The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLIII-B4-2022
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIII-B4-2022, 9–15, 2022
https://doi.org/10.5194/isprs-archives-XLIII-B4-2022-9-2022
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIII-B4-2022, 9–15, 2022
https://doi.org/10.5194/isprs-archives-XLIII-B4-2022-9-2022
 
01 Jun 2022
01 Jun 2022

DATABASE STORAGE AND TRANSPARENT MEMORY LOADING OF BIG SPATIAL DATASETS IMPLEMENTED WITH THE DUAL HALF-EDGE DATA STRUCTURE

P. Boguslawski1, P. Balak1, and C. Gold2 P. Boguslawski et al.
  • 1Institute of Geodesy and Geoinformatics, Wroclaw University of Environmental and Life Sciences, Wroclaw, Poland
  • 2Faculty of Computing, Engineering and Science, University of South Wales, Pontypridd, United Kingdom

Keywords: data structures, 3D modelling, DBMS, big data, dual half-edge, city model, BIM

Abstract. 3D spatial models covering big areas, such as cities, are widely developed in recent years. Loading of a whole model from a hard drive into a computer memory is often not possible due to big amount of data and memory size limitations. Optimisation techniques based on spatial indexing, such as tiling, are applied in order to load at least a part of a model as soon as possible, while the remaining parts are collected in the background. It is especially useful in visualisation of cities. A similar idea is proposed for a transparent loading of a model implemented with the dual half-edge (DHE) data structure and stored in a database. The existing DHE-based solutions require the whole model to be present in the memory, which is a considerable limitation in case of models covering big areas and including detailed representations of city objects, such as buildings and their interiors. The prototype mechanism developed in this work includes loading and unloading of model parts at a level of single edges as well as model tiling. This allows for spatial analysis without complete loading a big amount of data into memory.