A Symbol-based Bar Code Decoding Algorithm

We investigate the problem of decoding bar codes from a signal measured with a hand-held laser-based scanner. Rather than formulating the inverse problem as one of binary image reconstruction, we instead determine the unknown code from the signal. Our approach has the benefit of incorporating the symbology of the bar code, which reduces the degrees of freedom in the problem. We develop a sparse representation of the UPC bar code and relate it to the measured signal. A greedy algorithm is proposed for for decoding the bar code from the noisy measurements, and we prove that reconstruction through the greedy algorithm is robust both to noise and to unknown parameters in the scanning device. Numerical examples further illustrate the robustness of the method.