Bubble volume and bubble surface area are key parameters that affect the working performance of bubble columns. Due to their non-intrusive characteristics and high spatial resolution, image processing techniques can be used to measure bubble volume. However, bubbles in practical application are not always perfectly spherical or ellipsoidal. In homogenous flow regimes where bubble interaction is weak, the bubble geometry can be classified into three types: round cap, ellipsoidal, and disk-shaped. For large and severely distorted bubbles, it is difficult to employ existing image processing methods for measuring the bubble volume. This study describes an image analysis technique that was developed to obtain the volume and surface area of large and severely deformed bubbles. This proposed image processing method is based on a differential segmentation concept and possesses better adaptability because it considers both the bubble geometry and deflection angle of the bubble. The developed image processing technique is verified using the results of a bubble collection experiment. By making use of the developed image processing algorithm, the volume and surface area of discrete bubbles have been calculated. The relationship between bubble volume distribution and gas flow rate is discussed and the characteristics of bubble surface area variance are shown. Moreover, the influences of bubble volume and surface area on the bubble rise velocity are analyzed. (C) 2018 Elsevier Ltd. All rights reserved.