Comparison_sizes.png
v026v027.png
[PIPE2D-516] Different results from different size images with same physical parameters Created: 28/Feb/20 Updated: 12/Mar/20 Resolved: 12/Mar/20 |
|
| Status: | Done |
| Project: | DRP 2-D Pipeline |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Normal |
| Reporter: | ncaplar | Assignee: | ncaplar |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Attachments: |
Comparison_sizes.png
v026v027.png
|
| Story Points: | 6 |
| Description |
|
I have been noticed that there is an inconsistency between the results that I create on different sized images - e.g., the results differ if I create 20x20 or 50x50 images although I have used all the same parameters. That obviously should not happen.... The differences seem to be most pronounced in the ``scattering'' component around the central spot. This is manifested in the image below which shows:
Top left: 20x20 cut from the larger 50x50 image Top right: 20x20 image Bottom left: full 50x50 image Bottom right: (20x20 cut from the larger 50x50 image-20x20 image)/(20x20 image)
This ticket is to investigate and fix this problem.
|
| Comments |
| Comment by ncaplar [ 03/Mar/20 ] |
|
In the previous version of the code (internal version 0.26) the pupil image was apodized with the same factor (Gaussian filter with sigma=0.75 pixels), even though the pupil image size changes as the function of the size of the final image. I have now implemented (in the internal version of the code 0.27) that the apodization scales as a function of the scale of the pupil image. Having said that, I have not removed the effect. With this is I have basically changed the +- size of the effect and decreased its magnitude. I will continue the investigation. |
| Comment by ncaplar [ 06/Mar/20 ] |
|
The ``solution'' which I decide to pursue it to analyze all images with the same size for the corresponding pupil image. In this way the optical PSF created will always have the same properties and this problem will be avoided. There will be some overhead and slowdown in the generating spped, because we will be using larger arrays for pupil description in case of focused images, but I think this is acceptable. |
| Comment by ncaplar [ 12/Mar/20 ] |
|
I am at the moment generating the new solutions using the improvements specified in this ticket. I should probably open a new ticket to capture this. |