ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., III-5, 121-128, 2016
http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/III-5/121/2016/
doi:10.5194/isprs-annals-III-5-121-2016
 
06 Jun 2016
3D CENTRAL LINE EXTRACTION OF FOSSIL OYSTER SHELLS
A. Djuricic1,2, E. Puttonen3,4, M. Harzhauser2, O. Mandic2, B. Székely5, and N. Pfeifer1 1TU WIEN, Department of Geodesy and Geoinformation, Vienna University of Technology, Vienna, Austria
2Geological Paleontological Department, Natural History Museum Vienna, Vienna, Austria
3Finnish Geospatial Research Institute in the National Land Survey of Finland, Department of Remote Sensing and Photogrammetry, Finland
4Centre of Excellence in Laser Scanning Research, Geodeetinrinne 2, Masala 02431, Finland
5Department of Geophysics and Space Science, Eötvös Loránd University, Hungary
Keywords: TLS, point cloud, DSM, Delaunay triangulation, Voronoi diagram, Bézier curve, Crassostrea gryphoides Abstract. Photogrammetry provides a powerful tool to digitally document protected, inaccessible, and rare fossils. This saves manpower in relation to current documentation practice and makes the fragile specimens more available for paleontological analysis and public education. In this study, high resolution orthophoto (0.5 mm) and digital surface models (1 mm) are used to define fossil boundaries that are then used as an input to automatically extract fossil length information via central lines. In general, central lines are widely used in geosciences as they ease observation, monitoring and evaluation of object dimensions. Here, the 3D central lines are used in a novel paleontological context to study fossilized oyster shells with photogrammetric and LiDAR-obtained 3D point cloud data. 3D central lines of 1121 Crassostrea gryphoides oysters of various shapes and sizes were computed in the study. Central line calculation included: i) Delaunay triangulation between the fossil shell boundary points and formation of the Voronoi diagram; ii) extraction of Voronoi vertices and construction of a connected graph tree from them; iii) reduction of the graph to the longest possible central line via Dijkstra’s algorithm; iv) extension of longest central line to the shell boundary and smoothing by an adjustment of cubic spline curve; and v) integration of the central line into the corresponding 3D point cloud. The resulting longest path estimate for the 3D central line is a size parameter that can be applied in oyster shell age determination both in paleontological and biological applications. Our investigation evaluates ability and performance of the central line method to measure shell sizes accurately by comparing automatically extracted central lines with manually collected reference data used in paleontological analysis. Our results show that the automatically obtained central line length overestimated the manually collected reference by 1.5% in the test set, which is deemed sufficient for the selected paleontological application, namely shell age determination.
Conference paper (PDF, 2576 KB)


Citation: Djuricic, A., Puttonen, E., Harzhauser, M., Mandic, O., Székely, B., and Pfeifer, N.: 3D CENTRAL LINE EXTRACTION OF FOSSIL OYSTER SHELLS, ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., III-5, 121-128, doi:10.5194/isprs-annals-III-5-121-2016, 2016.

BibTeX EndNote Reference Manager XML