基于边缘定位的西林瓶柱面标签展开算法研究  

Research on Vial Label Unwrapping Algorithm Based on Edge Location

作　　者：姚立权 张萌[1] 史立峰[1] 朱虹[1] YAO Li-quan;ZHANG Meng;SHI Li-feng;ZHU Hong(Liaoning Equipment Manufacturing Vocational and Technical College,Shenyang 110161,China)

机构地区：[1]辽宁装备制造职业技术学院,辽宁沈阳110161

出　　处：《仪表技术与传感器》2021年第9期105-109,115,共6页Instrument Technique and Sensor

基　　金：辽宁省自然科学基金项目(2020-MS-361)。

摘　　要：为解决因柱面畸变导致的西林瓶标签全表面成像困难的问题,提出了基于边缘定位的西林瓶标签平面展开算法,得到一幅连续且无畸变的西林瓶标签平面展开图。依据柱面标签图像中的边缘确定西林瓶标签的空间位姿,建立标签的空间模型并确定像素点的最优观测相机,拼接融合相邻标签图像,最终得到西林瓶标签的平面展开图,在执行过程中不需要额外的辅助设备及复杂的标定步骤。改变西林瓶的空间位置并多次实验,结果表明得到一幅分辨率为850×250的柱面标签平面展开图仅需345.6 ms,实现了西林瓶柱面标签的快速全表面成像。In order to solve the problem that the surface of the vial label is difficult to image due to cylindrical distortion,a planar unwrapping algorithm for vial label based on edge positioning was proposed.A continuous and distortion-free vial lable plane unwrapping picture was obtained.The best observation camera for the model was determined after building the spatial model,and the adjacent images were stitched and fused together to finally obtain a plane unwrapping image of the vial label.During the process of the implementation,additional auxiliary equipment and complex calibration steps were not needed.The spatial position of the vial label was changed and several experiments were performed.The experimental results show that it takes only 345.6 ms to obtain a cylindrical lable plane unwrapping image with a resolution of 850×250,which can achieve rapid full-surface imaging of the vial label.

关 键 词：机器视觉 西林瓶标签 柱面畸变 边缘提取 全表面成像 位姿估计 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]