Aim: To segment the following image representing the pore space of a Silicon Carbide. The size of the image is 50*50*30 voxels.

 
VRML VRML.gz
 

 

Method 1: Use a segmentation algorithm based on the local aperture.

       
VRML VRML.gz VRML VRML.gz VRML VRML.gz VRML VRML.gz
We first select seeds: maximum balls located in the pore space. The second step consist in a growing region algorithm. Each color correspond to a different pore. The graph representing connection between the pores. Number indicates the local coordinance.

 

Method 2: This algorithm segments the pore space on a watershed-like algorithm

       
VRML VRML.gz VRML VRML.gz VRML VRML.gz VRML VRML.gz
Seed selected on local maxima of the aperture map. Segmentation of the image representing apertures. Segmentation of the image representing pores. Each color correspond to a different pore. The graph representing connections in the pore network.

Distribution of connexion:

 

Method 3: Segmentation of the pore space using the watershed algorithm.

     
VRML VRML.gz VRML VRML.gz VRML VRML.gz
Segmentation of the pore space using watershed algorithm on the distance map. Local apertures in the pore space. Connections in the pore network.

Distribution of connexion:

       
VRML VRML.gz VRML VRML.gz VRML VRML.gz VRML VRML.gz

 


VRML plugin for Windows: cortona

All images were made using OpenDX and DXSoil.


up

last update: 07.12.2012