Principally, it might be possible but under certain conditions. If you want to compare different conditions/materials/… with each other, you need a kind of standardized setup. So first, I would try to figure out a possibility if you can build yourself a setup / device to take the image always with the exact same distance to the wall (?).
The size indication (e.g. the cm indicating band you have in your first image) needs to be in all images as well. This will at leas enable a somewhat more accurate image scaling in ImageJ.
Last but not least, the lighting is very important. To be able to extract the bubbles from the image their outlines need to be somehow visible / detectable. So, I would play with the lighting conditions e.g. try with multiple light sources from multiple directions or only one direction. If you can see a border around the bubbles you might have a good chance in extracting and measuring them.
Also important is, that you take the image directly perpendicular to the wall. Other angles might distort the final measurement.
Hope that helps to start with the setup improvement (if possible). Thereafter, post again and we’ll see, what might be possible to achieve image processing wise.
Also, try to avoid .JPG images if possible since they have pixel artifacts and will make the analysis more difficult.