Advanced image processing of wire-mesh sensor data for two-phase flow investigation

Two-phase gas-liquid flows, the simultaneous flow of a gas and liquid in a pipe, are commonly found in several industrial activities, among them during the extraction and processing of crude oil. Wire-mesh sensors are flow imaging devices which are able to generate images of phase distribution of two-phase flows within a pipe cross section at high spatial and temporal resolutions. Appropriated image processing algorithms are however necessary in order to extract important flow parameter from raw data. In this paper, we describe a bubble identification algorithm based on three-dimensional image segmentation. Results of bubbles velocity and flow rate measurement based on the development algorithm are validated against reference models showing good agreement.