Updating cadastre data is a modern challenge in many countries, including Slovenia. Although the use of drones in land registration is not yet common, it appears to be a promising technology that could support data collection for real estate cadastres. The purpose of the assignment is to review publications in the field of UAV photogrammetry in the cadastre and review the factors that influence the quality of cadastral data collection based on UAV photogrammetry. In the experimental part of the task, we analyzed and compared the results of capturing data on buildings and roads, which were captured in different ways on the basis of data obtained with the optical sensors of remote-controlled aircraft. We used already processed UAV photogrammetric data captured in the Kompolje study area. First, a comparison is made of the data on the perimeter of the buildings, which are covered manually on the basis of the photogrammetric point cloud, with those already registered in the building cadastre. The following is a comparison between the manually captured building perimeter data and the data obtained from the classified point cloud. In the following, we also compared the data on the above sea level of the ridge obtained by manual recording with the above sea level of the ridge registered in the building cadastre. Finally, a comparison is shown between the road data that we captured manually and the road data that we obtained using an automatic rule-based image classification process. We found positional differences between the data of the manual survey of buildings based on the photogrammetric point cloud and the data of the building cadastre. Data on buildings in the cadastre are entered on the basis of mass photogrammetric coverage of buildings from aerial images, and the photogrammetric point cloud was created from data with better spatial resolution. When comparing the results of manual data acquisition of buildings and roads with the data obtained by automatic point cloud classification by visual interpretation, we did not notice any major deviations. However, there are still cases where deviations occur - there are several reasons. Among other things, the shadows of trees and other objects with shadows represent a big problem for automatic data capture.
|
