A 3D Point Cloud Deep Learning Approach Using LiDAR to Identify Ancient Maya Archaeological Sites

Research demonstrates that deep learning models using 3D point cloud data from LIDAR outperform those relying solely on 2D data for identifying ancient Maya archaeological sites, such as the UNESCO World Heritage Site of Copán in Honduras. This approach significantly reduces data preparation requirements while achieving higher classification accuracy. The study also explores the effectiveness of data augmentation strategies for small 3D datasets, offering a more efficient and precise method for analyzing complex anthropogenic landscapes. By leveraging advanced object recognition algorithms, this work advances automated archaeological site detection, addressing the challenges posed by vast LIDAR datasets.

Richards-Rissetto, H., Newton, D., & Al Zadjali, A. (2021). A 3D POINT CLOUD DEEP LEARNING APPROACH USING LIDAR TO IDENTIFY ANCIENT MAYA ARCHAEOLOGICAL SITES. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, VIII-M-1-2021, 133–139. https://doi.org/10.5194/isprs-annals-viii-m-1-2021-133-2021