Dots Analysis project for industrial use for a total begginer

Hello everyone,

I started in September a engineer school in apprenticeship with a paper crafting company. One of my first mission is to find a way to do an automatic counting of “dots” on the paper. I ear about ImageJ few years ago, for scientific analysis, and I wanted to check if it can be used in my case.

I already tried to use directly ImageJ for counting dots on different types and colours of papers and it worked, but now I have a bigger problem. I don’t know anything about java programming.

What I want to do is more or less that :

  • The operator set up the samples
  • The operator double click on a icon on the computer
  • The camera take the picture
  • ImageJ treat the picture
  • A window appear with the result ( number and size of the dots )
  • The operator validate, the results are sent to an excel file and the picture is erased, everything close.

I have to make it really easy to use, some operators are not used to work with computers, if it’s just a 2 clicks operations it’s ok, but if they have to search in menus to launch a macro or things like that, they are lost ( and they are not going to put any effort to understand it or do it right ).

For all that, I think I must create a Java application to do action outside of ImageJ ( taking the picture and erasing the picture, sending data to an already existing excel file ), and inside ( picture analysis ).

I tried to find an answer by myself before, but without programming knowledge and not being totally used to English, it’s sometimes hard to understand what the guides are saying and if it concern my problem. If everything I said is possible, I can take few weeks/months to learn how to program, but I cannot afford wait this long just for understanding if it’s possible to do it.

So do you think it’s possible? If yes, what guides do you suggest for learning how to do that?

If I’m totally wrong and it’s impossible, I can still make a 3 clicks process, 1 for taking the picture, 1 for treating it and saving results and 1 for erasing picture and showing result in middle of screen. If done like that, can it work?

Thanks for reading all of that and for your advices, and sorry for my bad english and understanding of ImageJ.

Hello @PaulMontoya37 and welcome to the ImageJ forum!

Your task sounds definitely like something that can be done in ImageJ. It would help if you can post here some image examples so other users and developers can recommend you the easiest solutions based on the level of difficulty involved.

Cheers!

Hi @PaulMontoya37

I would use KNIP for this task. I created a small workflow as a starting point (download (28.9 KB)):

Store all images in a certain directory and load the files with List Files. In the gray meta-node “Extract Dots” you can perform the segmentation. The segmentation has to be adapted to your use-case, currently just a threshold is applied. Then for each image a csv-file with the extracted data is saved to the input directory. Finally the processed image is deleted.

The visual inspection is not part of the workflow, because it will slow down the analysis. And if your image acquisition is always the same and you found the segmentation parameters, the segmentation-results will not change. This means, that you can check more images in less time.

Instead of writing the results to a csv file we could perform a segment analysis and trigger actions based on the outcome (if too many points are found, send an email).

2 Likes