Intelligent cloud management high-definition card printer system

Abstract

The invention discloses an intelligent cloud management high-definition card printer system, and relates to the technical field of card printers, the intelligent cloud management high-definition card printer system comprises a processor, a database and a printing module, the processor is connected with the database through signal transmission, and the printing module is connected with the processor through signal transmission. The printing module comprises an information acquisition assembly and a printing assembly, the information acquisition assembly is used for acquiring material information and environment information, the printing assembly is used for printing cards, and the processor controls the printing module to start by receiving a primary starting signal and a secondary starting signal. The processor constructs a color compensation module through information collected by the information collection assembly to output a color compensation value to adjust the printing assembly. By designing and constructing the color compensation model, the function of high-precision card printing is realized, the problems of dependence on off-line measurement, color gamut deviation and response lag are solved, the color deviation can be compensated in real time, and the quality and specialty of card printing are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of card printers, and in particular to an intelligent cloud-managed high-definition card printer system. Background Art

[0002] my country's annual demand for ID cards exceeds 1 billion, and the high-end, high-resolution printer market has a compound annual growth rate of 20%. Existing equipment's printing speed, resolution, and durability are unable to meet the anti-counterfeiting technology and output quality requirements of high-security scenarios such as finance and government affairs. Traditional ID card printers are limited in efficiency, accuracy, and functional scalability, making it difficult to meet the large-scale and high-security demand for paper cards. The integration of the Internet of Things, cloud computing, and artificial intelligence technologies is driving the evolution of ID card printing towards integrated automatic recognition, intelligent parameter adjustment, and automated information processing.

[0003] Since the difference in light reflectivity of metal cards will amplify color deviations, the professionalism and aesthetics of printing will be affected if the color output characteristics of the printer cannot be calibrated. At the same time, there is a lack of braking and recognition capabilities for the ink absorption characteristics of the card material. Special processes such as lamination and hot stamping will change the color rendering effect, and color compensation parameters cannot be dynamically adjusted according to material characteristics. In terms of electromagnetic compatibility, it is necessary to ensure that it is not affected by external electromagnetic interference to avoid distortion of card printing caused by electromagnetic interference.

[0004] Patent CN107203355B discloses a fully enclosed ID card production system and method. The above patent achieves the advantages of outward openness, ease of use, high confidentiality, low cost and fast printing when printing ID cards.

[0005] The above patent transmits the printing data of the external terminal to the control host through the data communication interface, and the control host drives the card printer to print the card. After the card printer finishes printing, the control host drives the logistics packaging module to package the printed card. After the card is packaged, the control host drives the inkjet printer to print the logistics information on the outer packaging of the card and then send it out of the logistics package export. The entire card production process is recorded by the monitoring camera, and the video data is stored in the control host and transmitted to the external terminal in real time through the data communication interface, making it more convenient for users to use, but there is room for improvement in the accuracy of card printing.

[0006] To this end, this application proposes an intelligent cloud-managed high-definition card printer system for high-precision card printing. Summary of the Invention

[0007] The purpose of the present invention is to provide an intelligent cloud-managed high-definition card printer system to solve the technical problem of insufficient card printing accuracy raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent cloud-managed high-definition card printer system, comprising a processor, a database, and a printing module, wherein the processor is connected to the database via signal transmission, the printing module is connected to the processor via signal transmission, the printing module comprises an information acquisition component and a printing component, the information acquisition component is used to collect material information and environmental information, the printing component is used to print cards, and the processor controls the printing module to start by receiving a primary start signal and a secondary start signal;

[0009] The information acquisition component transmits the collected information to the processor through signal transmission. The processor maps RGB to the CIE XYZ standard space through bilinear interpolation to eliminate the color gamut deviation caused by device dependence. By introducing a single nonlinear compensation function based on the Gamma2.2 curve, the quantization error of the dark area is optimized, and a band energy weighting factor is introduced for UV ink. The excitation efficiency is dynamically adjusted according to the measured intensity of the 365nm and 395nm light sources. After constructing a 3x3 color compensation matrix, the BRDF material reflection model is integrated. Based on deep neural network mapping, the input layer inputs standard XYZ values, temperature and humidity, and UV intensity, the hidden layer is a 3x256ReLU unit, and the output layer outputs RGB compensation values ​​through a linear activation function to construct a color compensation model. The processor compensates the output RGB compensation values ​​to the original parameters and then completes the card printing through the printing component.

[0010] Preferably, the information acquisition component includes a spectroradiometer, a multispectral imaging sensor, a UV intensity sensor, a temperature and humidity sensor, a laser ranging sensor and a surface roughness sensor;

[0011] The spectroradiometer is used to measure the spectral power distribution of the visible light band from 300nm to 700nm and analyze the RGB three-channel intensity values. The multispectral imaging sensor collects CIE LAB space raw data. The UV intensity sensor is used to collect ultraviolet intensity in the 365nm and 395nm bands. The temperature and humidity sensor is used to collect ambient temperature and humidity information. The laser ranging sensor is used to measure the distance to the target object. The surface roughness sensor feedbacks the specular reflection coefficient based on laser confocal microscopy technology.

[0012] Preferably, the printing component includes a print head unit, an ink delivery unit, an automatic card feeder and an electric card puncher. The print head unit adopts thermal sublimation transfer technology and performs high-resolution printing through a built-in semiconductor temperature regulation chip. The ink delivery unit integrates cyan, magenta, yellow and black ink pipelines and a dynamic diffusion compensation valve, which can automatically adjust the ink flow according to temperature and humidity data. The automatic card feeder and the electric card puncher are used for batch throughput of cards to be printed.

[0013] Preferably, the processor is connected to the authentication module through signal transmission, and the authentication module includes a face acquisition component and a fingerprint acquisition component, the face acquisition component is used to collect user face information, and the fingerprint acquisition component is used to collect user fingerprint information;

[0014] After receiving user information from the face and fingerprint collection components, the processor encrypts the user data using the AES-256 algorithm and transmits it to the database. It then decrypts the data in a secure environment for feature processing. For abnormalities in user information collection, the processor performs hierarchical compensation on the user information. Primary compensation involves invoking a GAN for repair and injecting controllable Laplace noise to protect privacy, while also enhancing grayscale stretching. Deep compensation involves comparing historical data using the Mahalanoobis distance and combining it with a dynamic threshold formula to monitor feature drift. When the cumulative difference exceeds the threshold, incremental learning updates the template, correcting for historical data deviations. Covariance analysis is used to analyze cross-modal feature similarity. Tensor decomposition technology is used to map the user's facial 3D feature points and fingerprint ridge topology to a unified vector space. Hyperparameters are initialized using a convolutional neural network, and dynamic weight adjustment is integrated. The cross-entropy loss function optimizes the target and the Adam optimizer is used for training. Evaluation indices are independently set, and cross-validation is performed through block iterative training to construct a user recognition model. After user information recognition is complete, the information is transmitted to the processor as a primary start signal.

[0015] Preferably, the face collection component includes: a camera and an ID card reader, the camera is used to collect user face information, the ID card reader is used to automatically extract user identity information, and the processor performs verification and comparison after receiving information transmitted by the camera and the ID card reader;

[0016] The fingerprint collection component is a capacitive sensor, which integrates a glass cover protective layer and anti-static circuit. It builds a three-dimensional fingerprint model based on the capacitance changes caused by the concave and convex texture. The full-width sensor covers 256 levels of grayscale.

[0017] Preferably, the processor is connected to the consumables monitoring module through signal transmission. The consumables monitoring module includes a card monitoring component, an ink monitoring component, and a print head monitoring component. The card monitoring component is used to monitor the number and status of cards to be printed, the ink monitoring component is used to monitor the number and status of unused ink, and the print head monitoring component is used to monitor the working status of the print head unit. The consumables monitoring module transmits information about the card, ink, and print head unit to the processor as a secondary start signal.

[0018] The card monitoring component includes a paper sensor and a material identifier. The paper sensor is used to detect the card status, and the material identifier is used to identify the card type.

[0019] The ink monitoring component includes a liquid level sensor, a viscosity sensor, and a conductivity probe. The liquid level sensor is used to detect the remaining amount of ink cartridges, while the viscosity sensor and conductivity probe are used to detect the ink atomization characteristics.

[0020] The print head monitoring component includes a piezoelectric sensor and a temperature sensor. The piezoelectric sensor is used to detect the vibration frequency of the print head unit, and the temperature sensor is used to detect the operating temperature of the print head unit.

[0021] Preferably, the information acquisition module and the print head monitoring component are connected to the cleaning module through signal transmission. The cleaning module includes a card cleaning component and a print head maintenance component. The cleaning module cleans the device after receiving the signals transmitted by the information acquisition module and the print head monitoring component, and feeds the information back to the processor. The processor quantifies the cleaning process of the cleaning module and converts the cleaning cycle into a cleaning attenuation factor and incorporates it into the band energy weight factor.

[0022] The card cleaning assembly includes a cleaning card and a cleaning roller. The cleaning card is used to clean the internal channels and sensors of the device, and the cleaning roller is used to remove dust from both sides of the card.

[0023] The print head maintenance kit includes an IPA cleaning tool and a sticky roller. The IPA cleaning tool is used to dissolve ink and remove dust, and the sticky roller is used to absorb debris.

[0024] Preferably, the band energy weight factor is ,in is the excitation intensity compensation factor, is the extrusion excitation energy intensity of UV ink in the 300nm~400nm band, and are the measured intensities of the 365nm and 395nm UV light sources, respectively, which are used to adjust the color rendering efficiency of UV ink in the 300nm to 400nm band. λ is the cleaning attenuation factor. ,in For cleaning cycle.

[0025] Preferably, the bilinear interpolation algorithm is , where X is the target color score, used to represent a component of the CIE XYZ standard space, n is the number of reference points, and i is the summation index variable. The contribution weights of the RGB channels to each component of the CIE XYZ standard space are used to eliminate the mapping deviation caused by device color gamut differences. 、 and The device input value for the i-th reference point represents the intensity of the red, green, and blue channels respectively.

[0026] Preferably, the dynamic threshold formula is: ,in is the dynamic threshold for adaptive drift monitoring, and N is the number of accumulated samples.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. This invention achieves high-precision card printing by designing and constructing a color compensation model, solving the problems of reliance on offline measurement, color gamut deviation, and response lag. It can compensate for color deviation in real time, improving the quality and professionalism of card printing.

[0029] 2. This invention achieves the function of accurately identifying user identities by designing and building a user identification model, solving the problems of user privacy leakage and resource waste, reducing the occurrence of abnormal user identity identification, and improving the stability and security of card printing;

[0030] 3. The present invention incorporates a consumables monitoring module to enable timely adjustment and replenishment of consumables, solving the issues of response lag and printing template unification caused by static threshold warnings. It can also avoid color deviations caused by environmental changes, adjust printing parameters according to different materials, and reduce resource loss.

[0031] 4. The present invention realizes the function of quickly locating and handling faults by designing a cleaning module, an information collection component and a print head monitoring component, solving the problems of long troubleshooting time, short equipment service life and poor printing effect. It can timely identify the cause of the fault, accurately locate the fault location, improve the user experience, and optimize the equipment maintenance process. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a schematic diagram of the high-definition card printer system of the present invention;

[0033] Figure 2 This is a schematic diagram of the printing module of the present invention;

[0034] Figure 3 This is a schematic diagram of the authentication module of the present invention;

[0035] Figure 4 This is a schematic diagram of the composition of the consumables monitoring module of the present invention;

[0036] Figure 5 This is a schematic diagram of the cleaning module of the present invention. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] Example 1: Please refer to Figure 1 and Figure 2 An intelligent cloud-managed high-definition card printer system includes a processor, a database, and a printing module. The processor is connected to the database via signal transmission, and the printing module is connected to the processor via signal transmission. The printing module includes an information acquisition component and a printing component. The information acquisition component is used to collect material information and environmental information. The printing component is used for card printing. The processor controls the printing module to start by receiving a primary start signal and a secondary start signal.

[0039] The information acquisition component transmits the collected information to the processor through signal transmission. The processor maps RGB to the CIE XYZ standard space through bilinear interpolation to eliminate the color gamut deviation caused by device dependence. By introducing a single nonlinear compensation function based on the Gamma2.2 curve, the quantization error of the dark area is optimized, and a band energy weighting factor is introduced for UV ink. The excitation efficiency is dynamically adjusted according to the measured intensity of the 365nm and 395nm light sources. After constructing a 3x3 color compensation matrix, the BRDF material reflection model is integrated. Based on deep neural network mapping, the input layer inputs standard XYZ values, temperature and humidity, and UV intensity, the hidden layer is a 3x256ReLU unit, and the output layer outputs RGB compensation values ​​through a linear activation function to construct a color compensation model. The processor compensates the output RGB compensation values ​​to the original parameters and then completes the card printing through the printing component;

[0040] The information acquisition component includes a spectroradiometer, a multispectral imaging sensor, a UV intensity sensor, a temperature and humidity sensor, a laser ranging sensor, and a surface roughness sensor;

[0041] The spectroradiometer is used to measure the spectral power distribution of the visible light band from 300nm to 700nm and analyze the RGB three-channel intensity values. The multispectral imaging sensor collects CIE LAB space raw data. The UV intensity sensor is used to collect ultraviolet intensity in the 365nm and 395nm bands. The temperature and humidity sensor is used to collect environmental temperature and humidity information. The laser ranging sensor is used to measure the distance to the target object. The surface roughness sensor is based on laser confocal microscopy technology to feedback the specular reflection coefficient.

[0042] The band energy weight factor is ,in is the excitation intensity compensation factor, is the extrusion excitation energy intensity of UV ink in the 300nm~400nm band, and are the measured intensities of the 365nm and 395nm UV light sources, respectively, which are used to adjust the color rendering efficiency of UV ink in the 300nm to 400nm band. λ is the cleaning attenuation factor. ,in For cleaning cycle;

[0043] The bilinear interpolation algorithm is , where X is the target color score, used to represent a component of the CIE XYZ standard space, n is the number of reference points, and i is the summation index variable. The contribution weights of the RGB channels to each component of the CIEXYZ standard space are used to eliminate the mapping deviation caused by device color gamut differences. 、 and The device input value for the i-th reference point represents the intensity of the red, green and blue channels respectively;

[0044] Furthermore, when the user uses the card printer to print the card, the spectroradiometer is used to measure the spectral power distribution of the visible light band of 300nm to 700nm and analyze the RGB three-channel intensity value, the multispectral imaging sensor is used to collect CIE LAB space original data, the UV intensity sensor is used to collect the ultraviolet intensity of the 365nm band and 395nm band, the temperature and humidity sensor is used to collect the environmental temperature and humidity information, the laser ranging sensor is used to measure the distance of the target object to assist the BRDF material reflection model, which is used to calculate the incident angle and reflection path of the light, and optimize the reflectivity parameters of UV ink at different printing distances. The surface roughness sensor is used to feedback the mirror reflection coefficient based on laser confocal microscopy technology and input it into the BRDF model. The processor obtains the information of the information collection component during the operation of the card printer and uses the bilinear interpolation algorithm. The RGB value is mapped to the CIE XYZ standard color space, which is defined based on the visual characteristics of the human eye. The color gamut deviation dependent on the device is eliminated through linear conversion, the color of different devices is unified, and the color distortion caused by hardware differences is avoided. Then, a single nonlinear compensation function is introduced based on the Gamma2.2 curve. By dynamically adjusting the gradient slope of the dark area, the highlight area is compressed while the value distribution of the dark area is expanded, and the color level break caused by quantitative stratification is reduced. According to the measured intensity of the light source in the 365nm band and the 395nm band, the formula Calculate the band energy weight factor to ensure the color consistency under different UV bands, so as to adapt to the characteristics of UV light sources and improve the color accuracy of special inks. By constructing a 3x3 color compensation matrix, the dispute and cross-coupling ratio of the RGB channels are adjusted by matrix multiplication. Based on the microsurface theory, BRDF describes the scattering behavior of optical fiber on the surface of the material. The reflected light intensity is calculated by the normal distribution function and the geometric attenuation term, so that the compensated color conforms to the material reflection law of the real world, simulates the optical properties of the surface material of the object, and enhances the physical reality of the color in three-dimensional space. The color compensation model built based on the deep neural network receives the standard XYZ value, temperature and humidity at the input layer. Color and UV intensity are used as environmental perception variables. The hidden layer is a 3-layer fully connected network with 256 ReLU units in each layer. By learning nonlinear mapping relationships, it is used for feature abstraction and complex pattern fitting. The output layer outputs RGB compensation values ​​through a linear activation function, which is directly superimposed on the original signal to complete the correction. The training optimizes the parameters through the loss function to construct a color compensation model. This can avoid the color gamut deviation problem caused by device differences caused by relying on ICC curves or basic color matching in existing technologies. At the same time, temperature and humidity are integrated into the color compensation model as environmental parameters, avoiding the problem of material color deviation caused by the delayed response of environmental parameters in existing technologies, thereby improving the quality and professionalism of card printing.

[0045] Example 2: Please refer to Figure 1 and Figure 3 , an intelligent cloud management high-definition card printer system, the processor is connected to the authentication module through signal transmission, the authentication module includes a face acquisition component and a fingerprint acquisition component, the face acquisition component is used to collect user face information, and the fingerprint acquisition component is used to collect user fingerprint information;

[0046] After receiving user information from the face and fingerprint collection components, the processor encrypts the user data using the AES-256 algorithm and transmits it to the database. It then decrypts the data in a secure environment for feature processing. For abnormalities in user information collection, the processor performs hierarchical compensation on the user information. Primary compensation involves calling a GAN for repair and injecting controllable Laplace noise to protect privacy, while also enhancing grayscale stretching. Deep compensation involves comparing historical data using the Mahalanoobis distance and combining it with a dynamic threshold formula to monitor feature drift. When the cumulative difference exceeds the threshold, incremental learning updates the template, correcting historical data deviations. Covariance analysis is used to analyze cross-modal feature similarity, and tensor decomposition technology is used to map the user's facial 3D feature points and fingerprint ridge topology to a unified vector space. Hyperparameters are initialized through a convolutional neural network, and dynamic weight adjustment is integrated. The cross-entropy loss function optimizes the target and the Adam optimizer is used for training. Evaluation indices are independently set, and cross-validation is performed through block iterative training to build a user recognition model. After user information recognition is complete, the information is transmitted to the processor as a primary start signal.

[0047] The dynamic threshold formula is: ,in is the dynamic threshold for adaptive drift monitoring, and N is the number of accumulated samples;

[0048] Furthermore, after completing the user face and fingerprint collection, the data is immediately encrypted using the ASE-256 algorithm. The ASE-256 algorithm uses a symmetric key mechanism to divide the data into 128-bit blocks for multiple rounds of transformation to ensure confidentiality and integrity during transmission. The encrypted data is transmitted to the database and decrypted in a secure isolated environment. The user's face image is pre-processed by grayscale, denoising, and normalization to extract contours and facial features. The user's fingerprint image includes ridge enhancement and eigenvalue matching to reduce the impact of noise and improve the accuracy of subsequent feature extraction. When the user information collection is abnormal, that is, the image is blurred or missing, the data is repaired and privacy is protected through a hierarchical compensation mechanism to ensure model robustness. The primary compensation is generated The adversarial network simulates the real data distribution through the generator-discriminator architecture and repairs the missing or damaged parts of the image. In this process, sensitive information is blurred by injecting controllable Laplace noise to prevent privacy leakage. At the same time, the injected noise intensity is controllable, which can avoid user information distortion caused by excessive noise injection. It also enhances grayscale stretching, adjusts the image contrast histogram, expands the grayscale range, and improves uneven lighting or low contrast problems. Depth compensation is achieved by comparing historical data using the Mahalanobis distance, calculating the multidimensional statistical distance between the current feature and the historical data set, identifying abnormal offsets, and detecting feature drift through a dynamic threshold formula. The threshold is dynamically adjusted according to the data distribution. When the cumulative difference exceeds the threshold The online learning mechanism can update the model template and adapt to the new data stream. At the same time, it analyzes the feature correlation through covariance analysis and adjusts the deviation to reduce the error. The deep compensation uses statistical methods to deal with long-term drift to ensure the adaptability of the model. The processed user face feature points and fingerprint ridge topology are mapped to a unified vector space, and the correlation between the two is quantified. Through high-order tensor decomposition, the different modal data are reduced to a shared vector space, retaining key structural information. The spatial features are extracted through the convolutional neural network, and the hyperparameters are initialized with preset values ​​to accelerate subsequent training. During training, the weights of each layer are adjusted by integrating dynamic weight adjustment to adapt to changes in feature importance. The cross entropy loss function is used as the loss function for classification tasks, which can minimize The difference between the predicted probability distribution and the true label is calculated through Adam combined with dynamic and adaptive learning rates, the exponential weighted average of the gradient is calculated and the deviation is corrected, the parameters are efficiently converged, the data is loaded in blocks, and mini-batch gradient descent is performed to save memory and accelerate training. By splitting the data set for training and verification, the generalization ability of the model is evaluated to prevent overfitting, which can avoid user privacy leakage and resource waste, and improve the stability and security of card printing. After completing user identity compensation, the processor receives user information for identity confirmation, and the completed identity confirmation is transmitted to the processor as a first-level start signal. The processor waits for the second-level start signal and then starts the printing module to print the card, optimizing the energy consumption of the equipment and avoiding energy waste.

[0049] Example 3: Please refer to Figure 1 、 Figure 2 and Figure 4 An intelligent cloud-managed high-definition card printer system, the printing components include a print head unit, an ink delivery unit, an automatic card feeder, and an electric card puncher. The print head unit uses thermal sublimation transfer technology and a built-in semiconductor temperature regulation chip for high-resolution printing. The ink delivery unit integrates cyan, magenta, yellow, and black ink pipelines and a dynamic diffusion compensation valve that can automatically adjust the ink flow according to temperature and humidity data. The automatic card feeder and electric card puncher are used to batch process cards to be printed.

[0050] The processor is connected to the consumables monitoring module through signal transmission. The consumables monitoring module includes a card monitoring component, an ink monitoring component, and a print head monitoring component. The card monitoring component is used to monitor the number and status of cards to be printed, the ink monitoring component is used to monitor the number and status of unused ink, and the print head monitoring component is used to monitor the working status of the print head unit. The consumables monitoring module transmits information about the card, ink, and print head unit to the processor as a secondary start signal;

[0051] The card monitoring component includes a paper sensor and a material identifier. The paper sensor is used to detect the card status, and the material identifier is used to identify the card type.

[0052] The ink monitoring component includes a liquid level sensor, a viscosity sensor, and a conductivity probe. The liquid level sensor is used to detect the remaining amount of ink cartridges, while the viscosity sensor and conductivity probe are used to detect the ink atomization characteristics.

[0053] The print head monitoring component includes a piezoelectric sensor and a temperature sensor. The piezoelectric sensor is used to detect the vibration frequency of the print head unit, and the temperature sensor is used to detect the operating temperature of the print head unit.

[0054] Furthermore, before the user completes identity authentication and prints the card, the paper sensor automatically feeds the card slot through infrared beam scanning, and monitors the number of cards to be printed and their physical status in real time. When a missing card or an abnormal card is detected, the card feeding is immediately stopped and an alarm is triggered. The material identifier emits a near-infrared spectrum, distinguishes the card type according to the reflection characteristics, and automatically pre-matches the printing parameters to prevent transfer failure caused by material incompatibility. When the paper sensor detects a card jam, the electric card punch automatically rotates in the opposite direction for 0.5 seconds to release the pressure. At the same time, an audible and visual alarm prompts manual intervention. The liquid level sensor measures the remaining amount of the ink cartridge through capacitance changes. When the amount is lower than the 10% threshold, it pushes a replenishment warning to the management personnel. The viscosity sensor and the conductivity probe analyze the ink characteristics in real time. When the viscosity is greater than 30cP, it indicates that the ink is polymerized, triggering the dynamic diffusion compensation valve to increase the flow rate. When the conductivity fluctuates by 5% or more, it indicates that the ink contains impurities and needs to be filtered. The normal frequency range of the piezoelectric sensor at the print head unit is 120Hz-150Hz. If it is lower than 100Hz, it indicates that the print head unit is clogged and needs to be cleaned. If the frequency suddenly rises to more than 180Hz, it indicates mechanical wear and needs to prompt the management personnel to replace it. After receiving the secondary start signal transmitted by the consumables monitoring module, the processor determines that the device can print cards according to the secondary start signal. The processor sends The command is sent to the automatic card feeder, which drives the rollers through a stepper motor to separate the stacked blank cards one by one. The optical sensor detects the position of the cards to ensure that the cards are accurately fed into the printing area. The ink delivery unit delivers ink to the ribbon through the four-color pipeline according to the printing command. The ink flow is adjusted in real time through dynamic diffusion compensation. The ribbon absorbs the ink and covers the surface of the card. The semiconductor temperature regulation chip controls the temperature in different areas. The dye in the low temperature zone of 120℃-170℃ sublimates into gas and penetrates the card layer. The gaseous dye in the high temperature zone of 200℃-230℃ condenses and solidifies to form a permanent image. The print head unit scans line by line at a resolution of 12000dpi. After the single color is completed, the next layer is superimposed. The finished cards are moved to the stamping area, and the electric card punching machine is used to complete the trimming or punching of the cards. The temperature sensor is used to ensure the temperature stability of the high-temperature and low-temperature zones to ensure the complete printing of the card. It can dynamically adjust the ink flow to avoid passive response problems such as color deviation and clogging caused by environmental changes. At the same time, it can adjust the printing parameters according to different materials, and collect information through the information collection component during the card printing process to transmit it to the processor. The processor builds a color compensation model to perform color compensation to improve the card printing effect. When the processor receives the secondary start signal transmitted by the consumables monitoring module and confirms that the card printing cannot be completed, it cancels the operation of the printing module and notifies the management personnel to perform equipment maintenance.

[0055] Example 4: Please refer to Figure 1 and Figure 5, an intelligent cloud-managed high-definition card printer system, wherein the information acquisition module and the print head monitoring component are connected to the cleaning module through signal transmission. The cleaning module includes a card cleaning component and a print head maintenance component. The cleaning module cleans the device after receiving the signals transmitted by the information acquisition module and the print head monitoring component, and feeds the information back to the processor. The processor quantifies the cleaning process of the cleaning module and converts the cleaning cycle into a cleaning attenuation factor and incorporates it into the band energy weight factor;

[0056] The card cleaning assembly includes a cleaning card and a cleaning roller. The cleaning card is used to clean the internal channels and sensors of the device, and the cleaning roller is used to remove dust from both sides of the card.

[0057] The print head maintenance kit includes an IPA cleaning tool and a sticky roller. The IPA cleaning tool is used to dissolve ink and remove dust, and the sticky roller is used to absorb debris.

[0058] The band energy weight factor is ,in is the excitation intensity compensation factor, is the extrusion excitation energy intensity of UV ink in the 300nm~400nm band, and are the measured intensities of the 365nm and 395nm UV light sources, respectively, which are used to adjust the color rendering efficiency of UV ink in the 300nm to 400nm band. λ is the cleaning attenuation factor. ,in For cleaning cycle;

[0059] Furthermore, before and after card printing, a cleaning card composed of a microfiber coating and directional airflow holes is passed through the inside of the device at a speed of 0.5m / s to remove fingerprints or oil stains from each sensor window, and negative ions are released through the surface coating to neutralize the residual static electricity on the card. The cleaning roller is made of silicone material and combined with vacuum adsorption to remove particulate impurities at a speed of 100rpm, and the print head is cleaned with an IPA cleaning tool and a sticky wheel. In addition to the information collected by the print head cleaning component and the information collection module, if there is still a problem with the card printed after the color compensation output by the color compensation model is adjusted and the printing parameters are adjusted, the test results of the print head monitoring component are combined to report to the printer. Managers push maintenance information to complete equipment maintenance, which enables them to promptly identify the cause of the fault, accurately locate the fault location, and shorten equipment downtime. At the same time, the processor quantifies the cleaning process of the equipment by the cleaning module, and converts the cleaning cycle into a cleaning attenuation factor after calculation, which is then incorporated into the calculation of the band energy weight factor. By using the cleaning attenuation factor to reflect the loss of ultraviolet transmission efficiency, it ensures that the energy weight can more accurately match the current equipment status, reduces the compensation deviation under high-fluctuation working conditions, and adjusts the contribution weights of 365nm short-wave ultraviolet light and 395nm long-wave ultraviolet light through the cleaning attenuation factor, which can avoid over-compensation and color distortion of the finished card printing.

[0060] Example 5: Please refer to Figure 1 and Figure 3 , an intelligent cloud-managed high-definition card printer system, the face acquisition component includes: a camera and an ID card reader, the camera is used to collect user facial information, the ID card reader is used to automatically extract user identity information, and the processor performs verification and comparison after receiving the information transmitted by the camera and ID card reader;

[0061] The fingerprint acquisition component is a capacitive sensor, which integrates a glass cover protective layer and anti-static circuit. It builds a three-dimensional fingerprint model based on the capacitance changes caused by the concave and convex texture. The full-width sensor covers 256 levels of grayscale.

[0062] Furthermore, after the camera completes the collection of user face information and user fingerprint collection, and the processor completes the construction of the user recognition model, the user ID card information is obtained through the ID card reader, and the user identity can be further confirmed by obtaining the information in the database. When the user face information or fingerprint information is incomplete, it can be supplemented by the user identity information stored in the database. Combined with primary compensation and depth compensation, the compensation error is reduced, thereby improving the accuracy of card printing. The integration of liveness detection and chip two-way verification can reduce the risk of identity fraud. At the same time, the card printing can be completed by the user independently, thereby improving the efficiency of card printing.

[0063] Working principle: After the camera and capacitive sensor complete the user's face and fingerprint collection, the information is transmitted to the processor. The processor encrypts the data through the ASE-256 algorithm, transmits the encrypted data to the database, decrypts it in a secure and isolated environment, and performs pre-processing such as grayscale, denoising and normalization on the user's face image, extracts the contour and facial features, and performs ridge enhancement and eigenvalue matching on the user's fingerprint image. When the user information collection is abnormal, that is, the image is blurred or missing, the data is repaired and privacy is protected through a hierarchical compensation mechanism. The primary compensation is to generate an adversarial network through a generator-discriminator architecture to simulate the real data distribution, repair missing or The damaged image part is blurred by injecting controllable Laplace noise in this process to prevent privacy leakage. At the same time, the injected noise intensity is controllable, which can avoid user information distortion caused by excessive noise injection. It also enhances grayscale stretching and adjusts the image contrast histogram to expand the grayscale range and improve uneven lighting or low contrast problems. Depth compensation is achieved by comparing historical data with the Mahalanobis distance, calculating the multidimensional statistical distance between the current feature and the historical data set, identifying abnormal offsets, and detecting feature drift through the dynamic threshold formula. The threshold is dynamically adjusted according to the data distribution. When the cumulative difference exceeds the threshold The online learning mechanism can update the model template and adapt to the new data stream. At the same time, it can adjust the deviation to reduce the error by analyzing the feature correlation through covariance analysis. The deep compensation uses statistical methods to deal with long-term drift to ensure the adaptability of the model. The processed user face feature points and fingerprint ridge topology map are mapped to a unified vector space, and the correlation between the two is quantified. Through high-order tensor decomposition, the different modal data are reduced to a shared vector space, retaining key structural information. The spatial features are extracted through the convolutional neural network. The preset values ​​of the hyperparameters are initialized to accelerate the subsequent training. During the training, the weights of each layer are adjusted by integrating dynamic weight adjustment to adapt to the feature weight. The cross-entropy loss function, as the loss function for classification tasks, can minimize the difference between the predicted probability distribution and the true label. Adam combines dynamic and adaptive learning rates to calculate the exponentially weighted average of gradients and correct deviations, efficiently converge parameters. Data is loaded in blocks, and mini-batch gradient descent is performed to save memory and accelerate training. The generalization ability of the model is evaluated by training and validating the split data set to prevent overfitting. At the same time, the user's ID card information is obtained through the ID card reader, and the information in the database can be obtained to further confirm the user's identity. The completion of user identity confirmation is used as the first-level start signal.

[0064] After the user completes identity authentication and before the card is printed, the paper sensor automatically scans the card feed slot through an infrared beam, monitoring the number, remaining amount and physical condition of the cards to be printed in real time. If a missing card or an abnormal card is detected, the card feed is immediately stopped and an alarm is triggered. The material identifier emits near-infrared light to distinguish card types based on reflection characteristics, automatically pre-matching printing parameters to prevent transfer failure caused by material incompatibility. When the paper sensor detects a card jam, the electric card punch automatically rotates in the opposite direction for 0.5 seconds to release pressure, and an audible and visual alarm prompts manual intervention. The liquid level sensor measures the remaining amount of the ink cartridge through capacitance changes. When the amount falls below the 10% threshold, it prompts the management staff to resupply. The viscosity sensor and conductivity probe analyze the ink characteristics in real time. When the viscosity is greater than 30cP, it indicates ink polymerization, triggering the dynamic diffusion compensation valve to increase the flow rate. Conductivity fluctuations of 5% or more indicate that the ink contains impurities and requires a filtration cycle. The normal frequency range of the piezoelectric sensor in the print head unit is 120Hz-150Hz. If it is lower than 100Hz, it indicates that the print head unit is clogged and needs to be cleaned. A sudden rise in frequency to greater than 180Hz indicates mechanical wear, prompting management personnel to replace the equipment's consumables. The consumables monitoring module uses the monitoring information on the equipment's consumables as a secondary start signal. After receiving both the primary and secondary start signals, the processor sends a command to the automatic card feeder. The automatic card feeder uses a stepper motor to drive rollers to separate the stacked blank cards one by one. An optical sensor detects the card position to ensure accurate delivery to the printing area. The ink delivery unit delivers ink to the ribbon through four-color pipes according to the print instructions. Dynamic diffusion compensation adjusts the ink flow in real time. The ribbon absorbs the ink and coats the card surface. A semiconductor temperature control chip controls the temperature in different zones. In the low-temperature zone (120°C-170°C), the dye sublimates into a gaseous state and penetrates the card layer. In the high-temperature zone (200°C-230°C), the gaseous dye condenses and solidifies, forming a permanent image. The print head unit scans line by line at a resolution of 12,000 dpi. After a single color is completed, the next layer is overprinted. The printed cards move to the stamping area, where an electric puncher trims or punches the cards. Temperature sensors ensure stable temperatures in both the high and low temperature zones.

[0065] When the user uses the card printer to print the card, the processor obtains the information of the information collection component during the working process of the card printer and uses the bilinear interpolation algorithm to calculate the value of the card. The RGB value is mapped to the CIE XYZ standard color space, which is defined based on the visual characteristics of the human eye. The color gamut deviation dependent on the device is eliminated through linear conversion, the color of different devices is unified, and the color distortion caused by hardware differences is avoided. Then, a single nonlinear compensation function is introduced based on the Gamma2.2 curve. By dynamically adjusting the gradient slope of the dark area, the highlight area is compressed while the value distribution of the dark area is expanded, and the color level break caused by quantitative stratification is reduced. According to the measured intensity of the light source in the 365nm band and the 395nm band, the formula Calculate the band energy weight factor to ensure the color consistency under different UV bands, so as to adapt to the characteristics of UV light sources and improve the color accuracy of special inks. By constructing a 3x3 color compensation matrix, the dispute and cross-coupling ratio of the RGB channels are adjusted through matrix multiplication. Based on the microsurface theory, BRDF describes the scattering behavior of optical fiber on the surface of the material. The reflected light intensity is calculated through the normal distribution function and the geometric attenuation term, so that the compensated color conforms to the material reflection law in the real world, simulates the optical properties of the surface material of the object, and enhances the physical reality of the color in three-dimensional space. The color compensation model constructed based on the deep neural network receives the standard XYZ value, temperature, humidity and UV intensity as environmental perception variables at the input layer. The hidden layer is a 3-layer fully connected network with 256 ReLU units in each layer. By learning nonlinear mapping relationships, it is used for feature abstraction and complex pattern fitting. The output layer outputs the RGB compensation value through the linear activation function, which is directly superimposed on the original signal to complete the correction. The training optimizes the parameters through the loss function to construct the color compensation model.

[0066] Before and after card printing, a cleaning card consisting of a microfiber coating and directional airflow holes is passed through the interior of the device at a speed of 0.5m / s to remove fingerprints or oil stains from each sensor window. Negative ions are released through the surface coating to neutralize residual static electricity on the card. The cleaning roller is made of silicone material and uses vacuum adsorption to remove particulate impurities at a speed of 100rpm. The print head is cleaned with an IPA cleaning tool and a sticky dust wheel. In addition to the information collected by the print head cleaning component and the information collection module, if there are still problems with the card printed after the color compensation output by the color compensation model and the adjusted printing parameters, the maintenance information will be pushed to the management personnel based on the detection results of the print head monitoring component to complete the maintenance of the equipment.

[0067] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An intelligent cloud-managed high-definition card printer system, characterized by: The system comprises a processor, a database and a printing module. The processor is connected to the database via signal transmission, and the printing module is connected to the processor via signal transmission. The printing module comprises an information acquisition component and a printing component. The information acquisition component is used to collect material information and environmental information, and the printing component is used to print cards. The processor controls the activation of the printing module by receiving a primary activation signal and a secondary activation signal. The information acquisition component transmits the collected information to the processor through signal transmission. The processor maps RGB to the CIE XYZ standard space through bilinear interpolation to eliminate the color gamut deviation caused by device dependence. By introducing a single nonlinear compensation function based on the Gamma 2.2 curve, the quantization error of the dark area is optimized, and a band energy weighting factor is introduced for UV ink. The excitation efficiency is dynamically adjusted according to the measured intensity of the 365nm and 395nm light sources. After constructing a 3x3 color compensation matrix, the BRDF material reflection model is integrated. Based on deep neural network mapping, the input layer inputs standard XYZ values, temperature and humidity, and UV intensity, the hidden layer is a 3x256ReLU unit, and the output layer outputs RGB compensation values ​​through a linear activation function to construct a color compensation model. The processor compensates the output RGB compensation values ​​to the original parameters and then completes card printing through the printing component.

2. The intelligent cloud-managed high-definition card printer system according to claim 1, characterized in that: The information acquisition component includes a spectroradiometer, a multispectral imaging sensor, a UV intensity sensor, a temperature and humidity sensor, a laser ranging sensor, and a surface roughness sensor; The spectroradiometer is used to measure the spectral power distribution of the visible light band from 300nm to 700nm and analyze the RGB three-channel intensity values. The multispectral imaging sensor collects CIE LAB space raw data. The UV intensity sensor is used to collect ultraviolet intensity in the 365nm and 395nm bands. The temperature and humidity sensor is used to collect ambient temperature and humidity information. The laser ranging sensor is used to measure the distance to the target object. The surface roughness sensor feedbacks the specular reflection coefficient based on laser confocal microscopy technology.

3. The intelligent cloud-managed high-definition card printer system according to claim 1, characterized in that: The printing assembly includes a print head unit, an ink delivery unit, an automatic card feeder and an electric card puncher. The print head unit adopts thermal sublimation transfer technology and performs high-resolution printing through a built-in semiconductor temperature regulation chip. The ink delivery unit integrates cyan, magenta, yellow and black ink pipelines and a dynamic diffusion compensation valve, which can automatically adjust the ink flow according to temperature and humidity data. The automatic card feeder and electric card puncher are used for batch throughput of cards to be printed.

4. The intelligent cloud-managed high-definition card printer system according to claim 1, characterized in that: The processor is connected to the authentication module through signal transmission. The authentication module includes a face acquisition component and a fingerprint acquisition component. The face acquisition component is used to collect user face information, and the fingerprint acquisition component is used to collect user fingerprint information. After receiving user information from the face and fingerprint collection components, the processor encrypts the user data using the AES-256 algorithm and transmits it to the database. It then decrypts the data in a secure environment for feature processing. For abnormalities in user information collection, the processor performs hierarchical compensation on the user information. Primary compensation involves invoking a GAN for repair and injecting controllable Laplace noise to protect privacy, while also enhancing grayscale stretching. Deep compensation involves comparing historical data using the Mahalanoobis distance and combining it with a dynamic threshold formula to monitor feature drift. When the cumulative difference exceeds the threshold, incremental learning updates the template, correcting for historical data deviations. Covariance analysis is used to analyze cross-modal feature similarity. Tensor decomposition technology is used to map the user's facial 3D feature points and fingerprint ridge topology to a unified vector space. Hyperparameters are initialized using a convolutional neural network, and dynamic weight adjustment is integrated. The cross-entropy loss function optimizes the target and the Adam optimizer is used for training. Evaluation indices are independently set, and cross-validation is performed through block iterative training to construct a user recognition model. After user information recognition is complete, the information is transmitted to the processor as a primary start signal.

5. The intelligent cloud-managed high-definition card printer system according to claim 4, characterized in that: The face collection component includes: a camera and an ID card reader, the camera is used to collect user face information, the ID card reader is used to automatically extract user identity information, and the processor performs verification and comparison after receiving information transmitted by the camera and the ID card reader; The fingerprint collection component is a capacitive sensor, which integrates a glass cover protective layer and anti-static circuit. It builds a three-dimensional fingerprint model based on the capacitance changes caused by the concave and convex texture. The full-width sensor covers 256 levels of grayscale.

6. The intelligent cloud-managed high-definition card printer system according to claim 1, characterized in that: The processor is connected to the consumables monitoring module through signal transmission. The consumables monitoring module includes a card monitoring component, an ink monitoring component, and a print head monitoring component. The card monitoring component is used to monitor the number and status of cards to be printed, the ink monitoring component is used to monitor the number and status of unused ink, and the print head monitoring component is used to monitor the working status of the print head unit. The consumables monitoring module transmits information about the card, ink, and print head unit to the processor as a secondary start signal; The card monitoring component includes a paper sensor and a material identifier. The paper sensor is used to detect the card status, and the material identifier is used to identify the card type. The ink monitoring component includes a liquid level sensor, a viscosity sensor, and a conductivity probe. The liquid level sensor is used to detect the remaining amount of ink cartridges, while the viscosity sensor and conductivity probe are used to detect the ink atomization characteristics. The print head monitoring component includes a piezoelectric sensor and a temperature sensor. The piezoelectric sensor is used to detect the vibration frequency of the print head unit, and the temperature sensor is used to detect the operating temperature of the print head unit.

7. The intelligent cloud-managed high-definition card printer system according to claim 6, characterized in that: The information acquisition module and the print head monitoring component are connected to the cleaning module through signal transmission. The cleaning module includes a card cleaning component and a print head maintenance component. The cleaning module cleans the device after receiving the signals transmitted by the information acquisition module and the print head monitoring component, and feeds the information back to the processor. The processor quantifies the cleaning process of the cleaning module and converts the cleaning cycle into a cleaning attenuation factor and incorporates it into the band energy weight factor. The card cleaning assembly includes a cleaning card and a cleaning roller. The cleaning card is used to clean the internal channels and sensors of the device, and the cleaning roller is used to remove dust from both sides of the card. The print head maintenance kit includes an IPA cleaning tool and a sticky roller. The IPA cleaning tool is used to dissolve ink and remove dust, and the sticky roller is used to absorb debris.

8. The intelligent cloud-managed high-definition card printer system according to claim 7, characterized in that: The band energy weight factor is ,in is the excitation intensity compensation factor, is the extrusion excitation energy intensity of UV ink in the 300nm~400nm band, and are the measured intensities of the 365nm and 395nm UV light sources, respectively, which are used to adjust the color rendering efficiency of UV ink in the 300nm to 400nm band. λ is the cleaning attenuation factor. ,in For cleaning cycle.

9. The intelligent cloud-managed high-definition card printer system according to claim 1, characterized in that: The bilinear interpolation algorithm is , where X is the target color score, used to represent a component of the CIE XYZ standard space, n is the number of reference points, and i is the summation index variable. The contribution weights of the RGB channels to each component of the CIE XYZ standard space are used to eliminate the mapping deviation caused by device color gamut differences. 、 and The device input value for the i-th reference point represents the intensity of the red, green, and blue channels respectively.

10. The intelligent cloud-managed high-definition card printer system according to claim 4, characterized in that: The dynamic threshold formula is: ,in is the dynamic threshold for adaptive drift monitoring, and N is the number of accumulated samples.

Images