S. Ezekiel, G. Greenwood, and D. Pazzaglia (USA)
Barcode, Wavelet, Energy, Identification, Correlation and Energy Spectrum
In this paper, we present a wavelet-based method for automatic barcode character detection. Barcodes are widely used in a wide array of applications. In order to facilitate barcodes, users must have a method for scanning a barcode. The barcode scanner scans and identifies the characters present in the barcode. Barcode scanners only work if the barcode image is recognizable. In the event of image distortion, it will fail to recognize the characters. Our method overcomes this problem, by reliably identifying the characters using multiresolution analysis. This analysis, removes any existing noise by convoluting various filters. We also, apply morphological operators to fill the gaps that are caused during the noise filtering process. Once these gaps have been filled, we extract the characters. Each character is then compared with a predefined dictionary of characters by using two measures: correlation, and multiresolution approximate coefficient energy to find a matching character. Finally, we display the best matched character. The result suggests that this method is effectively capable of being applied to a broad range of barcodes. Since this method is simple, efficient, and has a real-time response, it can be implemented in embedded systems.
Important Links:
Go Back
