List of Coordinates in PictureSnapApp

[bgarcia]

I have a question about a new capability in PictureSnapApp. It concerns applying the same translation that is applied when one calibrates an image to instrument stage coordinates TO a set of points (x, y text).
 
I’m not really even sure this is usable by the typical probe community, but it would help us out.
 
OK, so here’s the typical scenario wherein we’re using PictureSnapApp:
 
Session1:
An automated SEM finds a large (1000s) number of samples/locations. Then we use PictureSnapApp to go back and make a short list of some interesting locations, which them get sent on to SIMS where they use PictureSnapApp to find these same locations. Sample is removed from SEM, used by SIMS, then put into storage.
 
Session2, at some later point in time:
SIMS asks is there are any more sample locations to analyze from stored sample,. This means that we could put the sample back in the SEM, call up the PictureSnapApp image, and see the Session 1 locations. We can’t find new locations, however, because the looong list of locations from Session 1 is in terms of the previous coordinate system. So while the PictureSnapApp image/point-list can be recalibrated every time with a new coord system, the list of points generated outside of PictureSnapApp from SEM Session 1 no longer apply.
 
My question is, would it be possible to somehow convert an old x-y-coordinate to the current coordinate system in PictureSnapApp, as text? Then I could type in the old x-y, get a new x-y, then type it in the SEM stage and zoom off.

[John Donovan]

Hi Ben,
I just want to make sure I understand what you are asking.  So to make sure we are on the same page you:

1. Put the sample in the SEM, load an image of the sample, calibrate the image using two (or ideally three) reference points, then you annotate the image with points of interest on the sample.

2. Then you put the sample in the SIMS instrument and they load the calibrated image from the SEM instrument using the previously calibrated image file and the ACQ file from the SEM containing all the annotated positions.  Then your SIMS people re-calibrate the image from the SEM, again using two (or ideally three) reference positions (which could be the same reference positions or different reference positions) now that the sample is in the SIMS instrument.   

By the way, does your SIMS instrument  have a "half-cartesian" stage?  That is, x-axis is cartesian, but the y-axis is "anti-cartesian"?

3. Then they look at your list of annotated positions, but they only analyze some of the positions that you originally annotated.

4. Then they give the sample back to you, but then at some later time you want them to go and analyze the other positions that weren't analyzed previously.

5. So you give them the sample again, they put it back in the SIMS instrument, they re-calibrate the image again (since it may have rotated or shifted since the previous time it was in the SIMS instrument). And then they go to the other annotated positions that they didn't do from the previous SIMS session.

I don't see they problem.  What am I missing?

[bgarcia]

1. Put the sample in the SEM, load an image of the sample, calibrate the image using two (or ideally three) reference points, then you annotate the image with points of interest on the sample.

2. Then you put the sample in the SIMS instrument and they load the calibrated image from the SEM instrument using the previously calibrated image file and the ACQ file from the SEM containing all the annotated positions.  Then your SIMS people re-calibrate the image from the SEM, again using two (or ideally three) reference positions (which could be the same reference positions or different reference positions) now that the sample is in the SIMS instrument.   

3. Then they look at your list of annotated positions, but they only analyze some of the positions that you originally annotated.

^All of the above is correct!

Here's the rest:

4. Then they give the sample back to you, but then at some later time you want them to go and analyze the other positions that weren't analyzed previously.

So now I've got a list of locations from the 1st analysis, but when I put the sample back in the SEM for a 2nd analysis, the original list of locations I have (from my SEM) no longer apply since the image in PicSnap has the be recalibrated (because the sample doesn't go in the SEM the exact same, each time, so once the sample comes off the stage all location data outside of PicSnap is rendered highly inaccurate).

Clear as mud?

[John Donovan]

1. Put the sample in the SEM, load an image of the sample, calibrate the image using two (or ideally three) reference points, then you annotate the image with points of interest on the sample.

2. Then you put the sample in the SIMS instrument and they load the calibrated image from the SEM instrument using the previously calibrated image file and the ACQ file from the SEM containing all the annotated positions.  Then your SIMS people re-calibrate the image from the SEM, again using two (or ideally three) reference positions (which could be the same reference positions or different reference positions) now that the sample is in the SIMS instrument.   

3. Then they look at your list of annotated positions, but they only analyze some of the positions that you originally annotated.

^All of the above is correct!

Here's the rest:

4. Then they give the sample back to you, but then at some later time you want them to go and analyze the other positions that weren't analyzed previously.

So now I've got a list of locations from the 1st analysis, but when I put the sample back in the SEM for a 2nd analysis, the original list of locations I have (from my SEM) no longer apply since the image in PicSnap has the be recalibrated (because the sample doesn't go in the SEM the exact same, each time, so once the sample comes off the stage all location data outside of PicSnap is rendered highly inaccurate).

Clear as mud?

Kind of.

When you say "all location data outside of PicSnap is rendered highly inaccurate" what positions are you talking about?

If you just re-load the sample into the SEM, and then re-calibrate the original image from the first SEM session using the same reference positions, all the analysis positions will still be there and they will all get re-calibrated also, so you can now just go to the ones you skipped the first time.

PictureSnapApp is not just used to transfer points between instruments, but also to re-find positions after the sample was removed from the instrument and re-loaded again into the *same* instrument.

I'm sure I'm missing something important so feel free to call me at Probe Software if you want to chat about it.

[John Donovan]

1. Put the sample in the SEM, load an image of the sample, calibrate the image using two (or ideally three) reference points, then you annotate the image with points of interest on the sample.

2. Then you put the sample in the SIMS instrument and they load the calibrated image from the SEM instrument using the previously calibrated image file and the ACQ file from the SEM containing all the annotated positions.  Then your SIMS people re-calibrate the image from the SEM, again using two (or ideally three) reference positions (which could be the same reference positions or different reference positions) now that the sample is in the SIMS instrument.   

3. Then they look at your list of annotated positions, but they only analyze some of the positions that you originally annotated.

^All of the above is correct!

Here's the rest:

4. Then they give the sample back to you, but then at some later time you want them to go and analyze the other positions that weren't analyzed previously.

So now I've got a list of locations from the 1st analysis, but when I put the sample back in the SEM for a 2nd analysis, the original list of locations I have (from my SEM) no longer apply since the image in PicSnap has the be recalibrated (because the sample doesn't go in the SEM the exact same, each time, so once the sample comes off the stage all location data outside of PicSnap is rendered highly inaccurate).

Clear as mud?

Kind of.

When you say "all location data outside of PicSnap is rendered highly inaccurate" what positions are you talking about?

If you just re-load the sample into the SEM, and then re-calibrate the original image from the first SEM session using the same reference positions, all the analysis positions will still be there and they will all get re-calibrated also, so you can now just go to the ones you skipped the first time.

PictureSnapApp is not just used to transfer points between instruments, but also to re-find positions after the sample was removed from the instrument and re-loaded again into the *same* instrument.

I'm sure I'm missing something important so feel free to call me at Probe Software if you want to chat about it.

Just to follow up on this topic, the issue was:

1. Ben was utilizing another app (actually DTSA-2 from Nicholas Ritchie) to digitize the many thousands of points on his Tescan instrument. 

2. He would then document a few dozen of these points in PictureSnapApp, and then take the PictureSnapApp files to his SIMS where they would re-calibrate the digitized image to the SIMS stage and locate those dozens of points for SIMS analysis.

3. The problem occurs when they decide to put the sample back into the SEM. Because although re-calibrating the SEM image in PictureSnapApp on the SEM accurately re-calibrates the previously documented dozens of points, of course that doesn't affect the other thousands of points digitized using DTSA-2 (which were *not* documented in PictureSnapApp).

I suggested that he post a question in the DTSA-2 board on this forum and see if Nicholas has a fiducial calibration method in DTSA-2, *or* if there is a way to export the points from DTSA-2 into a format that can be read by the TestFid app which comes free with CalcZAF and is really nice for coordinate transformation of large position datasets:

https://probesoftware.com/smf/index.php?topic=46.msg256#msg256

Although most people use these three fiducial .POS files for digitizing standard mounts, of course they can also be utlized for digitizing unknown samples as well.  The .POS file format is documented in the PFE Reference manual.

[John Donovan]

And to follow up on Ben's dilemma on using a third party app to digitize stage coordinates outside of PictureSnapApp, we do distribute the TestFid application, which can transform thousands and thousands of points using three fiducial (reference) marks, in the free CalcZAF distribution.

It's nowhere near as fancy as PictureSnapApp, but if you can create an input file for it, it can be pretty handy.  There's also no menu for it, but you can locate the executable in the C:\Program Files (x86)\Probe Software\Probe for EPMA folder and make a "shortcut" to it if you decide to use it regularly.

Here is a screen snap of the application after it loads:



After you load a selected .POS input file, you get this dialog where you can update the coordinates of the three fiducial marks that were utilized to create the digitized point file:



After that (here I just clicked OK in this case), you get the transformed points based on the offset and rotation:



Also detailed information on the .POS file format can be found here:

https://probesoftware.com/smf/index.php?topic=46.msg9615#msg9615