Calculating total twin boundary length in map

[niels cautaerts]

I'm trying to estimate the twin density (twinning area/total volume) in my material through a simple stereological formula:

N_L = 2/pi * L / S

Here N_L is the twin density, L is the total twin boundary length as measured from the EBSD map and S is the scan area size.
I calculate S somewhat awkwardly with:

S = abs(ebsd.prop.x(end))*abs(ebsd.prop.y(end));

The problem is I don't really know how to best calculate L. I have the boundaries I define as twins stored in the variable gbtwinrec. How do I get the total length in scan units?

I tried:

L = sum(gbtwinrec.segmentSize)

But there seems to be something wrong with this since the values in segmentSize are all integers. Also, when I look at segmentSize in the boundaries for all my grains for instance, all values are either 0 or a huge number, and id's repeat in segmentID suggesting that if I sum all segmentSizes I'll get repetitions.

What is the correct way of doing this?

[ruediger Kilian]

Hi Niels,
boundary.segLength is the length of the boundary in scan units. Smoothing of boundaries might in some cases be a good idea because otherwise boundary segments might be just steps of 0 and 90 degree (in case of a square grid) and hence one might overestimate the length.

Hope this helps.
Cheers,
Rüdiger