A robust localization method for industrial Data Matrix Code [articol]

Abstract

This paper provides a localization solution for Data Matrix Codes dotted on different materials in different orientations. Knowing the real world size of the Data Matrix pattern and using the parameters of the industrial camera of the recognition system, the developed method can locate the exact position and the orientation of the pattern in an image. We use an adaptive threshold method for the image binarization, in order to be independent of illumination variations or nonuniform background. While the Data Matrix pattern being composed only of dots, it is very difficult to recognize the pattern. To overcome this, we are using morphological operators to transform the pattern in a solid square. The size of the modules and the distance between them as well as the size and the orientation of the data matrix pattern are estimated out of the image. The presented algorithm was tested with very good results for Data Matrix Codes dotted on different materials and different angles