Abstract: In order to solve the problem of low decoding rate of data matrix image by irregular deformation,this paper proposed an adaptive localization method by edge gradient feature.The algorithm can significantly improve the accuracy of localization.The algorithm consists of the following three steps.Firstly,according to the contour feature of DM code,it locates “L” shape to get the position of the code.Secondly,it estimates the number of code blocks by detecting the dashed line and locates mapping points by affine transformation.Finally,local adjusting is performed to correct irregular deformation by edge gradient feature.In consequence,the original code can be reconstructed.In experiments,we tested decoding by a variety of data matrix code images.The results show that this algorithm can be applied to images in high distortion and decoding accuracy.The decoding accuracy can be greatly improved from 93% to 98%.

Key words: 2D Barcode,Irregular deformation,Structure extraction,Local feature,Edge gradient