The Design And Realization Of Real-time Recognition System For Multi-material Barcode

With the advent of the Internet and the era of large data,barcode as an efficient and convenient information storage technology is widely used in people's daily life and industrial production.At the same time,the barcode recognition based on image processing has become a hot spot in recent years since its broad applicable range and high recognition rate.However,due to the diversity of bar code materials and types,many existing bar code reading systems have many deficiencies,such as poor applicability,low recognition rate,slow processing speed,etc.In order to solve the above problems,this paper proposes a set of multi-material barcode real-time recognition system based on image processing,which can quickly identify the barcode on metal plate,PCB circuit board and commodity packaging.The innovation of this paper is:(1)Developed different preprocessing programs for barcode on different materials;(2)Proposed an improved threshold algorithm to improve the threshold effect of bar codes on metal parts;(3)Proposed a decode strategy based on the image pyramid,which can greatly reducing the barcode recognition calculation;(4)Optimized and integrated the mainstream barcode decoding libraries Zbar and libdmtx.The work of this paper is mainly for the following aspects:(1)In order to realize the barcode image segmentation on the PCB board,extracted the S component in HSV space to divide the barcode and the background area;(2)In order to solve the problems that the OTSU method cannot deal with the detail and Bernsen algorithm is sensitive to noise,proposed an improved threshold algorithm to eliminate the artifacts of the high reflection characteristics of metal parts;(3)High-resolution barcode images have the problem of information redundancy,resulting in a large amount of barcode recognition,so proposed an improved decoding strategy based on image pyramid,which can reduce the computational complexity and improve the decoding speed by replacing the original image with lower resolution image;(4)Designed and built the software platform based on Visual Studio 2010 integrated development environment,and run the identification system in the Windows system successfully.The experimental results show that the system can effectively identify the barcodes on metal plates,PCB boards and printed publications.The recognition speed can meet the needs of industrial production,and has the advantages of wide application range and high efficiency etc.