Wednesday, 27 April 2022

8:45-10:00Workshop session 0: Gentle introduction in the archaeological interpretation of lidar derived data
10:30-11:45Workshop session 0: Gentle introduction in the archaeological interpretation of lidar derived data
13:30-14:00Introduction to the meeting
14:00-16:00Keynote 1 by Benjamin Štular and Edisa Lozić
Airborne lidar data processing in the context of contemporary landscape archaeology
Keynote 2 by Žiga Kokalj
Machine learning, airborne laser scanning and archaeology
Keynote 3 by Damian Evans
A global overview of technical, ethical and other issues in the acquisition and use of lidar data
16:00-17:00coffee break and poster session
17:00-18:30Invited lecture 1 by Laure Nuninger, Philip Verhagen and Catherine Fruchart
Linking theories, past practices, and archaeological remains of movement through ontological reasoning
Invited lecture 2 by Rachel Opitz
Re-thinking the landscape through the lens of archaeological ALS
Invited lecture 3 by Cinzia Bettineschi and Luigi Magnini
Fly now, shovel later – an analytical perspective for UAV prospection in archaeology

Thursday, 28 April 2022

8:45-10:00Introductory discussion about workshops
10:00-10:30coffee break
10:30-12:30Workshop session 1
13:30-16:30Field trip afternoon (2 choices) – a walk or the cave
17:00-19:00Workshop session 2
19:30-21:00International dinner

Friday, 29 April 2022

8:45-10:45Workshop session 3
10:45-11:15coffee break
11:15-12:30Round table discussion and closing session


Gentle introduction in the archaeological interpretation of airborne lidar derived data

Benjamin Štular, Edisa Lozić, Elise Fovet
In this introductory hands-on workshop, participants will be guided step-by-step through the basic workflow for using lidar data in a local-scale project based on visual, expert interpretation. The main steps of the process will be explained, from accessing point cloud data to creating a record of your archaeological interpretation of the features you identified using the Open LiDAR Toolbox, an open source QGIS plugin. After you have completed this workshop, you will be familiar with the key concepts and practical skills involved in:

  • Data acquisition,
  • Point cloud data classification,
  • DEM Interpolation,
  • DEM visualisation,
  • geodatabase setup, and
  • visual archaeological interpretation.

Practical exercises will focus on archaeology-specific concerns relevant for point cloud classification, DEM interpolation and data visualisation. Because this is a hands-on workshop, participants are expected to be familiar with a GIS package (either ArcGIS or QGIS) and to bring their laptops with software installed (QGIS, Open LiDAR Toolbox plugin). If you lack prior GIS experience, it is possible to participate as an observer. In this case, you can expect to become familiar enough with the workflow to be able to contribute to a lidar project in collaboration with technical co-workers.

Deep Learning: Workflow

Wouter Verschoof-van der Vaart, Lucy Killoran, Žiga Kokalj
This workshop will explore the latest developments in the implementation of Convolutional Neural Networks (CNNs), the main algorithms used in Deep Learning (DL). A CNN is a feature extractor and image-classifier that learns to generalize from given examples (i.e. a large set of labelled images) rather than relying on a human operator to set parameters or formulate rules.
First, we will briefly recap and build on the introduction to Deep Learning keynote. Then, in the workshop, we will practically demonstrate that in order to successfully detect particular archaeological traces in the landscape, different processing steps are needed. We will try our hands on detecting hollow roads (medieval cart tracks) in lidar data from the Veluwe, in the central part of the Netherlands. Subsequently we will use a recently developed workflow called Carcassonnet to detect these cart tracks. Participants are encouraged to bring their own lidar data containing hollow roads.

Mindful creation of learning sets

Rachel Opitz, Dave Cowley, Laure Nuninger, Alexandre Guyot
This workshop focuses on the challenges of creating learning sets for ML-driven archaeological lidar projects workflows. The role of learning sets in ML and basic principles for their creation will be reviewed. The workshop will focus on discussions of how to define what you’re trying to detect, how to approach ‘cleaning’ datasets created based on the historic environment record or pre-existing survey data, working with point and polygon definitions of feature location, and the reuse of existing learning sets from different regions. The workshop will include a practical critique of example learning sets as an exercise.

Archaeological applications of UAV-borne lidar

Elise Fovet, Carine Calastrenc, Nicolas Poirier, Cinzia Bettineschi
This workshop introduces the specific technical requirements for using UAV (drone) platforms to acquire lidar data and the methods used for the processing of these datasets. The workshop will provide an overview of the key steps, from the planning of field missions to the creation and interpretation of DTMs. The quality of the lidar data and the derived elevation models will be discussed in detail (point density, spacing, accuracy). In this workshop’s hands-on exercises, participants will compare UAV lidar datasets with airborne lidar datasets in a series of exercises designed to illustrate the differences in data quality and build their familiarity with UAV lidar data characteristics.


Airborne lidar data processing in the context of contemporary landscape archaeology

Benjamin Štular, Edisa Lozić
The use of topographic airborne lidar data has become an essential part of archaeological prospection. From the outside, processing airborne lidar data may seem like a tedious and repetitive task that is completely objective and most of the scientific community’s attention is focused on archaeological interpretation. Therefore, all too often archaeologists rely on outside experts to process the airborne lidar data, and sometimes even the interpretive mapping is outsourced to non-archaeologists. As a result, DTM visualisations, for example, are usually accepted as “lidar images”: as facts rather than information (“facts” sensu David L. Clarke).
In this talk, we will present our personal and somewhat subjective view of the dichotomy between facts and “facts” – i.e., the data-information-knowledge pyramid – in the  context of landscape archaeology and airborne lidar data processing. In this way, we hope to initiate the discussion necessary for the lidar community to place data processing in the context of contemporary landscape archaeology.

Machine learning, airborne laser scanning and archaeology

Žiga Kokalj
As with spatial analysis and site recording in geographic information systems in the late 1980s and discovery and monitoring with remote sensing technologies and techniques in the 00s, archaeological and heritage management practices are slowly warming to the new tide of technological development of advanced machine learning techniques that support large-scale studies. While basic research on even the simplest problems, such as input data, is ongoing, encouraging results covering diverse geographic regions and archaeological features, mostly involving aerial laser scanning data, are sparking a vibrant community seeking to make the methods more easily understood and available. There are even start-up companies offering machine learning results as a product.
In this talk, I will present the current state of research, some available tools, and future prospects.

A global overview of technical, ethical and other issues in the acquisition and use of lidar data

Damian Evans
Aerial lidar acquisitions over wide areas are becoming increasingly common worldwide, and in many cases, the acquisitions are being driven by research programs in archaeology and related fields. The mechanics of these projects, and the life histories of the datasets they generate, are of interest not only for what they reveal about the landscapes of the places in question, but also because of the broader implications for the ways in which we acquire, use, and share archaeological lidar data. Some challenges are obvious – for example the persistent social and geographic inequalities in coverage, access, and the financial or technical capacity to acquire and use lidar data. Others are subtler and more complex, such as the delicate balance between principles of reproducibility and open science and the interests of specific groups in retaining sovereignty over three-dimensional facsimiles of cultural and natural heritage. This talk seeks to illustrate and explore many of these tensions, drawing on case studies worldwide, in order to promote consideration and discussion of the various issues involved.

Linking theories, past practices, and archaeological remains of movement through ontological reasoning

Laure Nuninger, Philip Verhagen, Catherine Fruchart
Over the past decade, the amount of information available to archaeologists has grown dramatically. The rapid acquisition of observational data, the creation of digital data based on its interpretation, and the digitisation of complementary data sources has played a significant role in this “information explosion”. New methods for knowledge creation are therefore needed for the present data-rich research context. This presentation focuses on the study of movement, drawing on rich digital datasets. Through a series of case studies, it aims to assess how researchers have identified, conceptualized, and linked the material traces describing diverse movement processes at a regional scale. It explores the construction of ontologies which enable us to explicitly relate material elements, identified in the observed landscape, to the knowledge or theory that explains their role and their relationships within movement processes. It concludes by illustrating how the breakdown of implicit conceptual references into explicit, logical chains of reasoning, describing basic entities and their relationships, allows the use of these constituent elements to reconstruct, analyze, and compare movement practices from the bottom up.

Re-thinking the landscape through the lens of archaeological ALS

Rachel Opitz
Research on long-term trajectories of landscapes, past and future, is quintessentially transdisciplinary. ALS data is drawn into diverse applications and methods, percolating across domain boundaries and building connections. In this paper, I reflect on the context, practices, and partial successes of archaeological uses of ALS to study the slippery entity of ‘landscapes’. In doing so, I will highlight the dense interconnections between contemporary public conversation around sustainability, the environment and landscape change, received ideas about the landscape, and new perspectives created through archaeological ALS-based research. I will then consider how we might position our archaeological ALS work to be more connected to work in other disciplines, agencies and communities invested in understanding the landscape.

Fly now, shovel later – an analytical perspective for UAV prospection in archaeology

Cinzia Bettineschi, Luigi Magnini
The rapid development of Unmanned Aerial Vehicles (UAVs) technology combined with advanced data processing techniques is paving new ways for documenting, investigating, and monitoring archaeological landscapes in their spatio-temporal complexity. Under certain conditions, UAV imagery can achieve these goals even better than satellite-based or traditional aerial remote sensing. This is due to the great flexibility in day/time/seasonality of the acquisitions, the ability to achieve extremely high spatial, spectral, and topographic resolution, and the availability of native multisensor configurations.
This contribution will discuss the current potential and future perspectives of UAV-based archaeological survey, including intelligent sensing, multiscale data integration, 3D point cloud modelling and processing, through a series of case studies, particularly from the Italian Alps.