Feature Extracting, Correcting And Locating Of Industrial Data Matrix Barcode

Data Matrix bar is a typical kind of two-dimension code,by the means of laser marking,lattice identification,code-spurting,printing,they can be marked on all kinds of materials.With the advantages of miniature and high density,Data Matrix bar is widely used in the area of industrial production.However,with the feature of serious wear,low contrast,the identification of Data Matrix bar is always difficult.To solve the factors of adverse environments and poor shooting conditions which caused terrible impact to the rapid and accurate identification of Data Matrix bar code from complex images,we design the algorithm below:1) the significant amount of right-angle corners of the images and the inner characteristic of the existence of bright and dark corners being taken into account,we propose a novel self-adaption corner detection algorithm based on the mask algorithm theory. The algorithm uses the smaller mask, which will just sort out the corners that may be the right-angle vertexes and use circle mask as a filter, the characteristics of the Data Matrix images are considered when choosing the shape and the range of the mask, the threshold of local binarization as well as the criteria of the filters.2)with the contraction of typical clustering algorithm,we choose to use the affinity propagation algorithm,which is more suitable for the identification of Data Matrix bar.Also,we combine the feature of the bar into the algorithm,which became the algorithm of semi-supervised affinity propagation.This algorithm focus on the feature of Data Matrix bar,which can bring much benefits to the clustering.3)we use the algorithm,which combines both algorithms of masking and edge detection,to detect the “L” shape of Data Matrix bar.To achieve the purpose of correcting,we use the algorithm of bilinear interpolation as well as typical algorithm of image rotating.Experimental results indicate that in the sample of 160 Data Matrix bar images,the even accuracy is of right-angle corners detection is 92%,the accuracy of the clustering is 96%,the locating and correcting of the bar is accurate.Conclusion:The algorithm has the ability of detecting and correcting when dealing with the Data Matrix bars that in complex condition.