Position verification

Hi all,
I am currently working on a small project on how to find the center of a ‘blop’ on an image, using ImageJ 1.46r and standardize the method in order for all the users to have reproducible results.But first let me go through a bit of background.
In Radiotherapy Physics departments a fundamental test is the verification of position. There are multiple ways to verify the position of an radioactive source. The radioactive source normally is transferred through special guide tubes to various positions in an appropriate applicator. We are interested in the positions were the source stopped. We nominally program the source to stop in specific positions. To visualize, the stopping positions we place a special film which changes color when it is irradiated. The film is called Gafchromic film.
We tape the special applicators with a piece of film and we mark the tip of the applicator. After the irradiation of the film we try to find the center of each stopping position ‘blop’ and measure from the tip of the applicator. The following image depicts the above mentioned. The black outline is the shape of the applicator and the grey-ish oval ‘blops’ are the stopping positions of the radiation source.

I normally use Image-Color-Split Chanel to use the red channel. I am using the the 5cm line as a reference and set the scale, then I draw a line on the tip of the applicator. Furthermore, I create another line going through all the ‘blops’ and use the line profile option to find the stopping positions. The following image depicts the above mentioned.

My problem and question is that as you can see is a bit noisy and cannot really tell which part is center of the blop. I would appreciate any help or solution.


Apply a filter to the image (Gaussian blur) to smooth the signal before you make a line profile.
This will smooth the curves of your ‘blops’ to detect your peaks (center).

This is little bit like the tree rings example and find the tree growth rings for counting.

There is also a peak finder tool available to get the center (peak) automatically:



Please make also sure not to work with JPEG images (which you apparently used) which introduces artifacts and introduces errors in your measurements (off channel hint from @Herbie).