Angewandte Geographische
Metrology for AeroSpace (MetroAeroSpace). IEEE, Padua, Italy, pp. 589–592.
Informationsverarbeitung XII: Beiträge Zum AGIT-Symposium Salzburg 2000.
Costa, G., Feitosa, R., Fonseca, L., Oliveira, D., Ferreira, R., Castejon, E., 2010.
Herbert Wichmann Verlag, Heidelberg, 9783879073498, pp. 12–23.
Knowledge-based interpretation of remote sensing data with the InterImage system:
Baatz, M., Hoffmann, C., Willhauck, G., 2008. Progressing from object-based to object-
major characteristics and recent developments. In: Proceedings of the GEOBIA 2010:
oriented image analysis. In: Blaschke, T., Lang, S., Hay, G.J. (Eds.), Object-Based
Geographic Object-Based Image Analysis, Ghent, Belgium, 29 June–2 July 2010.
Image Analysis. Springer Berlin Heidelberg, Berlin, Heidelberg, 9783540770572, pp.
Cowley, D.C., 2015. Aerial photographs and aerial reconnaissance for landscape studies.
29–42.
In: Chavarría Arnau, A., Reynolds, A. (Eds.), Detecting and Understanding Historic
Bennett, R., Cowley, D., De Laet, V., 2014. The data explosion: tackling the taboo of
Landscapes, PCA Studies, SAP, Società Archeologica S. r.l: Mantova,
automatic feature recognition in airborne survey data. Antiquity 88, 896–905.
9788887115994, pp. 37–66.
https://doi:10.1017/S0003598X00050766.
Crutchley, S., 2015. Using airborne lidar for interpreting archaeological landscapes. In:
Benz, U.C., Hofmann, P., Willhauck, G., Lingenfelder, I., Heynen, M., 2004. Multi-re-
Chavarría Arnau, A., Reynolds, A. (Eds.), Detecting and Understanding Historic
solution, object-oriented fuzzy analysis of remote sensing data for GIS-ready in-
Landscapes, PCA Studies, SAP, Società Archeologica s.r.L: Mantova. 9788887115994,
formation. ISPRS J. Photogrammetry Remote Sens. 58, 239–258. https://doi:10.
pp. 67–92.
1016/j.isprsjprs.2003.10.002.
Davis, D.S., 2018. Object-based image analysis: a review of developments and future
Bertalanffy, L. van, 2003. General system theory: foundations, development, applications,
directions of automated feature detection in landscape archaeology. Archaeol.
Rev. In: 14. Paperback Print. Braziller, New York9780807604533, .
Prospect. 1–9. https://doi.org/10.1002/arp.1730.
Bescoby, D.J., 2006. Detecting Roman land boundaries in aerial photographs using Radon
Davis, D.S., Sanger, M.C., Lipo, C.P., 2018. Automated mound detection using lidar and
transforms. J. Archaeol. Sci. 33, 735–743. https://doi.org/10.1016/j.jas.2005.10.
object-based image analysis in Beaufort County, South Carolina. SE. Archaeol. 1–15.
012.
https://doi:10.1080/0734578X.2018.1482186.
Bettineschi, C., 2018. Archaeometric Study of Egyptian Vitreous Materials from Tebtynis:
Davis, D.S., Lipo, C.P., Sanger, M.C., 2019. A comparison of automated object extraction
Integration of Analytical and Archaeological Data. PhD Thesis. University of Padova.
methods for mound and shell-ring identification in coastal South Carolina. J.
Bewley, R.H., Crutchley, S.P., Shell, C.A., 2005. New light on an ancient landscape: lidar
Archaeol. Sci.: Report 23, 166–177. https://doi.org/10.1016/j.jasrep.2018.10.035.
survey in the Stonehenge World Heritage Site. Antiquity 79, 636–647. https://doi:10.
De Guio, A., 2015. Cropping for a better future: vegetation indices in archaeology. In:
1017/S0003598X00114577.
Chavarría Arnau, A., Reynolds, A. (Eds.), Detecting and Understanding Historic
Bishr, Y., 1998. Overcoming the semantic and other barriers to GIS interoperability. Int. J.
Landscapes, PCA Studies, SAP, Società Archeologica s.r.L: Mantova. 9788887115994,
Geogr. Inf. Sci. 12, 299–314. https://doi.org/10.1080/136588198241806.
pp. 109–152.
Blaschke, T., Burnett, C., Pekkarinen, A., 2004. New contextual approaches using image
De Guio, A., Betto, A., Migliavacca, M., Magnini, L., 2013. Mountain fossil landscapes and
segmentation for object-based classification. In: Jong, S.M. de, Meer, F.D. van der
the “archaeology of us”: an object/pattern/scenery recognition experiment. In: Lugli,
(Eds.), Remote Sensing Image Analysis: Including the Spatial Domain. Springer,
F. (Ed.), Ethnoarchaeology: Current Research and Field Methods: Conference
Dordrecht, 9781402025594, pp. 211–236 Remote sensing and digital image pro-
Proceedings, Rome, Italy, 13th-14th May 2010. BAR International series,
cessing.
Archaeopress, Oxford, England, 9781407310831, pp. 241–247.
Blaschke, T., Hay, G.J., Kelly, M., Lang, S., Hofmann, P., Addink, E., Queiroz Feitosa, R.,
De Guio, A., Magnini, L., Bettineschi, C., 2015. GeOBIA approaches to remote sensing of
van der Meer, F., van der Werff, H., van Coillie, F., Tiede, D., 2014. Geographic
fossil landscapes: two case studies from Northern Italy. In: Traviglia, A. (Ed.), Across
object-based image analysis – towards a new paradigm. ISPRS J. Photogrammetry
Space and Time: Papers from the 41st Conference on Computer Applications and
Remote Sens. 87, 180–191. https://doi:10.1016/j.isprsjprs.2013.09.014.
Quantitative Methods in Archaeology: Perth, 25-28 March 2013, Computer
Böck, S., Immitzer, M., Atzberger, C., 2016. Automated segmentation parameter selection
Applications and Quantitative Methods in Archaeology. Amsterdam University Press,
and classification of urban scenes using open-source software. In: Proceedings of the
Amsterdam, 9789089647153, pp. 45–53.
GEOBIA 2016: Solutions & Synergies, Enschede, The Netherlands, 14–16 September
De Laet, V., Paulissen, E., Waelkens, M., 2007. Methods for the extraction of archae-
2016.
ological features from very high-resolution Ikonos-2 remote sensing imagery, Hisar
Boer, A. de, 2007. Using pattern recognition to search LIDAR data for archeological sites.
(southwest Turkey). J. Archaeol. Sci. 34, 830–841. https://doi:10.1016/j.jas.2006.
In: Figueiredo, A., Leite Velho, G. (Eds.), The World Is in Your Eyes. CAA2005.
09.013.
Computer Applications and Quantitative Methods in Archaeology. Proceedings of the
de Matos-Machado, R., Toumazet, J.-P., Bergès, J.-C., Amat, J.-P., Arnaud-Fassetta, G.,
33rd Conference, Tomar, March 2005. CAA Portugal, Tomar, pp. 245–254.
Bétard, F., Bilodeau, C., Hupy, J.P., Jacquemot, S., 2019. War landform mapping and
Bosco, A., Barbarino, M., Valentini, R., D'Andrea, A., 2015. Low-cost surveys of the domus
classification on the Verdun battlefield (France) using airborne LiDAR and multi-
of stallius eros in pompeii. In: ISPRS - International Archives of the Photogrammetry,
variate analysis: war landform mapping and classification on the Verdun battlefield.
Remote Sensing and Spatial Information Sciences XL-5/W4, pp. 187–192. https://
In: Earth Surface Processes and Landforms, . https://doi.org/10.1002/esp.4586.
doi.org/10.5194/isprsarchives-XL-5-W4-187-2015.
De Reu, J., Trachet, J., Laloo, P., De Clercq, W., 2016. From low cost UAV survey to high
Brophy, K., Cowley, D. (Eds.), 2005. From the Air: Understanding Aerial Archaeology.
resolution topographic data: developing our understanding of a medieval outport of
Stroud, Tempus9780752431307, .
bruges: from low cost UAV survey to high resolution topographic data. Archaeol.
Burigana, L., Magnini, L., 2018. Image processing and analysis of radar and lidar data:
Prospect. 23, 335–346. https://doi.org/10.1002/arp.1547.
new discoveries in Verona southern lowland (Italy). STAR. Sci. Technol. Archaeol.
Dehn, M., Gärtner, H., Dikau, R., 2001. Principles of semantic modeling of landform
Res. 3, 490–509. https://doi:10.1080/20548923.2018.1426273.
structures. Comput. Geosci. 27, 1005–1010. https://doi:10.1016/S0098-3004(00)
Burigana, L., Griggio, E., Marra, A., Magnini, L., 2017. Modelli digitali di elevazione:
00138-2.
structure from Motion e trattamenti del rilievo su piccola e larga scala. In: Deotto, G.,
Devereux, B.J., Amable, G.S., Crow, P., Cliff, A.D., 2005. The potential of airborne lidar
Bettineschi, C., Magnini, L., Toninello, L. (Eds.), Progetto Horus: Visioni Dall’alto
for detection of archaeological features under woodland canopies. Antiquity 79,
Dello Spazio Archeologico, Padova UP: Padova. 9788869381089, .
648–660. https://doi.org/10.1017/S0003598X00114589.
Caie, P.D., Zhou, Y., Turnbull, A.K., Oniscu, A., Harrison, D.J., 2016. Novel histopatho-
Doneus, M., Briese, C., Fera, M., Janner, M., 2008. Archaeological prospection of forested
logic feature identified through image analysis augments stage II colorectal cancer
areas using full-waveform airborne laser scanning. J. Archaeol. Sci. 35, 882–893.
20
https://doi:10.1016/j.jas.2007.06.013. Aoyama, K., Yonenobu, H., 2017. Archaeological application of airborne LiDAR with
Doneus, M., Verhoeven, G., Atzberger, C., Wess, M., Ruš, M., 2014. New ways to extract object-based vegetation classification and visualization techniques at the lowland
archaeological information from hyperspectral pixels. J. Archaeol. Sci. 52, 84–96. Maya site of ceibal, Guatemala. Rem. Sens. 9, 563. https://doi:10.3390/rs9060563.
https://doi:10.1016/j.jas.2014.08.023. Jahjah, M., Ulivieri, C., Invernizzi, A., Parapetti, R., 2007. Archaeological remote sensing
Downum, C., Price, L., 1999. Applied archaeology. Hum. Organ. 58, 226–239. https:// application pre-post war situation of Babylon archaeological site—Iraq. Acta
doi.org/10.17730/humo.58.3.61383586jj524v25. Astronaut. 61, 121–130. https://doi.org/10.1016/j.actaastro.2007.01.034.
Drăguţ, L., Blaschke, T., 2006. Automated classification of landform elements using ob- Knoth, C., Nüst, D., 2017. Reproducibility and practical adoption of GEOBIA with open-
ject-based image analysis. Geomorphology 81, 330–344. https://doi:10.1016/j. source software in docker containers. Rem. Sens. 9, 290. https://doi:10.3390/
geomorph.2006.04.013. rs9030290.
Ecognition (Trimble Inc.). Körting, T., Fonseca, L., Câmara, G., 2013. GeoDMA—geographic data mining analyst.
Eisank, C., Drăguţ, L., Blaschke, T., 2011. A generic procedure for semantics-oriented Comput. Geosci. 57, 133–145.
landform classification in object-based image analysis. In: Hengl, T., Evans, I.S., Kramer, I.C., 2015. An Archaeological Reaction to the Remote Sensing Data Explosion:
Wilson, J.P., Gould, M. (Eds.), Geomorphometry: Redlands, CA, USA 2011, (7-9 Reviewing the Research on Semi-automated Pattern Recognition and Assessing the
Sept.), pp. 125–128. Potential to Integrate Artificial Intelligence. MSc thesis. University of Southampton.
Feuchtinger, A., Walch, A., Dobosz, M., 2016. Deep tissue imaging: a review from a Lambers, K., Traviglia, A., 2016. Automated detection in remote sensing archaeology: a
preclinical cancer research perspective. Histochem. Cell Biol. 146, 781–806. https:// reading list. AARGnews 53, 25–29.
doi:10.1007/s00418-016-1495-7. Lamotte, A., Masson, E., 2016. Arché-OBIA: un concept d’analyse quantitative d’images
Forbes, D.J., 2017. Complexity and geographic scale. In: Quattrochi, D.A. (Ed.), numériques appliqué aux bifaces du gisement de Gouzeaucourt (Nord, FR). Notae
Integrating Scale in Remote Sensing and GIS. Routledge, London, 9781482218268, Praehistoricae 36, 121–130.
pp. 11–40. Lasaponara, R., Masini, N. (Eds.), 2012. Satellite Remote Sensing, Remote Sensing and
Freeland, T., Heung, B., Burley, D.V., Clark, G., Knudby, A., 2016. Automated feature Digital Image Processing. vol. 16 Springer Netherlands, Dordrecht9789048188000, .
extraction for prospection and analysis of monumental earthworks from aerial LiDAR Lasaponara, R., Masini, N., 2013. Satellite synthetic aperture radar in archaeology and
in the Kingdom of Tonga. J. Archaeol. Sci. 69, 64–74. https://doi:10.1016/j.jas.2016. cultural landscape: an overview. Archaeol. Prospect. 20, 71–78. https://doi.org/10.
04.011. 1002/arp.1452.
Gamanya, R., De Maeyer, P., De Dapper, M., 2009. Object-oriented change detection for Lasaponara, R., Masini, N., 2018. Space-based identification of archaeological illegal
the city of Harare, Zimbabwe. Expert Syst. Appl. 36, 571–588. https://doi.org/10. excavations and a new automatic method for looting feature extraction in desert
1016/j.eswa.2007.09.067. areas. Surv. Geophys. 39, 1323–1346. https://doi.org/10.1007/s10712-018-9480-4.
Gerisch, M., Smettan, J., Ebert, S., Athelogou, M., Brand-Saberi, B., Spindler, N., Mueller, Leonardi, G., Balista, C., 1992. Linee di Approccio al Deposito archeologico. In: Leonardi,
W.C., Giri, S., Bader, A., 2018. Qualitative and quantitative analysis of cardiac pro- G. (Ed.), Proceedings of the International Conference “Formation processes and ex-
genitor cells in cases of myocarditis and cardiomyopathy. Front. Genet. 9. https:// cavation methods in archaeology: perspectives”, Padua 15-27 July 1991, Saltuarie
doi:10.3389/fgene.2018.00072. dal laboratorio del Piovego 3, Padova, pp. 75–99.
Ghazouani, F., Farah, I.R., Solaiman, B., 2018. Semantic remote sensing scenes inter- Lillesand, T.M., Kiefer, R.W., Chipman, J.W., 2004. Remote Sensing and Image
pretation and change interpretation. In: Thomas, C. (Ed.), Ontology in Information Interpretation, fifth ed. Wiley, New York9780471152279.
Science. InTech, pp. 205–224. https://doi.org/10.5772/intechopen.72730. Lucas, G., 2012. Understanding the Archaeological Record. Cambridge University Press,
Gibril, M.B.A., Bakar, S.A., Yao, K., Idrees, M.O., Pradhan, B., 2017. Fusion of Cambridge, New York9781107010260.
RADARSAT-2 and multispectral optical remote sensing data for LULC extraction in a Magnini, L., 2017. Remote Sensing e Object-Based Image Analysis: metodologie di ap-
tropical agricultural area. Geocarto Int. 32, 735–748. https://doi:10.1080/ proccio per la creazione di standard archeologici. PhD Thesis, University of Padova.
10106049.2016.1170893. Magnini, L., Bettineschi, C., De Guio, A., 2017. Object-based shell craters classification
Gonzalez, R.C., Woods, R.E., Schmid, S.R., Battiato, S., Stanco, F., 2008. Elaborazione from lidar-derived sky-view factor. Archaeol. Prospect. 24, 211–223. https://doi:10.
Delle Immagini Digitali. Pearson Prentice Hall, Milano9788871925066. 1002/arp.1565.
Graham, S., Weingart, S., 2015. The equifinality of archaeological networks: an agent- Masson, A., Lamotte, A., 2018. Arché-OBIA: vestiges lithiques au Paléolithique et ex-
based exploratory lab approach. J. Archaeol. Method Theory 22, 248–274. https:// ploitation quantifiée de leurs images numériques. Hors-série du CRAHN (Centre de
doi:10.1007/s10816-014-9230-y. Recherches Archéologique de Haute-Normandie), pp. 153–165.
Grippa, T., Lennert, M., Beaumont, B., Vanhuysse, S., Stephenne, N., Wolff, E., 2016. An Menze, B.H., Ur, J.A., 2012. Mapping patterns of long-term settlement in Northern
open-source semi-automated processing chain for urban OBIA classification. In: Mesopotamia at a large scale. Proc. Natl. Acad. Sci. Unit. States Am. 109, E778–E787.
Proceedings of the GEOBIA 2016: Solutions & Synergies, Enschede, The Netherlands, https://doi.org/10.1073/pnas.1115472109.
14–16 September 2016. Menze, B.H., Ur, J.A., Sherratt, A.G., 2006. Detection of ancient settlement mounds: ar-
Gruber, T.R., 1993. A translation approach to portable ontology specifications. Knowl. chaeological survey based on the SRTM terrain model. Photogramm. Eng. Rem. Sens.
Acquis. 5, 199–220. https://doi:10.1006/knac.1993.1008. 72, 321–327.
Guyot, A., Hubert-Moy, L., Lorho, T., 2018. Detecting neolithic burial mounds from lidar- Moriarty, C., Cowley, D.C., Wade, T., Nichol, C.J., 2019. Deploying multispectral remote
derived elevation data using a multi-scale Approach and machine learning techni- sensing for multi-temporal analysis of archaeological crop stress at Ravenshall, Fife,
ques. Rem. Sens. 10, 225. https://doi.org/10.3390/rs10020225. Scotland. Archaeol. Prospect. 26 (1), 33–46. https://doi:10.1002/arp.1721.
Hanson, W.S., 2008. The future of aerial archaeology (or are algorithms the answer?). In: Nash, D.T., Petraglia, M.D., 1987. Natural Formation Processes and the Archaeological
Advances on Remote Sensing for Archaeology and Cultural Heritage Management: Record. Society for American Archaeology, BAR, Oxford, England9780860544555.
Proceedings of the 1st International EARSeL Workshop CNR, Rome, September 30- Opitz, R.S., Cowley, D. (Eds.), 2013. Interpreting Archaeological Topography: Airborne
October 4, 2008. Aracne, Rome, 9788854820302, pp. 47–50. Laser Scanning, 3D Data and Ground Observation. Occasional Publication of the
Hanson, W.S., 2010. The future of aerial archaeology in Europe. Photo Interprétation.Eur. Aerial Archaeology Research Group. Oxbow Books, Oxford, UK , Oakville,
J. Appl Remote sens. 46 (1), 3–11. CT9781842175163, .
Harris, E.C., 1997. Principles of Archaeological Stratigraphy, vol. 2 Acad. Pr, Pal, N.R., Pal, S.K., 1993. A review on image segmentation techniques. Pattern Recogn.
London9780123266514 ed., 3. print. 26, 1277–1294. https://doi:10.1016/0031-3203(93)90135-J.
Hawkins, E.D., Duarte, D., Akinduro, O., Khorshed, R.A., Passaro, D., Nowicka, M., Parcak, S.H., 2009. Satellite Remote Sensing for Archaeology. Routledge, London, New
Straszkowski, L., Scott, M.K., Rothery, S., Ruivo, N., Foster, K., Waibel, M., York9780415448772.
Johnstone, R.W., Harrison, S.J., Westerman, D.A., Quach, H., Gribben, J., Robinson, Pregesbauer, M., Trinks, I., Neubauer, W., 2014. An object oriented approach to auto-
M.D., Purton, L.E., Bonnet, D., Lo Celso, C., 2016. T-cell acute leukaemia exhibits matic classification of archaeological features in magnetic prospection data. Near
dynamic interactions with bone marrow microenvironments. Nature 538, 518–522. Surf. Geophys. 12. https://doi.org/10.3997/1873-0604.2014014.
https://doi:10.1038/nature19801. Quintus, S., Day, S.S., Smith, N.J., 2017. The efficacy and analytical importance of
Hay, G.J., Castilla, G., 2008. Geographic Object-Based Image Analysis (GEOBIA): a new manual feature extraction using lidar datasets. Adv. Archaeol. Pract. 5, 351–364.
name for a new discipline. In: Blaschke, T., Lang, S., Hay, G.J. (Eds.), Object-based https://doi:10.1017/aap.2017.13.
Image Analysis: Spatial Concepts for Knowledge-Driven Remote Sensing Sanger, M.C., 2015. Determining depositional events within shell deposits using computer
Applications, Lecture Notes in Geoinformation and Cartography. Springer, Berlin, vision and photogrammetry. J. Archaeol. Sci. 53, 482–491. https://doi.org/10.1016/
9783540770572, pp. 75–89. j.jas.2014.10.026.
Hay, G.J., Marceau, D., Dube, P., Bouchard, A., 2001. A multiscale framework for land- Schiffer, M.B., 1972. Archaeological context and systemic context. Am. Antiq. 37,
scape analysis: object-specific analysis and upscaling. Landsc. Ecol. 16, 471–490. 156–165. https://doi:10.2307/278203.
Heidrich, A., Schmidt, J., Zimmermann, J., Saluz, H.P., 2013. Automated segmentation Schneider, A., Takla, M., Nicolay, A., Raab, A., Raab, T., 2015. A template-matching
and object classification of CT images: application to in vivo molecular imaging of approach combining morphometric variables for automated mapping of charcoal kiln
avian embryos. Int. J. Biomed. Imaging 2013, 1–10. https://doi:10.1155/2013/ sites: automated mapping of charcoal kiln sites. Archaeol. Prospect. 22, 45–62.
508474. https://doi.org/10.1002/arp.1497.
Hein, I., Rojas-Domínguez, A., Ornelas, M., D'Ercole, G., Peloschek, L., 2018. Automated Schuetter, J., Goel, P., McCorriston, J., Park, J., Senn, M., Harrower, M., 2013.
classification of archaeological ceramic materials by means of texture measures. J. Autodetection of ancient Arabian tombs in high-resolution satellite imagery. Int. J.
Archaeol. Sci.: Report 21, 921–928. https://doi.org/10.1016/j.jasrep.2017.12.032. Remote Sens. 34, 6611–6635. https://doi.org/10.1080/01431161.2013.802054.
Hofmann, P., Marschallinger, R., Unterwurzacher, M., Zobl, F., 2013. Marble provenance Sevara, C., Pregesbauer, M., 2014. Archaeological feature classification: an object or-
designation with object based image analysis: state‐of‐the‐art rock fabric character- iented approach. South Eastern European Journal of Earth Observation and
ization from petrographic micro-graphs. Austrian. J. Earth Sci. 106 (2), 40–49. Geomatics 3 (2s), 139–144.
Husserl, E., Moran, D., 2001. Logical Investigations, International Library of Philosophy. Sevara, C., Pregesbauer, M., Doneus, M., Verhoeven, G., Trinks, I., 2016. Pixel versus
Routledge, London, New York9780415241892. object — a comparison of strategies for the semi-automated mapping of archae-
Inomata, T., Pinzón, F., Ranchos, J.L., Haraguchi, T., Nasu, H., Fernandez-Diaz, J.C., ological features using airborne laser scanning data. J. Archaeol. Sci.: Report 5,
21
485–498. https://doi:10.1016/j.jasrep.2015.12.023. DEMs for archaeological predictive mapping. J. Archaeol. Sci. 39, 698–703. https://
Smeulders, A.W.M., Worring, M., Santini, S., Gupta, A., Jain, R., 2000. Content-based doi:10.1016/j.jas.2011.11.001.
image retrieval at the end of the early years. IEEE Trans. Pattern Anal. Mach. Intell. Verhagen, P., Drăguţ, L., 2013. Discovering the Dutch Mountains. An experiment with
22, 1349–1380. https://doi:10.1109/34.895972. automated landform classification for purposes of archaeological predictive mapping.
Smith, B., 2001. Fiat objects. Topoi 20 (2), 131–148. In: Contreras Cortés, F., Farjas, M., Melero, F.J. (Eds.), CAA2010: Fusion of Cultures:
Stek, T.D., 2016. Drones over Mediterranean landscapes. The potential of small UAV's Proceedings of the 38th Annual Conference on Computer Applications and
(drones) for site detection and heritage management in archaeological survey pro- Quantitative Methods in Archaeology, Granada, Spain, April 2010. BAR international
jects: a case study from Le Pianelle in the Tappino Valley, Molise (Italy). J. Cult. series, Archaeopress, Oxford, 9781407311081, pp. 213–216.
Herit. 22, 1066–1071. https://doi:10.1016/j.culher.2016.06.006. Weinberg, G.M., 1975. An Introduction to General Systems Thinking. Wiley, New
Tapete, D., Cigna, F., 2017. Trends and perspectives of space-borne SAR remote sensing York9780471925637.
for archaeological landscape and cultural heritage applications. J. Archaeol. Sci.: Wilgocka, A., Rączkowski, W., Kostyrko, M., Ruciński, D., 2016. Sherlock Holmes' or Don
Report 14, 716–726. https://doi:10.1016/j.jasrep.2016.07.017. Quixote‘s certainty? Interpretations of cropmarks on satellite imageries in archae-
Toumazet, J.-P., Vautier, F., Roussel, E., Dousteyssier, B., 2017. Automatic detection of ological investigation. In: Fourth International Conference on Remote Sensing and
complex archaeological grazing structures using airborne laser scanning data. J. Geoinformation of the Environment (RSCy2016), . https://doi:10.1117/12.2241802.
Archaeol. Sci.: Report 12, 569–579. https://doi.org/10.1016/j.jasrep.2017.03.012. Wiseman, J., El-Baz, F. (Eds.), 2007. Remote Sensing in Archaeology, Interdisciplinary
Traviglia, A., 2011. Integrated archaeological investigations for the study of the Greater Contributions to Archaeology. Springer, New York9780387444536, .
Aquileia area. In: CEUR Workshop Proceedings, vol. 806. CEUR-WS.org, pp. C1–C11. Witharana, C., Civco, D.L., 2014. Optimizing multi-resolution segmentation scale using
Traviglia, A., Torsello, A., 2017. Landscape pattern detection in archaeological remote empirical methods: exploring the sensitivity of the supervised discrepancy measure
sensing. Geosciences 7, 128. https://doi:10.3390/geosciences7040128. Euclidean distance 2 (ED2). ISPRS J. Photogrammetry Remote Sens. 87, 108–121.
Traviglia, A., Cowley, D., Lambers, K., Sciences, A., 2016. Finding common ground: https://doi:10.1016/j.isprsjprs.2013.11.006.
human and computer vision in archaeological prospection. AARGnews 53, 11–24. Witharana, C., Ouimet, W.B., Johnson, K.M., 2018. Using LiDAR and GEOBIA for auto-
Trier, Ø.D., Pilø, L.H., 2012. Automatic detection of pit structures in airborne laser mated extraction of eighteenth–late nineteenth century relict charcoal hearths in
scanning data: automatic detection of pits in ALS data. Archaeol. Prospect. 19, southern New England. GIScience Remote Sens. 55, 183–204. https://doi:10.1080/
103–121. https://doi.org/10.1002/arp.1421. 15481603.2018.1431356.
Trier, Ø.D., Larsen, S.Ø., Solberg, R., 2009. Automatic detection of circular structures in Wood, W.R., Johnson, D.L., 1978. A survey of disturbance processes in archaeological site
high-resolution satellite images of agricultural land. Archaeol. Prospect. 16, 1–15. formation. Adv. Archaeol. Method Theor. 1, 315–381.
https://doi.org/10.1002/arp.339. Wuest, B., Zhang, Y., 2009. Region based segmentation of QuickBird multispectral ima-
Trier, Ø.D., Cowley, D.C., Waldeland, A.U., 2018. Using deep neural networks on airborne gery through band ratios and fuzzy comparison. ISPRS J. Photogrammetry Remote
laser scanning data: results from a case study of semi-automatic mapping of ar- Sens. 64, 55–64. https://doi:10.1016/j.isprsjprs.2008.06.005.
chaeological topography on Arran, Scotland. Archaeol. Prospect. https://doi.org/10. Zhang, X., Du, S., 2016. Learning selfhood scales for urban land cover mapping with very-
1002/arp.1731. high-resolution satellite images. Rem. Sens. Environ. 178, 172–190. https://doi:10.
Van De Kerchove, R., Hanson, E., Wolff, E., 2014. Comparing pixel-based and object- 1016/j.rse.2016.03.015.
based classification methodologies for mapping impervious surfaces in Wallonia Zhang, X., Du, S., Ming, D., 2018. Segmentation scale selection in geographic object-based
using ortho-imagery and LIDAR data. In: Proceedings of the GEOBIA 2014: image analysis. In: He, Y., Weng, Q. (Eds.), High Spatial Resolution Remote Sensing:
Advancements, Trends and Challenges, Thessaloniki, Greece, 22 May 2014. Data, Analysis, and Applications. Taylor & Francis, Boca Raton, FL, 9781498767682,
Verhagen, P., Drăguţ, L., 2012. Object-based landform delineation and classification from pp. 201–228.
22

Contento

Uploaded by

Contento

Uploaded by

Journal of Archaeological Science 107 (2019) 10–22

Contents lists available at ScienceDirect

Journal of Archaeological Science

Theory and practice for an object-based approach in archaeological remote T

1. Introduction interpretation of the remotely sensed imagery, even in the archae-

Figure 6. ArchaeOBIA levels of application. (a) regional-level: classiﬁcation of

For this reason, we propose a conceptual revision of the term. In

Fig. 7. Chart representing the general workﬂow of an ArchaeOBIA project.

Hoﬀmann et al. (2013) 2013 micrographs/not reported Microscopic

You might also like