This paper proposes a personalized packaging solution integrating digital printing technology and computer image fusion for the problems of low data collection efficiency, insufficient image processing accuracy and high printing customization cost in traditional packaging design. Based on big data analysis, a consumer preference model is constructed, and a K-mean clustering algorithm is used to extract packaging design features. The RTV model and GrabCut algorithm realize image smoothing and accurate segmentation, and combine with digital printing technology to complete high-precision variable data output. In the performance test, after 50 frames of target images are processed by this paper’s image processing algorithm, the mean value of Y-value of all target images after processing is 0.957, and the information deviation is always controlled within 5μrad. The average airtightness pass rate of this paper’s solution reaches 99.987%, and the control group is reduced by more than 0.4% compared with this paper’s solution. The top five satisfaction rankings in formal practice account for three of the desired demand attributes, and the basic demand attributes have lower satisfaction rankings except for local characteristics.
