Flow velocity measurement plays an important role in the process detection of solid-liquid two-phase flow. The twin-plane electrical resistance tomography (ERT) sensor has been increasingly addressed to measure and compute flow velocity in multiphase flow. The cross-correlation (CC) principle based on a pair of ERT sensors is the most used method for flow velocity measurement. As for the dredging engineering, there are various CC-based methods, and each of them often has a particular applicable range. Therefore, the discovery of these ranges is a vital issue under noisy conditions and complex flow patterns in practice. A mismatched algorithm will lead to the low accuracy in the velocity estimation process. In this paper, three improved CC-based methods are compared, and their particular applicable ranges are recovered by experiments.