Automatic acquisition and identification method and device for cigarette 32-bit bar code

Abstract

The invention discloses an automatic acquisition and identification method and device for a cigarette 32-bit bar code, and the method comprises the following steps: 1, displaying a positioning grid through a scanning interface of image scanning equipment, and positioning a to-be-shot cigarette bar code; 2, when the cigarette enters the positioning grid position, shooting a cigarette bar code image; 3, carrying out binarization processing on the cigarette bar code image by adopting a processing method of fusing color features of the laser character image and stroke width features of laser characters; 4, extracting characters in each positioning grid by adopting character fixed width information and a vertical projection character cutting method; 5, adopting an optical character recognition algorithm to automatically recognize the extracted characters; and 6, comparing and identifying the identified cigarette bar code characters with the character representative information records stored in the database, and identifying cigarette information represented by the cigarette bar code characters. The precision and efficiency of cigarette bar code recognition can be greatly improved.

Description

technical field [0001] The invention relates to the technical field of automatic identification of cigarette barcodes, in particular to a method and device for automatic collection and identification of 32-bit barcodes of cigarettes. Background technique [0002] There are 32 Arabic numerals and English characters printed by laser on the surface of the stick packaging of Chinese cigarettes, which are used to record the origin, date, type and other information of the cigarettes. Tobacco monopoly agencies often need to identify these characters on the cover of cigarette packs to classify and trace cigarettes. Since laser coding cannot affect the visual effect of the outer packaging of cigarettes, laser coding is generally small and light white characters. These characters are light in color and small in size, and it is difficult and time-consuming to recognize them with the naked eye. At present, monopoly personnel mainly transcribe the confiscated cigarette barcodes in the ...

AI Technical Summary

Problems solved by technology

These characters are light in color and small in size, and it is difficult and time-consuming to recognize them with the human naked eye

At present, monopoly personnel mainly transcribe the confiscated cigarette barcodes in the market manually, resulting in a heavy workload for monopoly personnel, difficult identification, and low efficiency caused by manual transcription, making it difficult to avoid errors. staff work problems

[0003] First, the existing cigarette barcode collection system is generally a coded character image enlargement collection platform, which is inconvenient to carry and still requires manual identification and recording

These systems cannot help staff obtain barcode information quickly and conveniently

[0004] Second, in the field of character recognition, artificial intelligence character recognition methods have been used in various document scanning scenarios, such as automatic recognition of printed documents, automatic recognition of street and road character signs, etc. The recognition of laser coding has not yet been studied

[0005] Compared with the existing artificial intelligence character recognition applications, the characteristics of cigarette laser barcode recognition include: first, the characters are light in color and are easily disturbed by light; on the other hand, the characters are small in size and difficult to recognize with human eyes; third, The characters are on the surface of the cigarette package, which is easily interfered by the color pattern of the cigarette package, making it difficult and error-prone to recognize by manual and machine; fourth, the position of the laser printed barcode is not fixed, and it is difficult to use a fixed device for image acquisition; fifth, the laser barcode is It is 32 bits, each barcode represents specific information

Images