Using pyptv for tracking of a single particle?

[sarah barnová]

Hello,

My name is Sarah and I am trying to use pyptv for tracking of a single particle motion in air with 2 cameras. I use pyptv on Windows.

I've spent some time trying to undertstand and get pyptv working. So far, I think I've understood the calibration but I still find [0] correspondences. I was wondering whether the tracking algorithm is actually appropriate for tracking single particle. The particle is similar to a 3mm long cylinder with <1mm diameter. I did not find any post about single particle tracking and I ask myself whether pyptv might be useful for my work. Put another way, is there some intrinsic reason that hinders the possibilities to track a single particle in a video?

Thanks for any advice!

Sarah Barnova

[Alex Liberzon]

There should be no problem, on opposite, it should be much easier to work with a single particle, in air. Just share somewhere a working folder with calibrations, parameters, and some images (not all, at least 5 for each camera) to check what's wrong.

When you work with 2 cameras only it's important to check the box "Include Pairs" - as you usually we're interested in 4- or 3- camera detections.

--
You received this message because you are subscribed to the Google Groups "openptv" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openptv+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openptv/96980514-713e-4090-a41b-51ed274651a5n%40googlegroups.com.

[sarah barnová]

Hello Alex,

Thank you very much for your response.

Here is my working directory: https://github.com/sarah-ba/test_cavity_S.git

test_cavity_S is entirely based on your example test_cavity. I progressively changed photos, orientation files and parameters for calibration or tracking.

The cameras stayed in the same position for the calibration images and for the video images.

I proceed in this order: I do the calibration. Then I apply Highpass filter, click on Image and coord and finally on Correspondances. I get [0] correspondances even though my calibration seems to work. I increased the Tol. band but it did not help. 

1.) I don't understand whether my calibration is not good enough (maybe my initial guess should fit more the centroids) or there is some other problem.

2.) I am not sure what the "Frame 123456789" means. Can you please explain me the purpose of files cam1.#10001_targets, cam1.10001_targets and cam1.123456789_targets ?

Thank you for your time,

Sarah

[Alex Liberzon]

Hi,

there was only one thing that I think confused the system - the glass vector z component was set to -999. I changed it to 1.0 - with the index of refractions all equal to 1, this vector has no meaning, but it's used in the calculations, so with -999 it breaks something in the imaging path.

Best regards

Alex

[Alex Liberzon]

You've got a pull request from my fork on Github.

[Alex Liberzon]

am not sure what the "Frame 123456789" means. - this is s temporary file created by any set of images in the manual processing, i.e. when we do Start - > Highpass - > Detect - > Correspondence = the result has no actual frame number so we use just this temporary value. One day, somebody will have a very long sequence with an actual frame of 123456789 and then the code will break ;)

[sarah barnová]

Thank you very much!! :)

Best regards,

Sarah