Accurate measurement of bubble size in digital images can prove challenging in conditions that involve significant overlapping of bubbles and/or images with a combination of in and out of focus bubbles. The method employed in the current study uses the circular Hough transform to identify circular features in the image. This method can handle mildly non-circular (i.e. non-spherical) bubbles as well as non-trivial overlapping. Several image pre-processing steps are employed to improve the method’s accuracy in detecting in-focus bubbles along with their radii, reduce both spurious detection of out-of-focus bubbles and non-bubble features, and reduce non-detection of in-focus bubbles. Pre-processing steps include histogram equalization and illumination variation reduction via filtering. In this paper, results from the algorithm are compared against a manual detection of bubbles in digital images under conditions of varying bubble diameter and bubble number density.