BARCODE PROCESSING

摘要

A blurred barcode image is processed by providing an image representation thereof comprising grayscale values. The image representation is deconvoluted using a candidate motion kernel to get a deconvoluted representation. A barcode similarity measure is calculated for the deconvoluted representation to indicate how close the distribution of the grayscale values of the deconvoluted representation is to an optimal distribution for a barcode image. The kernel provision, deconvolution and measure calculation are repeated for different candidate kernels and the candidate kernel resulting in a deconvoluted representation that is closest to a barcode image as determined based on the barcode similarity measures is selected. The selected kernel is used for deconvoluting the blurred barcode image to get a deblurred barcode image that can be read and decoded.