Point Clouds Segmentation of Mixed Scenes with Archeological Standing Remains: A Multi-Criteria and Multi-Scale Iterative Approach

Abstract

The integration of Airborne Laser Scanning (ALS) survey into archaeological research and cultural heritage management has substantially added to our knowledge of archaeological remains in forested areas, and is changing our understanding of how these landscapes functioned in the past. Further, the results of ALS-based surveys of woodlands can now potentially be incorporated into micro-regional and landscape scale studies which rely on survey data, making an important contribution to our understanding of settlement patterns and land use – but doing so requires us to recognize and manage a host of biases inherent in ALS-based survey data. While many types of archaeological remains manifest as micro-topography, several important classes of features commonly appear as standing remains. The identification of these remains is important for archaeological prospection surveys based on ALS data. Standing structures in mixed scenes with vegetation are not well addressed by standard classification approaches. In this paper we propose an approach to the identification of these structures in the point cloud based on multi-scale measures of roughness, and measures of local density and normal orientation. We demonstrate this approach using discrete-return ALS data collected in the Franche-Comté region of France.