Format: Paper presentations with discussion
Convenors:
Jarrad Kowlessar, Archaeology, Flinders University, Australia, jarrad.kowlessar@flinders.edu.au
Andrea Jalandoni, Griffith University, Australia, a.jalandoni@griffith.edu.au
Mark Willis, Sacred Sites Research, Inc., USA, willis.arch@gmail.com
Rock art research has been significantly transformed by the integration of digital archaeology, fostering groundbreaking advancements in how we record, analyse, and interpret ancient markings. Across the globe, new technologies and methodologies in data science and remote sensing are expanding the boundaries of rock art research, offering insights previously unattainable through traditional approaches.
This session will explore the diverse array of digital techniques employed in the study of rock art, including advanced visualisation tools, virtual and augmented reality applications, machine learning, geospatial analyses, and cutting-edge 3D recording methods. These approaches have not only enhanced the precision and scope of rock art documentation but have also initiated innovative pathways for interpretation, preservation, and public engagement.
We invite papers that showcase comprehensive analyses achieved through digital methods, as well as presentations that introduce concise and novel approaches to the recording, displaying, and analysing of rock art. Submissions may highlight case studies, comparative methodologies, or the adaptation of emerging technologies to address specific challenges in rock art research. This session also aims to facilitate discussions on the theoretical and practical implications of digital methods, their role in cross-disciplinary collaborations, and the ethical considerations surrounding the use of these technologies in archaeological practice.
By bringing together researchers and heritage professionals who apply a diverse range of digital tools and techniques, this session aims to foster a dynamic exchange of ideas, share best practices, and drive innovation in rock art studies. We hope to deepen our understanding of rock art’s cultural importance while also advancing the capabilities, accessibility, and ethical frameworks of digital archaeology.
We welcome contributions from archaeologists and heritage professionals who are integrating digital methods into rock art research, whether through experimental approaches or proven methodologies. This session invites papers as well as presentations focused on specific methods or techniques which advance rock art research and recording. We also invite presenters to a short group discussion at the end of the session.
Papers:
Testing the Value and Limits of Photogrammetry and GIS to Record Ultrafine Palaeolithic Engravings
Alfredo Sanchez-Hernandez, Pre-EINA Research Group, Facultat de Ciències Humanes i Socials, Departament d’Història, Geografia i Art, Universitat Jaume I, Spain
Inés Domingo, Secció de Prehistòria i Arqueologia, ICREA-Universitat de Barcelona, Spain
Dídac Roman, Pre-EINA Research Group, Facultat de Ciències Humanes i Socials, Departament d’Història, Geografia i Art, Universitat Jaume I, Spain
A recent review of the Palaeolithic portable art collection from Matutano cave (Castellón, Spain) has drawn attention to the need to find more accurate and objective methods of documenting very fine prehistoric engravings. This conclusion was reached following the discovery that a significant proportion of the pieces categorised as engravings were, in fact, natural marks. It is therefore imperative that improvements are made to the accuracy of recording techniques in order to differentiate between anthropogenic and natural marks on rock surfaces. To address this issue, a pioneering method initially developed by Jalandoni and Kottermair (2019) was tested. This approach involves the use of Digital Elevation Models (DEM) derived from photogrammetry, enabling the precise measurement of engraving grooves at both centimetre and millimetre scales. This facilitates comprehensive and objective assessments of their depth and width. The topographic position index is then applied and vectorised in GIS, facilitating the automated extraction of engravings and minimizing subjectivity in interpretation.
While this technique has proven effective for well-defined engravings, its applicability to very fine incisions, such as those of the Cova Matutano collection, remained untested. This communication critically evaluates the potential and limitations of the method when adapted for the documentation of subtle engravings.
The Importance of Data Annotation in the Study of Rock Art at Injalak Hill, West Arnhem Land, Northern Territory, Australia
Benedict Dyson and Andrea Jalandoni, Griffith University, Australia
Accurate data annotation is essential for training deep learning models that can identify, classify, and segment images (e.g., SAM2 or Mask RCNN). Pre-trained models are particularly beneficial in heritage research, as they require less data for fine-tuning to specific use cases. However, effective fine-tuning depends on the availability of precisely annotated datasets. In this work, we employ semi-automated rock art annotation using the prompt-based model DEXTR, combined with decorrelation and denoising algorithms. This integrated approach enhances annotation accuracy by making rock art motifs more apparent and reducing noise from the rock surface. The resulting per-pixel annotations will be used to fine-tune an instance segmentation model, enabling automated data annotation at Injalak Hill. We will explore the analytical potential of fully annotated datasets, which can facilitate finely detailed, motif-level analyses unattainable through traditional methods. These analyses may reveal spatiotemporal patterns in rock art placement, establish site-wide chronologies, and support more fine-grained typological studies. Additionally, this paper addresses the ethical considerations of applying AI models to Aboriginal rock art, including mitigation strategies at Injalak Hill and the implications of sharing AI models trained on Indigenous data.
Using Digital Archaeology to Assist Indigenous Communities Engage with Their Rock Art in New Ways
Andrea Jalandoni, Place, Evolution, Rock Art, Heritage Unit (PERAHU); Australian Research Centre for Human Evolution (ARCHE); Griffith Centre for Social and Cultural Research (GCSCR), Griffith University
Digital methods, such as 3D modelling and DStretch, are already standard (or should be) in rock art research because they produce more accurate recordings and are more cost-effective than traditional methods. In recent years there have been significant innovations through assimilating methods from other disciplines (e.g., data science and remote sensing) to create more and better data. The data serves as a baseline to monitor against past and future change and helps archaeologists make better interpretations. These advances not only improve research potential but have also helped Indigenous and non-Indigenous people to engage with their cultural heritage more tangibly.
In this presentation I will showcase a few of my community-led or partnered digital archaeology projects: completed, in progress, and aspirational. They include assisting Chamoru in the Northern Mariana Islands of Tinian and Aguiguan document their cultural heritage; working with the Bininj in western Arnhem Land to create a digital twin of Injalak Hill (AU) to tease out spatial questions about the rock art and reveal ancient decision-making; and helping a Gooreng Gooreng artist from Bundaberg (AU) put his rock art back together.
Virtual Reality for Gunbalanya: Turning Digital Archaeological Data into Immersive Experiences
Calum Farrar and Andrea Jalandoni, Griffith University, Australia
The ever-expanding range of digitised cultural heritage from Gunbalanya, just a few hours east of Darwin, requires more creative approaches to incorporate different kinds of data and its context. Rock art and bark paintings have been the main focus of these digitisation efforts as important aspects of local cultural life in the present and deep into the past. Virtual Reality (VR) headsets are being produced with greater and more diverse capabilities, expanding the ranges of options available to archaeologists for placing digital copies of sites and artefacts into immersive, interactive environments. Initial use of VR in Gunbalanya has focused on interactive 3D models of some of the 100+ year old bark paintings held in the Melbourne Museum, allowing remote access to a famous collection. Following the positive feedback from the Gunbalanya community, particular its practicing artists, development has continued on new ways to interact with digitised cultural heritage and situate it in the unique environments and narratives of the region. This presentation exhibits the latest developments, future directions and ties in with a practical demonstration in the Art at a Crossroads Creative Exhibit.
Contribution to a New Method for Deciphering the Gender of Hand Stencils
Jean-Michel Chazine, Centre de Recherche et de Documentation Pluridisciplinaire (CREDO) Aix-Marseille Université (AMU), Centre National de la Recherche Scientifique (CNRS), Marseille, France
Benjamin Pillot, UMR Espace-Dev IRD, Montpellier, France
Between 1992-2005, a French-Indonesian research team discovered and assessed 2000 hand stencils in East Kalimantan caves. We initially developed the ©kalimain1.0 software and used the 2D:4D digit ratio to try determine the gender of their creators. However, this method faced limitations, asking for a more advanced approach based on new morphometrics data. Thus, a new upgraded software (©kalimain 2.0) was built by B. Pillot upon a deep neural network trained with the recent morphometrics findings from Nelson et al. (2017). Overall, it enables us to: 1) sketch out hand outlines that are then processed by the underlying model, 2) retrieve hand dimensions, and 3) estimate gender accordingly. To unleash the full potential of ©kalimain 2.0 for sex prediction, we eventually analysed an unrelated collection of 100 contemporary handprints gathered by Mackie (2014). The stencils from this dataset were traced using a spray technique close to the general stencilling process. We show that a well calibrated model using machine learning is able to accurately distinguish between males and females from unknown hand stencils. In the case of the Mackie dataset, ©kalimain 2.0 is actually able to classify more than 90% of the stencils successfully.
Digitally Excavating Rock Art: Can Digital ‘Archaeological Excavation’ of Rock Art Provide Insights into Past Environmental, Social and Cultural Conditions and Human Interactions from Djulirri and Red Lily Lagoon, Arnhem Land, Northern Territory
Dr Samuel Dix, Archaeology, Flinders University; Biosis Pty, Ltd, Australia
Emily Ward, Dr Daryl Wesley and Dr Jarrad Kowlessar, Archaeology, Flinders University, Australia
In archaeological excavation, stratification theory is universally used in understanding the history and land disturbances that make up an archaeological record. Commonly, these are defined by material being added, removed, excavated, filled in, and so forth. This stratification is the primary tool archaeologists use in the understanding of site formation. Recent studies have demonstrated that analysing the superimposition of rock art is similar to the archaeological excavation process. Digital techniques now allow researchers to virtually ‘excavate’ rock art sites, applying the principles of traditional archaeology to uncover insights into past ecological conditions, social dynamics, and human interactions. Using examples from Djulirri and Red Lily Lagoon in Arnhem Land, Northern Territory, this paper demonstrates how digital excavation of rock art reveals important interpretations and narratives from its stratigraphy, similar to traditional ‘dirt archaeology’.
Gulamada Project: Utilising Digital Archaeology for Rock Art Research and Management in the Blue Mountains
Lauren Roach, Wayne Brennan, and Dr Amy Mosig Way, The University of Sydney, Australia
The Blue Mountains area in NSW Australia has had limited research performed in terms of Aboriginal cultural heritage sites. There are many sites not recorded, and those which are often lack management. In an area prone to bushfires, it is important to have an accurate inventory of sites which can be appropriately managed by community and professionals. The Blue Mountains consists of a vast network of sites within a cultural landscape, including rock shelter sites with pigment art, and platform rock engravings
Our Gulamada Project emphasises a collaboration of science and culture; utilising scientific and digital technologies to address cultural concerns. We are building an accessible, online geospatial data collection and management system named the ‘Coolamon’ in partnership with local Aboriginal groups. The system will ultimately be managed by these local custodians for use in ongoing research and management.
The Blue Mountains is a difficult, vast area to traverse due to thick bush and steep terrain. During fieldwork we are undertaking drone-based pre-survey flights to assist with identification of rock shelters and sandstone platforms which are targeted during pedestrian surveys. The team is producing 3D photogrammetry models, including aerial models of engraving platforms and SLR models of rock art shelters.
Digital Archives and Rock Art—A Case Study in Putting an African Rock Art Archive Online
Benjamin W. Smith and Tyler M Wilce, School of Social Sciences, University of Western Australia, Australia
The Peter Genge Collection at the University of Western Australia contains some 12,000 digital files relating to rock art sites, mostly in south-west Zimbabwe, Africa. The records in the collection include: site cards; photographs; 35mm slides; sketches; notebook entries; and other ephemera. The collection was put together in the 1980s and 1990s and contains the only recordings of many sites, including of rock paintings that have since been damaged or destroyed. The collection is of enormous social and scientific significance. Our challenge will be familiar to many people, how can we bring this collection online and make it searchable so as to meet the aspirations of many different audiences? At the same time, we need to suppress some aspects of the collection to honour cultural sensitivities and protect the sites. In this paper we discuss the approaches we have taken to digitising this collection, and to rebuilding it as an online searchable digital archive.
An Immersive Perspective: A Data Driven Approach to Virtual Cultural Landscape Simulation for Rock Art Sites
Jarrad Kowlessar, Daryl Wesley and Ian Moffat, Archaeology, Flinders University, Australia
A major challenge of rock art research is to understand and experience rock art in the context of its original inscription. This often requires the synthesis of a multitude of data to develop archaeological models of past landscapes and communicate these models in a way that mirrors the practical exploration of place yet can occur in a digital space. This paper addresses this challenge using a Virtual Cultural Landscape approach, thereby extending the models of past landscapes previously created from separate digital data sources, and combining them into immersive visualisation, providing a means of meaningful engagement with rock art and its past landscape. This approach aims to leverage the interpretive benefits of phenomenological engagement with space, to better interpret and communicate the combined results derived from the multiple methodological approaches. This paper will also explore different modes of engaging with virtual cultural landscape models viewing including virtual production facilities, virtual reality and mixed reality. The paper presents research conducted in Red Lily Lagoon study area in Kunwinjku country in Western Arnhem land Australia.
Stable Confusion? Exploring the Potential of Artificial Intelligence in Rock Art Research
Mark Willis, Sacred Sites Research, USA
Recently, Artificial Intelligence (AI) has garnered significant attention for its remarkable ability to generate visually striking and artistic images from little more than brief text prompts. What was once reserved for those with access to powerful computers and technical expertise is now widely available to the general public, enabling even those without specialised knowledge to harness its capabilities. This advancement raises intriguing questions about how these tools might be applied to the study and documentation of rock art. Specifically, could AI potentially uncover patterns in rock art styles that human researchers might overlook, or could it independently verify characteristics that define specific artistic styles? This paper aims to explore the potential applications of AI in the field of rock art by training an AI model to recognise, replicate, and analyse various styles of prehistoric and historic rock art. In addition to examining the technical possibilities, the paper will also delve into the moral and ethical implications of using AI in cultural and archaeological contexts, considering the impact on authenticity, preservation, and cultural sensitivity.
Digital Access: Managing Disaster Through Sustainable Documentation at Fels Cave, Vanuatu
Chris Ballard, Dept of Pacific and Asian History, School of Culture, History and Language, College of Asia & the Pacific, The Australian National University, Australia
Meredith Wilson, independent researcher
Terhi Nurmikko-Fuller, The Australian National University, Australia
Kit Nelson, The Australian National University, Australia
Edson Willie, Vanuatu Cultural Centre, Vanuatu
Richard Matanik, Lelema World Heritage Committee
This paper describes the context for a photogrammetric documentation of the interior of the chamber cave of Fels, on the island of Lelepa in Vanuatu. Fels is one of the largest caves in the Pacific and has one of the richest records of rock art in the region. As the location of the death of the 16th-century chief Roi Mata, it is also one of the key places in the UNESCO World Heritage cultural landscape of Chief Roi Mata’s Domain. The documentation project was part of a precautionary program, in light of a local history of earthquakes and cave instability, to determine the safety of visitors to the cave. The documentation was completed in September 2024 and was followed by the 17 December earthquake which devastated Efate and Port Vila, and also destroyed the entrance and all access to Fels Cave. While our ability to return to the cave remains in question, we are exploring the scope for digital access to the cave and its rich record of rock art, for the community, visitors and researchers, and developing approaches and technologies that place future monitoring and documentation in the hands of the custodial community and the Vanuatu Cultural Centre.
How Portable Digital Technology is Transforming How We Engage with Rock Art on Flinders Island Group, Cape York Peninsula
Olivia Arnold, The University of Queensland, Australia
Digital technology has transformed rock art recording and analysis by providing fast and effective tools. In 2022, the Flinders Island Group Rock Art Project (FIGRAP) commenced its first fieldwork, integrating portable LiDAR to overcome time constraints and site accessibility challenges. At the time, portable iPhone LiDAR was a novel method recommended for documenting small (i.e. ca. 3x3m) spaces and untested with Australian rock art. However, FIGRAP successfully applied this new technology to sites ranging in length from 5m to 30m. Over two fieldwork seasons, the team visited 12 sites on three islands, and recorded 1,469 painted motifs. The use of digital technology facilitated participation from Traditional Owners, including, importantly, those unable to physically access some sites. After each fieldwork day we shared the LiDAR and DStretch-processed images plus drone footage with Elders, which sparked valuable discussions around site interpretation and management. These digital methods have introduced innovative ways to analyse, share, and experience rock art but strengthened collaboration with Traditional Owner and external stakeholder groups. This paper explores opportunities that have emerged from integrating these technologies, including developments of digital exhibitions that have the potential to contribute to the long-term management and protection of the islands’ culturally significant rock art.
Re-examining Australia First Nations Rock Art Archives with Digital Technologies
Madeleine A. Kelly, ARC Centre of Excellence for Australian Biodiversity and Heritage, Monash University, VIC 3800, Australia; College of Science and Engineering, Flinders University, Australia
This paper discusses the application of modern digital technologies in the analysis of an archive of rock art recordings from the 1980s and 1990s. I present a systematic archaeological investigation of a rock shelter in Wardaman Country, Northern Australia, employing visual enhancement techniques on high-quality film photographs and conducting a comprehensive examination of rock art conventions and superpositions using the Harris Matrix program. The analysis demonstrates how a detail and comprehensive examination of superpositions using Harris Matrices, discussed in light of ontological and epistemological perspectives of Wardaman Elders and Traditional Owners, offers greater insight into First Nations negotiations of cultural and pastoral responsibilities during early Western colonisation. Furthermore, the paper argues that the use of archival rock art recordings facilitates research in locations that are otherwise difficult to access and generates opportunities for cross-generational dialogue between researchers and First Nations communities, which illuminates historical archaeological practices in Australia.