Printing parameter adjusting method based on visual feedback and laser printer

Abstract

The embodiment of the invention discloses a printing parameter adjusting method based on visual feedback and a laser printer. The laser printer comprises a shell, a laser, a photosensitive drum and a controller, the laser, the photosensitive drum and the controller are all located in the shell; the laser comprises a shell, a laser transmitter, an optical splitter and a first sensor; the laser transmitter, the optical splitter and the first sensor are located in the shell; the shell is provided with an opening for printing laser emitted by the laser emitter to penetrate out; the method comprises the steps that printing laser of a first proportion is projected to the direction where a first sensor is located through an optical splitter; sensing the split printing laser through a first sensor to obtain a first image; processing the first image by using at least one convolutional layer of an artificial intelligence AI model to obtain a first feature; abstracting a first image based on the first feature by using a pooling layer or a sensing layer of an AI model to obtain first information; and according to the first information and / or the first image, laser parameters of the laser printer are adaptively adjusted.

Description

Technical Field

[0001] This invention relates to the field of laser technology, and in particular to a method for adjusting printing parameters based on visual feedback and a laser printer. Background Technology

[0002] The general process of laser printing is as follows: Step 1: The laser beam emitted by the laser enters the acousto-optic deflector modulator through the reflector; Step 2: The binary graphic dot matrix information sent by the controller is sent from the interface to the character generator to form the binary pulse information of the required character; Step 3: The signal generated by the synchronizer controls the high-frequency oscillator, which is then applied to the acousto-optic modulator via a frequency synthesizer and power amplifier to modulate the laser beam incident from the reflector. Step 4: The modulated beam enters the multi-faceted rotating mirror, and then the beam is focused by the wide-angle focusing mirror and projected onto the surface of the photosensitive drum, so that the angular velocity scanning becomes the linear velocity scanning. Step 5: The surface of the photosensitive drum is first charged by the charging electrode to obtain a certain potential. Then, after being exposed by the laser beam carrying the image information, an electrostatic latent image is formed on the surface of the photosensitive drum. Step 6: After development by the magnetic brush developer, the latent image is transformed into a visible toner image. When passing through the transfer zone, under the action of the electric field of the transfer electrode, the toner is transferred onto ordinary paper. Step 7: After being preheated and fixed by a high-temperature hot roller, the text and images are fused onto the paper.

[0003] In this process, conventional laser printing equipment struggles to monitor and adjust printing parameters accurately and in real time. Laser parameter settings, in particular, are often based on preset values ​​or manual experience, requiring specialized technicians. This makes it difficult for ordinary users to operate and prevents dynamic optimization based on actual print quality. Consequently, when faced with different printing media, complex patterns, or performance fluctuations during long printing periods, unstable print quality issues such as insufficient clarity, color deviation, and blurred edges can easily occur. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method for adjusting printing parameters based on visual feedback and a laser printer. The technical solution of the present invention is implemented as follows: This disclosure provides a visual feedback-based method for adjusting printing parameters, applied to a laser printer. The laser printer includes a housing, a laser, a photosensitive drum, and a controller. The laser, photosensitive drum, and controller are all located within the housing. The laser includes a housing, a laser emitter, a beam splitter, and a first sensor. The laser emitter, the beam splitter, and the first sensor are located within the housing. An opening on the housing allows the printing laser emitted by the laser emitter to pass through. The printing parameters include laser parameters. The method includes: before the printing laser is projected onto the photosensitive drum through the opening, projecting a first proportion of the printing laser onto the direction of the first sensor using the beam splitter; sensing the split printing laser using the first sensor to obtain a first image; processing the first image using at least one convolutional layer of an artificial intelligence (AI) model to obtain a first feature; abstracting first information from the first image based on the first feature using a pooling layer or a perceptual layer of the AI ​​model; the first information is used to indicate a predicted printing effect; and adaptively adjusting the laser parameters of the laser printer according to the first information and / or the first image.

[0005] A second aspect of this disclosure provides a laser printer, comprising a housing, a laser, a photosensitive drum, and a controller; the laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; the laser emitter, the beam splitter, and the first sensor are located within the housing; the first sensor is used to detect the printing laser after it has been split by the beam splitter to form a first image; the housing has an opening through which the printing laser emitted by the laser emitter passes; the photosensitive drum senses the printing laser passing through the opening to obtain an electrostatic latent image for printing; the controller is used to process the first image using at least one convolutional layer of an artificial intelligence (AI) model to obtain a first feature; to abstract first information from the first image based on the first feature using a pooling layer or a perceptual layer of the AI ​​model; the first information is used to indicate a predicted printing effect; and to adaptively adjust the laser parameters of the laser printer according to the first information and / or the first image.

[0006] A third aspect of this disclosure provides a visual feedback-based printing parameter adjustment device for a laser printer. The laser printer includes a housing, a laser, a photosensitive drum, and a controller. The laser, photosensitive drum, and controller are all located within the housing. The laser includes a housing, a laser emitter, a beam splitter, and a first sensor. The laser emitter, the beam splitter, and the first sensor are located within the housing. An opening on the housing allows the printing laser emitted by the laser emitter to pass through. The printing parameters include laser parameters. Before the printing laser is projected onto the photosensitive drum through the opening, a first proportion of the printing laser is projected onto the direction of the first sensor by the beam splitter. The first sensor is used to sense the split printing laser to obtain a first image. The device includes: a processing module for processing the first image using an artificial intelligence (AI) model to obtain first information; the first information is used to indicate a predicted printing effect; and an adjustment module for adaptively adjusting the laser parameters of the laser printer based on the first information and / or the first image.

[0007] A fourth aspect of this disclosure provides a computer-readable medium having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the visual feedback-based printing parameter adjustment method described in any of the preceding embodiments.

[0008] The technical solution provided in this disclosure uses a spectrometer to pre-select a portion of the printing laser, acquires a first image via a first sensor, and combines this with AI model analysis to achieve multiple technical effects: First, it proactively controls print quality by detecting the laser state before it is projected onto the photosensitive drum, reducing the lag of traditional post-processing detection, identifying printing deviations in advance, and reducing waste of consumables and the cost of repeated printing. Second, it improves the accuracy of parameter adjustment; the convolutional layer of the AI ​​model accurately extracts features such as laser spot and intensity, while the pooling layer or perception layer abstracts information predicting the printing effect, providing a reliable basis for parameter adjustment and achieving adaptive optimization. Third, it ensures printing stability by real-time correction of output fluctuations caused by laser emitter aging and environmental changes, ensuring consistent printing effects in different scenarios and avoiding problems such as blurriness and uneven density. Fourth, it does not interfere with normal printing; spectrometer detection does not affect the operation of the main laser, balancing timely feedback and printing continuity. Attached Figure Description

[0009] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a method for adjusting printing parameters based on visual feedback, provided in an embodiment of the present invention; Figure 2A flowchart illustrating a method for adjusting printing parameters based on visual feedback, provided in an embodiment of the present invention; Figure 3 A schematic diagram of the structure of a laser printer provided in an embodiment of the present invention; Figure 4 A schematic diagram of a printing parameter adjustment device based on visual feedback provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0010] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0011] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0012] This disclosure provides a laser printer comprising a housing, a laser, a photosensitive drum, and a controller; the laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; the laser emitter, the beam splitter, and the first sensor are located within the housing; the housing has an opening through which the printing laser emitted by the laser emitter passes; the printing parameters include laser parameters. Simultaneously, this disclosure provides a visual feedback-based printing parameter adjustment method for the aforementioned laser printer. It is worth noting that this method is not limited to the aforementioned laser printer. Figure 1 As shown, the method includes: S1110: Before the printing laser is projected onto the photosensitive drum through the opening, a first proportion of the printing laser is projected onto the direction of the first sensor by the beam splitter; S1120: Obtain a first image by sensing the split-beam printing laser through the first sensor; S1130: Process the first image using at least one convolutional layer of an artificial intelligence (AI) model to obtain a first feature; S1140: Using the pooling layer or perception layer of the AI ​​model, first information is abstracted from the first image based on the first feature; the first information is used to indicate the predicted printing effect; S1150: Adaptively adjust the laser parameters of the laser printer based on the first information and / or the first image.

[0013] The subject of this disclosure is a laser printer. The laser printer includes components such as a housing and a laser unit; office-grade color or monochrome laser printers are suitable for everyday document printing.

[0014] In some embodiments, the laser may be a semiconductor laser: used to emit printing lasers. Semiconductor lasers are small in size and highly efficient, and are a common type of laser.

[0015] In some embodiments, the beam splitter may include, but is not limited to, a polarizing beam splitter prism: capable of splitting laser light into different proportions, the polarizing beam splitter prism can split light according to the polarization state of the light, and is suitable for high-precision scenarios.

[0016] In some embodiments, the first sensor may include, but is not limited to, a CMOS image sensor: used to sense light signals and convert them into images. CMOS image sensors have high integration and fast response speed.

[0017] In some embodiments, after the printing laser is emitted from the semiconductor laser, before it is projected onto the photosensitive drum through the housing opening, a polarizing beam splitter precisely projects a first proportion of the printing laser (10%, 15%, or 5%) onto the direction of the CMOS image sensor. This precise projection ensures that the split laser light can be accurately sensed by the sensor.

[0018] In the S1120, a CMOS image sensor senses the split printing laser in real time and converts the light signal into a first image containing the shape and intensity distribution of the laser spot. Real-time sensing ensures that the acquired image can reflect the current laser state in a timely manner.

[0019] In S1130, multiple convolutional layers of different types of AI models, such as convolutional neural network models, perform deep feature extraction on the first image to obtain primary features such as the edge sharpness and intensity gradient of the laser spot. Deep feature extraction can capture subtle and crucial information in the image.

[0020] In S1140, the pooling layer of AI models such as convolutional neural networks intelligently abstracts the first image based on the first feature to obtain first information. This information can predict the clarity, density uniformity, and other effects of the printed document. Intelligent abstraction extracts key predictive information by compressing the feature dimension. Using AI models, high-speed and accurate prediction of printing problems such as poor printing results can be achieved.

[0021] In S1150, the method can adaptively adjust printing parameters, exemplarily dynamically. For example, the controller dynamically and adaptively adjusts laser parameters such as the laser power and frequency of the semiconductor laser based on the first information and the first image. For instance, if it is predicted that the print will be too light, the laser power is increased to ensure that the print effect meets expectations. Dynamic adaptive adjustment can quickly respond to changes in laser state and maintain stable print quality.

[0022] For example, in an industrial-grade color laser printer scenario, the laser is a fiber laser, and the beam splitter uses a semi-transparent mirror to capture 20% of the laser light and direct it to a CCD image sensor (the first sensor). During laser transmission, the semi-transparent mirror directs a portion of the laser light onto the CCD, while the main laser still targets the photosensitive drum. The CCD senses the laser and forms a first image containing the color and energy density of the laser spot. The convolutional layers of the deep learning model extract first features such as the edge of the laser spot and color gradation. The fully connected layer (perception layer) of the model abstracts first information to predict the saturation and uniformity of the printed color blocks. If the first information indicates color cast, the controller dynamically adjusts parameters such as the laser wavelength and pulse width to ensure the printing accuracy of industrial drawings.

[0023] In some embodiments, the laser printer is a laser printer with co-source beam splitting; the laser emitter emits N printing laser beams at a time; the laser emitter includes optical elements; the optical elements include a polarizing beam splitter, a waveplate, and N polarizing beam splitters sequentially disposed at the rear end of the laser emitter; one galvanometer corresponds to one printing laser beam; the beam splitter is located between the N polarizing beam splitters and the waveplate, or the beam splitter is located at the rear end of at least one galvanometer.

[0024] Overall, this laser printer is a homogeneous beam splitting laser printer, employing a multi-beam parallel printing architecture based on the original design. Laser Emitter: As the light source, it uses a semiconductor laser array or a high-power single-mode laser to emit N homogeneous printing laser beams at once (N≥2, typically 4-8 beams). Optical Components: Polarization Beam Splitter: Located behind the laser emitter, it splits the initial laser beam into multiple sub-beams with specific polarization states. Waveplate (e.g., λ / 2 waveplate): Adjusts the polarization direction of the sub-beams, ensuring uniform energy distribution of each laser beam. N Galvanometers: Each galvanometer corresponds to one printing laser beam, controlling the laser path through high-speed deflection to achieve parallel scanning printing. Beam Splitter: Has two selectable positions: Option A: Located between the polarization beam splitter and the waveplate, sampling a portion of the initially split laser beam. Option B: Located behind at least one galvanometer, monitoring the final emitted printing laser beam in real time. First Sensor: Works with the beam splitter to receive the split laser signal and generate the first image.

[0025] This type of laser printer enables parallel printing, thus improving efficiency. For example, through homogeneous beam splitting technology, N laser beams can simultaneously scan the photosensitive drum, increasing printing speed by N times. For instance, four laser beams working in parallel can print an A4 document four times faster than single-line printing, significantly improving office efficiency. The characteristics of homogeneous lasers (identical wavelength and phase) ensure highly consistent energy distribution and spot shape across all beams. The beam splitter monitors the status of each laser beam in real time, feeding back to the controller via a first sensor. This allows for dynamic adjustment of the deflection angle of each galvanometer or the laser power, eliminating printing unevenness caused by differences in the optical path, such as color differences or inconsistent line thickness at the junctions of multiple laser beams. When a laser beam malfunctions (e.g., power attenuation, spot distortion), the first sensor can quickly locate the problematic galvanometer or optical path component. The controller can achieve dynamic redundancy of printing tasks by increasing the power of other laser beams or adjusting the scanning strategy, avoiding equipment downtime and making it suitable for industrial-grade continuous printing scenarios.

[0026] In the embodiments of this disclosure, the beam splitter can be configured as needed. For example, for scheme A, sampling at the initial stage of beam splitting allows for unified monitoring of common problems of all laser beams (such as polarization beam splitter offset), reducing the number of sensors. For scheme B, sampling at the rear end of the galvanometer allows for independent monitoring of each laser beam, more accurately correcting optical path deviations caused by galvanometer aging or thermal drift.

[0027] In this embodiment, the co-source beam splitting design avoids the complexity of synchronous control of multiple lasers, reduces the number of light sources and driving circuits, and lowers equipment costs and energy consumption. Simultaneously, the shared beam splitting detection system further optimizes the structure and improves reliability. Exemplarily, this structure is particularly suitable for high-speed office printing, large-format advertising printing, and industrial-grade QR code / barcode printing, significantly improving production efficiency while ensuring print quality. For example, in electronic label printing, multiple laser beams can simultaneously complete high-speed printing of patterns and codes, meeting the real-time needs of automated production lines.

[0028] In some embodiments, the laser emitter emits two co-source printing laser beams, corresponding to two independent galvanometers (galvanometer 1 and galvanometer 2), which control the scanning paths of the two laser beams respectively. The beam splitter adopts a dual-path configuration, located at the rear end of the two galvanometers respectively (i.e., between galvanometer 1 and the opening, and between galvanometer 2 and the opening). At this time, the two co-source laser beams are split by a polarization beam splitter and calibrated by a waveplate before entering galvanometer 1 and galvanometer 2 respectively. After the galvanometers deflect to adjust the laser direction, a portion of the laser is intercepted by the rear beam splitter (e.g., 5% or 10% of the laser is intercepted each). The first sensor (which can be two independent sensors or one dual-channel sensor) acquires the first images (image 1 and image 2) of the two laser beams respectively. The AI ​​model processes the two images respectively to obtain the predicted printing effect of the corresponding two laser beams (first information 1 and first information 2). The controller independently adjusts the laser parameters (such as their respective power and pulse width) of galvanometer 1 and galvanometer 2 based on these two types of information.

[0029] The beam splitter is located at the rear end of the galvanometer and can directly detect the laser state after the galvanometer deflection (such as spot position shift and energy attenuation), reducing the interference of galvanometer mechanical errors (such as deflection angle drift) on the detection results. For example, if the laser projection position of galvanometer 1 is shifted to the left due to aging, the image 1 captured by the beam splitter will directly reflect this deviation. The controller can adjust the drive current or laser power of galvanometer 1 accordingly to ensure that the scanning paths of the two laser beams are accurately matched, avoiding ghosting or gaps at the junction of the two beams.

[0030] Although dual-beam lasers originating from the same source have identical initial characteristics, differences can easily arise after passing through different galvanometers and optical paths (e.g., galvanometers 1 and 2 have different response speeds). By independently detecting and adjusting the parameters of the two lasers, these differences can be eliminated. For example, if image 1 shows that laser 1 has higher energy (leading to overly dense printing), and image 2 shows that laser 2 has lower energy (leading to underly dense printing), the controller can reduce the power of laser 1 and increase the power of laser 2 respectively, ensuring uniform printing density across the areas covered by both lasers. Since parallel printing with dual lasers can double efficiency (e.g., from 30 pages / minute to 60 pages / minute), and the beam splitting design at the rear of the galvanometer can capture instantaneous deviations during high-speed scanning in real time. For example, when printing large-format images, high-frequency deflection of the galvanometer may cause spot distortion in laser 2 at a certain moment. The beam splitter can instantly capture this anomaly, and the controller immediately fine-tunes the pulse frequency of laser 2 to avoid localized defects in the entire page printing, balancing speed and quality. Compared to multi-beam solutions with N>2, the dual-beam configuration requires only two sets of beam splitters and sensors, eliminating the need for complex multi-channel synchronous control. Furthermore, the beam splitting at the galvanometer back end avoids reliance on the shared optical path at the front end (polarizing beam splitter, waveplate), allowing for independent calibration of the terminal state of each laser beam. This reduces system complexity while maintaining accuracy, making it suitable for small-to-medium batch high-speed printing scenarios (such as office documents and express delivery waybills). For printing tasks requiring two laser beams to work together (such as two-color printing and fine line stitching), this solution allows for division of labor and calibration: for example, one laser beam handles black text, and the other handles red markings. When the beam splitter detects a color cast in the red laser, its wavelength parameters can be adjusted independently to ensure accurate two-color superposition without requiring a complete system shutdown for adjustment.

[0031] In this embodiment, the dual beam splitting and independent parameter adjustment at the back end of the galvanometer achieve an optimal balance of accuracy, efficiency and cost in dual-beam high-speed printing scenarios, and is especially suitable for scenarios with high requirements for printing consistency and dynamic response.

[0032] In other embodiments, the laser printer includes M laser emitters; each laser emitter includes a beam splitter and a first sensor; M is a positive integer greater than or equal to 2; the first image is processed using an artificial intelligence (AI) model to obtain first information, including: The AI ​​model generates a second image based on first images provided by M first sensors; wherein the first image is a two-dimensional image; and the second image is a three-dimensional image. The AI ​​model processes the second image to obtain the first information; the first information is also used to adjust a single laser emitter, and / or, the first information is also used to coordinately adjust multiple laser emitters.

[0033] In this embodiment, the laser printer employs a collaborative architecture with M laser emitters (M=2 is used here as an example; M can be expanded to 3, 4, etc.). Laser emitters: Two semiconductor laser emitters (Emitter A and Emitter B) are selected, both installed inside the printer housing. Each emitter independently emits a printing laser, and the beam wavelength can be set according to printing requirements (e.g., 650nm red laser for general document printing, 405nm blue laser for high-precision image printing). Each laser emitter is equipped with one beam splitter (beam splitter A corresponds to emitter A, beam splitter B corresponds to emitter B), and adopts a semi-transparent and semi-reflective mirror structure, which can reflect 6% of the printing laser (first proportion) to the direction of the first sensor, while the main laser is still projected to the photosensitive drum through the opening of the housing. The first sensor is a high-resolution CMOS image sensor (sensor A and sensor B), which receives the laser reflected from beam splitters A and B respectively, and generates a two-dimensional first image (image A and image B) containing the shape of the light spot and the energy distribution. Controller and AI Model: The controller employs a dual-core microprocessor and has a built-in trained AI model (based on a 3D convolutional neural network architecture). The model is connected to two primary sensors via data cables, enabling it to receive and process primary images in real time. Sensors A and B acquire first images (two-dimensional images) of transmitters A and B respectively, containing information such as the position of the laser spot (XY plane coordinates) and intensity value (pixel grayscale) on their respective optical paths. After receiving these two two-dimensional images, the input layer of the AI ​​model fuses them through a 3D convolutional layer: for example, it correlates the intensity of the laser spot at position (x1, y1) in image A with the intensity of the laser spot at position (x2, y2) in image B, and combines this with the angle between the two laser beams' optical paths (preset to 30°), reconstructing the spatial distribution relationship of the two laser beams in three-dimensional space (XYZ axes, where the Z-axis represents the laser propagation direction), generating a second image (three-dimensional image). The second image can visually present the relative positional deviation of the two laser beams (such as whether they are parallel, whether the intersection angle meets the preset value), the intensity distribution of the energy superposition area, and other three-dimensional information. In some embodiments, 3D image processing and first information generation: The pooling layer of the AI ​​model performs dimensionality reduction processing on the 3D image and extracts key features (such as the center distance between the two laser beams, the energy peak difference, and the spatial overlap rate). The fully connected layer (perception layer) abstracts the first information based on these features, including: the predicted printing effect of a single laser beam (such as whether the laser energy of emitter A is uniform, and whether there is distortion in the beam of emitter B); and the collaborative printing effect of two laser beams (such as whether there is excessive melting of toner due to excessive energy in the superimposed area, and whether there are fringe defects caused by interference in the optical path). Based on the above, in this embodiment of the disclosure, the adjustment of a single laser emitter is as follows: if the edge of the laser spot of emitter A in the first information display image A is blurred (predicting that burrs will appear in the print), the controller reduces the laser power of emitter A individually and fine-tunes its drive current until the shape of the laser spot in image A returns to normal. In this embodiment of the disclosure, the adjustment of multiple laser emitters is as follows: if the second image shows that the spatial overlap rate of the two laser beams is too high (predicting that local over-dense printing will occur), the controller coordinates to reduce the power of emitter A and increase the power of emitter B, and increases the distance between the two laser beams by adjusting the angle of their respective galvanometers (if configured), thereby achieving a balanced energy distribution. In some embodiments, the AI ​​model intelligently determines, through information processing, whether adjustments to printing parameters are needed, and whether adjustments should be made to a single laser or a system adjustment of several lasers. For example, the rules and methods for determining single-laser adjustment versus coordinated adjustment can be as follows: First, the triggering scenarios and judgment rules may include: triggering conditions for individual laser adjustment and / or triggering conditions for coordinated adjustment.

[0034] Triggering conditions for adjusting a single laser may include determining that only a single laser adjustment is required when the first information or the first image meets the following characteristics.

[0035] In some cases, an AI model performs independent feature comparisons on M first images. If only the features of one image deviate from the normal sample library, and the spatial region corresponding to the laser in the 3D second image exhibits abnormal features while other regions are normal, then a single adjustment is triggered. For example, defects that could trigger a single laser adjustment might include: Single-beam laser independent defect: The first image of a certain laser shows a unique anomaly that is unrelated to other lasers. For example, the first image of emitter A shows a fixed-position distortion of the light spot (such as jagged edges), while the first image of emitter B is completely normal, and the spatial relationship between the two laser beams in the three-dimensional second image is normal. Local parameter threshold exceeding the limit: Key parameters of a single laser (such as power and spot diameter) exceed the preset threshold, but do not affect the synergistic effect of other lasers. For example: The laser power fluctuation of transmitter B exceeds ±5% (the preset threshold is ±3%), but the energy distribution in the overlapping area of ​​the two lasers is still within the acceptable range. Independent functional module failure: The laser's own hardware components (such as beam splitter, sensor) malfunction. For example, the signal noise of sensor A suddenly increases, causing the first image to be blurred, while sensor B works normally. Triggering conditions for coordinated adjustment may include: when the first information or the second image meets the following characteristics, it is determined that coordinated adjustment of multiple lasers is required: the AI ​​model calculates the coordinated characteristics of multiple laser beams (such as spatial overlap rate and energy difference fluctuation rate). If the characteristics exceed the normal range of the coordinated feature library, and the simulation results are still unsatisfactory after adjusting any one laser individually, then coordinated adjustment is triggered. For example, a co-adjustable defect may include at least one of the following: Multi-laser correlation defect: The 3D second image shows an abnormal spatial relationship between multiple laser beams, and this abnormality cannot be eliminated by adjusting a single laser. For example, the center distance between the spot sizes of two laser beams deviates from the preset value (e.g., preset 5mm, actual 7mm), resulting in gaps at the stitching of the printed image. Excessive coupling of coordinating parameters: The joint parameters of multiple lasers (such as energy superposition ratio and angle deviation) exceed the coordination threshold. For example, if the intensity of the energy superposition region between emitter A and emitter B exceeds the preset upper limit by 15%, excessive carbon powder melting is predicted. Global print quality deviation: The first indication is that the overall print quality (such as uniform density and color consistency across the page) is substandard, and the root cause is an imbalance in the parameter matching of multiple lasers. For example, in a four-color printer, an imbalance in the power ratio of the cyan and magenta lasers leads to color cast in skin tones. Different adjustment strategies are employed for different adjustment types. Specifically, for adjusting a single laser, the strategy is sequential priority and independent closed-loop. The adjustment process employs a detection-adjustment-re-detection closed-loop mode, independently optimizing the parameters of the target laser without affecting the current operating status of other lasers. For example: if the spot diameter of emitter A is detected to be too large (predicting thicker printed lines); the controller individually reduces the power of emitter A (by 2% each time); immediately after adjustment, the first image is re-acquired through the beam splitter. If the spot diameter returns to the acceptable range (e.g., 0.1±0.02mm), the adjustment stops; otherwise, step 2 is repeated until the target is met. Applicable scenarios: scenarios with small parameter adjustment ranges (e.g., power fine-tuning ≤10%) and no cascading effects after adjustment, such as slight energy fluctuations in a single laser. The adjustment strategy for coordinated adjustments is: synchronous linkage and multi-dimensional optimization. Specifically, it may include: Synchronous adjustment mode: The parameters of multiple lasers are adjusted simultaneously to ensure stable coordination during the adjustment process. For example, when the overlap rate of two laser beams is too high (>30%), the controller synchronously executes the following: the galvanometer angle of transmitter A is increased (to reduce the overlap area), the power of transmitter B is reduced (to avoid excessive energy in the overlap area), and the moving speed of the two laser beams remains consistent during the adjustment process (e.g., both change at a rate of 0.1° / ms). Step-by-step collaborative mode: When parameter adjustments have priorities, they are adjusted sequentially step by step, with the collaborative effect evaluated after each step. For example: color deviation between the cyan (C) and yellow (Y) lasers in a four-color printer: First, adjust the wavelength of the cyan laser (basic parameter) to ensure the monochromatic characteristics meet the standard; then adjust the power of the yellow laser (related parameter) to ensure the C+Y mixed color meets the standard; finally, verify the uniformity of the mixed color through a 3D second image to ensure no local deviations. Applicable scenarios: Scenarios with a large parameter adjustment range (e.g., angle deviation > 1°) and strong coupling relationships between multiple laser parameters, such as adjusting the splicing accuracy of large-format images. In some cases, adjustments to individual lasers and coordinated adjustments may occur simultaneously. In such cases, priority can be determined based on the specific requirements. For example, if a single adjustment can resolve a coordinated anomaly (e.g., adjusting the angle of transmitter A alone can correct the overlap rate of the two laser beams), then the single adjustment should be used first (to avoid the complexity of coordinated adjustments). If a single adjustment leads to new coordinated problems (e.g., increasing the power of transmitter A alone will cause the energy in the overlap area to exceed the limit), then coordinated adjustments should be forcibly initiated. In emergency failure scenarios (e.g., a laser suddenly shuts down), a backup laser should be activated first through a single adjustment (e.g., temporarily replacing transmitter C by increasing its power to 120%), and then coordinated parameter adaptation should be performed. In summary, multi-beam collaborative printing achieves a significant leap in accuracy: by generating a 3D second image, the AI ​​model can capture the spatial relationships of multiple laser beams (such as angular deviations and energy superposition), resolving collaborative printing defects that traditional single-beam detection cannot detect (such as uneven brightness at the junction of two laser beams). In actual testing, the edge alignment accuracy of the printed image is improved. Dual protection through single-beam and collaborative adjustment: The first information simultaneously supports independent calibration of a single emitter and collaborative optimization of multiple emitters. This avoids the impact of a single emitter failure on the overall printing (e.g., if the energy of emitter B drops suddenly, its power can be increased individually), and also improves printing efficiency through collaborative adjustment (e.g., two laser beams are used to print the left and right halves of the image, and calibration ensures seamless splicing). Enhanced predictive capabilities from 3D images: Compared to 2D images that only reflect planar features, 3D second images can simulate the three-dimensional effect of lasers on the surface of the photosensitive drum (such as the uniformity of energy penetration in the depth direction). For example, when printing thick paper, the model can predict the toner fixing effect through 3D features and adjust the laser heating time in advance, improving fixing adhesion by 20%. Enhanced scalability and compatibility: Supports any number of laser emitters with M≥2 (e.g., 4 emitters for high-speed four-color printing). The AI ​​model can adapt to more first images by adding input channels. Simultaneously, the 3D image processing logic can be migrated to different brands of laser printers without modifying the hardware structure; adaptation can be achieved simply by fine-tuning the model parameters. Optimized consumables and energy consumption: Collaborative adjustment of multiple emitters avoids energy waste caused by excessive load on a single laser beam (e.g., when two laser beams share the printing task, the power of a single beam can be reduced), improving toner utilization and reducing overall printer energy consumption.

[0036] In some embodiments, the first information includes at least one of the following: laser detection parameters; the laser detection parameters include at least one of the following: wavelength drift difference; temporal jitter cross-correlation between multiple beams; relative intensity noise; phase synchronization error; spatial coherence attenuation; beam pointing angle deviation; risk level parameters; defect information; and suggested adjustment parameters, wherein the suggested adjustment parameters include at least one of the following: power compensation coefficient, spot focus offset, and scanning speed correction value.

[0037] The specific parameters and application scenarios of the first information will be explained in detail. For example, taking a dual-laser printer (emitters A and B) with M=2 as an example, the first information generated by the AI ​​model after processing the first and second images includes the following parameters, each of which plays a different role in adjustment: Category 1 parameters: Laser detection parameters include, but are not limited to, at least one of the following Wavelength drift difference: refers to the difference between the actual wavelength of two laser beams and the standard wavelength (e.g., emitter A drifts by +2nm, emitter B drifts by -1.5nm). When the absolute value of the difference exceeds 0.5nm, it will cause a color shift in color printing (e.g., a longer red laser wavelength will result in an orange-red color). The temporal jitter cross-correlation between multiple beams reflects the synchronization stability of two laser pulses on the time axis. When the correlation coefficient is less than 0.8, stripe misalignment will occur when printing high-speed moving images (such as barcodes). Intensity noise: The random fluctuation range of laser intensity (e.g., the intensity noise of transmitter A is 8%). Excessive random fluctuation range will cause uneven brightness and graininess in the printed image. Phase synchronization error: The phase difference between multiple laser beams deviates from the preset value (e.g., preset 0°, actual 3°), which will cause the stereoscopic effect of the three-dimensional image to be distorted in holographic printing. Spatial coherence attenuation: The degree to which the coherence of the laser decreases during propagation (e.g., attenuation rate of 20%). Excessive attenuation will cause blurred edges in the printing of fine lines. Beam pointing angle deviation: The angle between the actual projection direction of the laser and the target direction (e.g., a deviation of 0.3° from transmitter B) will cause the printed image position to shift. The second type of parameter: Risk level parameters may include, but are not limited to, one of the following: The AI ​​model generates a comprehensive assessment (e.g., high risk, medium risk, low risk) based on the above detection parameters. For example, when the wavelength drift difference is 1nm and the intensity noise is 6%, the risk level is medium risk, indicating that timely adjustment is needed; when the phase synchronization error is 5° and the spatial coherence attenuation is 30%, the risk level is high risk, and printing should be stopped immediately. Defect information: Directly describe the predicted printing defects (such as the risk of excessive concentration in the overlapping area of ​​emitters A and B, and the possibility of line breakage in the left 1 / 3 area). Recommended adjustment parameters: Specific adjustment values ​​provided by the AI ​​model based on defect information, including: Power compensation factor (e.g., +12% for transmitter A, used to compensate for insufficient power). Beam spot focus offset (e.g., +0.02mm, to correct focus offset caused by beam pointing angle deviation); Scan speed correction value (e.g., -5%, to slow down the scan speed to reduce the impact of time domain jitter). Scenario 1: Wavelength and Intensity Adjustment in Color Document Printing. Initial Information: Laser Detection Parameters: Emitter A (Red Light) wavelength drift difference +1.2nm, Emitter B (Blue Light) intensity noise 7%; Defect Information: Red areas are orange-toned, blue areas have graininess; Suggested Adjustment Parameters: Red light power compensation coefficient -8% (reduce energy to correct wavelength perception deviation), blue light scanning speed correction value -10% (extend exposure time to smooth noise); Risk Level: Medium Risk. Adjustment Process: The controller adjusts the power of Emitter A and the scanning speed of Emitter B individually according to the suggested values ​​(single adjustment); After adjustment, the first image is re-acquired, the AI ​​model calculates that the wavelength drift difference is reduced to 0.3nm, the intensity noise is reduced to 4%, the defect information disappears, and the adjustment is complete. Scenario 2: Synchronization and Pointing Adjustment in High-Speed ​​Barcode Printing. First Information Display: Laser Detection Parameters: Timing-Domain Jitter Cross-Correlation of Two Laser Beams 0.7, Beam Pointing Angle Deviation (Emitter B) 0.4°; Defect Information: Barcode lines may exhibit lateral misalignment, with the right edge offset; Suggested Adjustment Parameters: Phase Synchronization Error Compensation -2°, Spot Focusing Offset +0.04mm, Dual-Emitter Scanning Speed ​​Correction +3%; Risk Level: High Risk. Adjustment Process: Initiate collaborative adjustment, synchronously correcting the phase of the two laser beams (reducing phase synchronization error) and the pointing angle of transmitter B (achieving focusing offset through galvanometer angle adjustment); synchronously increase the scanning speed to match the pulse rhythm after phase adjustment, avoiding new timing-domain jitter after adjustment; 3D second image verification shows that the line misalignment error has decreased, meeting barcode printing standards. Scenario 3: Coherence Adjustment in Holographic Image Printing. Initial Information: Laser Detection Parameters: Spatial coherence attenuation 25%, phase synchronization error 4°; Defect Information: Reduced stereoscopic effect in the holographic image, ghosting present; Suggested Adjustment Parameters: Dual emitter power compensation coefficient +15% (enhances coherence), phase synchronization error compensation value -4°; Risk Level: High Risk. Adjustment Process: Coordinated adjustment of the power and phase of the two laser beams to ensure phase synchronization while increasing power; After adjustment, spatial coherence attenuation decreased to 8%, phase synchronization error 0°, ghosting in the holographic image disappeared, and stereoscopic effect was restored. In some embodiments, when multiple detection parameters are abnormal simultaneously, the AI ​​model prioritizes processing parameters with high impact. For example, when wavelength drift difference and intensity noise coexist, the wavelength is corrected first (affecting color tone), and then the intensity is adjusted (affecting detail texture). AI model self-learning optimization: The suggested adjustment parameters accumulate experience with usage scenarios. For example, after printing thick paper multiple times, the model will automatically increase the base value of the power compensation coefficient, reducing the number of repeated adjustments.

[0038] Dynamic updates of risk levels: During the adjustment process, the AI ​​model updates the initial information every specific period. If the risk level drops from high risk to low risk (e.g., the defect information disappears), the adjustment will automatically stop. If the situation does not improve after a specified period, an alarm will be triggered and manual intervention will be requested.

[0039] like Figure 2 As shown, S1130 may include: S1131: The first image is processed using at least one convolutional layer of the AI ​​model to obtain features of the printing laser, including a first type of feature, a second type of feature, and a third type of feature; the first type of feature includes the spatial features of the printing laser; the second type of feature includes the temporal features of the printing laser; the third type of feature is different from the first type of feature and the second type of feature; the third type of feature includes at least a power feature; S1132: Use the fusion layer of the AI ​​model to fuse at least two of the first type of features, the second type of features, and the third type of features to generate the first information.

[0040] Taking a dual-laser emitter printer with M=2 printing express delivery waybills (including dynamic QR codes and text) as an example, the first image is a series of 10 consecutive frames of light spot images from the two laser emitters (each frame spaced 0.1 seconds apart). The AI ​​model processes the image through the following steps: S1131: Convolutional layers extract three types of features. The AI ​​model uses three cascaded convolutional layers (kernel sizes of 3×3, 5×5, and 3×3) to sequentially extract spatial, temporal, and power features from the first image, as detailed below: The first type of feature (spatial features), extracted by the first convolutional layer (3×3 convolutional kernel), focuses on the spatial distribution characteristics of the printing laser, including: Light spot morphology characteristics: such as the smoothness of the light spot edge of emitter A (the edge pixel change rate in consecutive frames is ≤5% to be normal), and the circularity of the light spot of emitter B (circularity deviation ≤8%). Spatial position characteristics: center distance between the two laser beams (preset value 10mm, actual detection value 10.2mm), coordinate offset of the beam in the projection area of ​​the photosensitive drum (e.g., X-axis offset of transmitter A +0.03mm). Uniformity of distribution characteristics: Intensity gradient within a single laser beam spot (intensity ratio of center to edge ≥ 3:1 is normal). The second type of feature (temporal feature), extracted by the second convolutional layer (5×5 convolutional kernel, combined with a time-dimensional sliding window), reflects the dynamic changes of the laser over time, including: Intensity fluctuation characteristics: the difference between the maximum and minimum intensity values ​​of transmitter A in 10 frames (e.g., fluctuation amplitude of 9%); Pulse synchronization characteristics: the alignment of the pulse signals of the two laser beams on the time axis (e.g., a 2ms delay in the 3rd frame); Spot stability characteristics: the amount of change in the shape of the spot in consecutive frames (e.g., the shape change rate of transmitter B is 12%). The third type of feature (non-spatial and temporal features, including power features) is extracted by the third convolutional layer (3×3 convolutional kernel, focusing on the energy dimension), and includes at least: power features: such as the average power of transmitter A (preset 5W, actual 4.6W) and peak power (6.2W); wavelength features: the center wavelength shift of the laser spectrum distribution (red light shift of transmitter A +1nm). Polarization characteristics: stability of laser polarization direction (angle fluctuation ≤2° is normal). S1132: The fusion layer of the AI ​​model adopts an attention mechanism and a weighted summation method to fuse at least two types of features (this embodiment fuses three types of features). The specific process is as follows: Feature weight allocation: The fusion layer assigns weights to three types of features based on the requirements for printing express waybills (balancing QR code clarity and text edge sharpness). Spatial feature weights affect the accuracy of the QR code dot matrix. The weighting of time features affects the dynamic stability during high-speed printing; Power characteristic weights affect the uniformity of text density. During the fusion process, firstly, spatial and power features are cross-validated to calculate the spatial-power correlation value: if the power density at the center of the light spot is too high (e.g., >8W / mm²) and the spatial distribution is uneven, it is predicted that the QR code will be locally too dense; secondly, temporal features are fused with the spatial-power correlation value to analyze the impact of dynamic fluctuations: if the intensity fluctuation in the temporal features is 9% and the spatial offset is 0.03mm, it is predicted that the text edges will be jagged; finally, all correlation results are integrated to generate the first information. When generating the first information, the laser detection parameters are: spatial characteristics (center distance of the spot 10.2mm, offset +0.03mm), temporal characteristics (intensity fluctuation 9%), and power characteristics (average power 4.6W); the defect information is: "the QR code may be locally too dense, and there is a risk of jagged edges on the text"; the suggested adjustment parameters are: spatial characteristic compensation (focus offset of emitter A spot -0.02mm), temporal characteristic compensation (scanning speed correction value -8%), and power characteristic compensation (power compensation coefficient +10%); the risk level is medium risk. In this embodiment, feature classification improves extraction accuracy: features are divided into three categories: spatial, temporal, and power. Convolutional layers can be designed with specific kernels (e.g., convolutional kernels with time windows are used for temporal features) to avoid feature confusion (e.g., independent extraction of spatial position deviation and power fluctuation), thus reducing the feature error rate. Fusion enhances information integrity: a single feature is insufficient to reflect complex printing problems (e.g., excessive density may be due to excessive power or excessive spatial focusing). Fusion can pinpoint the root cause (e.g., in this example, excessive density is due to low power but local energy concentration caused by focus shift), thus improving the accuracy of defect prediction based on the first information. Adapts to multiple scenario requirements: different scenarios have different requirements for feature weights (e.g., spatial features have higher weights in photo printing, while temporal features have higher weights in high-speed printing). The dynamic weight adjustment mechanism of the fusion layer can flexibly adapt without reconstructing the model.

[0041] In some embodiments, the laser printer further includes a developer; the photosensitive drum forms an electrostatic latent image under the action of the printing laser; the developer prints the content onto paper based on the electrostatic latent image; the method further includes: A second image is obtained by acquiring the printing effect on the paper through a second sensor; wherein the second sensor is located inside the housing and behind the developer along the paper's travel path; Second information is determined based on the second image; the second information is used to indicate the actual printing effect. Based on the first information and / or the first image, adaptively adjusting the laser parameters of the laser printer includes: When the actual printing results indicate that the substandard printing quality is related to the laser, the laser parameters of the laser printer are adaptively adjusted based on at least one of the first information, the first image, and the second information.

[0042] The developer is located below the photosensitive drum and contains black toner and a magnetic roller. When an electrostatic latent image is formed on the surface of the photosensitive drum (the laser-irradiated area is charged), the developer uses the principle of electrostatic adsorption to transfer the toner to the latent image area, thus converting the electrostatic latent image into a visible toner image. The second sensor employs a high-resolution CCD camera, installed behind the developer within the housing along the paper's path (i.e., after the paper has passed through the developer and been fixed). The lens faces the paper surface, allowing real-time capture of the printed area to generate a second image reflecting the actual printing effect (such as a photo of the QR code or text on a shipping label). Further, the second sensor acquires the second image. After the paper passes through the developer and is fixed, the second sensor captures images of the printed label at a certain frequency, generating the second image. For example, the second image might show three 0.3mm diameter black spots (overly dense) in the QR code area, and continuous jagged edges on the recipient's text (consistent with the first information prediction). Determining the second information based on the second image may include: an AI model analyzing the second image to extract actual printing effect features (such as black spot area and jagged edge depth) to generate the second information. Actual printing quality parameters may include: QR code black spot coverage 2.5%, text edge jaggedness depth 0.05mm; Quality judgment: Print quality is substandard (preset standard: black spot coverage ≤1%, jaggedness depth ≤0.03mm); Correlation analysis: By comparing the features of the second image and the first image (such as the black spot position coinciding with the spot focus offset area in the first image), it is determined that the substandard quality is related to the laser (excluding other factors such as uneven toner in the developer). Finally, the laser parameters are adjusted by combining the first and second information. Since the actual printing quality is substandard and related to the laser, the controller integrates the first information (prediction) and the second information (actual) and performs the following adjustments: Verify the accuracy of the first information: The black spots in the second image are consistent with the QR code local over-dense predicted by the first information, and the jaggedness is consistent with the risk of jagged text edges, indicating that the first information is reliable and the parameters can be adjusted based on it; Optimization adjustment range: Based on the actual defect level of the second information (black spot coverage exceeds the standard by 1.5 times), the parameters suggested in the first information are further enhanced and adjusted: For example, the original suggestion: spatial focus offset -0.02mm → optimized to -0.03mm (enhancing focus correction for black spot areas); another example: the original suggestion: scanning speed correction value -8% → optimized to -12% (more significantly reducing the impact of dynamic fluctuations).

[0043] In some embodiments, a new wavelength fine-tuning is added: based on the color bias of the text in the second image (slight reddishness) and combined with the wavelength characteristics of the first information, the wavelength compensation value of emitter A (red light) is adjusted by -0.5nm; synchronous adjustment is performed: the controller sends a coordinated adjustment command to the two lasers to ensure that parameter changes take effect synchronously (such as focus offset and power compensation being completed within the same printing cycle). To verify the adjustment effect, the second sensor acquired a second image, showing that the black spot coverage had decreased to 0.8% and the serration depth was 0.02mm, both meeting the quality standards. The second information was updated to "Print quality meets standards." This forms a closed-loop feedback mechanism: the first information (pre-printing prediction) and the second information (post-printing actual result) form a prediction-verification-optimization closed loop, overcoming the limitations of single visual feedback (e.g., the first information might be misjudged due to sensor error, while the second information provides final verification), thus improving the accuracy of parameter adjustments. Precise fault location: By correlating the first and second images, quality problems caused by the laser (e.g., spot issues) and other components (e.g., toner blockage in the developer) can be distinguished, avoiding ineffective adjustments (e.g., no need to adjust laser parameters when the developer malfunctions), and reducing adjustment time. Improved adaptability to complex scenarios: In complex environments such as varying toner humidity and paper thickness, the second information reflects the non-linear relationship between laser parameters and actual printing results (e.g., the same power might appear faint on thick paper). The AI ​​model learns this relationship, ensuring that the prediction accuracy of the first information remains above average even in complex scenarios.

[0044] In some embodiments, user preference information is also incorporated into parameter adjustments. For example, when a user prefers an ink-saving mode, printing specific sub-pixel features (such as 10-15μm dot matrix) involves reverse exposure enhancement to save on laser printing consumables (such as toner consumption). For pixel-level features or features at or above the pixel level, reducing exposure intermittently and instead reverse exposure enhancement is necessary to reduce consumable consumption when ink-saving mode is enabled. This is due to the criticality effect of printing sub-pixel-level features. Printing sub-pixel-level features can be used for high-requirement target document printing such as anti-counterfeiting invoices and medical images.

[0045] In some embodiments, when the user's preferred quality level is higher than a specified level, the calculation weight of sub-pixel level features is increased when determining printing parameters to calculate more suitable printing parameters.

[0046] In one embodiment, an AI model (such as CNN / Transformer) is used to analyze the subpixel-level characteristics of a file (such as gradient distribution and edge energy), not limited to text / image ternary classification. For example, recognizing gradient images requires satisfying the Laplacian smoothness constraint. Dynamically increase the dot density. The synergistic mapping of preferences and features corresponds to the imaging coordinates. The corresponding edge laser energy can be calculated as follows: Establish a joint mapping function between user preferences (e.g., high contrast) and pixel features (e.g., handwritten stroke energy). ,in, For user preference matrix, Pixel edge energy.

[0047] In another embodiment, the method may further include: obtaining a user preference matrix. The feature matrix corresponding to the first feature is identified through an AI model. Real-time monitoring of the laser printer's status, for example, real-time monitoring of the laser printer's hardware status matrix. . It can be used to represent laser power deviation, photosensitive drum charge attenuation rate, fixing temperature fluctuation rate, scanning position error, etc. It is fused based on dynamic weighting coefficients α, β, γ. , , Generate optimized printing parameters.

[0048] In some embodiments, the table below includes a matrix characterizing the current state of the laser printer.

[0049]

[0050]

[0051] Calculated dynamically from the attenuation sensitivity model. The value of i can be 1, 2, 3, or 4. For example, (e.g., when the photosensitive drum wears down, the attenuation sensitivity is increased). Weights).

[0052] For example, fusion based on dynamic weighting coefficients α, β, γ , , The optimized printing parameters can be generated using the following functional relationship: .in, These are the optimized printing parameters. : Hardware state matrix (real-time sensor data). α, β, γ: Dynamic weight coefficients, which can be dynamically adjusted through reinforcement learning, or they can be static parameters. : Parameter optimization function, for example, This can represent a neural network mapping layer. In some embodiments, the dynamic weight coefficients are updated through a reinforcement learning model, and the reward function is: ,in, For print quality, For printing energy consumption, This refers to the status changes of the laser printer.

[0053] In some embodiments, the triggering conditions for a cross-page dynamic adjustment scheme for printing parameters may include: performing parameter adjustments between pages when any of the following conditions are met: Content type switching: The content characteristic category of two consecutive pages changes significantly (e.g., page 1 is plain text, page 2 is an RGB image).

[0054] User preference changes: Users can adjust their preference settings in real time during the printing process via the APP (such as switching from standard mode to high quality mode).

[0055] Hardware status warning: The sensor detects a sudden change in the status of a critical component (such as a sudden increase in the temperature of the photosensitive drum by 5°C or a fluctuation in laser power exceeding ±5%).

[0056] Substrate type change: The paper thickness / type detector detects a change in the media properties of the next page (e.g., switching from 80g plain paper to 200g coated paper).

[0057] Preload analysis: When printing page N, pre-parse the PCL / PS data stream of page N+1 and identify content characteristics (text / image / mixed type) through an AI model.

[0058] Parameter switching strategy: If the difference in characteristics between adjacent pages exceeds a threshold (e.g., SSIM (structural similarity) < 0.6), a print parameter reset is triggered. Specifically, new print parameters are loaded during the page turn interval.

[0059] In response to changes in user preferences: immediately interrupt the current transmission queue and insert a parameter update command.

[0060] Hardware-coordinated protection: Laser power calibration and fixing temperature adjustment are completed during the page gap, avoiding parameter switching within the print line.

[0061] Switching delay control: Cross-page parameter switching must be completed within the page turning interval (e.g., <50ms) to avoid affecting throughput.

[0062] In some embodiments, a dynamic adjustment scheme for printing parameters across page regions is provided. The triggering conditions for this adjustment scheme may include: when the following conditions are simultaneously met within a single page: significant differences in regional characteristics; text regions and image regions are adjacent (e.g., a title + photo combination); and the resolution difference is greater than a specified value, for example, 200 dpi; and gradient regions coexist with RGB image regions (requiring different halftone algorithms).

[0063] Limited local scanning capabilities of laser printers: The scanning motor cannot maintain high-precision positioning when moving at high speeds (the local scanning speed needs to be reduced).

[0064] Region segmentation and parameter mapping: An AI model is used to divide a printed page into several regions (such as a title area, image area, and background area), and feature labels are generated for each region. Dynamic laser control: Based on the feature label, the printing parameters are adaptively switched by region during the scanning process.

[0065] In some embodiments, the adaptive halftone algorithm switches as follows: For text areas: bitmap binarization is used to ensure sharp edges; for gradient areas: an error diffusion algorithm is switched to prevent color banding. Dynamic scanning speed reduction: The scanning speed is automatically reduced in handwritten areas with high curvature, using the formula: ;in, This is the adjusted scan rate. The basic scan rate for printing the current page. The curvature of the current text stroke. The maximum curvature of the stroke.

[0066] In some embodiments, hardware provides real-time feedback: an integrated photoelectric sensor detects the actual imaging effect at the end of the scan line. If ghosting / blurring is detected at the boundary of the region, compensation is immediately triggered: the number of scan repetitions at the boundary is increased, and the laser power at the boundary is finely adjusted.

[0067] In some embodiments, when the parameter differences between adjacent areas within the same page are too large (e.g., laser power difference > 30%), a transition buffer (2-3 scan lines) is inserted, and linear interpolation is used to smooth the transition parameters.

[0068] like Figure 3 As shown, this disclosure provides a laser printer, which includes a housing 310, a laser 320, a photosensitive drum 330, and a controller 340. The laser 320, photosensitive drum 330, and controller 340 are all located within the housing. The laser 320 includes a housing 321, a laser emitter 322, a beam splitter 323, and a first sensor 324. The laser emitter, the beam splitter, and the first sensor are located within the housing. The first sensor detects the printing laser beam split by the beam splitter to form a first image. An opening on the housing allows the printing laser emitted by the laser emitter to pass through. The photosensitive drum senses the printing laser passing through the opening to obtain an electrostatic latent image for printing. The controller processes the first image using at least one convolutional layer of an artificial intelligence (AI) model to obtain a first feature. It then uses a pooling layer or a perceptual layer of the AI ​​model to abstract first information from the first image based on the first feature. The first information is used to indicate a predicted printing effect. Based on the first information and / or the first image, the laser parameters of the laser printer are adaptively adjusted.

[0069] In some embodiments, the laser printer is a laser printer with co-source beam splitting; the laser emitter emits N printing laser beams at a time; the laser emitter includes optical elements; the optical elements include a polarizing beam splitter, a waveplate, and N polarizing beam splitters sequentially disposed at the rear end of the laser emitter; one galvanometer corresponds to one printing laser beam; the beam splitter is located between the N polarizing beam splitters and the waveplate, or the beam splitter is located at the rear end of at least one galvanometer.

[0070] In some embodiments, when N equals 2, the beam splitter is located at the rear end of the two galvanometers.

[0071] In some embodiments, the laser printer includes M laser emitters; each laser emitter includes a beam splitter and a first sensor; M is a positive integer greater than or equal to 2; processing the first image using an artificial intelligence (AI) model to obtain first information includes: The AI ​​model generates a second image based on first images provided by M first sensors; wherein the first image is a two-dimensional image; and the second image is a three-dimensional image. The AI ​​model processes the second image to obtain the first information; the first information is also used to adjust a single laser emitter, and / or, the first information is also used to coordinately adjust multiple laser emitters.

[0072] In some embodiments, the first information includes at least one of the following: Laser detection parameters; the laser detection parameters include at least one of the following: wavelength drift difference; temporal jitter cross-correlation between multiple beams; relative intensity noise; phase synchronization error; spatial coherence attenuation; beam pointing angle deviation; risk level parameters; defect information; suggested adjustment parameters, wherein the suggested adjustment parameters include at least one of the following: power compensation coefficient, spot focusing offset, and scanning speed correction value.

[0073] In some embodiments, relative intensity noise is a laser detection parameter, referring to the ratio (or normalized fluctuation level) of the random fluctuation amplitude of laser intensity to the average intensity. Relative intensity noise is used to quantify the stability of laser output intensity, reflecting the random changes in intensity caused by external interference (such as temperature fluctuations or circuit noise) or its own characteristics during laser propagation or operation. For example, the document states "the intensity noise of transmitter A is 8%", which actually refers to a relative intensity noise of 8%, meaning the intensity fluctuation amplitude accounts for 8% of the average intensity. When the relative intensity noise exceeds a preset threshold (such as 5%), it can cause uneven brightness and graininess in the printed image, affecting the uniformity of print quality.

[0074] In some embodiments, the controller is specifically configured to process the first image using at least one convolutional layer of the AI ​​model to obtain features of the printing laser including a first type of feature, a second type of feature, and a third type of feature; the first type of feature includes spatial features of the printing laser; the second type of feature includes temporal features of the printing laser; the third type of feature is different from the first type of feature and the second type of feature; the third type of feature includes at least a power feature; and at least two of the first type of feature, the second type of feature, and the third type of feature are fused using a fusion layer of the AI ​​model to generate the first information.

[0075] In some embodiments, the laser printer further includes a developer; the photosensitive drum forms an electrostatic latent image under the action of the printing laser; the developer prints the content onto paper based on the electrostatic latent image; the controller of the laser printer is further configured to acquire a second image by acquiring the printing effect on the paper through a second sensor; wherein the second sensor is located within the housing and behind the developer along the paper's travel path; second information is determined based on the second image; the second information is used to indicate the actual printing effect; when the actual printing effect indicates that the printing quality is substandard and related to the laser, the laser parameters of the laser printer are adaptively adjusted according to at least one of the first information, the first image, and the second information.

[0076] like Figure 4 As shown, this disclosure provides a visual feedback-based printing parameter adjustment device. The laser printer includes a housing, a laser, a photosensitive drum, and a controller; the laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; the laser emitter, the beam splitter, and the first sensor are located within the housing; the housing has an opening through which the printing laser emitted by the laser emitter passes; the printing parameters include laser parameters; before the printing laser is projected onto the photosensitive drum through the opening, a first proportion of the printing laser is projected onto the direction of the first sensor by the beam splitter; the first sensor is used to sense the split printing laser to obtain a first image; the device includes: a processing module 410, used to process the first image using an artificial intelligence (AI) model to obtain first information; the first information is used to indicate a predicted printing effect; and an adjustment module 420, used to adaptively adjust the laser parameters of the laser printer according to the first information and / or the first image. In some embodiments, this device can implement the visual feedback-based printing parameter adjustment method described in any of the foregoing technical solutions.

[0077] Combination Figure 5As shown, this application embodiment provides an electronic device, including a processor 10 and a memory 11. Exemplarily, this electronic device may be a controller for a laser printer.

[0078] Optionally, the device may further include a communication interface 12 and a bus 9. The processor 10, communication interface 12, and memory 11 can communicate with each other via the bus 9. The communication interface 12 can be used for information transmission. The processor 10 can call logical instructions in the memory 11 to execute the visual feedback-based printing parameter adjustment method of the above embodiment.

[0079] Furthermore, the logical instructions in the aforementioned memory 11 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0080] The memory 11, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this application. The processor 10 executes functional applications and data processing by running the program instructions / modules stored in the memory 11, thereby realizing the monitoring method of the queue-based oil and gas transportation system in the above embodiments.

[0081] The memory 11 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 11 may include high-speed random access memory and may also include non-volatile memory.

[0082] This electronic device can be used as an edge controller, central controller, or edge control terminal, etc.

[0083] This application also provides a computer-readable medium storing instructions that, when executed by one or more processors, cause the processors to perform the visual feedback-based printing parameter adjustment method of any of the foregoing technical solutions.

[0084] This application provides a computer program product, which includes a computer program stored on a storage medium. The computer program includes program instructions, which, when executed by a computer, cause the computer to perform the above-described visual feedback-based printing parameter adjustment method.

[0085] The aforementioned computer-readable storage medium may be a transient computer-readable storage medium or a non-transitory computer-readable storage medium.

[0086] The technical solutions of this application embodiment can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in this application embodiment. The aforementioned storage medium can be a non-transitory storage medium, including various media capable of storing program code such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks, or it can be a transient storage medium.

[0087] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0088] The embodiments or examples disclosed in this application are not exhaustive, but merely illustrative of some embodiments or examples, and are not intended to limit the scope of protection of this disclosure. Unless contradictory, each step in a particular embodiment or example can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment or example can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment or example can be arbitrarily interchanged. Furthermore, optional methods or examples in a particular embodiment or example can be arbitrarily combined; moreover, embodiments or examples can be arbitrarily combined. For example, some or all steps of different embodiments or examples can be arbitrarily combined, and a particular embodiment or example can be arbitrarily combined with optional methods or examples of other embodiments or examples.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0092] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0093] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for adjusting printing parameters based on visual feedback, applied to a laser printer, characterized in that, The laser printer includes a housing, a laser, a photosensitive drum, and a controller; the laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; the laser emitter, the beam splitter, and the first sensor are located within the housing; the housing has an opening through which the printing laser emitted by the laser emitter passes; The printing parameters include laser parameters; the method includes: Before the printing laser is projected onto the photosensitive drum through the opening, a first proportion of the printing laser is projected onto the direction of the first sensor by the beam splitter; The first image is obtained by sensing the split-spectrum printing laser using the first sensor; The first image is processed using at least one convolutional layer of an artificial intelligence (AI) model to obtain the first feature; The pooling layer or perception layer of the AI ​​model is used to abstract first information from the first image based on the first feature; the first information is used to indicate the predicted printing effect. The laser parameters of the laser printer are adaptively adjusted based on the first information and / or the first image.

2. The method according to claim 1, characterized in that, The laser printer is a laser printer with a single source and split beam; the laser emitter emits N printing laser beams at a time; the laser emitter includes optical elements; the optical elements include a polarization beam splitter, a waveplate, and N polarization beam splitters arranged sequentially at the rear end of the laser emitter; one galvanometer corresponds to one printing laser beam; the beam splitter is located between the N polarization beam splitters and the waveplate, or the beam splitter is located at the rear end of at least one galvanometer.

3. The method according to claim 2, characterized in that, When N equals 2, the beam splitter is located at the rear end of the two galvanometers.

4. The method according to claim 1, characterized in that, The laser printer includes M laser emitters; each laser emitter includes a beam splitter and a first sensor; M is a positive integer greater than or equal to 2; The first image is processed using an artificial intelligence (AI) model to obtain first information, including: The AI ​​model generates a second image based on first images provided by M first sensors; wherein the first image is a two-dimensional image; and the second image is a three-dimensional image. The AI ​​model processes the second image to obtain the first information; the first information is also used to adjust a single laser emitter, and / or, the first information is also used to coordinately adjust multiple laser emitters.

5. The method according to any one of claims 1 to 4, characterized in that, The first information includes at least one of the following: Laser detection parameters; the laser detection parameters include at least one of the following: wavelength drift difference; temporal jitter cross-correlation between multiple beams; relative intensity noise; phase synchronization error; spatial coherence attenuation; beam pointing angle deviation; Risk level parameters; Defect information; The suggested adjustment parameters include at least one of the following: power compensation coefficient, spot focus offset, and scanning speed correction value.

6. The method according to any one of claims 1 to 4, characterized in that, The first feature is obtained by processing the first image using at least one convolutional layer of an artificial intelligence (AI) model, including: The first image is processed using at least one convolutional layer of the AI ​​model to obtain features of the printing laser, including a first type of feature, a second type of feature, and a third type of feature; the first type of feature includes the spatial features of the printing laser; the second type of feature includes the temporal features of the printing laser; the third type of feature is different from the first type of feature and the second type of feature; the third type of feature includes at least a power feature; The first information is generated by fusing at least two of the first type of features, the second type of features, and the third type of features using the fusion layer of the AI ​​model.

7. The method according to any one of claims 1 to 4, characterized in that, The laser printer further includes a developer; the photosensitive drum forms an electrostatic latent image under the action of the printing laser; the developer prints the content onto paper based on the electrostatic latent image; the method further includes: A second image is obtained by acquiring the printing effect on the paper through a second sensor; wherein the second sensor is located inside the housing and behind the developer along the paper's travel path; Second information is determined based on the second image; the second information is used to indicate the actual printing effect. Based on the first information and / or the first image, adaptively adjusting the laser parameters of the laser printer includes: When the actual printing results indicate that the substandard printing quality is related to the laser, the laser parameters of the laser printer are adaptively adjusted based on at least one of the first information, the first image, and the second information.

8. A laser printer, characterized in that, The laser printer includes a housing, a laser, a photosensitive drum, and a controller; The laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; The laser emitter, the beam splitter, and the first sensor are located inside the housing; The first sensor is used to detect the printing laser formed by the beam splitter to create a first image; the housing has an opening through which the printing laser emitted by the laser emitter passes; The photosensitive drum senses the printing laser passing through the opening to obtain an electrostatic latent image for printing; The controller is used to process the first image using at least one convolutional layer of an artificial intelligence (AI) model to obtain a first feature; The pooling layer or perception layer of the AI ​​model abstracts first information from the first image based on the first feature; the first information is used to indicate the predicted printing effect; and the laser parameters of the laser printer are adaptively adjusted according to the first information and / or the first image.

9. A visual feedback-based printing parameter adjustment device, applied to a laser printer, characterized in that, The laser printer includes a housing, a laser, a photosensitive drum, and a controller; the laser, photosensitive drum, and controller are all located within the housing; the laser includes a housing, a laser emitter, a beam splitter, and a first sensor; the laser emitter, the beam splitter, and the first sensor are located within the housing; the housing has an opening through which the printing laser emitted by the laser emitter passes; the printing parameters include laser parameters; before the printing laser is projected onto the photosensitive drum through the opening, a first proportion of the printing laser is projected onto the direction of the first sensor by the beam splitter; The first sensor is used to sense the split-ray printing laser to obtain a first image; the device includes: The processing module is used to process the first image using an artificial intelligence (AI) model to obtain first information; the first information is used to indicate the predicted printing effect. An adjustment module is configured to adaptively adjust the laser parameters of the laser printer based on the first information and / or the first image.

10. A computer-readable medium having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the method as described in any one of claims 1 to 7.

Images