Multi-symbology and multiple 1D/2D barcodes extraction framework

Abstract

Image-based barcode recognition technique is a robust and extendable approach for versatile 1D/2D barcodes reading. Most of methods discussed in literature may either work for single 1D/2D barcode or rely on finding the unique finder pattern. Multi-symbology barcode extraction is a practical issue and yet challenging issue. Extended from our preliminary investigation and for realistic consideration, this work proposes a general segmentation framework to achieve extraction of real barcodes under complex background when multiple types of symbology appear in the same snapshot for 1D barcodes, 2D barcodes, or both co-exist. The proposed algorithm has three main steps: background small clutters elimination, potential barcodes segmentation and barcode verification. The whole algorithm combines several image processing methods, namely, image subtraction, Gaussian smoothing filtering, morphological operation, connected component labeling and iterative thresholding. Experimental results indicate that the proposed approach can segment multiple barcodes from the complex background with acceptable accuracy. © 2011 Springer-Verlag Berlin Heidelberg.