Abstract

The rich archaeological heritage of the Veluwe, a large forested area in the central Netherlands, has been investigated for decades through field survey and excavations of selected sites, e.g. prehistoric burial mounds. A recently released LiDAR-derived elevation model of the entire Netherlands offers the opportunity to study the archaeological landscape of the Veluwe on a regional scale. However, this requires new, efficient and effective methods for the archaeological analysis of large remote sensing datasets.

Since 2017 we have been developing WODAN , the first automated detector of multiple classes of archaeological objects in LiDAR data. This algorithm is based on Deep Learning, which requires labeled input data for training. In order to acquire such data and to compare the performance of computational methods to those of human operators, we also conducted a Citizen Science project in which hundreds of volunteers mapped the same classes of archaeological objects in LiDAR data in Zooniverse, an online platform for Citizen Science projects.

In this talk I will present the current state of the Heritage Quest project and discuss opportunities for a fruitful integration of Data Science and Citizen Science approaches for the benefit of archaeological landscape studies.