RE: [Dinamica EGO] Re: Isometry

[Hermann Rodrigues]

Hi, Joshua,

Isometry can be very tricky. In Dinamica, the isometry is a probability multiplier. It is used to multiply the probabilities of cells during the patch construction phase. That's why using 1 as the isometry does not affect the patch shape. Using 0, forces the patch to spread and a value greater than 1 forces the patch to contract. This value is not normalized, so it's effects depends on the probability values in your probability map.

And the isometry is not limited to 2. You can use a greater value as well. However, since the value is a probability multiplier, the corresponding probability will be clamped when the multiplication results in value greater than the probability upper limit.

Regards,

Hermann

Enviado pelo meu Windows Phone

---

From: Joshua Plisinski via Dinamica EGO
Sent: ‎9/‎29/‎2014 12:55 PM
To: dinami...@googlegroups.com
Subject: [Dinamica EGO] Re: Isometry

Does anyone have any insight on to how this is calculated within dinamica?  I am currently stuck using perimeter / area ratio, but I know that dinamica is calculating "isometery" in a different way.  Any help would be great!

On Thursday, August 7, 2014 12:13:32 PM UTC-4, Joshua Plisinski wrote:

Please,

How exactly does Dinamica calculate patch Isometry?  What is the math behind the metric?  I understand that the value can vary from 0-2 and that I could experiment and try to mimic the patches in my historic landscapes, but I would much rather calculate patch isometry directly from my historic patches.

Thanks in advance,

Josh

--
You received this message because you are subscribed to the Google Groups "Dinamica EGO" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dinamica-ego...@googlegroups.com.
To post to this group, send email to dinami...@googlegroups.com.
Visit this group at http://groups.google.com/group/dinamica-ego.
For more options, visit https://groups.google.com/d/optout.

[jplis...@acechildren.org]

Thank you very much for your reply Hermann.  So there is basically no way to calculate "isometry" from historical change patches?  The only way to achieve accurate results is to do multiple simulations with different isometry values and compare the results?

Thanks,

Josh

[Hermann Rodrigues]

Hi Josh,

As far as I know, there is no automatic way to calibrate the isometry parameter from a set of historical images. The only way I know is using trial and error.

Best,

----------------------------------------------------------------------
Hermann Rodrigues
her...@csr.ufmg.br
Skype: hermann.rodrigues
Centro de Sensoriamento Remoto / UFMG
http://www.csr.ufmg.br/dinamica

[jplis...@acechildren.org]

Thanks Hermann, your previous response was very helpful.

What is the neighborhood rule for the isometry multiplier?  Is it applied for every cell in a patch as the patch grows or is it applied only to the last cell to be changed?

Thanks,

Josh

On Tuesday, October 14, 2014 1:26:25 PM UTC-4, Hermann Rodrigues wrote:

[Hermann Rodrigues]

Josh,

During the patch formation process, every time Dinamica selects a cell, it puts a sliding window centered on that cell and collects the cell neighbors. Then, it selects another cell, among all cells selected previously, and center the sliding window on that cell again, repeating the process.

When a cell that was collected by a previous patch formation process iteration is collected again, its cell probability is multiplied by the isometry factor, possibly  increasing or decreasing the chance of selecting that cell in the future.

Best,

----------------------------------------------------------------------
Hermann Rodrigues
her...@csr.ufmg.br
Skype: hermann.rodrigues
Centro de Sensoriamento Remoto / UFMG
http://www.csr.ufmg.br/dinamica

[Joshua Plisinski]

Greetings Herrmann,

Thanks a lot for the clarification.  I will be sure to send you some of our results to add to the website when we are done with our project.

Sincerely,

Josh

--
You received this message because you are subscribed to a topic in the Google Groups "Dinamica EGO" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dinamica-ego/zsV6FY9FAj8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dinamica-ego...@googlegroups.com.

[Hermann Rodrigues]

Keep us posted, Josh!

Best,

----------------------------------------------------------------------
Hermann Rodrigues
her...@csr.ufmg.br
Skype: hermann.rodrigues
Centro de Sensoriamento Remoto / UFMG
http://www.csr.ufmg.br/dinamica