Research And Development Of2D Bar Code Real-time Recognition Technology With Complex Context

With information age developing rapidly, the increasing amount of information isrequired for storage, one-dimensional bar code has been unable to meet. The two-dimensionalbar code has the advantages of a large data capacity, the error correction ability and highreliability to make two-dimensional barcode gradually accepted. The two-dimensionalbarcode image actually collected has various kinds of problems such as uneven illumination,complex background interference, slope to bring great difficulties to identify. How to identifythe two-dimensional bar code accurately and quickly has become an important area ofresearch. This paper focuses on the identification method of Data Matrix two-dimensionalbarcode with complex background interference.In order to overcome the complex background interference, uneven illumination, barcodeslope, combine with machine vision and image processing to realize to quickly identify DataMatrix bar code. Quick recognition method is proposed based on the chain code tracking,straight-line segment extraction. The method locates the barcode region by means of the "L"shape characteristics of the Data Matrix barcode. This paper’s main work includes:(1) pretreat Data Matrix barcode. First, use a median filter algorithm to filter the noisypicture. Then, use Sobel operator to extract image gradient information. Finally, usenon-maxima suppression method and Otsu threshold segmentation method to extract theimage edge information.(2) Locating Data Matrix barcode. Firstly, the image edge points are tracked and chaincodes are recorded based on chain code tracking method. Secondly, line segment whosestraight line similarity is below the threshold is removed with fast straight line segmentextraction method. The fracture line is combined with the linear connection method and theunreliable line is excluded by length discrimination. Thirdly, locate the Data Matrix barcodewith the characteristics of the barcode. Eventually, rotate and sample barcode.(3) Identify the Data Matrix barcode. Correct and decode Data Matrix barcode,according to the Data Matrix barcode encoding rules.(4) Use Matlab simulation to test the correctness of the algorithm. Implement therecognition system in C language. Debug the program and optimize the program. Testalgorithm function module and the performance of the identification system such asrecognition rate testing, reliability testing, real-time testing.For640*480size image, actual test shows that the recognition method can quickly andaccurately locate two-dimensional bar code area with the host computer running time of about40ms, the lower machine recognition time of around400ms, the recognition rate of99.39%.The proposed method is ready to meet the requirements of industrial field with strongreal-time ability and high reliability. Data Matrix two-dimensional bar code recognitionalgorithm has actual application value and broad prospects for development.