Column Label Stitching Based on Cylinder Correction and Weighted Fusion
WANG Min
TIAN Zikang
CAO Xiaomeng
CAO Ran
WU Jia
Abstract：In order to address the problem of incomplete images and severe deformation and ghosting in the stitching of cylindrical labels on penicillin bottles,a cylindrical label image stitching method based on cylindrical correction and weighted fusion was proposed.Firstly,the curved label image was unfolded using a cylindrical correction algorithm,and the overlapping area of the label image was roughly extracted using a phase correlation algorithm.Then,the speeded up robust features(SURF)algorithm and random sample consumus(RANSAC)algorithm were used to extract feature points in the overlapping area and eliminate mismatched points,respectively.Finally,combining the dynamic programming suture search criteria and nonlinear weighted fusion algorithm,seamless stitching of cylindrical label images of penicillin bottles was achieved.The experimental results show that this algorithm not only accelerates the speed of image registration,but also effectively avoids the generation of ghosting and stitching lines.It can effectively unfold label images of curved surfaces,greatly improve image quality,and contribute to the extraction of cylindrical label information.
Keywords：cylinder correctionphase correlationSURFimage stitchingvial
Publication Date：2025-10-25
Online Publishing Date：2025-11-17(First online date of this platform, not the publication date of the document)
Pages：7( 26-31,39 )