Semi-automatic Detection of Cultural Heritage in Lidar Data

This paper presents new methods for the semi-automatic detection of some kinds of cultural heritage in forested areas in Norway, and reports on a work in progress. Some areas have a large number of old pitfall traps that were used for deer hunting 2000-500 years ago. Other areas have a large number of iron production sites that were in use 1400-700 years ago. These two kinds of cultural heritage manifest themselves as pits in the terrain. We have developed methods for the automatic detection of such pits in lidar data with at least 5 emitted pulses per m. We are now extending the automatic detection methods to locate grave mounds, stone fences, and old roads in the lidar data. Experience from on-going archaeological field work clearly demonstrates the benefits of combining automatic detection methods with visual inspection of the lidar data to achieve a map of possible cultural heritage remains, before the actual field survey. The field work can be performed much more efficiently. Since the archaeological feature candidates have already been geo-referenced and measured in the lidar data, the field work is reduced to accepting or rejecting the candidates. Thus, a much larger number of archaeological features can be mapped per day.