The method of combining the results of localization algorithms for character and bar code labels

The article shows lack of versatility in existing methods of localization of symbolic markings. A new method is proposed that would combine the results of several different methods using probability theory and fuzzy logic to improve accuracy of the results. The coefficient of localization accuracy for the algorithm is introduced. It regulates the degree of influence that the algorithm is going to have on the resulting outcome. The article presents the results of the experimental study that uses two localization algorithms on an image with industrial marking.