| Geoscientists frequently encounter the problem of analyzing and visualizing 3D geological features. Three-dimensional geographic information systems (3D GIS) offer these capabilities, however, a major obstacle in the development of 3D GIS is the large quantity of data associated with 3D representations. Data structures based on recursive decomposition modelling, such as octrees, provide a means of efficiently storing 3D representations while supporting fast geoprocessing algorithms. Tests were conducted to determine the advantages and disadvantages of various octree forms for representing geological features. Regular, grouped and linear octrees were constructed and tested using identical data sets. The regular octrees were the least efficient in terms of data storage. Linear octrees provided the most substantial data compaction but problems with data retrieval were encountered. The grouped octrees were found to offer the best combination of properties. They were able to store the representations efficiently yet supported fast geoprocessing functions. |
PDF Full Text Request