Dr. Claudine Gravel-Miguel Presents Terrain Model Work At SAA Meeting

Dr. Claudine Gravel-Miguel Presents Terrain Model Work At SAA Meeting

Dr. Claudine Gravel-Miguel Presents Terrain Model Work At SAA Meeting

Dr. Claudine Gravel-Miguel, a research scientist with the Center for Applied Fire and Ecosystem Science (CAFES) at the New Mexico Consortium recently presented her work at the 89th Annual Society for American Archaeology (SAA) Meeting in New Orleans, Louisiana.

The goal of SAA Meetings is to advance archaeology, preservation, and practice. The SAA empowers its members to understand humanity’s past through ethically-based scientific and humanistic investigation, to promote preservation of archaeological resources and cultural heritage through support of legislation and education, to create collaboration between the profession and descendant communities, and to assist professional growth.

Gravel-Miguel’s presentation at this year’s meeting was titled: “Combining aerial LiDAR and Deep Learning to detect archaeological features in the Piedmont National Wildlife Refuge, Georgia”.

A growing number of archaeologists are using LiDAR-derived high-resolution Digital Terrain Models (DTM) to detect and document archaeological features. Early adopters of this method used visualizations to manually detect archaeological features; however, recent technological advances provide new tools that can considerably increase the efficiency and effectiveness of archaeological feature detection in DTMs.

In this presentation Gravel-Miguel presented her team’s methodology to detect archaeological features in images derived from LiDAR data, with a case study focused on the Piedmont National Wildlife Refuge.

Their team uses Deep Learning Convolutional Neural Network (CNN) models – a subset of the larger Machine Learning (ML) toolset – to quickly detect archaeological features over large areas. CNNs specialize in the analysis of visual imagery, making them particularly useful for the detection of visual features.

They trained their U-Net models to detect the presence and location of archaeological terraces in images of a landscape.

Most models performed very well and the best one was even able to detect terraces the team had overlooked. This demonstrates that this work can help agencies to identify and better protect the cultural resources that are on their land.

Other scientists who worked on this project include Grant Snitker with CAFES, Jayde Hirniak with CAFES and Arizona State University and Katherine Peck with CAFES and the University of New Mexico.