Abstract: Two-phase flow phenomena are commonly observed in boiling water reactor (BWR), high-power-density pressurized water reactor (PWR), and other liquid-cooled nuclear reactors. For example, a certain amount of steam is generated in the coolant channels of BWR cores. Although modern PWRs do not permit nucleate boiling in the entire core average channels, subcritical or even saturated nucleate boiling is allowed in the hottest channels. In steam generators, the quantity of steam produced far exceeds within the reactor core. During reactor accidents, particularly loss-of-coolant accidents (LOCAs), the entire primary circuit system transitions into a two-phase flow condition. Traditional methods such as high-speed camera observation, probe detection, γ-ray and X-ray imaging are unsuitable for observing two-phase flow due to their limitations in penetrating non-transparent pipelines, invasiveness, and shallow penetration depth, respectively. However, neutron imaging technology can effectively capture two-phase flow information through metal pipelines without disrupting the flow. Previous studies using neutron imaging have primarily focused on qualitative analysis, with limited systematic quantitative research. Moreover, the luminous flux after neutron conversion is far less than that of optical imaging, most image processing methods and techniques applicable to optical imaging will not be suitable for neutron imaging. Therefore, a neutron visualization study on air-water two-phase flow in vertical circular pipes was presented and a comprehensive image processing method for neutron imaging in two-phase flow was introduced in this paper. The neutron imaging two-phase flow image processing methods were mainly divided into three steps. The first step was preprocessing, which requires normalizing the two-phase flow images, full water images, empty water images, and dark field images that have been captured, and then performing noise reduction processing on the processed images. The second step was void fraction calculation, which involves calculating the void fraction information by bringing the processed full water images, empty water images, and dark field images into the two-phase flow images. The third step was gas phase velocity calculation, which tracks the gas bubbles using the average cross-sectional void fraction value and calculates the gas phase velocity. Finally, the results of the void fraction calculations were compared with those obtained from the empirical correlation proposed by Soviet scholar Armand. Additionally, the position tracking of the average cross-sectional void fraction was compared with theoretical values to derive gas phase velocity. An error analysis was conducted, identifying primary sources of error as bubble aggregation, neutron imaging image errors, and flow fluctuations caused by pumps. The overall error is relatively small, thereby validating the accuracy and reliability of the method presented in this paper. The application of neutron imaging technology in two-phase flow research was primarily investigated in the study. The successful implementation of neutron imaging in air-water two-phase flow demonstrates its feasibility in this field. Future research will focus on addressing hydrogen scattering issues that significantly impact neutron imaging accuracy and exploring neutron visualization of high-temperature and high-pressure water two-phase flow.