A flexible image analysis method for measuring bubble parameters
Sheng Zhong,Xu Zou,Zhibing Zhang,Hongzhou Tian
DOI: https://doi.org/10.1016/j.ces.2015.10.033
IF: 4.7
2016-01-01
Chemical Engineering Science
Abstract:Mass transfer modeling is important to design and scale-up gas–liquid reactors in the research of gas–liquid mass transfer intensification. The parameters of bubbles in reactors, including size, shape and dynamics, are the keys in mass transfer modeling. The digital image analysis technique is convenient and efficient and has been widely used to measure bubble parameters. In this paper, we have proposed a flexible method for measuring the bubble parameters, which consists of establishing a template database for difform bubbles, splitting contours for every overlapped single bubble in bubble swarm, and reconstructing the segmented bubbles. The contour splitting algorithm is based on a new concave point extraction algorithm proposed in this paper. Based on the template database that we establish, the proposed reconstruction method can reconstruct segmented bubbles with different shapes and measure their important parameters effectively, such as area, perimeter, major/minor axis, roundness, etc. Experimental results show that for dense bubble images, our proposed method can segment and reconstruct bubbles accurately from overlapping bubbles and reliable in most cases.