Show simple item record

dc.rights.licenseCC-BY-NC-ND
dc.contributor.advisorExterne beoordelaar - External assesor,
dc.contributor.authorRiesner, Christian
dc.date.accessioned2022-05-11T00:00:39Z
dc.date.available2022-05-11T00:00:39Z
dc.date.issued2022
dc.identifier.urihttps://studenttheses.uu.nl/handle/20.500.12932/41559
dc.description.abstractThe Jewish cemetery in The Hague is a listed national monument of The Netherlands and is a document of the Jewish heritage in the area. The site is occupied by more than 10.000 people buried there. Only 2.860 can be located on the ground as they are marked with a gravestone on their surface top. The spatial positions of the remaining unmarked burial locations are of high interest to cemetery management and the Jewish community. As a physical inventory is not possible, non-invasive methods are preferred. Ground Penetrating Radar (GPR) is a technology that has been successfully applied to detect sub-surface features. A survey with multiple sensors, including GPR, was undertaken on two sample areas of the cemetery. The conventional method to analyse the collected data sets is the interpretation of vertical and horizontal visualisations. This method required expert knowledge. An alternative approach is to integrate the GPR data into a Geographic Information System (GIS) in the form of a 3D-point cloud. The data assessment was supported by applying the deep learning algorithm, PointCNN, that performs an automated classification of the points to segment the burial locations as features of interest. In this research, various classification models were created, applied, and evaluated to their usability. Furthermore, additional data sets were combined with the classified point clouds to incorporate more information to identify the unmarked burials. The applied method highlights the advantages of a GIS for this spatial analytical task and reveals some vulnerable points in the application of deep learning. Therefore, the workflow in this research project will have to undergo adaptations as recommended at the end of the report.
dc.description.sponsorshipUtrecht University
dc.language.isoEN
dc.subjectGround-penetrating radar (GPR) data were used in the 3D point cloud representation to be integrated into a GIS for the automated classification via deep-learning methodology.
dc.titleIntegrating Ground Penetrating Radar data in a GIS to detect unmarked graves via a Convolutional Neural Network framework for 3D point cloud interpretation - A case study at the Jewish Cemetery in The Hague, The Netherlands
dc.type.contentMaster Thesis
dc.rights.accessrightsOpen Access
dc.subject.keywordsGIS; GPR; point cloud; deep-learning; grave; classification; ArcGIS; detection; subsurface; model
dc.subject.courseuuGeographical Information Management and Applications (GIMA)
dc.thesis.id3797


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record