Skip to content

Faulty defocus compensation due to slow first images #113

@iverks

Description

@iverks

In our setup, for whatever reason the first two images take a long time to acquire, generally 1200 and 1000 ms each.

When the first few defocus frames arrive, the calculation acquisition_time = (t_start - t0) / (i - 1) from line 301 in cred/experiment.py overestimates the real acquisition time and does not properly skip a frame until sufficiently many frames have been taken to reduce the average time.

One way to solve it could be to use a more robust estimator, like the median frame time instead of the mean.

Example

Acquisition times. Here you can see that the first two frames take a long time to acquire for whatever reason.
Image

Time between captures. Here you can see that even though the defocus frames 10 and 20 take a long time, the frames 11 and 21 are not skipped.
Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions