Onset energy detection

[Gozdriov]

Hi Francisco, 

Thanks for your reply, trying to get the derivative to work I found on the documentation

estimate_elastic_scattering_threshold

which might do what I want with the correct parameters....???

best

Angel

On Wednesday, July 3, 2013 5:03:50 PM UTC+1, Francisco wrote:

Hi Gozdriov,

No, we don't have edge detection. The closest thing that we have is find_peaks1D_ohaver. It could roughly work for edge detection when applied to the diffentiated spectrum. Are you thinking in something more sophisticated?

By the way, could you open a new thread for any future discussion that is not related to the original subject of this one ("Not working as a library"). In this way other users will have an easier time finding the information.

Best,

Francisco

2013/7/3 Gozdriov <gozd...@gmail.com>

Hi Francisco,

Is there a way to get the onset value for a

<EELSSpectrum, title: SP01, dimensions: (29, 83, 1440)>

something similar to estimate_zero_loss_centre but for the starting of an edge and that returns a value for each spectrum???

Thanks

Best

Gozdriov

[Francisco de la Peña]

Hi Angel,

Yes, you're right, I forgot about this one. Although these methods might work in simple cases, they'll probably fail in the presence of fine structure, overlapping and delayed maximum edges. I'm not aware of any robust solution to this problem but if there is it would be nice to implement it in Hyperspy.

Best,

Francisco

2013/7/4 Gozdriov <gozd...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "hyperspy-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hyperspy-user...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Magnus Nord]

Hey,

I've been working a little bit on this earlier, using the following technique:

- Find the lowest point right before the ionization edge.
- Find the highest point in the fine structure (for example in the Manganese edge in LSMO this would be the L3 top point)
- Then do the following:
-- some_good_fraction*(highest_value-lowest_value)+lowest_value = intensity_value, where some_good_fraction is in the 0.05 - 0.2 range
-- Find the energy(intensity_value) which is between energy(lowest_value) and energy(highest_value)

Energy(intensity_value) is then your onset energy.

This procedure currently requires some manual input, and I don't really know how robust it is. But I don't really know about any other methods...

I have some code for this locally, but it is currently a horrible mess. I'll probably reimplement it in a better fashion and merge it with hyperspy sometime in the future when I need it for my own work.

Magnus