基于图像区域和骨架特征的条形码印刷缺陷检测算法  

Algorithm of Barcode Printing Defect Detection Based on Region and Skeleton Features of Image

作　　者：胡丹[1] 黄辉[1] 

机构地区：[1]台州科技职业学院机电与模具工程学院,浙江台州318000

出　　处：《工业控制计算机》2024年第1期126-128,131,共4页Industrial Control Computer

摘　　要：条形码广泛用于商品的流通,为提高条形码质检的准确性和效率,提出了基于图像区域特征和骨架特征的条形码印刷缺陷检测算法。相机采集条形码图像后,利用模板匹配定位条形码,按照仿射变换参数进行图像矫正;然后,通过阈值分割、连通域、闭运算等处理,得到条码ROI待检区域;最后,利用区域最小外接矩形长度特征、区域矩形度特征、骨架特征先后检测条符,识别缺陷位置并存入新区域。生产现场测试表明,该方案的检测成功率为99.2%,平均耗时51.982 ms;实验仿真数据还验证所提算法适用不同角度的图像。与其他方法相比,该算法在识别精度、运行速度和鲁棒性方面均具有优势,能获得较好的检测效果。Barcode is widely used in commodity circulation.In order to improve the accuracy and efficiency of barcode quality inspection,an algorithm of barcode printing defect detection based on region and skeleton features of image is proposed.After the camera collects the barcode image,it uses the template to match and locate the barcode,and carries out image correction according to the affine transformation parameters.Then,through threshold segmentation,connected domain,closed operation and other processing,the barcode ROI area is obtained.Finally,using the minimum circumscribed rectangle length feature of the region,the rectangle degree feature of the region,and the skeleton feature to successively detect the stripes,identify the defect location and put it in the new region.The production site test shows that the detection success rate of this scheme is 99.2%,and the average time is 51.98 ms.Experimental simulation data also verify that the proposed algorithm is applicable to images from different angles.Compared with other methods,the algorithm in this paper has advantages in recognition accuracy,running speed and robustness,and achieves better detection results.

关 键 词：缺陷识别 区域特征 骨架特征 EAN-13条码 

分 类 号：TS896[轻工技术与工程] TP391.44[自动化与计算机技术—计算机应用技术]