Theory and practice for an object-based approach in archaeological remote sensing
Graphical abstract
Introduction
In recent years, archeological research has seen an increasing number of remote sensing (RS) applications with the use of new sensors and data types, such as multi/hyper-spectral imagery (Traviglia, 2011; Lasaponara, Masini, 2012; Doneus et al., 2014; Agapiou et al., 2014; De Guio, 2015; Moriarty et al., 2019), radar (Wiseman, El-Baz, 2007; Lasaponara, Masini, 2013; Chen et al., 2016; Tapete, Cigna, 2017; Burigana, Magnini, 2018) and LiDAR data (Bewley et al., 2005; Devereux et al., 2005; Doneus et al., 2008; Challis et al., 2011; Opitz, Cowley, 2013), that have joined the classic aerial photographs. However, these innovations had only little impact on the traditional photointerpretation, which remains essentially a work for the human operator via visual inspection (Brophy, Cowley, 2005; Cowley, 2015; Crutchley, 2015; Wilgocka et al., 2016; Quintus et al., 2017).
The reduction of the instrumental costs and the exponential increase in the volume of datasets of the last few years prompts for an overall revision of the methods traditionally used in archaeology (Bennett et al., 2014). In this context, the automation or semi-automation of image analysis seems to offer an opportunity to speed up and grant better reproducibility for the classification and the subsequent interpretation of the remotely sensed imagery, even in the archaeological field. This approach has a long history in the domains of environmental, material and biomedical sciences (Heidrich et al., 2013; Caie et al., 2016; Feuchtinger et al., 2016; Hawkins et al., 2016), but its potential is yet to be fully exploited for systematic research on cultural heritage.
As previously pointed out, the number of papers dealing with object, pattern and scenery recognition (OPSR) of archaeological contexts is still very limited and includes applications of template matching, machine learning, convolutional neural network (CNN), custom algorithms and object-based methods (see Traviglia, Torsello, 2017 and Davis, 2018 for a general overview on the topic). In this paper, we will focus our discussion on object-based image analysis (OBIA or GeOBIA, with a geographic connotation) which was described as “an evolving paradigm with specific tools, software, methods, rules, and language (that) is increasingly being used in studies which need to conceptualize and formalize knowledge representing location based reality” (Blashke et al., 2014).
In order to promote the interoperability of the rule-sets, it is necessary to make explicit what it is implicit in the classification and interpretation process. For this purpose, we propose a theoretical framework aimed at formalizing expert archaeological knowledge using ontologies (i.e. formal, explicit specifications of a shared conceptualization, according to Gruber, 1993). Moreover, we introduce the concept of Diachronic Semantic Models (DhSM), developed to better explain the long-term evolution of the landscape in machine-readable language (§ 4.1).
As object-based applications make their way into archaeological practice, it becomes increasingly important to find a shared language and a common protocol of investigation, ideally passing from operational practice to operational routine. In this paper, we suggest a general workflow for OBIA applications in archaeology built from a wide range of published and unpublished case-studies to ease the comparability of data. Finally, we argue that there is a growing urgency to find a common way for publishing rule-sets and rule-set libraries to be semi-automatically or automatically implemented for archaeological investigations. This topic is of general interest for the OBIA community, but it should be stressed with a particular emphasis in view of further increasing the role of object-based applications in archaeology, as most of the operators have a humanistic background and, consequently, longer learning times in the development of customized rule-sets.
Section snippets
The OBIA approach
RS imagery are composed of pixels (or voxels, in a 3D coordinate system), whose dimension is a function of the sensor used and of the parameters employed for the acquisition (Gonzalez et al., 2008). While pixel-based classifications rely only on the information contained in each single pixel (Lillesand et al., 2004), the basic entity of OBIA is represented by image-objects (sometimes also called image-segments) (Hay et al., 2001; Blaschke et al., 2004). An image-object is “a discrete region of
Practical issues
Despite the strengths and opportunities offered by OBIA, a series of drawbacks and weakness must be taken into account when considering the slow emergence of this method in archaeological RS. The criticism of conservative archaeologists has significantly slowed down in the last few years as the technological advances in the field of computer-aided OPSR have demonstrated their significant contribution to the research (Bennett et al., 2014); however, a series of practical issues is still
Conclusions and perspectives
OBIA is a growing trend in archaeological RS. In this paper, we offered a synthesis of its basic principles and discussed the most relevant papers dealing with archaeological case studies. Moreover, we showed the practical problems which still preclude a wider diffusion of the method among the archaeological RS community and the field operators. It was argued that some of these issues can be overcome by introducing a theoretical framework able to formalize expert knowledge. The incompleteness
Author contributions
Conceptualization, L.M. and C.B.; methodology, L.M. and C.B.; software, L.M.; validation, C.B. and L.M.; investigation, L.M.; data curation, L.M.; writing—original draft preparation, C.B.; writing—review and editing, L.M.; visualization, C.B.; project administration, L.M.
Funding
This research was carried out within the framework of the postdoctoral grant “Remote diagnostic for monitoring endangered archaeological landscapes” funded by the project Horus 2.0 (Department of Cultural Heritage, University of Padova).
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgments
Numerous colleagues have contributed to the debate on the issues presented in this paper, but major thanks are due to Armando De Guio for the constant inspiration and scientific support. The authors would also like to thank the anonymous referees whose critical reading helped in improving the accepted draft of the manuscript.
References (128)
- et al.
Combining image analysis and modular neural networks for classification of mineral inclusions and pores in archaeological potsherds
J. Archaeol. Sci.
(2014) - et al.
Advances in Geographic Object-Based Image Analysis with ontologies: a review of main contributions and limitations from a remote sensing perspective
ISPRS J. Photogrammetry Remote Sens.
(2013) - et al.
Multi-resolution, object-oriented fuzzy analysis of remote sensing data for GIS-ready information
ISPRS J. Photogrammetry Remote Sens.
(2004) Detecting Roman land boundaries in aerial photographs using Radon transforms
J. Archaeol. Sci.
(2006)- et al.
Geographic object-based image analysis – towards a new paradigm
ISPRS J. Photogrammetry Remote Sens.
(2014) An approach to the automatic surveying of prehistoric barrows through LiDAR
Quat. Int.
(2017)- et al.
A comparison of automated object extraction methods for mound and shell-ring identification in coastal South Carolina
J. Archaeol. Sci.: Report
(2019) - et al.
Methods for the extraction of archaeological features from very high-resolution Ikonos-2 remote sensing imagery, Hisar (southwest Turkey)
J. Archaeol. Sci.
(2007) - et al.
Principles of semantic modeling of landform structures
Comput. Geosci.
(2001) - et al.
Archaeological prospection of forested areas using full-waveform airborne laser scanning
J. Archaeol. Sci.
(2008)
New ways to extract archaeological information from hyperspectral pixels
J. Archaeol. Sci.
Automated classification of landform elements using object-based image analysis
Geomorphology
Automated feature extraction for prospection and analysis of monumental earthworks from aerial LiDAR in the Kingdom of Tonga
J. Archaeol. Sci.
Object-oriented change detection for the city of Harare, Zimbabwe
Expert Syst. Appl.
A translation approach to portable ontology specifications
Knowl. Acquis.
Automated classification of archaeological ceramic materials by means of texture measures
J. Archaeol. Sci.: Report
Archaeological remote sensing application pre-post war situation of Babylon archaeological site—Iraq
Acta Astronaut.
GeoDMA—geographic data mining analyst
Comput. Geosci.
A review on image segmentation techniques
Pattern Recogn.
Determining depositional events within shell deposits using computer vision and photogrammetry
J. Archaeol. Sci.
Evaluating the potentials of sentinel-2 for archaeological perspective
Rem. Sens.
Remotely acquired, not remotely sensed: using lidar as a field survey tool
Towards an ontological approach for classifying remote sensing images
Crystals and phase transitions in protohistoric glass materials
Phase Transitions
Multiresolution segmentation — an optimization approach for high quality multi-scale image segmentation
Progressing from object-based to object-oriented image analysis
The data explosion: tackling the taboo of automatic feature recognition in airborne survey data
Antiquity
General system theory: foundations, development, applications, Rev
Archaeometric Study of Egyptian Vitreous Materials from Tebtynis: Integration of Analytical and Archaeological Data
New light on an ancient landscape: lidar survey in the Stonehenge World Heritage Site
Antiquity
Overcoming the semantic and other barriers to GIS interoperability
Int. J. Geogr. Inf. Sci.
New contextual approaches using image segmentation for object-based classification
Automated segmentation parameter selection and classification of urban scenes using open-source software
Using pattern recognition to search LIDAR data for archeological sites
Low-cost surveys of the domus of stallius eros in pompeii
Image processing and analysis of radar and lidar data: new discoveries in Verona southern lowland (Italy). STAR
Sci. Technol. Archaeol. Res.
Modelli digitali di elevazione: structure from Motion e trattamenti del rilievo su piccola e larga scala
Novel histopathologic feature identified through image analysis augments stage II colorectal cancer clinical reporting
Oncotarget
Drones in archaeology. State-of-the-art and future perspectives: drones in archaeology
Archaeol. Prospect.
Application of Hough Forests for the detection of grave mounds in high-resolution satellite imagery
Object-oriented Analysis of Remote Sensing Images for Land Cover Mapping: Conceptual Foundation and a Segmentation Method to Derive a Baseline Partition for Classification
Image objects and geographic objects
A generic toolkit for the visualization of archaeological features on airborne LiDAR elevation data: visualizing archaeological features in airborne LiDAR
Archaeol. Prospect.
A space view of radar archaeological marks: first applications of COSMO-skymed X-band data
Rem. Sens.
Horus — a drone project for visual and IR imaging
Knowledge-based interpretation of remote sensing data with the InterImage system: major characteristics and recent developments
Aerial photographs and aerial reconnaissance for landscape studies
Using airborne lidar for interpreting archaeological landscapes
Object-based image analysis: a review of developments and future directions of automated feature detection in landscape archaeology
Archaeol. Prospect.
Cited by (26)
Archaeological site segmentation of ancient city walls based on deep learning and LiDAR remote sensing
2024, Journal of Cultural HeritageAuto-identification of linear archaeological traces of the Great Wall in northwest China using improved DeepLabv3+ from very high-resolution aerial imagery
2022, International Journal of Applied Earth Observation and GeoinformationLiDAR predictive modeling of Pacific Northwest mound sites: A study of Willamette Valley Kalapuya Mounds, Oregon (USA)
2021, Journal of Archaeological Science: ReportsMapping burial mounds based on UAV-derived data in the Suusamyr Plateau, Kyrgyzstan
2020, Journal of Archaeological ScienceCitation Excerpt :The methods available so far are mostly based on LiDAR (Light Detection and Ranging) datasets (Trier et al., 2012; Trier et al., 2015; Freeland et al., 2016; Sevara et al., 2016; Guyot et al., 2018; Davis et al., 2019a; Lambers et al., 2019; Verschoof-van der Vaart and Lambers, 2019; Verschoof-van der Vaart et al., 2020), high-resolution satellite imagery (Caspari et al., 2014) or Google Earth Imagery (Caspari and Crespo, 2019). In this regard, Magnini and Bettineschi (2019) recently suggested the adequacy of UAV-derived data for OBIA applications. OBIA approaches offer the possibility of integrating statistical and/or semantic rules within the classification scheme (Blaschke et al., 2014).
Defining what we study: The contribution of machine automation in archaeological research
2020, Digital Applications in Archaeology and Cultural HeritageCitation Excerpt :For databases to be compatible, metadata must be thorough (and codified) to enable researchers to locate and compare different information based on search-terms. As such, the development of ontologies – or shared concepts – within both database creation and different computer analysis techniques, have permitted for better sharing of information and improved performance of data analysis (Arvor et al., 2013, 2019; Binding et al., 2008; Rajbhandari et al., 2019; also see Magnini and Bettineschi, 2019). Thus, for opponents of machine learning in archaeology, it should be noted that the argument here does not advocate such approaches as a panacea for archaeological practice.