Archaeological Machine Learning: Using Machine Learning to Supplement Field Mapping
Topics: Spatial Analysis & Modeling
,
,
Keywords: Machine Learning, Artificial Intelligence, Mapping, Archaeology, Remote Sensing
Session Type: Virtual Paper Abstract
Day: Tuesday
Session Start / End Time: 3/1/2022 11:20 AM (Eastern Time (US & Canada)) - 3/1/2022 12:40 PM (Eastern Time (US & Canada))
Room: Virtual 21
Authors:
Leila Character, Department of Geography and the Environment, University of Texas at Austin
Tim Beach, Department of Geography and the Environment, University of Texas at Austin
Cody Schank, Department of Geography and the Environment, University of Texas at Austin
Takeshi Inomata, School of Anthropology, University of Arizona
Agustin Ortiz JR, Underwater Archaeology Branch, Naval History and Heritage Command
Adam Rabinowitz, Department of Classics, University of Texas at Austin
Sheryl Luzzader-Beach, Department of Geography and the Environment, University of Texas at Austin
Thomas Garrison, Department of Geography and the Environment, University of Texas at Austin
,
,
Abstract
This project entails creating a series of supervised machine learning models to predict the locations of archaeological features using remotely sensed imagery. Combined with field mapping, this method offers an efficient and fast approach to comprehensive archaeological site mapping. The project began in 2018 with the goal of creating a targeted method of finding cave entrances at Maya archaeological sites located in the dense tropical forests of Guatemala and Belize. In 2019, we used a random forest classifier, airborne laser scanning (ALS) data, and a training dataset of known caves to successfully identify several previously undocumented caves in northwestern Belize. Two of these caves contained archaeological materials. Building on this work, modeling has been expanded to include other types of potentially hidden and obscured features that colleagues are interested in studying. These include ancient Maya archaeological features in Guatemala and Mexico, shipwrecks off the coast of the United States, and ancient burial mounds in Romania. The models for the archaeological features take ALS, sonar, and multispectral imagery as input, are based on existing convolutional neural network architectures, and make use of transfer learning. These models can be used to create more accurate maps of archaeological features to aid management objectives, study patterns across the landscape, and find new features. Such models can easily be adjusted to identify other types of features and accept different types of imagery as input. This work seeks to make machine learning methods accessible to non-computer scientists interested in study, management, and conservation of archaeological heritage.
Archaeological Machine Learning: Using Machine Learning to Supplement Field Mapping
Category
Virtual Paper Abstract
Description
This abstract is part of a session. Click here to view the session.
| Slides
