Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Renishaw reader calculates the scale for the microscope image incorrectly #301

Closed
Asharits opened this issue Aug 13, 2024 · 7 comments · Fixed by #327
Closed

Renishaw reader calculates the scale for the microscope image incorrectly #301

Asharits opened this issue Aug 13, 2024 · 7 comments · Fixed by #327
Labels
type: bug Something isn't working

Comments

@Asharits
Copy link

Describe the bug

In the Renishaw reader, the scale for the microscope image (WHTL) is being calculated incorrectly for hyperspectral maps. This causes the mapped area (red box) to show up in the wrong area.

To Reproduce

I verified this by comparing images to mapped area in the WiRE versus what Hyperspy is plotting for multiple datasets.

Expected behavior

I am not sure where in the reader this is being calculated but it is an easy fix. In the original metadata for the microscope image, FocalPlaneXYOrigins contains the origin for X and Y. FocalPlaneXResolution and FocalPlaneYResolution contains the length of the image. These values just need to be added to the origin values and then divided by the image shape to obtain the correct scale.

I tested this out on some of my datasets and it fixed the scale and the red box was now in the correct place.

Python environment:

  • RosettaSciIO version: 0.6
  • Python version: 3.11.8
  • HyperSpy version: 2.1.1
@Asharits Asharits added the type: bug Something isn't working label Aug 13, 2024
@ericpre
Copy link
Member

ericpre commented Aug 13, 2024

Are you saying that the offset and scale of the microscope image are incorrect but the hyperspectral maps are correct?

The scale and offset of the microscope image are calculated in

rosettasciio/rsciio/utils/image.py

Lines 42 to 70 in e499110

def _parse_axes_from_metadata(exif_tags, sizes):
# return of axes must not be empty, or dimensions are lost
# if no exif_tags exist, axes are set to a scale of 1 per pixel,
# unit is set to None, hyperspy will parse it as a traits.api.undefined value
offsets = [0, 0]
fields_of_views = [sizes[1], sizes[0]]
unit = None
if exif_tags is not None:
# Fallback to default value when tag not available
offsets = exif_tags.get("FocalPlaneXYOrigins", offsets)
# jpg files made with Renishaw have this tag
fields_of_views = exif_tags.get("FieldOfViewXY", fields_of_views)
unit = FocalPlaneResolutionUnit_mapping[
exif_tags.get("FocalPlaneResolutionUnit", unit)
]
axes = [
{
"name": name,
"units": unit,
"size": size,
"scale": fields_of_views[i] / size,
"offset": offsets[i],
"index_in_array": i,
}
for i, name, size in zip([1, 0], ["y", "x"], sizes)
]
return axes

And what it does seems to be consistent with your description of how it should be calculated.

As I remember, the data that I have doesn't have this issue. Can you share files and screenshots?

@Asharits
Copy link
Author

Asharits commented Aug 14, 2024 •
edited
Loading

Hmm, I wonder why it isn't working for me. Below is a screenshot of the same dataset from WiRE and in Hyperspy. You can see the map area does not match up.

WiRE Screenshot
Screenshot 2024-08-14 141404

Hyperspy Screenshot
Screenshot 2024-08-14 111531

@Asharits
Copy link
Author

To me it looks like an issue with the way scale is calculated using FieldOfViewXY. Maybe there is a difference depending on the version of WiRE? I am on 5.6

@Asharits
Copy link
Author

Asharits commented Oct 8, 2024

@ericpre Looking back at this I misspoke. You only need to change fields_of_views to be equal to (FocalPlaneXResolution, FocalPlaneYResolution). Doing this yields the correct map area for me.

Figure_2-7_map_test_Signal

@ericpre
Copy link
Member

ericpre commented Oct 19, 2024

@Asharits, can you share a file?

@Asharits
Copy link
Author

Yes, here is a file: 532nm_1800 map.zip

Here is the screenshot of the file from WiRe:
WiRe Screenshot

Here is what Hyperspy plots before the fix:
Figure_532nm_1800_map_Signal

Here is it after:
Figure_532nm_1800_map_Signal2

@ericpre ericpre mentioned this issue Oct 24, 2024
Merged
5 tasks
@ericpre
Copy link
Member

ericpre commented Oct 24, 2024

@Asharits, thank you for sharing the file.
Your suggested changes seems to be correct. The data that I used as reference to parse the calibration, have been acquired with WiRE 5.2. The data saved with the two different version of software (WiRE 5,2 and 5.6) doesn't have the same metadata: in case of WiRE 5.2 FieldOfViewXY are the same values as (FocalPlaneXResolution, FocalPlaneYResolution) and therefore doesn't have the issue. For data saved with WiRE 5.6, these values are different!

I will do some check when I have a chance but it all looks consistent. I made a PR with the fix: #327.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
type: bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Fix scale in white field image in renishaw reader ericpre/rosettasciio
2 participants
@ericpre @Asharits