In this work, an imaging processing technique was used to study bubble distribution and evolution in the carbon nanotube (CNT) growth process for the first time. The Fractal analysis method is used to analyze the degree of gas-solid mixing and particles movement. A superposition imaging method was proposed and used to investigate the effect of bubbles on gas-CNT particle mixing. Using a tapered fluidized bed reactor (TFBR) without a distributor, we found that bubbles play different roles in particle mixing at different stages of CNT growth in tapered fluidized bed reactor (TFBR) without a distributor. The degree of mixing of gas-CNT particles and the degree of particle movement are related to the growth time of the CNT. In the growth process of agglomerated multi-walled CNT (@A), strong gas-solid contact dominates in the 1-5 min and 15-20 min windows. Stable fluidization of @A plays an important role in 35-40 min and 45-50 min windows. The degree of dispersion of bubbles of vertical array multi-walled CNT (@A) decreases over time. Importantly, efficient heat and mass transfer are ensured in the three time periods before 15.28 min. After 35.28 min, the @V growth process exhibits stable fluidization. Based on the probability density and joint probability density of bubble aspect ratio and shape factor, we found that different growth time of CNT differentially influences the bubble characteristics in the bed. Additionally, the distribution tendency of @A particle and particle productivity were strikingly similar, confirming that the image processing technique and fractal analysis method adopted in this study are feasible and reliable.