Barcode Image Identification Based on Maximum a Posterior Probability

Abstract

This paper presents a new method robust to noise for vision-based barcode recognition. We propose an effective and stable barcode recognition technique based on the maximum a posterior probability of observation signals. For robust recognition of barcode image mixed with noise signals, first, the center and angle of the barcode area are extracted without rotation, and second, the barcode recognition techniques using the likelihoods based on multiple barcode scan lines is proposed. This proposed method shows superior performance as a result of its application to various barcode images.