Automated Archaeological Feature Detection Using Deep Learning on Optical UAV Imagery: Preliminary Results

This communication article provides a call for unmanned aerial vehicle (UAV) users in archaeology to make imagery data more publicly available while developing new application facilitate the use of common deep learning algorithm (mask region-based convolutional neural network; Mask R-CNN) instance segmentation. The intent is provide specialists with GUI-based tool that can apply annotation used training network models, enable and development segmentation allow classification auto-discovery features. generic be variety settings, although was tested using datasets from United Arab Emirates (UAE), Oman, Iran, Iraq, Jordan. Current outputs suggest trained are able help identify ruined structures, is, structures such as burials, exposed building ruins, other surface features some degraded state. Additionally, qanat(s), or ancient underground channels having access holes, mounded sites, which have distinctive hill-shaped features, also identified. Other classes possible, helps their own training-based approach feature identification classes. To improve accuracy, we strongly urge greater publication UAV by projects open journal publications public repositories. something done fields now needed heritage archaeology. Our provided part given.