Research On The Quality Examination Of Commodity Barcode On Machine Vision

Commodity bar code is composed of bar,space and characters people could recognize.Used for identification of goods,it is usually printed on the packaging and plays a very important role in the circulation of commodities.By scanning the barcode people can quickly obtain the corresponding product information.To standardize the presenting pattern of the barcode,the government introduced Barcode for Commodity-Retail Commodity Numbering and Barcode Marking(GB 12904-2008)and Barcode for Commodity-Barcode Symbol Print Quality Verification(GB 18348-2008).They stipulate 13 judgment standards for commodity barcode quality inspection,which shall be subject to quality inspection and management of barcode inspection institutions at all levels in China.At present,the domestic barcode inspection institutions mainly adopt specific device for quality inspection,which mostly adopts the infrared laser scanning detection mechanism and can detect some parameters,while the rest of the test item and must be completed by manual operation.These methods are slow in detection speed and detection efficiency and easy to be affected by human factors.Therefore,this article looks forward to a new kind of detection method,combining the machine vision technology with the bar code quality inspection of goods.By collecting barcode image,processing and analyzing images,we can get all testing parameters and thus realize the integration and intelligent quality inspection.The most commonly used ean-13 commodity barcode was selected as the research object.The main work includes the following parts:1)The design of scanning facility.According to the testing requirement and the actual situation of the sample under test,the system completed the selection of camera,lens,lighting and other equipment,and structural design,which constitute the barcode image acquisition device.Automatic exposure is set through the camera interface,and a soft trigger acquisition mode is used to obtain barcode images or target images to be measured.2)Camera Calibration.According to the field of view,the checkerboard calibration graphs were designed,and target images were collected.Through the feature extraction procedure,the relationship between the image coordinate and scene coordinate mapping were researched.and the calibration parameters were determined.After that,the image distortion model was deduced,and provide a basis for image distortion correction and the calculation of the barcode size parameters.3)Process of the barcode image.For the collected bar code sample images,the camera calibration distortion model was firstly used to correct the distortion,effectively reducing the influence of lens error.Then,the image binarization was carried out using OTSU method,and the image horizontal and vertical directions were calculated using Sobel operator.Then,the barcode area was located by combining morphological processing and contour extraction.By using Hofer line detection,the deflection angle of bar code and horizontal direction can be measured,and then the image can be rotated.Finally,the image of barcode region and character region can be segmented by projecting features.4)Analysis and calculation of test parameters.The image of barcode region was processed as follows:similar edge distance measurement was used to decode;the optical parameters could be obtained by horizontal sampling and analyzing the barcode area.The barcode size parameters could be obtained by calculating the horizontal and vertical projection.The character area image was processed as follows:the character size parameters can be calculated through horizontal and vertical projection.The character area was segmented and normalized,and the digital character recognition was carried out through template matching,which can be compared with the result of decoding,and reference decoding parameters can be measured.Through the above analysis and calculation,all test data can be obtained.In the current study,100 commercial bar code samples were selected from the bar code quality supervision and inspection station of Shaanxi province.This method can complete the test successfully and obtain all the data of test items.Comparing with the current method,the method presented in this paper has slight differences in some detection data,but the detection results are consistent and correct,all of which conform to the national standards.Our method takes about 15 seconds to complete a test,while the traditional method takes about 15 minutes.The detection efficiency is greatly improved.The test results show that this method is simple to operate,fast to detect,accurate and reliable to improve the inspection efficiency and level of barcode quality.