A method and apparatus for precision printing

Abstract

The application discloses a kind of precision printing method and equipment, first in paper holographic positioning mark, set main triangle mark rough positioning, auxiliary triangle mark fine positioning, and the digital code mapping of both is stored in database;Printing color group is configured with laser pointer identifier, inertial measurement sensor and tension detection roller, and visual position, kinematic displacement and paper tension signal are synchronously collected;After signal fusion algorithm weighted processing, output real-time fusion positioning signal triggers printing, realizes pattern and positioning mark accurate superposition.Construction fault-tolerant mechanism: when visual signal fails due to mark contamination, etc., switch to sensor fusion mode, predict position based on pre-failure trajectory and paper characteristics, dynamically adjust printing phase or start compensation mechanism to correct.At the same time, collect multi-dimensional working condition data, analyze overprint accuracy trend, dynamically optimize fusion weight and printing parameters, form closed-loop control, effectively improve overprint accuracy and stability under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of printing technology, and more specifically to a precision printing method and equipment. Background Technology

[0002] Currently, with the accelerated intensification of my country's printing industry, printing companies are continuously expanding their production scale, deepening technological iteration, and intensifying market competition. Against this backdrop, the manufacturing processes in printing workshops are gradually moving towards digitalization and refinement, resulting in an exponential increase in the amount of production information data. However, in the fields of high-end packaging and anti-counterfeiting printing, especially when involving processes such as holographic positioning hot stamping and laser embossing, the persistent problem of inaccurate printing positioning remains a key bottleneck restricting product quality improvement.

[0003] To address the challenges of precise printing alignment, some solutions already exist in existing technologies. For example, Chinese invention patent application CN105599473A discloses a holographic positioning precision printing method, which achieves multi-color group registration alignment by pre-setting laser triangles on the paper surface and using a laser cursor recognition device for signal capture. This method improves the automation level of printing alignment to a certain extent.

[0004] However, the applicant still faces limitations in dealing with the complex and ever-changing production environment of modern printing, specifically in the following three aspects: First, the existing paper stretching process mainly relies on roller transmission and feedback from tension sensors for mechanical stretching. This control mode is essentially open-loop or simply closed-loop, lacking a deep understanding of the physical properties of paper. In actual production, the paper's stretching rate changes non-linearly due to differences in fiber characteristics and fluctuations in ambient temperature and humidity. Existing stretching equipment cannot dynamically adjust the stretching force and tension curve according to the paper material (such as coated paper, cardboard, film) and real-time environmental parameters, making it difficult to eliminate the matching error between the preset mark and the actual printed pattern.

[0005] Secondly, although existing technologies utilize photoelectric sensors to track printing plate seams, these systems are mostly "passive tracking" rather than "active identification." Faced with potential issues such as mark contamination, glare interference, or momentary printing jumps during the printing process, traditional photoelectric sensors lack intelligent anomaly detection logic and fault-tolerance mechanisms. Once tracking fails, the equipment often cannot quickly adapt and correct itself, easily leading to continuous production of defective products and increased material waste.

[0006] Finally, printing is a continuous process involving multiple printing units working together, with the states of each color group influencing each other. In existing technologies, registration errors in preceding color groups often fail to be transmitted to subsequent color groups in real time. Due to the lack of a unified industrial data collaboration platform, the control of registration accuracy exhibits an "island effect." Coupled with the cumulative factors such as long-term wear and tear of mechanical parts and slight differences between paper batches, the system lacks self-calibration and error compensation capabilities, leading to accuracy drift after long-term high-speed operation, making it difficult to guarantee the stability of product quality. Summary of the Invention

[0007] To overcome the shortcomings of traditional photoelectric sensors, which lack intelligent anomaly detection logic and fault tolerance mechanisms and are often unable to quickly adapt and correct themselves once tracking fails, this invention provides a precision printing method and device.

[0008] To solve the above problems, the present invention is implemented according to the following technical solution:

[0009] In a first aspect, the present invention provides a precise printing method, comprising the following steps: fabricating holographic positioning paper, including: forming holographic positioning marks on the paper, the holographic positioning marks comprising a primary triangular mark and an auxiliary triangular mark, the primary triangular mark being used for coarse positioning, the auxiliary triangular mark being used for fine positioning, and the two having a digitally encoded positional mapping relationship and stored in a database; configuring a laser cursor recognition device, an inertial measurement sensor, and a tension detection roller on the printing color group to respectively collect visual position signals, kinematic displacement signals, and paper tension signals; and employing a signal fusion algorithm to weightedly fuse the visual position signals, kinematic displacement signals, and tension signals, and outputting a real-time fused signal. Positioning signal; based on the real-time fused positioning signal, the printing action is triggered to make the printed pattern match the holographic positioning mark; an anomaly judgment and fault tolerance correction mechanism is constructed: when the laser cursor recognition device fails due to mark contamination, interference or loss, it immediately switches to the signal fusion mode based on inertial measurement sensor and tension detection roller, uses the position trajectory before failure and paper physical characteristics to predict the current position, and dynamically adjusts the printing phase or starts the compensation mechanism for adaptive correction according to the predicted deviation; collects multi-dimensional working condition data in the production process, analyzes the trend of registration accuracy change, dynamically optimizes signal fusion weight and printing parameters, and forms closed-loop control.

[0010] Preferably, the dynamic weight adjustment rule of the signal fusion algorithm is as follows: when the signal-to-noise ratio (SNR) of the visual position signal is greater than 30 dB, the visual weight W1 = 0.7, the IMU weight W2 = 0.25, and the tension compensation weight W3 = 0.05; when the SNR of the visual position signal is between 15 dB and 30 dB, W1 = 0.5, W2 = 0.4, and W3 = 0.1; when the SNR of the visual position signal is less than 15 dB, W1 = 0.3, W2 = 0.55, and W3 = 0.15; when the visual signal is lost, W1 = 0, W2 = 0.7, and W3 = 0.3.

[0011] Preferably, the step of collecting and analyzing multi-dimensional operating condition data during the production process includes: performing anomaly detection on the collected multi-dimensional operating condition data using an isolated forest model, identifying and removing discrete noise data; identifying negative data and abnormal data exceeding preset upper and lower thresholds according to preset anomaly rules; confirming abnormal time points based on abnormal data and discrete data, retrieving the theoretical parameters of the production work order corresponding to that time point for dynamic threshold comparison, and triggering a self-repair process if the threshold range is exceeded.

[0012] Preferably, the self-repair process includes: when the abnormal data is negative, taking the absolute value of the data; when the abnormal data is greater than a preset upper limit or less than a preset lower limit, updating the data to the valid data of the previous time point; when the processed data is still not within the theoretical threshold range, confirming the deviation direction, and adjusting the tension and phase parameters of the printing equipment according to the deviation direction until the data returns to the threshold range.

[0013] Preferably, before configuring the sensor, an adaptive positioning and splicing step is included: a positioning and splicing device is installed at the paper feed end, a high-precision photoelectric eye is used to track the seam of the paper, and a laser cursor recognition device is used to confirm the position of the triangular markers of two adjacent rolls of paper; when the triangular marker position deviation is detected to exceed the dynamic threshold, the splicing mechanism is controlled to adjust the paper feed position of the next roll of paper so that the distance error between adjacent triangular markers is controlled within ±10μm; after the splicing is completed, the splicing point is automatically marked, and the transition data within the preset time window before and after the splicing point is excluded during subsequent data analysis.

[0014] Preferably, before triggering the printing action, an intelligent data-driven stretching control step is also included: installing a paper positioning and stretching device at the paper feed of the first printing color group; collecting real-time data on workshop temperature and humidity, paper moisture content, and current stretching tension; inputting the cleaned data into the stretching parameter optimization model to dynamically calculate the current optimal stretching tension value, so as to minimize the synchronization error between the paper spacing and the printing plate circumference.

[0015] Preferably, the method further includes a cross-color group feedforward and feedback composite control step: real-time comparison of registration error data of adjacent printing color groups; when the registration deviation between color group N and color group N+1 exceeds 5μm, feedforward prediction of the compensation amount of color group N+1 based on the measured error of color group N; at the same time, feedback correction of the feedforward prediction value based on the measured error of color group N+1, obtaining the corrected compensation amount and converting it into a printing start phase adjustment command for color group N+1 for execution.

[0016] Preferably, the process further includes a parameter evolution step: at preset intervals, the comprehensive registration accuracy index of each printing color group is calculated based on the production information after cleaning. The comprehensive registration accuracy index includes the mean, standard deviation, and maximum deviation of the registration error. If the index is better than the historical best value, the current control parameter snapshot is saved as the optimal parameter template. The control parameters include the tensile tension value, the fusion weight coefficient, and the printing start phase compensation amount. Subsequent production continues to optimize based on the optimal parameter template, thereby achieving continuous evolution of the control parameters.

[0017] Secondly, the present invention also provides a precision printing device, including a frame and an unwinding unit, a positioning and splicing device, a paper positioning and stretching device, and multiple printing color groups arranged sequentially on the frame along the paper feeding direction. The positioning and splicing device includes a paper receiving platform, a photoelectric eye and a first laser cursor recognition device fixed above the paper receiving platform, and a splicing adjustment mechanism for adjusting the paper feeding position of the next roll of paper. The splicing adjustment mechanism includes a servo motor, a lead screw drive pair driven by the servo motor, a slide table carrying the roll seat of the next roll of paper, and a grating ruler for detecting the displacement of the slide table. The servo motor and the grating ruler are both electrically connected to a central controller to form a closed-loop position control. The paper positioning and stretching device includes at least one fixed roller and one movable stretching roller. The end of the movable stretching roller is connected to a piezoelectric ceramic actuator. The central controller receives the drive voltage output from the central controller to fine-tune the tensile force; the paper positioning and stretching device is also equipped with a sensor group for collecting workshop temperature and humidity, paper moisture content, and real-time tensile tension; each printing color group is equipped with a second laser cursor recognizer, an inertial measurement sensor, and a tension detection roller in contact with the paper, and the printing roller of each printing color group is connected to the registration adjustment motor via a phase adjustment mechanism; the central controller is electrically connected via a data bus to the photoelectric eye, the first laser cursor recognizer, the servo motor and the grating ruler of the positioning and splicing device, the sensor group and the piezoelectric ceramic actuator of the paper positioning and stretching device, and the second laser cursor recognizer, the inertial measurement sensor, the tension detection roller and the registration adjustment motor of each printing color group; the central controller is configured to execute any of the precision printing methods described in the first aspect.

[0018] Compared with the prior art, the beneficial effects of the present invention are:

[0019] This invention introduces a tension detection roller to monitor the paper's stress state in real time, and combines visual position signals with kinematic displacement signals, using a signal fusion algorithm for weighted processing. It can deeply perceive the differences in physical properties of paper under different materials (such as coated paper, cardboard, and film), as well as the nonlinear changes in stretch rate caused by temperature and humidity fluctuations, thereby dynamically adjusting the tensile force and tension curve to ensure high-precision matching between the preset mark and the actual printed pattern, fundamentally eliminating errors caused by the instability of paper's physical properties. The holographic positioning mark designed in this invention includes primary and secondary triangulation marks with digitally encoded mapping relationships, combined with kinematic data from an inertial measurement unit (IMU), giving the system anti-interference capabilities. When encountering mark contamination, reflective interference, or instantaneous tension jumps causing visual signal failure, the system can switch to a fusion prediction mode based on the IMU and tension detection roller within milliseconds, using the trajectory data before the failure to predict the current position and dynamically correct the printing phase. This effectively avoids continuous waste caused by tracking failure in existing technologies and significantly reduces material loss. This invention breaks down information silos between printing color groups. By collecting multi-dimensional operating data and performing trend analysis, it achieves real-time transmission and compensation of registration errors from preceding color groups to subsequent color groups. Simultaneously, the system can dynamically optimize signal fusion weights and printing parameters based on mechanical wear and batch-to-batch paper variations. This effectively suppresses precision drift caused by long-term high-speed operation, ensuring continuous stability and consistency of product quality, and meeting the demands for extreme precision in high-end packaging and anti-counterfeiting printing. Attached Figure Description

[0020] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein:

[0021] Figure 1 This is a schematic diagram of the equipment structure of a precision printing method and device according to the present invention;

[0022] Figure 2 This is a schematic diagram of the positioning and splicing device in the precision printing method and equipment of the present invention;

[0023] Figure 3 This is a flowchart of the positioning and splicing device of the present invention;

[0024] In the diagram: 1 – Unwinding unit;

[0025] 2 - Positioning and splicing device, 21 - Paper receiving platform, 22 - Photoelectric eye, 23 - First laser cursor recognizer, 24 - Splicing adjustment mechanism, 241 - Servo motor, 242 - Screw transmission pair, 243 - Slide table, 244 - Grating ruler;

[0026] 3 - Paper positioning and stretching device; 31 - Fixed roller; 32 - Movable stretching roller; 33 - Piezoelectric ceramic actuator;

[0027] 4 - Printing color group, 41 - Second laser cursor recognizer, 42 - Inertial measurement sensor, 43 - Tension detection roller, 44 - Printing roller, 45 - Phase adjustment mechanism, 46 - Registration adjustment motor. Detailed Implementation

[0028] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0029] like Figure 3 As shown, the precision printing method of the present invention includes the following steps: manufacturing holographic positioning paper, including: forming holographic positioning marks on the paper, the holographic positioning marks comprising a main triangular mark and an auxiliary triangular mark, the main triangular mark being used for coarse positioning, and the auxiliary triangular mark being used for fine positioning, and the two establishing a digitally encoded positional mapping relationship and storing it in a database; configuring a laser cursor recognition device, an inertial measurement sensor, and a tension detection roller on the printing color group to collect visual position signals, kinematic displacement signals, and paper tension signals respectively; and using a signal fusion algorithm to weightedly fuse the visual position signals, kinematic displacement signals, and tension signals to output a real-time fused positioning result. Position signal; based on the real-time fused positioning signal, the printing action is triggered to make the printed pattern match the holographic positioning mark; an anomaly judgment and fault tolerance correction mechanism is constructed: when the laser cursor recognition device fails due to mark contamination, interference or loss, it immediately switches to the signal fusion mode based on inertial measurement sensor and tension detection roller, uses the position trajectory before failure and paper physical characteristics to predict the current position, and dynamically adjusts the printing phase or starts the compensation mechanism for adaptive correction according to the predicted deviation; collects multi-dimensional working condition data in the production process, analyzes the trend of registration accuracy change, dynamically optimizes signal fusion weight and printing parameters, and forms closed-loop control.

[0030] In a preferred embodiment, the dynamic weight adjustment rule of the signal fusion algorithm is as follows: when the signal-to-noise ratio (SNR) of the visual position signal is greater than 30 dB, the visual weight W1 = 0.7, the IMU weight W2 = 0.25, and the tension compensation weight W3 = 0.05; when the SNR of the visual position signal is between 15 dB and 30 dB, W1 = 0.5, W2 = 0.4, and W3 = 0.1; when the SNR of the visual position signal is less than 15 dB, W1 = 0.3, W2 = 0.55, and W3 = 0.15; when the visual signal is lost, W1 = 0, W2 = 0.7, and W3 = 0.3.

[0031] In a preferred embodiment, the step of collecting and analyzing multi-dimensional operating condition data during the production process includes: performing anomaly detection on the collected multi-dimensional operating condition data using an isolated forest model, identifying and removing discrete noise data; identifying negative data and abnormal data exceeding preset upper and lower thresholds according to preset anomaly rules; confirming abnormal time points based on abnormal data and discrete data, retrieving the theoretical parameters of the production work order corresponding to that time point for dynamic threshold comparison, and triggering a self-repair process if the threshold range is exceeded.

[0032] In a preferred embodiment, the self-repair process includes: when the abnormal data is negative, taking the absolute value of the data; when the abnormal data is greater than a preset upper limit or less than a preset lower limit, updating the data to the valid data of the previous time point; when the processed data is still not within the theoretical threshold range, confirming the deviation direction, and adjusting the tension and phase parameters of the printing equipment according to the deviation direction until the data returns to the threshold range.

[0033] In a preferred embodiment, before configuring the sensor, an adaptive positioning and splicing step is included: a positioning and splicing device is installed at the paper feed end, a high-precision photoelectric eye is used to track the printing plate seam, and a laser cursor recognition device is used to confirm the position of the triangular markers of two adjacent rolls of paper; when the triangular marker position deviation is detected to exceed a dynamic threshold, the splicing mechanism is controlled to adjust the paper feed position of the next roll of paper so that the distance error between adjacent triangular markers is controlled within ±10μm; after the splicing is completed, the splicing point is automatically marked, and the transition data within a preset time window before and after the splicing point is excluded during subsequent data analysis.

[0034] In a preferred embodiment, before triggering the printing action, a smart data-driven stretching control step is also included: installing a paper positioning and stretching device at the paper feed of the first printing color group; collecting real-time data on workshop temperature and humidity, paper moisture content, and current stretching tension; inputting the cleaned data into the stretching parameter optimization model to dynamically calculate the current optimal stretching tension value, so as to minimize the synchronization error between the paper spacing and the printing plate circumference.

[0035] In a preferred embodiment, the method further includes a cross-color group feedforward and feedback composite control step: real-time comparison of registration error data of adjacent printing color groups; when the registration deviation between color group N and color group N+1 exceeds 5μm, feedforward prediction of the compensation amount of color group N+1 based on the measured error of color group N; and simultaneously correcting the feedforward prediction value based on the measured error of color group N+1, obtaining the corrected compensation amount and converting it into a printing start phase adjustment command for color group N+1 for execution.

[0036] In a preferred embodiment, the method further includes a parameter evolution step: at preset intervals, the comprehensive registration accuracy index of each printing color group is calculated based on the production information after cleaning. The comprehensive registration accuracy index includes the mean value, standard deviation, and maximum deviation of the registration error. If the index is better than the historical best value, the current control parameter snapshot is saved as the optimal parameter template. The control parameters include the tensile tension value, the fusion weight coefficient, and the printing start phase compensation amount. Subsequent production continues to optimize based on the optimal parameter template to achieve continuous evolution of the control parameters.

[0037] Specifically, this involves creating holographic positioning markers with a holographic effect on a paper substrate through molding or coating processes. These markers consist of a main triangular marker with thicker lines and a larger area, and auxiliary triangular markers with finer lines distributed around the main triangular marker. The main triangular marker is used for rapid coarse positioning during high-speed operation, while the auxiliary triangular markers are used for high-precision fine positioning. A unique digitally encoded positional mapping relationship is established between each set of main and auxiliary triangular markers. This relationship, along with paper batch information, is stored in the system database for subsequent processes to access.

[0038] At the paper feed end after the unwinding unit, the photoelectric eye of the positioning splicing device tracks the paper seam in real time, while the first laser cursor recognizer reads the holographic triangular mark positions of the current roll end and the next roll beginning. The central controller calculates the positional deviation between the triangular marks of the two rolls in real time and compares it with a dynamic threshold. Once the deviation exceeds the allowable range, the servo motor of the splicing adjustment mechanism is immediately controlled to drive the slide table to make a lateral fine adjustment to the new paper roll shaft seat through the lead screw transmission pair, and the displacement is fed back by the grating ruler to form a closed loop, ultimately controlling the distance error between adjacent triangular marks within ±10μm. After splicing, the central controller automatically marks the splicing point in the data stream. During subsequent statistical analysis and control, transitional data within a preset time window before and after the splicing point will be excluded to avoid splicing disturbances affecting the registration accuracy.

[0039] Before the paper enters the feed chute of the first printing unit, it passes through a paper positioning and stretching device. This device is equipped with a sensor array that collects real-time data on workshop temperature and humidity, paper moisture content, and the actual tension value of the current stretching section. The central controller has a built-in stretching parameter optimization model trained based on experimental data. This model can be selected using either the finite element method (FEM) or the cohesive force principle (CZM). The temperature, humidity, and moisture content data after cleaning are input into the model to dynamically calculate the optimal stretching tension value that minimizes the synchronization error between the paper's spacing and the printing plate's circumference. The piezoelectric ceramic actuator receives the driving voltage based on this value, pushing the movable stretching roller to perform micron-level displacement adjustment, ensuring the paper enters the printing area with ideal tension, laying the foundation for precise registration.

[0040] At each printing color group, a second laser pointer recognizer acquires the visual position signal of the holographic positioning target, an inertial measurement sensor acquires the kinematic displacement signal in the paper feed direction, and a tension detection roller acquires the real-time paper tension signal. The central controller uses a weighted fusion algorithm to fuse the three types of signals and outputs a real-time fused positioning signal. The weighting coefficients are dynamically adjusted based on the signal-to-noise ratio of the visual position signal.

[0041] When the signal-to-noise ratio is greater than 30dB, the visual weight W1=0.7, the inertial measurement weight W2=0.25, and the tension compensation weight W3=0.05;

[0042] When the signal-to-noise ratio is between 15dB and 30dB, W1=0.5, W2=0.4, and W3=0.1;

[0043] When the signal-to-noise ratio is less than 15dB, W1=0.3, W2=0.55, W3=0.15;

[0044] When the visual signal is completely lost due to marker damage or occlusion, W1=0, W2=0.7, W3=0.3.

[0045] Based on the fused positioning signal, the central controller triggers the printing actions of each color group's printing rollers, ensuring precise registration of the printed pattern onto the reference defined by the holographic positioning mark. Simultaneously, the system incorporates an anomaly detection and fault-tolerance mechanism: when the laser cursor recognition device for a certain color group fails due to holographic mark damage, strong light interference, or other reasons, it immediately switches to a fusion mode using the inertial measurement sensor and tension detection roller. Utilizing the pre-failure position trajectory and physical properties such as the paper's elastic modulus, a kinematic model predicts the current paper displacement. Based on the predicted deviation, the printing phase of that color group is dynamically adjusted or a compensation mechanism is activated for adaptive correction, ensuring that registration accuracy does not change abruptly.

[0046] During multi-color printing, the central controller compares the registration errors of adjacent color groups in real time. When the deviation between color group N and color group N+1 exceeds 5μm, based on the measured error of color group N, the compensation amount of color group N+1 is predicted by feedforward using a transfer function model. Simultaneously, the measured error of color group N+1 is collected, and the feedforward prediction value is corrected to obtain the final compensation amount. The central controller converts this compensation amount into a printing start phase adjustment command for color group N+1 and sends it to the registration adjustment motor, driving the plate roller phase adjustment mechanism to complete rapid correction and eliminate the registration deviation between color groups in its early stages.

[0047] During production, multi-dimensional operating condition data such as registration error, signal-to-noise ratio, IMU output, tension value, temperature and humidity, and motor current are continuously collected. First, an isolated forest model is used to detect anomalies in the data, identifying and removing discrete noise points. Then, negative data and abnormal data exceeding upper and lower thresholds are identified according to preset rules. Once an abnormal time point is confirmed, the theoretical parameters of the corresponding production work order at that moment are retrieved for dynamic threshold comparison. If the values ​​exceed the allowable range, a self-repair process is triggered: negative data is processed by taking the absolute value; data exceeding the limit is replaced with valid data from the previous time point; if the data is still not within the theoretical threshold after processing, parameters such as the printing equipment's tension setpoint and printing phase are compensated and adjusted according to the direction of deviation until the data returns to the normal range.

[0048] Every preset period, the central controller calculates the comprehensive registration accuracy index for each printing color group based on the cleaned production information, including the mean, standard deviation, and maximum deviation of the registration error. If the current index is better than the stored historical best value, a snapshot of the current control parameters is immediately saved as an optimal parameter template, containing information such as tensile tension values, fusion weight coefficients for each color group, and printing start phase compensation. Subsequent production continues iterative optimization from this optimal template, achieving continuous evolution of control parameters and long-term self-optimization of registration accuracy.

[0049] Example 2:

[0050] like Figures 1-2As shown, the precision printing equipment of the present invention includes a frame and an unwinding unit 1, a positioning and splicing device 2, a paper positioning and stretching device 3, and multiple printing color groups 4 arranged sequentially on the frame along the paper feeding direction. The positioning and splicing device 2 includes a paper receiving platform 21, a photoelectric eye 22 fixed above the paper receiving platform 21 and a first laser cursor recognition device 23, and a splicing adjustment mechanism 24 for adjusting the paper feeding position of the next roll of paper. The splicing adjustment mechanism 24 includes a servo motor 241, a lead screw transmission pair 242 driven by the servo motor 241, a slide 243 carrying the roll seat of the next roll of paper, and a grating ruler 244 for detecting the displacement of the slide 243. The servo motor 241 and the grating ruler 244 are both electrically connected to the central controller to form a closed-loop position control. The paper positioning and stretching device 3 includes at least one fixed roller 31 and a movable stretching roller 32. The end of the movable stretching roller 32 is connected to a piezoelectric ceramic actuator 33. The actuator 33 receives the drive voltage output from the central controller to fine-tune the tensile force; the paper positioning and stretching device 3 is also equipped with a sensor group for collecting workshop temperature and humidity, paper moisture content and real-time tensile tension; each printing color group 4 is equipped with a second laser cursor recognizer 41, an inertial measurement sensor 42 and a tension detection roller 43 in contact with the paper, and the plate roller 44 of each printing color group 4 is connected to the registration adjustment motor 46 through a phase adjustment mechanism 45; the central controller is electrically connected to the photoelectric eye 22, the first laser cursor recognizer 23, the servo motor 241 and the grating ruler 244 of the positioning and splicing device 2, the sensor group and the piezoelectric ceramic actuator 33 of the paper positioning and stretching device 3, and the second laser cursor recognizer 41, the inertial measurement sensor 42, the tension detection roller 43 and the registration adjustment motor 46 of each printing color group 4 through a data bus; the central controller is configured to execute any of the precision printing methods described in the first aspect.

[0051] First, holographic positioning paper is prepared. Holographic positioning markers with a holographic effect are formed on the substrate paper using molding or coating processes. These markers include primary and secondary triangular markers. The primary triangular markers have thicker lines and a larger area, used for rapid, rough positioning during high-speed operation. The secondary triangular markers are distributed around the primary triangular markers according to a preset geometric relationship, with finer lines, used for high-precision positioning. A unique positional mapping relationship with a digital code is established between each set of primary and secondary triangular markers. This mapping relationship, along with information such as paper batch and number, is pre-stored in a database. Before the printing operation begins, the prepared holographic positioning paper roll is installed onto the unwinding unit 1.

[0052] When a roll of paper is about to run out and a new roll needs to be connected, the positioning and splicing device 2 starts working. During the paper feeding process, the photoelectric eye 22 tracks the paper seam in real time, and the first laser cursor recognizer 23 simultaneously reads the positions of the holographic positioning marks at the end of the current roll and the beginning of the next roll. The central controller calculates the positional deviation between the two triangular marks and compares it with a preset dynamic threshold. When the positional deviation of the main triangular mark is detected to exceed the threshold, the central controller sends a command to the splicing adjustment mechanism 24. The servo motor 241 drives the lead screw transmission pair 242, which drives the new paper roll shaft seat to make a slight lateral adjustment through the slide table 243. The grating ruler 244 provides real-time feedback on the position of the slide table 243 to form a closed loop. The adjustment goal is to control the distance error between adjacent triangular marks within ±10μm after the two adjacent rolls of paper are spliced. After the splicing is completed, the central controller automatically marks the splicing point in the data stream. During subsequent data analysis, transitional data within a preset time window before and after the splicing point is excluded to avoid the impact of splicing disturbances on control accuracy and statistics.

[0053] After unwinding and splicing, the paper enters the paper positioning and stretching device 3, located before the paper feed of the first printing color group 4. A sensor array collects real-time data on workshop temperature and humidity, paper moisture content, and the actual tension value of the current stretching section. The central controller has a built-in stretching parameter optimization model, trained using a sample library established through experiments. During operation, real-time temperature, humidity, and moisture content data after cleaning are input into the model to dynamically calculate the optimal stretching tension value under the current conditions. The piezoelectric ceramic actuator 33 receives the driving voltage, pushing the movable stretching roller 32 to perform micron-level displacement, precisely adjusting the stretching force on the paper to minimize the synchronization error between the plate distance and the printing plate circumference before the paper enters the first color group, laying the foundation for subsequent overprinting.

[0054] At each of the four printing color groups, the second laser cursor recognizer 41 collects the visual position signal of the holographic positioning target, the inertial measurement sensor 42 collects the kinematic displacement signal of the paper in the paper feeding direction, and the tension detection roller 43 collects the real-time tension signal of the paper. The central controller uses a signal fusion algorithm to weight and fuse the visual position signal, kinematic displacement signal, and tension signal, and outputs a real-time fused positioning signal. The weighting coefficients are dynamically adjusted according to the signal-to-noise ratio of the visual position signal, and the specific rules are as follows:

[0055] When the visual position signal signal-to-noise ratio is greater than 30dB, the visual weight W1 = 0.7, the inertial measurement weight W2 = 0.25, and the tension compensation weight W3 = 0.05.

[0056] When the signal-to-noise ratio is between 15dB and 30dB, W1=0.5, W2=0.4, W3=0.1;

[0057] When the signal-to-noise ratio is less than 15dB, W1=0.3, W2=0.55, W3=0.15;

[0058] When the visual signal is completely lost due to damage or occlusion of the markings, W1=0, W2=0.7, W3=0.3.

[0059] After receiving the fusion positioning signal, the central controller triggers the printing action of the printing roller 44 to ensure that the printed pattern of each color group is precisely aligned with the coordinate reference defined by the holographic positioning mark.

[0060] An anomaly detection and fault correction mechanism was constructed during the printing process. When the laser cursor recognition device for a certain color group fails due to holographic mark damage, strong light interference, or signal loss, the system immediately switches to a signal fusion mode based on the inertial measurement sensor 42 and the tension detection roller 43. The central controller uses the effective position trajectory before the failure, combined with the physical properties of the paper such as elastic modulus and mass, to recursively predict the current paper displacement through a kinematic model, and dynamically adjusts the printing phase of that color group according to the prediction deviation, or activates a local compensation mechanism for adaptive correction to ensure that the registration accuracy does not change significantly.

[0061] To further improve the registration accuracy of multi-color groups, a cross-color group feedforward and feedback composite control strategy is introduced. The central controller compares the registration error data of adjacent printing color groups 4 in real time. When the registration deviation between color group N and color group N+1 exceeds 5μm, based on the measured error of color group N, a feedforward predicted compensation amount is generated through the identified inter-color group transfer function; simultaneously, the measured error of color group N+1 itself is collected, and the feedforward predicted value is corrected by feedback to obtain the corrected final compensation amount. The central controller converts this compensation amount into a printing start phase adjustment command for color group N+1 and sends it to the registration adjustment motor 46 of that color group, driving the plate roller 44 and the phase adjustment mechanism 45 to perform rapid phase adjustment, thereby eliminating the inter-color group registration deviation at the initial stage.

[0062] During production, multi-dimensional operating condition data is continuously collected, including overprinting errors of each color group, visual signal-to-noise ratio, inertial measurement output, tension value, temperature and humidity, and motor current. The central controller processes the data as follows:

[0063] First, an isolation forest model is used to detect anomalies in the multi-dimensional operating data, identifying and removing discrete noise data caused by instantaneous sensor fluctuations. Second, based on preset anomaly rules, invalid data with negative values ​​and abrupt changes exceeding statistical upper and lower thresholds are identified. Once an anomaly point is confirmed, the theoretical parameters of the corresponding production work order are immediately retrieved for dynamic threshold comparison. If the deviation exceeds the allowable range of the work order, a self-repair process is automatically triggered.

[0064] If the abnormal data is negative, then the absolute value of the data is taken directly.

[0065] If the abnormal data is greater than the preset upper limit or less than the preset lower limit, the current value will be updated with the valid data from the previous time point.

[0066] After processing, the data is checked again to see if it returns to the theoretical threshold range. If it still deviates, the central controller adjusts the tension setpoint, printing phase and other parameters of the printing equipment according to the direction of deviation until the data returns to the normal range.

[0067] At each preset production cycle or changeover interval, the central controller calculates the comprehensive registration accuracy index for each printing color group 4 based on the cleaned production information, including the mean, standard deviation, and maximum deviation of the registration error. If the current comprehensive index is better than the stored historical best value, a snapshot of the current control parameters is immediately saved to form an optimal parameter template. The saved control parameters include the tensile tension value of the paper positioning and stretching device 3, the signal fusion weighting coefficient of each color group, and the printing start phase compensation amount of each color group. Subsequent production continues to optimize and iterate from this optimal parameter template, realizing the continuous evolution and self-learning of the equipment control parameters, so that the registration accuracy remains at its optimal state throughout the entire life cycle.

[0068] Other structures of the precision printing method and equipment described in this embodiment are referred to in the prior art.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A precision printing method, characterized in that, Includes the following steps: The process of making holographic positioning paper includes: forming a holographic positioning mark on the paper, wherein the holographic positioning mark includes a main triangular mark and an auxiliary triangular mark, wherein the main triangular mark is used for coarse positioning and the auxiliary triangular mark is used for fine positioning, and the two are established with a digitally encoded position mapping relationship and stored in a database; A laser cursor recognition device, an inertial measurement sensor, and a tension detection roller are configured on the printing color group to collect visual position signals, kinematic displacement signals, and paper tension signals, respectively. The visual position signal, kinematic displacement signal and tension signal are weighted and fused using a signal fusion algorithm to output a real-time fused positioning signal; The printing action is triggered based on the real-time fusion positioning signal, so that the printed pattern is aligned with the holographic positioning mark; Construct an anomaly detection and fault tolerance correction mechanism: When the laser cursor recognition device fails to display a visual signal due to mark damage, interference, or loss, it immediately switches to a signal fusion mode based on an inertial measurement sensor and a tension detection roller. It uses the position trajectory before the failure and the physical characteristics of the paper to predict the current position and dynamically adjusts the printing phase or activates a compensation mechanism for adaptive correction based on the predicted deviation. Collect multi-dimensional operating condition data during the production process, analyze the trend of registration accuracy changes, dynamically optimize signal fusion weights and printing parameters, and form a closed-loop control.

2. The precision printing method according to claim 1, characterized in that, The dynamic weight adjustment rule of the signal fusion algorithm is as follows: When the signal-to-noise ratio of the visual position signal is greater than 30dB, the visual weight W1 = 0.7, the IMU weight W2 = 0.25, and the tension compensation weight W3 = 0.

05. When the signal-to-noise ratio of the visual position signal is between 15dB and 30dB, W1=0.5, W2=0.4, and W3=0.

1. When the signal-to-noise ratio of the visual position signal is less than 15dB, W1=0.3, W2=0.55, W3=0.15; When the visual signal is lost, W1=0, W2=0.7, W3=0.

3.

3. The precision printing method according to claim 1, characterized in that, The steps for collecting and analyzing multi-dimensional operating condition data during the production process include: Anomaly detection was performed on the collected multidimensional working condition data using the isolated forest model, and discrete noise data was identified and removed. The system identifies negative data and abnormal data that exceed preset upper and lower thresholds based on preset anomaly rules. Based on abnormal and discrete data, identify abnormal time points, retrieve the theoretical parameters of the production work order corresponding to that time point for dynamic threshold comparison, and if the threshold is exceeded, trigger the self-repair process.

4. The precision printing method according to claim 3, characterized in that, The self-healing process includes: When the abnormal data is negative, the absolute value of the data is taken. When the abnormal data is greater than the preset upper limit or less than the preset lower limit, the data is updated to the valid data of the previous time point. If the processed data is still outside the theoretical threshold range, the direction of deviation is identified, and the tension and phase parameters of the printing equipment are adjusted according to the direction of deviation until the data returns to the threshold range.

5. The precision printing method according to claim 1, characterized in that, Before the sensor configuration step, an adaptive positioning and stitching step is also included: A positioning and splicing device is installed at the paper feeding end, and a high-precision photoelectric eye is used to track the seam of the paper. At the same time, a laser cursor recognition device is used to confirm the position of the triangular mark of two adjacent rolls of paper. When the detected position deviation of the triangular marker exceeds the dynamic threshold, the splicing mechanism is controlled to adjust the paper feeding position of the next roll of paper so that the distance error between adjacent triangular markers is controlled within ±10μm. After the connection is completed, the connection point is automatically marked, and the transition data within the preset time window before and after the connection point is excluded during subsequent data analysis.

6. The precision printing method according to claim 1, characterized in that, Before triggering the printing action, a smart data-driven stretching control step is also included: Install a paper positioning and stretching device at the paper feed inlet of the first printing color group; Real-time data collection of workshop temperature and humidity, paper moisture content, and current tensile tension; The cleaned data is input into the stretching parameter optimization model to dynamically calculate the current optimal stretching tension value, thereby minimizing the synchronization error between the paper spacing and the printing plate circumference.

7. The precision printing method according to claim 1, characterized in that, It also includes a cross-color group feedforward plus feedback composite control step: Real-time comparison of registration error data between adjacent printing color groups; When the overprinting deviation between color group N and color group N+1 exceeds 5μm, the compensation amount of color group N+1 is predicted based on the measured error of color group N. Simultaneously, the feedforward prediction value is corrected based on the measured error feedback of color group N+1, and the corrected compensation amount is converted into the printing start phase adjustment command of color group N+1 and sent for execution.

8. The precision printing method according to claim 1, characterized in that, It also includes the parameter evolution step: At preset intervals, the comprehensive registration accuracy index of each printing color group is calculated based on the production information after cleaning. The comprehensive registration accuracy index includes the mean value, standard deviation and maximum deviation of the registration error. If the indicator is better than the historical best value, the current control parameter snapshot is saved as the optimal parameter template. The control parameters include tensile tension value, fusion weight coefficient and printing start phase compensation amount. Subsequent production will continue to optimize the parameters starting from the aforementioned optimal parameter template, thereby achieving continuous evolution of the control parameters.

9. A precision printing device, comprising a frame and an unwinding unit, a positioning and splicing device, a paper positioning and stretching device, and multiple printing color units arranged sequentially on the frame along the paper feeding direction, characterized in that... ； The laser cursor recognition device includes a first laser cursor recognizer and a second laser cursor recognizer; The positioning and splicing device includes a paper receiving platform, a photoelectric eye and a first laser cursor recognizer fixed above the paper receiving platform, and a splicing adjustment mechanism for adjusting the paper feeding position of the next roll of paper; the splicing adjustment mechanism includes a servo motor, a lead screw drive pair driven by the servo motor, a slide table that carries the roll seat of the next roll of paper, and a grating ruler for detecting the displacement of the slide table. The servo motor and the grating ruler are both electrically connected to the central controller to form a closed-loop position control. The paper positioning and stretching device includes at least one fixed roller and one movable stretching roller. The end of the movable stretching roller is connected to a piezoelectric ceramic actuator, which receives the driving voltage output by the central controller to fine-tune the stretching force. The paper positioning and stretching device is also equipped with a sensor group for collecting workshop temperature and humidity, paper moisture content and real-time tensile tension. Each of the printing color groups is equipped with a second laser cursor recognizer, an inertial measurement sensor, and a tension detection roller that contacts the paper. The printing roller of each printing color group is connected to the registration adjustment motor via a phase adjustment mechanism. The central controller is electrically connected via a data bus to the photoelectric eye, the first laser cursor recognizer, the servo motor and the grating ruler of the positioning and splicing device, the sensor group and the piezoelectric ceramic actuator of the paper positioning and stretching device, and the second laser cursor recognizer, the inertial measurement sensor, the tension detection roller and the registration adjustment motor of each printing color group. The central controller is configured to perform the precision printing method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 8.

Images