Search

CN-121973560-A - Label printing method and system based on visual identification detection

CN121973560ACN 121973560 ACN121973560 ACN 121973560ACN-121973560-A

Abstract

The invention relates to a label printing method and system based on visual identification detection, and belongs to the technical field of label printing and quality detection. The method forms closed-loop control by embedding visual detection into three key nodes before printing, during printing and after printing. The method comprises the steps of marking and pre-checking a medium reference before printing, capturing images in real time during printing, calculating deviation, dynamically compensating printing positions, and comprehensively judging and sorting the multi-dimensional quality of finished products after printing. The corresponding system comprises a vision acquisition module, an image processing module, a printing execution module and a control module. The invention realizes the transition from passive detection to active prevention, remarkably improves the consistency and production efficiency of label printing quality, and reduces the comprehensive cost.

Inventors

  • LIU JUN
  • Mo Rongyuan
  • SHEN ZUXING
  • Xi yongfeng

Assignees

  • 重庆赛帕斯汽车零部件股份有限公司

Dates

Publication Date
20260505
Application Date
20260129

Claims (10)

  1. 1. The label printing method based on visual recognition detection is characterized by comprising the following steps of: S1, before printing, standard calibration and medium state pre-detection are carried out, the initial state of a medium image of a label to be printed is collected, a medium geometric standard is extracted, medium defects are pre-detected, and the theoretical position of the label is obtained according to the medium geometric standard; S2, in printing, real-time dynamic calibration of visual guidance is performed, a finished label image is captured after synchronization or delay for a preset time in the printing process, the printing position of the label is obtained according to the finished label image, the real-time deviation between the printing position and the theoretical position is calculated, and the real-time deviation is fed back to a printing control system for dynamic compensation; And S3, after printing, comprehensively judging and classifying the quality of the finished product, detecting the multi-dimensional defects of the finished product label image to obtain a judging result, and triggering corresponding execution actions according to the judging result.
  2. 2. The method for printing labels based on visual recognition and detection according to claim 1, wherein in step S1, the extracted media geometric reference is specifically a physical boundary, interval or positioning mark in the media image is recognized by edge detection and contour analysis.
  3. 3. The method according to claim 1, wherein in step S2, the real-time deviation includes at least one of a translational deviation, a rotational deviation, and a scaling deviation, and the dynamic compensation is applied to the subsequent print content or the adjacent label of the current print job.
  4. 4. The method according to claim 1, wherein in step S3, the multi-dimensional defect detection includes both physical defect detection based on template alignment, content correctness checking based on OCR or OCV, and barcode quality classification based on international standards.
  5. 5. The label printing method based on visual recognition detection according to claim 1, wherein in the step S3, when the determination result is qualified, the execution action is normal conveyance, when the determination result is acceptable flaw, the execution action is recording mark, and when the determination result is serious flaw, the execution action is physical rejection.
  6. 6. A label printing system based on visual recognition detection for implementing a label printing method based on visual recognition detection as claimed in any one of claims 1 to 5, comprising: The visual acquisition module is used for acquiring a medium image and a finished label image of a label to be printed at a key node of a printing process; the image processing module is used for executing image processing, deviation calculation and defect detection analysis; the printing execution module is used for receiving the analysis result and controlling the printer to execute printing actions; the control module is connected with the vision acquisition module, the image processing module and the printing execution module through a control bus and is used for realizing data interaction and closed-loop control.
  7. 7. The label printing system of claim 6, wherein the trigger signal of the vision acquisition module is synchronized with a paper feed motor or a printhead actuation signal in the label printer.
  8. 8. The label printing system based on visual recognition detection of claim 6, wherein the image processing module integrates a template alignment algorithm, an OCR/OCV engine and a barcode analyzer.
  9. 9. A label printing system based on visual recognition detection according to claim 6, wherein the system is integrated within the label printer in an embedded form or works online with the printer as a stand-alone module through a communication interface.
  10. 10. The label printing system of claim 6, further comprising a data storage and learning module coupled to the control module for long-term recording of the inspection data and the deviation history data and for optimizing inspection model parameters and predictive maintenance.

Description

Label printing method and system based on visual identification detection Technical Field The invention relates to the technical field of label printing and quality detection, in particular to a label printing method and system based on visual identification detection. Background The label is used as a key component of product identification, logistics tracing and information carrier, and the printing quality directly influences the accurate reading of information and the management efficiency of products. Currently, the main label printing process generally takes "printing" and "quality detection" as two independent links. After printing, performing spot check by manual visual check or independent off-line detection equipment. This approach has significant drawbacks: The efficiency is low, the cost is high, the manual detection speed is low, the manual detection is easy to fatigue under high strength, the omission ratio is high, and the material circulation and time cost are increased when the offline detection is performed; the real-time performance and closed-loop control are lacking, namely, the detection result is lagged, and errors cannot be corrected in real time in the printing process, so that waste products are continuously generated; The detection dimension is single, the traditional method focuses on printing content (such as OCR verification), the comprehensive judgment capability of printing physical defects (such as burnt words, broken lines, stains and positioning deviation) is insufficient, and particularly, labels with high light reflection and complex curved surfaces or backgrounds are difficult to process. In the prior art, although some schemes attempt to introduce visual detection, the methods are limited to post-detection judgment or simple presence detection, and cannot deeply integrate visual information into a print control closed loop, so that the spanning from 'perceived defect' to 'real-time defect generation inhibition' cannot be realized. Disclosure of Invention The invention aims to solve the technical problem of providing a label printing method and a label printing system based on visual identification detection, which realize the full-flow, real-time and online quality monitoring and closed-loop control of a printing process, fundamentally improve the consistency and reliability of printing quality and reduce the rejection rate and labor cost. The technical scheme for solving the technical problems is as follows: a label printing method based on visual recognition detection comprises the following steps: S1, before printing, standard calibration and medium state pre-detection are carried out, the initial state of a medium image of a label to be printed is collected, a medium geometric standard is extracted, medium defects are pre-detected, and the theoretical position of the label is obtained according to the medium geometric standard; S2, in printing, real-time dynamic calibration of visual guidance is performed, a finished label image is captured after synchronization or delay for a preset time in the printing process, the printing position of the label is obtained according to the finished label image, the real-time deviation between the printing position and the theoretical position is calculated, and the real-time deviation is fed back to a printing control system for dynamic compensation; And S3, after printing, comprehensively judging and classifying the quality of the finished product, detecting the multi-dimensional defects of the finished product label image to obtain a judging result, and triggering corresponding execution actions according to the judging result. Optionally, in step S1, the extraction medium geometric reference is a physical boundary, space or positioning mark identifying the tag through edge detection and contour analysis. Optionally, in step S2, the real-time deviation at least comprises one of translational deviation, rotational deviation and scaling deviation, and the dynamic compensation is applied to the subsequent print content or adjacent labels of the current print job. Optionally, in step S3, the multi-dimensional defect detection comprises physical defect detection based on template comparison, content correctness checking based on OCR or OCV, and bar code quality grading based on international standards. Optionally, in step S3, when the determination result is qualified, the executing action is normal conveying, when the determination result is acceptable flaw, the executing action is recording mark, and when the determination result is serious flaw, the executing action is physical rejecting. A label printing system based on visual recognition detection for implementing a label printing method based on visual recognition detection as in any one of the above embodiments, comprising: The visual acquisition module is used for acquiring a medium image and a finished label image of a label to be printed at a key node of a printing proces