Calculate eigenvectors and eigenvalues from ODF

149 views
Skip to first unread message

Matteo

unread,
Aug 30, 2017, 4:01:21 AM8/30/17
to MTEX
Dear all,

I'm trying to calculate eigenvectors and eigenvalues from ODFs, but can't seem to find a dedicated function for it (or maybe I just can't find it). My objective is to try to replicate the Woodcock (1977) eigenvalues diagram to quantify the texture in my samples. Is there any way to do this in MTEX?

Thanks for the help
Cheers,

Matteo

Ralf Hielscher

unread,
Sep 1, 2017, 5:02:11 AM9/1/17
to MTEX
This is not yet directly implemented. However you can do the following

ori = odf.calcOrientation(100000)
[~,~,lambda,eigV] = mean(ori)

Then lambda gives you the eigen values.

Ralf.

Matteo

unread,
Sep 4, 2017, 3:22:26 AM9/4/17
to mtex...@googlegroups.com
Thank you very much Ralf. I don't understand why the outcome is a matrix 4x4 for eigV and four values for lambda (eigenvalues). Shouldn't eigenvectors and eigenvalues be only three?
Cheers
Matteo


------------------------

Matteo Demurtas
Ph.D. student
University of Padova
Department of Geosciences
via G. Gradenigo, 6
35131, Padova (Italy)

Skype: matteodemurtas

Fault rocks!


--
If you want to reduce the number of emails you get through this forum login to https://groups.google.com/forum/?fromgroups=#!forum/mtexmail, click "My membership" and select "Don't send me email updates". You can still get emails on selected topics by staring them.
---
You received this message because you are subscribed to a topic in the Google Groups "MTEX" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/mtexmail/RnTcDvUd0JA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to mtexmail+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/mtexmail.
For more options, visit https://groups.google.com/d/optout.

ruediger Kilian

unread,
Sep 5, 2017, 10:05:50 AM9/5/17
to mtex...@googlegroups.com
Hi Matteo,
for directions your will get 3 eigenvectors/values but for orientations there are 4 values. Roughly, for uniformly distributed orientations, you’ll get a relatively small largest eigenvalue, identical second und third values and a non-zero smallest value. For a “perfect" unimodal distribution, the largest eigenvalue is 1 and all others are 0. For a “perfect” fibre the two smallest eigenvalues values are 0. For real data, values will only approach 1 and 0.

In case you wanted to calulate eigenvectors/values for directions, here’s an example:
v=vector3d.rand(100) % some input
v=[v.x v.y v.z];
M=v.'*v;
[evec,eval]=eig(M,'vector');


Hope this helps,

Cheers,
Rüdiger
> To unsubscribe from this group and all its topics, send an email to mtexmail+u...@googlegroups.com.
> Visit this group at https://groups.google.com/group/mtexmail.
> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> If you want to reduce the number of emails you get through this forum login to https://groups.google.com/forum/?fromgroups=#!forum/mtexmail, click "My membership" and select "Don't send me email updates". You can still get emails on selected topics by staring them.
> ---
> You received this message because you are subscribed to the Google Groups "MTEX" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to mtexmail+u...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages