|
Poisson Surface Reconstruction |
|
|
|
|
Written by Administrator
|
|
Sunday, 25 May 2008 |
|
Our on-site geometry capture work in the Yucatan yielded many different datasets, depending on the type of spatial sampling implemented. The majority of the data exists as point data.
Many techniques have been presented over the years to extract isosurfaces from point data. An interesting recent approach is "Poisson Surface Reconstruction" presented at Eurographics Symposium on Geometry Processing (2006) by Michael Kazhdan, Matthew Bolitho and Hugues Hoppe.
In this approach, a set of unstructured, unindexed point data is used as input:
0.007703 -0.068122 0.002691 -0.071399 0.535275 -0.841655
-0.002067 -0.088884 -0.049808 -0.302592 -0.878297 0.370179
0.033268 -0.068050 -0.068345 -0.045432 -0.997169 0.059915
-0.005424 -0.082875 -0.054366 -0.819175 0.573542 0.001197
-0.002053 -0.061424 -0.012995 0.888578 0.410287 -0.205168
0.035102 0.013736 -0.074576 0.775357 -0.517500 -0.361959
-0.013898 0.077890 0.070396 -0.740077 -0.060540 0.669792
0.030195 -0.045183 -0.014970 -0.490700 0.770551 -0.406775
For details, see the "Poisson Surface Reconstruction" paper here: http://www.cs.jhu.edu/~misha/MyPapers/SGP06.pdf
|
|
Last Updated ( Tuesday, 05 August 2008 )
|
