Machine learning techniques for structural health monitoring of heritage buildings: A state-of-the-art review and case studies

This paper presents WODAN2.0, a workflow using Deep Learning for the automated detection of multiple archaeological object classes in LiDAR data from the Netherlands. WODAN2.0 is developed to rapidly and systematically map archaeology in large and complex datasets. To investigate its practical value, a large, random test dataset—next to a small, non-random dataset—was developed, which better represents the real-world situation of scarce archaeological objects in different types of complex terrain. To reduce the number of false positives caused by specific regions in the research area, a novel approach has been developed and implemented called Location-Based Ranking. Experiments show that WODAN2.0 has a performance of circa 70% for barrows and Celtic fields on the small, non-random testing dataset, while the performance on the large, random testing dataset is lower: circa 50% for barrows, circa 46% for Celtic fields, and circa 18% for charcoal kilns. The results show that the introduction of Location-Based Ranking and bagging leads to an improvement in performance varying between 17% and 35%. However, WODAN2.0 does not reach or exceed general human performance, when compared to the results of a citizen science project conducted in the same research area.

Combining Deep Learning and Location-Based Ranking for Large-Scale Archaeological Prospection of LiDAR Data from The Netherlands

This work presents a semi-automatic approach to the 3D reconstruction of Heritage-Building Information Models from point clouds based on machine learning techniques The use of digital information systems leveraging on three-dimensional (3D) representations in architectural heritage documentation and analysis is ever increasing For the creation of such repositories, reality-based surveying techniques, such as photogrammetry and laser scanning, allow the fast collection of reliable digital replicas of the study objects in the form of point clouds Besides, their output is raw and unstructured, and the transition to intelligible and semantic 3D representations is still a scarcely automated and time-consuming process requiring considerable human intervention More refined methods for 3D data interpretation of heritage point clouds are therefore sought after In tackling these issues, the proposed approach relies on (i) the application of machine learning techniques to semantically label 3D heritage data by identification of relevant geometric, radiometric and intensity features, and (ii) the use of the annotated data to streamline the construction of Heritage-Building Information Modeling (H-BIM) systems, where purely geometric information derived from surveying is associated with semantic descriptors on heritage documentation and management The “Grand-Ducal Cloister” dataset, related to the emblematic case study of the Pisa Charterhouse, is discussed

From the Semantic Point Cloud to Heritage-Building Information Modeling: A Semiautomatic Approach Exploiting Machine Learning

Sixty-six new archaeological sites have been discovered thanks to the combined use of different remote sensing techniques and open access geospatial datasets (mainly aerial photography, satellite imagery, and airborne LiDAR). These sites enhance the footprint of the Roman military presence in the northern fringe of the River Duero basin (Leon, Palencia, Burgos and Cantabria provinces, Spain). This paper provides a detailed morphological description of 66 Roman military camps in northwestern Iberia that date to the late Republic or early Imperial eras. We discuss the different spatial datasets and GIS tools used for different geographic contexts of varied terrain and vegetation. Finally, it stresses out the relevance of these novel data to delve into the rationale behind the Roman army movements between the northern Duero valley and the southern foothills of the Cantabrian Mountains. We conclude that methodological approaches stimulated by open-access geospatial datasets and enriched by geoscientific techniques are fundamental to understand the expansion of the Roman state in northwestern Iberia during the 1st c. BC properly. This renewed context set up a challenging scenario to overcome traditional archaeological perspectives still influenced by the cultural-historical paradigm and the pre-eminence of classical written sources.

Following the Roman Army between the Southern Foothills of the Cantabrian Mountains and the Northern Plains of Castile and León (North of Spain): Archaeological Applications of Remote Sensing and Geospatial Tools

Marina Khoroshiltseva

Papers

Machine Learning for Cultural Heritage: A Survey