|
|
|
| RI | Thesis Proposal | 16 May 2003 | |
Robotics Institute Thesis Proposal 16 May 2003
Place and Time |
Abstract |
Further Details |
Thesis Committee
Image and Range Coherent Surface Modeling
Pragyana Mishra
Robotics Institute
Carnegie Mellon University
| Place and Time |
NSH 3002
8:00 AM
| Abstract |
I propose a new framework for image and range coherent surface modeling, a unified approach for generating a surface model from 3D range data that is coherent with a sequence of images. Generating 3D models requires
approximating a particular surface by a mesh of simplicial elements like
triangles or tetrahedra. These elements approximate surface appearance with their corresponding texture maps. The proposed framework generates 3D models by aligning a 3D point-cloud and a set of 2D images and then
generating a mesh that best accounts for the given image sequence and 3D
range data.
The alignment minimizes a distance measure relating the statistics of two
randomly sampled data sets, in our case, a range map and intensity images.
Sensor data have regions best suitable for computing a statistic while parts of it are cluttered or unresolved. The framework combines statistics from separate local regions of data by exploiting the generalized Crofton's
theorem in integral geometry.
Once the 3D range-data is aligned with respect to a set of images, the
framework generates, refines, and fairs a mesh such that its projection into all of the images is as close as possible to the real image sequence. The error between the projected and the actual images is used to refine a mesh that closely approximates the surface being viewed.
Through this framework, I introduce the idea of EigenFairing, an
image-coherent mesh fairing technique that exploits the distance-from-feature-space of texture-mapped faces to incrementally fair a mesh to lie closer to the real surface. The resulting model in turn, helps
to better estimate local statistics of the surface for establishing
correspondence and alignment with images. This novel approach would yield the closest possible approximation of the surface, given a 3D point-set and an image sequence.
| Further Details |
A copy of the thesis proposal document can be found at http://www.cs.cmu.edu/~pragyan/proposal.pdf.
| Thesis Committee |
This page maintained by robotwebmaster@ri.cmu.edu.