Research On Barcode Recognition Technology In A Complex Background

As an information coding technology, barcode is simple, reliable and efficient, which has been widely used in commodity circulation, the postal system, library management and other various fields. Existing barcode laser scanners need to be aligned with barcode manually. Barcode can be detected, localized and recognized automatically using barcode image recognition technology, without human intervention. However, because of complex background, non-uniform illumination and rotation, barcode image recognition is difficult. To solve these problems, this paper will research on barcode recognition technology in a complex background. The main contents are as follows:(1) Barcode localization. Directional consistency of barcode is the most distinctive characteristic in barcode image under complex background. It can be used to detect image areas may contain barcode and eliminate most background interferences. Then the exact barcode area can be found according to other barcode characters. In this article, averaging squared gradients method is introduced to evaluate directional consistency of image area and estimate barcode orientation, which improves the accuracy and efficiency of barcode location algorithm. This algorithm can be applied to localize barcode in any orientation.(2) Barcode decoding. The accuracy of barcode decoding is influenced by binarization result greatly. To solve the non-uniform illumination interference in barcode image, an adaptive curve segment algorithm is proposed to binarize barcode, which is better than traditional image binarization algorithms. Because look-up table and the similar edge distance decoding method are sensitive to barcode binarization error, a barcode decoding algorithm based on similarity measure and a voting scheme of multiple scanlines is proposed to improve reliability.(3) A barcode recognition software system is developed based on barcode recognition algorithm in this paper. Many experiments show that this barcode recognition algorithm has good accuracy and reliability, and can be used in real-time applications.