The objective of this project is to demonstrate and validate the use of machine learning (ML) based predictive modeling (PRMML-CRS) to help DoD installations identify areas likely to contain historic artifacts and/or assets. This tool minimizes the need for extensive field surveys, reducing delays, lowering costs, and improving efficiency in construction and development projects. 

PRMML-CRS takes an innovative approach to cultural resources (CR)surveys by integrating ML into remote sensing to rapidly identify areas with a high probability of CR and provides that in Geographic Information Systems (GIS) format to allow CR managers easy access to the information to support informed decision-making. PRMML-CRS consists of four primary steps: 1) applying ML on satellite imagery to identify areas with high potential for CR, 2) using unmanned aircraft systems (UAS) to survey and collect high- resolution imagery/sensor data of these same areas, 3) using ML to identify CR in UAS-collected data, and 4) integrating data and survey results into GIS. The effectiveness of PRMML-CRS will be compared to the most common CR survey method, ground surveys, in terms of ability to detect CR, survey efficiency, and cost- effectiveness. Results from the project can show CR managers how remote sensing, UAS technology, and ML can be used to support CR management. Successful technology transition of PRMML-CRS technology will include the development of a GIS tool that will help CR managers to apply PRMML-CRS at their specific locations. Because PRMML-CRS is modular, it can be used at all installations, and individual installations can apply the steps that are relevant to their location, as time and funding permits. 

Most cultural resource surveys currently rely on boots-on-the-ground field crews, which are often limited to small-scale, project-specific surveys rather than comprehensive installation-wide assessments. PRMML-CRS enables a landscape-level evaluation of CR, narrowing focus to specific locations and significantly reducing the need for ground surveys. This tool can save tens of thousands of dollars per survey and shorten approval timelines by providing accurate CR inventories, accelerating consultations and land use assessments. The high return-on-investment (ROI) from this time savings translates directly into improved installation readiness, preventing costly delays to mission-critical activities and supporting faster, more efficient development.