After some tests I found that it was because the method, in big blobs, was called to many times one after another without returning.

I came up with this not recursive alternative algorithm, which should work:

1: Select one pixel of origin in the picture; this could be the first pixel (0, 0) in the picture.

2: Create a list containing only that pixel.

3: As long as this list not are empty, do:

3.1: Pick out the first pixel in the list (and remove it from the list).

3.2: If the pixel is background or out of bound; no more should be done (Continue to point 3).

3.3: Otherwise, the pixel is part of a blob, and the pixel now has to be marked as background.

3.4: The surrounding pixels now have to be added to the list, and the loop can continue.

4: Continue now by selecting another pixel in the picture, which not already have been visited and continue to point 2. If all pixels have been visited, all blobs are found.

No, I don’t plan on delving any deeper into blob detection at this point. I am looking into OpenCV, however, and if I make any progress, I will post it here.

Do we (here) ever get a version with use of calculus? )