How to get the coordinate of a sphere in a series of high speed photographs?

I’m totally new to imageJ. I’m using high speed photographs to get the shift and velocity of an falling steel sphere of a radius of around several centimeters.
So I wonder if I can get the sphere’s coordinate in the pixel axis automatically, so that I can calculate its moving velocity later?
Thank you very much your answer.

Hi Ian,

I think your tasks involves several processing steps (preprocessing, segmentation, centroid calculation). Each of these steps highly depends on the kind of images you are processing …, so if you could provide one or two example images, I think people can easier help you.

Have a nice weekend!


I’m too new to put two graphs, so here is one of my photograph.
The black sphere is the one that need to be measured.

which one? both of them?

The upper one, for the lower one is the shade. I will try to get rid of the shade using other backgrounds.

Does the analysis need to be fully automatic (mostly depending on the number of images you want to process) or can there be a manual step involved (like manually annotating the object?).

I need to analyze around 10 images at least for one drop and there would be around 10 drops for one experiment. It’s better be fully automatic. Currently, I read the coordinate of the downside of the sphere manually for several times. If manually annotating the object is faster than what I’m doing now?

I find the “Tracker” is specialized in motion picture analysis.
Thank you very much for your advice.

Hi @lan

Are the white spots reflections? And do they stay at this position or move down with the ball?

Or in other words: Is the ball the only moving object over all images?

Can you upload another image? Maybe the 8th image?

I bumped you up a trust level. You should be able to post more images now.

Most of the settings on this forum are still Discourse defaults. If they ever limit your ability to solve or discuss problems, please feel welcome to poke an admin.

1 Like

This is another pic.
Sorry to reply late.
The ball is the only moving object and the white spots are reflections of my lights. I’ll try to use frosted glass to eliminate them.

Thank you very much.
I’ll go to admins for help next time.

Hello @lan

I created this KNIME workflow as a solution for your problem.

With this workflow I get these results:

You can download the workflow here (1.9 MB).

Note: This workflow only works if the light comes from above, because I assume that the y coordinate of the centroid of the shadow is bigger than the y coordinate of the centroid of the steel sphere.


Since the background is fixed perhaps subtracting it could simplify the segmentation, ie create a difference image and apply a threshold.

I actually subtract a maximum projection of all images from every frame.