An Effective Method of 1-D Bar Code Image Identification

This paper proposes a new approach to achieve angle correction for images with arbitrary rotated bar code. A new projection curve based bar code reading method and a new global and local based Euclidean distance decoding algorithm are proposed. The proposed method is effective for identifying bar code images which are low-resolution, non-uniform illuminated, photographed remotely, noisy or slightly blurry, and it is especially suitable for recognizing low-resolution, severely noisy, and small-scale bar code image. Also, this method has the virtue of real-time, robustness and high recognition rate, etc. Experimental results using some international UPC-A bar code library show that the proposed algorithm delivers substantial improvements when compared with some latest methods.