SBIR-STTR Award

We propose a new method developed by our employees and collaborators for the digital representation of real world objects. This will result in an implicit surface urban terrain model that interpolates raw point cloud data. The method can handle complex topologies and geometries easily. Topology changes pose no difficulties. The key idea involves shrink wrapping an implicit surface around data points using the level set method. No assumptions about connectivity of these points is needed. New multiresolution techniques for speeding up the algorithm, reducing storage requirements, including information about the normals to the surface, weighting points according to their statistical significance, removing outliers, tensor voting, using dynamic visibility based interpolation, ENO interpolation and Level Set System's new, proprietary level set based compression will be developed and used here. New visual metrics involving Hausdorff distance for implicit surfaces will measure the accuracy and reliability of the resulting software package

Keywords:
LEVEL SET METHOD, IMPLICIT SURFACE, MULTIRESOLUTION, VISIBILITY,TENSOR VOTING, ENO INTERPOLATION, LEVEL SET BASED COMPRESSION, HAUSDORFF DISTANCE