This MATLAB demo computes the interpolatory SubdivisionSurfaces for a patch of the bathymetric data.

This is part of a tool that we use to explore interpolatory subdivision patch fitting decisions before implementing them as StreamingModules.

Install

Unzip the file Guinew.zip and point MATLAB to the directory. Data included in tempPadj.mat is from the bathymetric data supplied by Rob Anderson.

Instructions

In MATLAB, type demoInterpFit to run the demo. Data included in tempPadj.mat is from the bathymetric data supplied by Rob Anderson is loaded. Then you click the "find P" button to find the parameters of the interpolatory subdivision P; MATLAB prints the time; on my laptop, it takes about 10 seconds. (This is a prototype for the faster implementation as StreamingModules) Then you click the "refinement steps" button. To also see the original data click "show sample points".

Notes & Screenshots

Note:

1. Our triangular patch is chosen larger than the original data so that we do not have to correctly implement all boundary cases in this prototype. In particular, the spikes down to zero should be ignored.
2. There are finer details in the original points along the drop off that are not being represented at this scale.
3. There are outliers in the original data, which should be disregarded when they do not fit the patch well. This tool has been useful for assessing what parameters we should use to distinguish between points and outliers in StreamingModules.
4. For details on computation, here is a MATLAB notebook on speeding up Wenjie's original demo.

• Demo initialization:
  
• Fit to data points:
  

-- WenjieHe - 01 Jun 2008
to top