基于图像拼接技术的缺损二维条码恢复算法研究

Information Recovery of Defect 2-D Barcode Based on Image Mosaic

论文在刀具表面二维条码自动识读系统的基础上,针对部分刀具的柱形表面容易使条码图像产生扭曲变形的问题,通过建立曲面畸变模型逆向校正条码图像。同时,对于在同一视角下显示不完全的条码图像,利用配套的工装设备,提出采用图像拼接算法恢复原始条码图像的方法。首先,采用相位相关理论估计重叠区域,作为后续特征提取的范围约束;其次,提取相关区域的SIFT特征点,并利用近似最近邻算法实现SIFT特征匹配;最后,采用渐入渐出的平滑算法,使拼接处的图像光强平滑过渡,达到无缝拼接。实验表明,论文提出的算法不需要预建立相机标定,对于大曲率、大形变的条码具有较高的识读率。因此,该方法具有更强的实用性与通用性。

To resolve the problem of the distortion of the 2-D harcode on the cylindrical surface, surface distortion model is setted up to correct the barcodes based on the automatic reading system of the two-dimensional barcode in the surface of tools. Moreover, considering some barcodes cannot display in one image, the missing barcode image based on image mosaic is recovered. First, the overlapped areas with related phase as the range of the feature collecting is estimated. Second, the SIFT points of related areas are eollectted and the SIFT features with the algorithm of approximate nearest neighbor are matched. Third, the original images are revovered by using the step-in-step-out meth- od. The experiments show that the algorithm improves the efficiency of the 2-D barcode reading.