ID Barcode Recognition Technology Based On Image Processing

Barcode has been widely used and brings more convenience or even new experience into our daily life,it is a critically important way to improve the speed and accuracy of computer data entry in personal computer era.With the rapid development of Mobile Internet,it is also playing a irreplaceable role on mobile smart devices like smart phone etc.Hence,barcode recognition on basis of digital image processing has limitless applications in the future.However,barcode image captured in the practical environment usually suffer from various degrees of image distortion such as blurring,low-light,noise and illumination etc.,which will seriously damage the quality of barcode image.As a result,barcode localization and recognition are inclined to fail when barcode images encounter those situations.Therefore,it's a challenge work for designing an effective and robust barcode recognition algorithm to settle those ubiquitous problems.The main works of this thesis are presented as follows:Firstly,the structure of 1-D barcode and code strategy are presented and analyzed,and then commonly used decode algorithms are presented.To overcome the problems of interference of noise and illumination,noise detection algorithm,image denoising algorithms and illumination preprocessing algorithms are analyzed,furthermore,there advantages and disadvantages are presented to guide our future barcode localization and recognition.To tackle the mostly common blurring problem,we propose a useful barcode image deblurring algorithm,which uses an effective blurring detection algorithm to confirm whether or not there exists blurring in the barcode image,and then estimate the motion kernel of the blurred barcode image,in the end,an non-blind deblurring algorithm is used to recover the sharp barcode image.Secondly,an barcode localization algorithm is proposed to find barcode area under complicated environment,which first utilize straight line detection algorithm based on Hough Transform to find the lines of the barcode image that determines the coarse location of the barcode,and then another barcode localization algorithm based on color features is used to find out the candidate barcode areas of the barcode image,and finally select the candidate barcode area that include the most lines as final localized barcode.Various experimental results demonstrate that the proposed barcode localization algorithm is simple but efficient.Finally,an efficient and effective algorithm is proposed to recognize the barcode locating in the complicated environment.Numerous experiments are conducted on two popular barcode image databases to compare our proposed algorithm with the state-of-art barcode recognition algorithm such or commercial barcode recognition software based on digital image processing,experimental results has demonstrated that: For slightly and medium degraded barcode image,the proposed algorithms performs much better than the Gallo-Manduchi algorithm and the three state-of-art commercial barcode recognition software like DTK,DataSymbol and BcTester;For motion blurring barcode images,DTK,DataSymbol and BcTester barcode recognition software are lost their effectiveness but the proposed algorithm performs effectively;For the degrade barcode images that have generated by interference of noise,illumination and blurring caused by focus of camera,Gallo-Manduchi algorithm performs more effetively than the proposed algorithm.