Simultaneous Real-Time Segmentation of Diversified Barcode Symbols in Complex Background

Traditional barcode localization methods based on image analysis are sensitive to the types of the target symbols and the environment they are applied to. To develop intelligent barcode reading system used in industry, a real-time region based barcode segmentation approach which is available for various types of linear and two-dimensional symbols is proposed. The two-stage approach consists of the target region connection part by orientation detection and morphological operation, and the target region detection part by efficient contour based connected component labeling. The result of target location, which is represented by the coordinates of the upper left corner and the lower right corner of its bounding box, is robust to the barcode orientation, noise, and uneven environment illumination. The segmentation method is proved to work well in the real-time barcode reading system. In the experiments, 100% of the DATAMATRIX codes, 99.3% of the Code39 symbols and 98.7% PDF417 codes are corrected segmented.