A printing device and its visual positioning method for positioning a printing substrate
By working together with a visual positioning device and a moving component, automated and precise printing is achieved, solving the tedious and error-prone problems caused by manual pattern matching in existing technologies, and improving the automation level and production efficiency of inkjet printing equipment.
Patent Information
- Application Number
- CN202510635040.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-16
- Publication Date
- 2026-11-17
AI Technical Summary
Existing inkjet printing equipment requires manual pattern matching when printing on personalized substrates, which makes the process cumbersome and prone to errors, with low automation and difficulty in meeting the needs of high-efficiency production.
The system uses a visual positioning device to automatically acquire the spatial position information and printing information of the substrate, and achieves automated and precise printing through the cooperation of the first and second moving components.
It reduces the tediousness and error rate of manual operation, improves printing accuracy and production efficiency, adapts to substrates of different sizes or shapes, and enhances the stability and efficiency of large-scale production.
Smart Images

Figure CN123104246A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of printing technology, and in particular to a printing device and a visual positioning method for a printing substrate. Background Technology
[0002] Inkjet printing technology is a technique that uses tiny nozzles on a printhead to precisely eject ink droplets onto the surface of a printing medium to form images or text. This technology is widely used in office printing, industrial signage, packaging printing, and personalized customization. Traditional inkjet printing is typically designed for two-dimensional planar media, such as paper and plastic film. However, with the diversification of market demands, more and more industries require personalized printing on substrates of different materials and shapes; for example, custom patterns can be printed on the surfaces of glass, metal, ceramics, and wood to meet consumers' personalized needs. Taking customized mobile phone cases as an example, consumers can choose different patterns, colors, and even 3D effects to achieve a highly personalized product appearance.
[0003] However, in existing technologies, printing on large quantities of personalized substrates often requires matching a specific pattern to each medium. Since the shape, size, and required pattern of the substrate may vary, and their placement on the printing platform may also differ during printing, precise pattern matching and positioning are typically necessary before printing. Current processes generally rely on manual pattern matching, adjusting the pattern to be printed according to the specific size and position of each substrate to ensure precise alignment with the media surface, ultimately generating a printable file that the printing equipment can recognize. This process is not only time-consuming and labor-intensive but also prone to errors due to human intervention, leading to pattern misalignment, offset, or printing failure, thus affecting product quality and production efficiency. Furthermore, most existing inkjet printers have low levels of automation in mass personalized printing, making it difficult to meet the demands of high-efficiency production. Many traditional printing devices lack intelligent recognition and automatic adjustment functions, failing to automatically correct for the actual size and position of the substrate, requiring manual intervention, which increases production costs and error rates. Therefore, in the current market environment, how to improve the intelligence level of inkjet printing technology and enhance the automation and production efficiency of personalized printing has become an urgent problem for the industry to solve.
[0004] Chinese patent CN107020832B discloses an image printing system for IoT mobile phone cases. This system aims to address the problems of limited personalization options, high costs, and low efficiency associated with existing mobile phone case designs. The patent proposes an automated personalization method, utilizing an inkjet printer and a device for transmitting painted pattern information to achieve the painting of mobile phone cases. However, the system does not detail how it automatically matches the position of different mobile phone cases with their corresponding patterns, still potentially requiring manual intervention, a process that is cumbersome and prone to errors. Summary of the Invention
[0005] In view of this, embodiments of the present invention provide a printing device and a visual positioning method for visually positioning the substrate, in order to solve the problem that in the prior art, printing devices usually require manual matching of different personalized patterns with the positions of various substrates before printing, and then generating print files, which makes this process cumbersome and prone to errors.
[0006] In a first aspect, the present invention provides a printing apparatus employing visual positioning of a substrate, comprising a frame and a first moving component and a second moving component disposed on the frame.
[0007] The first movable component is provided with a printing component that can move relative to the first movable component, and a visual positioning device is also provided on one side of the printing component;
[0008] The second moving component is provided with a carrier component for holding the printing substrate, and the carrier component can enter the printing area covered by the printing component through the second moving component;
[0009] The carrier component is provided with an information identifier and at least one carrier container for accommodating the printing substrate. The information identifier includes printing information corresponding to the printing substrate medium.
[0010] The visual positioning device is used to obtain the position information of the substrate and, based on the information identifier, obtain the printing information;
[0011] The print file is transferred to the printing carriage where the printing component is located to realize printing, wherein the print file is generated from the printing information and the location information.
[0012] Preferably, the visual positioning device is mounted on one side of the printing carriage via a mounting bracket;
[0013] The visual positioning device includes a recognition component and a light source component spaced apart on top of the recognition component;
[0014] The light source assembly includes a filter disposed below the recognition assembly and a light source disposed on one side of the filter;
[0015] The frame is provided with a first expansion plate and a second expansion plate on both sides;
[0016] It also includes a printing area disposed between the first expansion plate and the second expansion plate;
[0017] The first moving component is mounted on the frame via the first expansion plate and the second expansion plate, and the printing component is provided with a printing interval distance from the printing area.
[0018] Preferably, the first moving component includes a guide rail frame and a slide groove disposed on the guide rail frame, and a first limiting member is provided on both sides of the guide rail frame;
[0019] The guide rail frame is provided with a first set of guide rails on both sides of the slide groove;
[0020] The first set of guide rails is equipped with a movable seat that is connected to the first set of guide rails via a slider;
[0021] The movable base is equipped with a moving part;
[0022] The first moving component drives the moving element to move within the slide groove via a moving cylinder, thereby causing the moving seat to move on the first set of guide rails;
[0023] The printing component is mounted on the movable base.
[0024] Preferably, the second moving component includes a second set of guide rails disposed on both sides of the printing area on the frame and a printing platform that moves based on the second set of guide rails;
[0025] The printing platform is equipped with positioning holes;
[0026] The carrier component is detachably mounted on the printing platform through the positioning hole.
[0027] Preferably, the frame is further provided with a third moving component;
[0028] The third moving component includes lifting parts disposed on both sides of the frame; the lifting parts on both sides of the frame are connected to the connecting shaft via couplings.
[0029] Preferably, the lifting part includes a worm gear transmission seat and a screw driven by the worm gear transmission seat;
[0030] The worm gear transmission base is also provided with a driven wheel;
[0031] The screw is connected to a lifting frame, which includes a connecting plate and a first guide rod and a second guide rod disposed on both sides of the connecting plate.
[0032] The first guide rod and the second guide rod move based on the guide sleeve disposed on the frame;
[0033] The frame is also equipped with a drive motor, and the output shaft of the drive motor is equipped with a drive pulley; the drive pulley is connected to the driven pulley via a belt.
[0034] The first guide rod, the second guide rod, and the other end of the screw are equipped with lifting plates;
[0035] The first extension plate and the second extension plate are respectively disposed on the lifting plates of the lifting parts on both sides of the frame, and move up and down based on the lifting plates.
[0036] Secondly, the present invention also provides a visual positioning method for a printing device, used for visually positioning a printing device as described above, the method comprising:
[0037] The substrate is placed inside the carrier container, and the information label is affixed to the carrier component;
[0038] The second moving component drives the printing platform to the preset printing position;
[0039] The printing component is driven by the first moving component to move towards the preset printing position;
[0040] The spatial location information of the substrate and the printing information in the information label are obtained through the visual positioning device.
[0041] The location information and the printing information are merged to generate a print file, which is then transmitted to the printing carriage for printing.
[0042] Preferably, obtaining the spatial location information of the substrate and the printing information in the information identifier through the visual positioning device includes:
[0043] Acquire an image of the substrate, and perform illumination compensation processing on the image of the substrate to obtain the target image;
[0044] Edge detection is performed on the target image to obtain the contour feature information of the substrate;
[0045] Geometric fitting is performed on the contour feature information to determine the boundary shape of the substrate;
[0046] Based on the boundary shape, and combined with a preset perspective transformation and posture calculation algorithm, the spatial position information of the substrate is determined, wherein the spatial position information includes position coordinates and posture angles.
[0047] The target image is input into a pre-trained identifier segmentation model to determine the identifier region;
[0048] The identification information of the identified area is extracted, and the printing information is determined based on the extracted identification information. The printing information includes the pattern to be printed, the color to be printed, and the path to be printed.
[0049] Preferably, the step of geometrically fitting the contour feature information to determine the boundary shape of the substrate includes:
[0050] Based on the contour feature information, an initial edge image corresponding to the target image is obtained;
[0051] The initial edge image is input into a pre-trained semantic segmentation model to obtain the semantic boundary mask of the substrate, wherein the semantic segmentation model is used to distinguish the substrate region and the background region in the image;
[0052] The initial edge image and the semantic boundary mask are subjected to boundary fusion processing to obtain the target boundary image;
[0053] Contour points are detected and extracted from the target boundary image, and the extracted contour point set is filtered to obtain a filtered contour point set, wherein the filtering includes removing outliers and repairing break points;
[0054] The selected set of contour points is subjected to geometric fitting to determine the boundary shape.
[0055] Preferably, the step of extracting identification information from the identification area and determining the printing information based on the extracted identification information includes:
[0056] A labeled region image is obtained, and the labeled region image is input into a visual coding model to determine the structural feature vector of the image modality. The visual coding model is an image coding network built based on a convolutional neural network.
[0057] Based on an end-to-end character detection and recognition network, text recognition is performed on the image of the identified area to obtain text content information;
[0058] The text content information is input into the text semantic encoder to extract the text vector of the text modality. The text semantic encoder is a pre-trained language model.
[0059] The structural feature vector and the text vector are subjected to multimodal fusion processing to obtain a fused feature vector. The fusion processing includes a cross-modal attention mechanism or a feature splicing fusion strategy.
[0060] The printing information is determined based on the fused feature vector and a preset multi-label classifier.
[0061] In summary, the beneficial effects of the present invention are as follows:
[0062] This invention provides a printing device and method for visually positioning the substrate. The method automatically acquires the spatial position information of the substrate and corresponding printing information through a visual positioning device, and then merges these information into a print file via a host computer, achieving automated and precise printing. Compared to existing technologies, this invention eliminates the need for manual pattern matching, reducing the tediousness and error rate of manual operation, and improving printing accuracy and production efficiency. Through the cooperation of the first and second moving components, the substrate can accurately enter the printing area, and real-time correction is performed by the visual positioning device to ensure that the pattern is accurately printed at the target position. Furthermore, this device is suitable for substrates of different sizes or shapes, and can simultaneously place substrates of different sizes or shapes within the carrier component for printing; it features high automation, intelligence, and strong adaptability, effectively improving the stability and efficiency of large-scale production. Attached Figure Description
[0063] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, and these are all within the protection scope of the present invention.
[0064] Figure 1 This is a schematic diagram of a printing device that uses visual positioning to locate the printing substrate.
[0065] Figure 2 This is a partial structural diagram of a printing device that uses visual positioning to locate the printing substrate.
[0066] Figure 3 This is a structural diagram of the first moving component;
[0067] Figure 4 This is a structural diagram of the printing component;
[0068] Figure 5 This is a structural diagram of the third moving component;
[0069] Figure 6 This is a structural diagram of the safety component;
[0070] Figure 7 This is a schematic diagram of the overall process of the visual positioning method for printing equipment;
[0071] Parts and component numbers in the diagram:
[0072] 100 - Frame, 110 - First expansion plate, 120 - Second expansion plate, 130 - Base frame, 140 - Drive motor, 141 - Drive wheel;
[0073] 200-First moving component, 210-Guide rail frame, 211-Slide groove, 212-First limiting component, 213-First set of guide rails, 214-Slider, 215-Moving seat, 216-Moving element;
[0074] 300-Second moving component, 310-Second set of guide rails, 320-Printing platform, 330-Carrying component, 331-Carrying container, 332-Information container, 340-Second limiting component;
[0075] 400-Third moving component, 410-Lifting part, 420-Worm gear transmission seat, 421-Screw, 422-Driven wheel, 430-Lifting frame, 431-Connecting plate, 432-First guide rod, 433-Second guide rod, 434-Guide sleeve, 435-Lifting plate, 440-Coupling, 450-Connecting shaft;
[0076] 500 - Printing Components;
[0077] 600 - Visual positioning device; 610 - Mounting bracket; 620 - Identification component; 630 - Light source assembly; 631 - Filter; 632 - Light source;
[0078] 700 - Maintenance workstation;
[0079] 800 - Safety component, 810 - Sensor, 821 - First link, 822 - Second link, 823 - Protective bar. Detailed Implementation
[0080] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In the description of the present invention, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements, but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, the element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Where there is no conflict, embodiments of the present invention and the various features thereof can be combined with each other, all of which are within the scope of protection of the present invention.
[0081] Example 1
[0082] Please see Figure 1 Embodiment 1 of the present invention provides a printing device that uses visual positioning of the printing substrate, including a frame 100 and a first moving component 200 and a second moving component 300 disposed on the frame 100; the frame 100 is the main structure of the device and is responsible for supporting the stable operation of the entire printing device;
[0083] The second moving component 300 is disposed on the rack 100, and the first moving component 200 is disposed on the second moving component 300 and can move based on the first moving component 200.
[0084] During the printing process, the first moving component 200 and the second moving component 300 can assist other parts of the printing equipment to move along a predetermined path, so as to accurately position and print the pattern during the printing process.
[0085] Please see Figure 1In this embodiment, a printing component 500 capable of moving based on the first moving component 200 is configured on the first moving component 200. Furthermore, a visual positioning device 600 is specifically provided on one side of the printing component 500.
[0086] More specifically, in the initial position of the printing equipment, the visual positioning device 600 is positioned on the printing component 500, closer to the second moving component 300. With this arrangement, when a printing job is about to begin, the printing component 500 moves towards a specific carrier component 330, preparing to print on the substrate. During this movement, the visual positioning device 600 approaches the carrier component 330 earlier than the printing component 500, allowing it to identify and process relevant information on the carrier component 330 before the actual printing job begins. This helps the printing component 500 to more accurately position and print the desired pattern. This advance identification capability greatly improves printing accuracy and efficiency, reducing printing errors caused by inaccurate positioning. Simultaneously, because the visual positioning device 600 can acquire information from the carrier component 330 in advance, the printing equipment can automatically adjust printing parameters, such as printing speed and printing pressure, to ensure optimal printing results. Therefore, this design of setting a visual positioning device 600 on the printing component 500 not only improves the degree of automation of printing, but also significantly improves the overall quality and efficiency of printing jobs.
[0087] The visual recognition pattern printing equipment of this invention is suitable for various substrates, including paper, plastic, metal, glass, ceramics, leather, and fabrics, and is widely used in packaging materials, electronic product casings, industrial nameplates, textiles, handicrafts, stationery and office supplies, and advertising display materials. This equipment can accurately identify the position and printing information of the substrate, achieving high-precision personalized printing. It is particularly suitable for customized production in industries such as food, pharmaceuticals, clothing, home decoration, and electronic products, improving production efficiency and printing quality to meet diverse market demands.
[0088] For example: cardboard boxes, plastic bags, labels, paper bags, metal packaging, etc., used for personalized packaging printing in the food, pharmaceutical, and cosmetic industries; as well as mobile phone cases, laptop shells, and headphone charging cases; or wood panels, ceramics, glass products, acrylic sheets, etc., used for home decoration and personalized gift making.
[0089] Please see Figure 1 and Figure 2In Embodiment 1 of the present invention, a carrier component 330 is disposed on the second moving component 300, the carrier component 330 being used to carry and hold the substrate. This arrangement of the carrier component 330 allows the substrate to be stably placed on the second moving component 300 and to move along with the second moving component 300. When the second moving component 300 moves within the printing device, the carrier component 330 moves into the printing area covered by the printing component 500. Thus, when the printing component 500 performs a printing operation, the substrate is in the correct printing position, ensuring the smooth progress of the printing process.
[0090] In Embodiment 1 of the present invention, the carrier component 330 is provided with at least one carrier container 331 for placing and holding the printing substrate. The carrier container 331 is designed to accommodate the printing substrate and ensure its stability and ease of operation.
[0091] During the design and manufacturing process, the number of carrier containers 331 can be flexibly adjusted according to actual needs. Each carrier container 331 includes a holding slot and a receiving box. For example, a dedicated carrier assembly 330 can be designed to meet specific production requirements; or various carrier containers 331 of different sizes can be set on the carrier assembly 330 so that, during printing operations, the most suitable carrier container 331 can be selected for use based on different printing tasks and material requirements. This design flexibility not only improves production efficiency but also adapts to diverse product design and production requirements, thus providing manufacturers with greater convenience and cost-effectiveness.
[0092] In addition, the carrier component 330 may also be provided with an information container 332, which corresponds to the printing information on the substrate within the carrier container 331. Information identification is set through the information container 332, which includes an information slot. The function of the information container 332 is to accurately record and identify the printing content on the substrate, thereby making the entire printing process more orderly and efficient.
[0093] In some embodiments, the information identifier is directly set in a position that is compatible with the carrier container 331, and during printing, the information identifier can be directly scanned and identified.
[0094] In some embodiments, the pattern information to be printed needs to be pre-stored in a specific information container 332 on the substrate contained in the carrier container 331. Once a print job is triggered, the visual positioning device 600 located in the printing device begins to function, recognizing and interpreting the pattern information in the information container 332. Subsequently, the host computer of the printing component 500 receives this information and generates the corresponding print file. These files are then transmitted to the printing carriage of the printing component 500, which is responsible for accurately printing the pattern information onto the substrate, thereby completing the entire printing process.
[0095] However, in this embodiment, by introducing a collaborative working mechanism between a visual positioning device 600 and a printing carriage, the tedious pre-printing preparation process can be effectively simplified and optimized, improving printing efficiency. Specifically, this visual positioning device 600 has the ability to automatically identify products of different sizes and can accurately capture the actual placement position of each phone case on the printing platform. Next, the visual positioning device 600 also reads the personalized pattern information stored in each information container 332. After receiving this information accurately matched by the visual positioning device 600, the printing carriage automatically adjusts its printing path and position to ensure that the pattern is printed accurately in the predetermined position on each product. In this way, not only are the tedious manual pattern matching steps eliminated, but the possibility of errors is also greatly reduced, making the process of mass printing personalized phone cases more efficient, accurate, and reliable.
[0096] Furthermore, the visual positioning device 600 is mounted on one side of the printing carriage via a mounting bracket 610, enabling the visual positioning device 600 to work in conjunction with the printing carriage, thereby improving overall work efficiency and accuracy.
[0097] Please see Figure 4The visual positioning device 600 includes a recognition component 620 and light source components 630 spaced apart on top of the recognition component 620. This structural design enables the visual positioning device 600 to obtain clearer and more accurate image information during image capture and processing. The recognition component 620 includes a camera, and the visual positioning device 600 also includes a processor and a memory, thus forming a complete image recognition and processing system. The processor is used to analyze and process the image data acquired by the recognition component 620, performing key algorithm operations such as image preprocessing, feature extraction, target recognition, and spatial positioning. The memory is used to store image data, processing results, and related algorithm models, supporting the system's rapid response and retention of historical data. Through the coordinated work of the camera, light source, processor, and memory, the visual positioning device 600 can still achieve high-precision and stable image capture and positioning recognition functions under different lighting conditions and complex backgrounds, significantly improving the overall positioning efficiency and reliability of the system.
[0098] The light source assembly 630 includes a filter 631 disposed below the recognition assembly 620 and a light source 632 disposed on one side of the filter 631. The function of the filter 631 is to filter out unwanted light, making the light passing through purer, while the light source 632 provides stable illumination for the visual positioning device 600, ensuring that the visual positioning device 600 can work normally in various working environments.
[0099] The light source 632 is used to illuminate the substrate on the carrier component. The light source component 630 as a whole also serves to shield the carrier component 330 from ambient light and reduce interference.
[0100] The filter 631 is typically made of optical glass, resin or other special optical materials and is coated to selectively transmit or block light of a specific wavelength.
[0101] The filter 631 can filter out ambient light interference, allowing only specific wavelengths of light to pass through, thereby enhancing the recognition component 620's ability to recognize specific colors or markings and improving detection accuracy. By selecting a suitable filter 631, target information (such as QR codes, characters, or specific logos) can be made clearer in the image, improving the recognition capability of the vision system. The filter 631 can remove unnecessary background light and stray light, improving the stability of the image acquired by the recognition component 620 and making the reading of printed information more accurate. The filter 631 is used in conjunction with the light source component 630. By filtering out ambient interference light and allowing only specific wavelengths of light to pass through, it helps optimize recognition conditions and ensures that the recognition component 620 works stably and accurately under different ambient lighting conditions.
[0102] Furthermore, please see Figure 1 and Figure 3 Due to the size requirements of the frame 100, the coverage area of the first moving component 200 may not be sufficient to completely cover the printing area on the frame 100; therefore, a first extension plate 110 and a second extension plate 120 are provided on both sides of the frame 100. The printing area on the frame 100 is located between the first extension plate 110 and the second extension plate 120; and the height of the first extension plate 110 and the second extension plate 120 is higher than the height of the printing area.
[0103] The first moving component 200 is disposed on the first expansion plate 110 and the second expansion plate 120. Since the first expansion plate 110 and the second expansion plate 120 are disposed at a height higher than the height of the printing area, the printing component 500 disposed on the first displacement component can have a safe printing interval distance with the printing area.
[0104] Furthermore, please see Figure 3 The first moving component 200 includes a guide rail frame 210 and a slide groove 211 disposed on the guide rail frame 210. First limiting members 212 are provided on both sides of the guide rail frame 210. A first set of guide rails 213 are provided on both sides of the slide groove 211. A movable seat 215 is provided on the first set of guide rails 213 and slidably connected to the first set of guide rails 213 via a slider 214. A mover 216 is provided on the movable seat 215 and is fixedly connected to the movable seat 215, allowing the mover 216 to move synchronously with the movable seat 215. The first moving component 200 drives the mover 216 to move within the slide groove 211 via a moving cylinder, thereby causing the movable seat 215 to move on the first set of guide rails 213. The printing component 500 is disposed on the movable seat 215.
[0105] The movable cylinder is disposed on one side of the guide rail frame 210, and the piston rod of the movable cylinder is fixedly connected to one side of the mover 216. When the piston rod of the movable cylinder extends or retracts, it drives the mover 216 to move away from or towards the movable cylinder, thereby driving the movable seat 215 to move, so that the printing assembly 500 can move on the first movable assembly 200 to achieve the action of moving closer to or away from the printing area.
[0106] The first limiting member 212 is used to limit the extreme position of the movable seat 215. The first limiting member 212 is a buffer limiter. When the movable seat 215 slides to its extreme position on the first set of guide rails 213, it will contact the buffer limiter, thereby preventing excessive movement of the movable seat 215 that could cause equipment damage or malfunction. The design of the buffer limiter can effectively absorb the impact force of the movable seat 215 at its extreme position, protecting the stable operation of the equipment. At the same time, the setting of the buffer limiter also improves the safety and reliability of the equipment.
[0107] Furthermore, please see Figure 2 The second moving component 300 includes a second set of guide rails 310 disposed on both sides of the printing area on the frame 100 and a printing platform 320 that moves based on the second set of guide rails 310; the printing platform 320 is provided with positioning holes; the carrying component 330 is detachably disposed on the printing platform 320 through the positioning holes. The printing platform 320 is also provided with mounting holes, and at least one pair of positioning holes are provided; the carrying component 330 is provided with through holes adapted to the positioning holes and mounting holes for positioning and installation with the printing platform 320, and for fixing the carrying component 330 to the printing platform 320. The printing area of the frame 100 is also provided with a base frame.
[0108] The second moving component 300 is driven by a cylinder, which is located at one end of the base frame; the piston rod of the cylinder is fixedly connected to the printing platform 320.
[0109] When the piston rod of the cylinder extends or retracts, it drives the printing platform 320 to move away from or towards the end of the cylinder, so that the printing platform 320 can carry the carrier component 330 into the printing area to prepare for printing.
[0110] Furthermore, second limiting members 340 are provided on both sides of the printing area; the second limiting members 340 are used to limit the displacement range of the printing platform 320 on the frame 100. The second limiting members 340 are configured as stacked limiting members, forming a multi-layered limiting structure to ensure the stability and accuracy of the printing platform 320 during movement. The stacked limiting member design allows the printing platform 320 to be gradually obstructed when approaching its limit position, thereby avoiding sudden, violent impacts and protecting the equipment from damage.
[0111] Furthermore, please see Figure 1 , Figure 2 and Figure 5The first moving component 200 is provided with a certain safe distance from the printing area, but this safe distance cannot be adjusted and can only print on substrates whose height does not exceed this safe distance; in order to enable the printing equipment to adapt to substrates of more heights and to enable printing to be carried out safely, a third moving component 400 is also provided on the frame 100.
[0112] The third moving component 400 is disposed on both sides of the frame 100. The first expansion plate 110 and the second expansion plate 120 are disposed on the third moving component 400 and can be raised and lowered based on the third moving component 400 to adjust the distance between the printing component 500 and the printing area, so as to cover substrates of more heights.
[0113] Furthermore, the third moving assembly 400 is designed to include lifting sections 410 disposed on both sides of the frame 100. These lifting sections 410 are key mechanical structures responsible for vertical movement. To ensure smooth and precise lifting movements of the lifting sections 410, a first extension plate 110 and a second extension plate 120 are respectively disposed on the lifting sections 410 on both sides of the frame 100. This design allows the first extension plate 110 and the second extension plate 120 to move accordingly with the lifting sections 410, thereby achieving the desired functional effect. In addition, to ensure coordinated movement of the lifting sections 410 on both sides of the frame 100, the lifting sections 410 are connected to the connecting shaft 450 via a coupling 440. The coupling 440 is a mechanical device capable of transmitting torque, ensuring that when the lifting section 410 on one side of the frame 100 moves, the lifting section 410 on the other side of the frame 100 can move synchronously. This synchronous movement mechanism is crucial for maintaining the stability and balance of the entire frame 100 structure, ensuring the smooth operation of the entire mechanical system.
[0114] Further, please see Figure 3 The lifting mechanism includes a worm gear transmission seat 420 and a screw 421 driven by the worm gear transmission seat 420. The lifting mechanism adopts a worm gear structure, which not only has a large load-bearing capacity but also maintains smooth operation and high precision. Furthermore, this design is maintenance-free, significantly reducing subsequent maintenance costs and workload.
[0115] Furthermore, the worm gear transmission seat 420 not only includes the basic worm gear structure, but also has a driven wheel 422 to ensure stability and efficiency during transmission. The screw 421 is fixedly connected to the lifting frame 430, which itself is composed of a connecting plate 431 and a first guide rod 432 and a second guide rod 433 on both sides of the connecting plate 431. This design makes the structure of the lifting frame 430 more stable and also facilitates precise lifting control. The movement of the first guide rod 432 and the second guide rod 433 is based on the guide sleeve 434 set on the frame 100. This design not only ensures the smooth movement of the guide rods, but also improves the durability and reliability of the entire lifting mechanism. The frame 100 is also specially equipped with a drive motor 140. The output shaft of the drive motor 140 is equipped with a drive wheel 141, which is connected to the driven wheel 422 on the worm gear transmission seat 420 via a belt, thereby realizing the effective transmission of motor power and ensuring the power source of the lifting mechanism.
[0116] At the other end of the lifting frame 430, the first guide rod 432, the second guide rod 433, and the other end of the screw 421 are all provided with lifting plates 435. The design of these lifting plates 435 enables the lifting frame 430 to bear more load and remain stable during the lifting process. The first extension plate 110 and the second extension plate 120 are respectively provided on the lifting plates 435 of the lifting parts 410 on both sides of the frame 100. They can move up and down based on the lifting plates 435. This extension design enables the entire lifting mechanism to adapt to the working requirements of different heights and sizes.
[0117] In use, the motor is started, and the worm gear transmission seat 420 is driven through the transmission connection between the driving wheel 141, the belt and the driven wheel 422. This causes the screw 421 to move up or down based on the worm gear transmission seat 420, thereby driving the first guide rod 432 and the second guide rod 433 to move based on the guide sleeve 434 through the connecting plate 431, thereby causing the lifting plate 435 to move up and down.
[0118] When the printing device is in use, the substrate and its printing information are placed in the carrier container 331 and the information container 332, respectively. The third moving component 400 adjusts the printing component 500 to a suitable and safe printing interval distance from the printing area. The carrier component 330 is then fixed onto the printing platform 320 through positioning holes and mounting holes. The second moving component 300 drives the printing platform 320 to the preset printing area. The first moving component 200 is then activated to drive the printing component 500 to move. At this time, the visual positioning device 600 first identifies the information carried on the carrier component 330 and feeds it back to the host computer of the printing component 500 to generate a print file. The printing carriage of the printing component 500 prints the substrate in the carrier component 330.
[0119] Example 2
[0120] Please see Figure 1 Embodiment 2 of the present invention provides a printing device that uses visual positioning of the substrate. Based on the printing device of Embodiment 1, it further includes a maintenance station 700 for collecting waste ink. During inkjet printing, the printhead may become clogged or dry, requiring regular printhead cleaning. The maintenance station 700 provides a dedicated area for printhead cleaning, ink flushing, and hydration to ensure stable print quality. The health of the printhead directly affects print quality; regular cleaning and maintenance can prevent ink from drying and clogging the nozzles, extend printhead lifespan, and reduce maintenance costs.
[0121] Furthermore, if waste ink is discharged directly into the printing area or inside the machine, it may contaminate the substrate or affect the normal operation of the equipment. Maintenance station 700 centrally collects waste, helping to keep the equipment clean and prevent ink leakage.
[0122] Because printed patterns vary, the printhead may need to release a certain amount of residual ink before color switching or printing to ensure that the newly printed colors are pure and uniform. Maintenance station 700 provides this function to prevent color mixing.
[0123] Furthermore, centralized collection of waste ink facilitates its disposal, reduces the environmental impact of indiscriminate waste ink discharge, and meets environmental protection requirements, thereby improving the sustainability of the equipment.
[0124] The maintenance station 700 is located on the other side of the printing component 500.
[0125] Example 3
[0126] Please see Figure 1 and Figure 6Embodiment 3 of the present invention provides a printing device that uses visual positioning of the substrate. Based on the printing device of Embodiment 1 or Embodiment 2, it further includes: a safety component 800; the safety component 800 includes a first link 821 and a second link 822 equipped with a sensor 810, and a protective rod 823 connecting the first link 821 and the second link 822.
[0127] At least one security component 800 is provided, located on the same side of the printing component 500 as the visual recognition component 620.
[0128] Furthermore, the safety components 800 are preferably configured as a pair, respectively located at both ends of the printing component 500.
[0129] When the protective rod 823 comes into contact with any structure on the printing equipment, the sensor 810 is activated and sends a signal, causing the printing carriage to stop moving and preventing it from colliding with other structures on the printing equipment, causing damage or affecting print quality. The sensor 810 can be a contact sensor. The protective rod 823 can be made of an elastic material, providing a certain degree of cushioning to reduce impact force in the event of a collision, protecting the printing equipment and the safety component 800 itself.
[0130] Example 4
[0131] Please see Figure 7 This invention provides a visual positioning method for a printing device based on any one of embodiments 1 to 3, for visually positioning a printing device as described above. The method includes:
[0132] The substrate is placed inside the carrier container, and the information label is affixed to the carrier component;
[0133] Specifically, the carrier container 331 is used to hold the printing substrate, and there is at least one such container on the carrier assembly 330. During printing, at least one printing substrate and its corresponding printing information are printed in the carrier container 331 on the carrier assembly 330.
[0134] When the visual positioning device 600 detects that the carrier container 331 has at least one printing substrate and the corresponding information container 332 has printing information, the visual positioning device 600 transmits the information to the host computer and controls the printing carriage to print the printing substrate according to the information identified by the visual positioning device 600.
[0135] When the visual positioning device 600 detects that none of the carrier containers 331 have any printing material, the visual positioning device 600 transmits the information to the host computer and controls the printing carriage to reset.
[0136] When the visual positioning device 600 detects that there is no printing information in all the information containers 332, the visual positioning device 600 transmits the information to the host computer and controls the printing carriage to reset.
[0137] When the visual positioning device 600 detects that only one of the carrier containers 331 contains a printing substrate and the corresponding information container 332 of the carrier container 331 does not contain printing information, the visual positioning device 600 transmits the information to the host computer and controls the printing carriage to reset.
[0138] When the visual positioning device 600 detects that only one of the information containers 332 contains printing information and the corresponding carrier container 331 does not contain any printing material, the visual positioning device 600 transmits the information to the host computer to control the printing carriage to reset.
[0139] Furthermore, the printed information in the information container 332 can be stored using QR codes, barcodes, RFID tags, text / number identifiers, and graphic markers, and corresponding information tags can be generated. The information tags are then placed in the information container 332 to facilitate identification by the visual positioning device 600.
[0140] These diverse information carriers make inputting printed information more flexible and convenient, meeting the needs of different application scenarios. For example, QR codes and barcodes, due to their ability to quickly identify and store large amounts of information, are particularly suitable for scenarios requiring efficient tracking and management of printed materials. RFID tags, with their contactless reading characteristics, can complete information reading without interfering with the placement of the printed material, improving the automation level of the printing process. Text / number labels and graphic markers are more suitable for occasions requiring intuitive information display, enabling operators to quickly understand the printed content.
[0141] The second moving component drives the printing platform to the preset printing position;
[0142] The printing component is driven by the first moving component to move towards the preset printing position;
[0143] The spatial location information of the substrate and the printing information in the information label are obtained through the visual positioning device.
[0144] The location information and the printing information are merged to generate a print file, which is then transmitted to the printing carriage for printing.
[0145] Specifically, the location information is obtained by scanning the carrier container 331 where the substrate is located using the visual positioning device 600, and obtaining its actual coordinate information on the printing platform. This process can employ technologies such as high-precision cameras, laser sensors 810, or structured light scanning to ensure accurate positioning of the substrate within the printing area. If a positional deviation of the substrate is detected, the host computer can calculate the offset and automatically compensate during the printing process. Simultaneously, the visual positioning device 600 scans the information container 332 corresponding to the substrate, reading the printing information stored in QR codes, barcodes, RFID tags, or other identification methods (information tags). This printing information includes: the pattern or text data to be printed; the printing format (such as resolution, color mode, etc.); specific anti-counterfeiting marks or batch numbers; information processing and printing file generation; the read location information and printing information are transmitted to the host computer (control system), which processes the data and integrates the pattern, text, or coded information into a file format that the printing carriage can execute (such as specific printing instructions). The host computer determines whether the carrier container 331 and the information container 332 correspond. It compares the spatial position information of the substrate (the carrier container 331's number) with the printing information in the information container 332 to ensure a one-to-one correspondence. Error handling: If there is substrate in the carrier container 331 but no printing information in the information container 332, the host computer controls the printing carriage to stop printing, displays an error message, and resets. If there is printing information in the information container 332 but no substrate in the carrier container 331, the host computer controls the printing carriage to stop printing, displays an error message, and resets. If a mismatch is detected between the carrier container 331 and the information container 332 (e.g., inconsistent numbers), the printing assembly 500 is reset, and the operator is prompted to check. Printing execution: Under the control of the host computer, the printing carriage moves above the corresponding substrate and performs precise alignment based on the previously acquired position information. The host computer or printer's processor merges the position information and the printing information to generate a print file. Inkjet printing or spraying is then performed according to the generated print file to complete the printing of the pattern on the substrate. During the printing process, the vision positioning device 600 continuously monitors the printing status, detects whether there are problems such as misalignment, ink shortage or deviation, and provides real-time feedback to the host computer to adjust the printing parameters or make corrections.
[0146] In one embodiment, the print file includes: information on the arrangement of the substrate in the carrier assembly. Since the placement of the substrate on the carrier assembly may have errors, the visual positioning device 600 needs to detect its specific position and feed the position information back to the host computer to ensure that the printing carriage can accurately align with the target area for printing. Different batches of carrier assemblies may have different substrate arrangements; this information can be used to dynamically adjust the printing path. Orientation information is also included to prevent printing deviations caused by incorrect substrate rotation, thus improving the pass rate. Some substrates are directional (such as printed labels, trademarks, QR codes, etc.). If the orientation is incorrect, the printed content may be inverted or the position may be inaccurate. When the visual positioning device 600 detects an orientation deviation, the host computer can automatically adjust the printing direction without manual intervention. Type information is also included, as different types of substrates (such as plastics, metals, paper, fabrics, etc.) have different ink absorption, adhesion, and curing methods, requiring the selection of appropriate printing parameters. This allows for adjustment of inkjet concentration based on material type, reducing ink waste and improving print quality. Size information is crucial because substrates from different batches may vary in size. The printing system needs to adapt to these dimensional changes to avoid printing beyond the boundaries or offset. If the substrate size is too small or exceeds the printing area, the system can automatically alert the system to prevent waste. Shape information adapts to substrates of different shapes (such as rectangles, circles, irregularly shaped workpieces, etc.) to prevent printing deviations caused by shape differences. This can be applied to personalized customization printing, such as electronic product casings, bottle caps, irregularly shaped metal plates, etc. Printing information includes pattern information, containing specific printing content such as graphics, text, barcodes, QR codes, etc. Color information ensures that the inkjet system prints according to preset colors; different media may require different color modes (CMYK / RGB). Printing path information determines the movement trajectory of the printing carriage, optimizes the inkjet sequence, reduces repetitive movements, and improves efficiency.
[0147] In one embodiment, obtaining the spatial location information of the substrate and the printing information in the information identifier through the visual positioning device includes:
[0148] Acquire an image of the substrate, and perform illumination compensation processing on the image of the substrate to obtain the target image;
[0149] Specifically, images of the substrate (such as a phone case) in the current working area are acquired through a visual positioning device. These are raw images, which are affected by factors such as ambient light changes, shadows, and reflections, potentially leading to inconsistent image contrast, color distortion, or blurred information in local areas. To improve the accuracy of subsequent image analysis, image enhancement networks based on histogram equalization, the Retinex algorithm, or deep learning can be used to perform illumination compensation processing, unifying the image's brightness and color distribution, ultimately outputting the target image.
[0150] The enhanced target image has higher clarity and consistency, significantly improving the accuracy of edge detection, segmentation and recognition, and is especially suitable for industrial environments with local occlusion and uneven lighting.
[0151] Edge detection is performed on the target image to obtain the contour feature information of the substrate;
[0152] Specifically, edge detection algorithms (such as Canny, Sobel, Laplacian, or deep edge learning models) are used to extract the edges of the substrate in the target image, generating a clear contour feature map. This step can be combined with a dynamic thresholding strategy to adaptively detect edge information of substrates with different shapes and materials (such as glossy and matte surfaces). High-quality edge extraction provides an accurate feature foundation for subsequent contour geometry fitting, while avoiding the problems of unstable identification of background interference and reflective areas in traditional edge extraction, thus improving the system's robustness in recognizing substrates under various complex backgrounds.
[0153] Geometric fitting is performed on the contour feature information to determine the boundary shape of the substrate;
[0154] Specifically, the extracted contour feature points are input into fitting algorithms, such as RANSAC line / polygon fitting, B-spline curve fitting, or deep learning-based boundary regression models, to achieve standardized reconstruction of the substrate edge shape. For irregular contours (such as customized phone cases), semantic segmentation auxiliary information can be added to enhance the fitting's adaptability to complex shapes. This achieves accurate boundary reconstruction and a unified standard format representation, providing a geometric basis for spatial pose calculation, image alignment, and automatic cropping of the pattern printing area, ensuring accurate alignment in subsequent printing operations.
[0155] Based on the boundary shape, and combined with a preset perspective transformation and posture calculation algorithm, the spatial position information of the substrate is determined, wherein the spatial position information includes position coordinates and posture angles.
[0156] Specifically, by matching boundary coordinate points with a preset reference model, the two-dimensional planar coordinates of the current substrate in image space are calculated. Furthermore, based on multi-point perspective transformation (such as monocular PnP, IPPE, or depth pose estimation networks), its actual pose in three-dimensional space, including X / Y / Z coordinates and rotation angles, is recovered. This solves the problem of positional offset and tilt caused by manual placement, achieves precise calibration of the printing path, avoids problems such as printed pattern offset and uneven rotation, and improves the quality and consistency of the finished product.
[0157] The target image is input into a pre-trained identifier segmentation model to determine the identifier region;
[0158] Specifically, semantic segmentation models (such as DeepLabv3, U-Net, SegFormer, etc.) or specially trained multi-task models are used to identify the location contours of marked regions such as barcodes / QR codes / image numbers in images. To improve the recognition accuracy in complex scenes, attention mechanisms and CRF post-processing can be introduced to eliminate non-marking interference elements. This allows for the rapid and stable extraction of barcode and other recognition regions, providing the location input for subsequent information extraction and avoiding recognition failures caused by non-standard pasting positions or contamination of the recognition region.
[0159] The identification information of the identified area is extracted, and the printing information is determined based on the extracted identification information. The printing information includes the pattern to be printed, the color to be printed, and the path to be printed.
[0160] Specifically, image OCR technology (such as PaddleOCR and EasyOCR) is used to extract text number information from the identification area, or multimodal recognition models (such as BLIP and LayoutLMv3) are used to perform image-text fusion recognition, identifying the image number or product type information from the identification. By querying the printing database or the corresponding relationship stored in the host computer, specific data such as the printing pattern, printing color, layer order, and printhead path planning corresponding to the identification are obtained. Through a closed-loop process from identification to printing task data generation, the automation and accuracy of printing tasks are improved, significantly reducing manual intervention and error rates, and meeting the needs for high-efficiency personalized printing.
[0161] In one embodiment, the step of geometrically fitting the contour feature information to determine the boundary shape of the substrate includes:
[0162] Based on the contour feature information, an initial edge image corresponding to the target image is obtained;
[0163] Specifically, after acquiring the target image of the substrate, an initial edge image is generated based on the contour feature information extracted in the previous step (e.g., the edge point set obtained through operators such as Canny and Sobel). This image is a binary image, with the edges highlighted and the remaining areas as background, which can roughly outline the outer contour and some detailed structures of the substrate. Gradient direction information can also be added to construct a more robust edge image representation, enhancing shape perception capabilities. This initial edge image serves as a structural guide for subsequent semantic segmentation, improving the segmentation model's ability to perceive boundary positions, avoiding common edge blurring or offset problems in semantic segmentation, and enhancing the overall accuracy of boundary extraction.
[0164] The initial edge image is input into a pre-trained semantic segmentation model to obtain the semantic boundary mask of the substrate, wherein the semantic segmentation model is used to distinguish the substrate region and the background region in the image;
[0165] Specifically, the initial edge image and the target image are input together into a pre-trained semantic segmentation model (such as U-Net). The model's ability to learn the semantic structure of the image is used to output a semantic boundary mask, which accurately separates the printing area from the background area. This mask is a pixel-level classification result, with each pixel labeled as "printed material" or "background," which has a significant advantage, especially when dealing with irregular shapes or occluded media.
[0166] By introducing high-precision semantic segmentation, the problems of traditional edge detection being sensitive to texture interference and having poor robustness to changes in illumination can be effectively avoided. The printing area can be identified from a global semantic perspective, improving the integrity and stability of boundary detection.
[0167] The initial edge image and the semantic boundary mask are subjected to boundary fusion processing to obtain the target boundary image;
[0168] Specifically, the initial edge image and the semantic boundary mask undergo boundary fusion processing. This fusion processing can be achieved through boundary overlap calculation between the edge image and the semantic mask, edge thinning, image morphological operations (such as erosion, dilation, and closing operations), or boundary reconstruction neural networks. This process combines the high-resolution structural information of traditional edge detection with the region discrimination capability of the semantic mask, thereby forming a smoother, more continuous, and semantically consistent target boundary image. Boundary fusion can remove isolated interference lines, fill in missing edges, and improve the coherence and geometric accuracy of the boundary contour, laying a high-quality image foundation for subsequent accurate contour point extraction and printing alignment, which is particularly suitable for industrial-grade high-precision printing scenarios.
[0169] Contour points are detected and extracted from the target boundary image, and the extracted contour point set is filtered to obtain a filtered contour point set, wherein the filtering includes removing outliers and repairing break points;
[0170] Specifically, contour tracking algorithms (such as the Suzuki-Abe algorithm) are used to extract contour points from closed or open boundaries in the target boundary image, generating a contour point set. Point cloud clustering, minimum distance filters, or topology-based rules are then applied to this set to remove noisy outliers and breakpoints. Breakpoints can be repaired using interpolation methods, shortest path search, or graph-based connection strategies to obtain a complete and smooth contour point sequence. This step ensures that the contour point set has good connectivity and accuracy, providing highly reliable input for boundary fitting, pose estimation, and subsequent printing path planning. It significantly enhances the system's robustness and recognition accuracy, especially under complex backgrounds or image defects.
[0171] The selected set of contour points is subjected to geometric fitting to determine the boundary shape.
[0172] In one embodiment, the step of extracting identification information from the identification area and determining the printing information based on the extracted identification information includes:
[0173] A labeled region image is obtained, and the labeled region image is input into a visual coding model to determine the structural feature vector of the image modality. The visual coding model is an image coding network built based on a convolutional neural network.
[0174] Specifically, after identifying the identification region (such as a printed label, QR code border, stamp pattern, etc.), this region is cropped from the target image and used as the input image for the identification region. This image is then input into a visual encoding model based on a convolutional neural network (CNN) (such as ResNet, EfficientNet, or MobileNet) to extract its structural feature vectors. These vectors preserve key information such as the geometric shape, texture structure, and color distribution in the image and are mapped into a multi-dimensional feature space as a structural semantic representation of the image modality.
[0175] This structural feature vector can be used for subsequent cross-modal alignment with text information, and can also support tasks such as fuzzy pattern matching and abnormal identifier detection. It provides a basis for recognizing non-textual printed information (such as graphic logos, stripes, and icons) in images, enhancing the comprehensiveness of overall identifier understanding.
[0176] Based on an end-to-end character detection and recognition network, text recognition is performed on the image of the identified area to obtain text content information;
[0177] Specifically, the labeled region image is input into an end-to-end text recognition network, such as CRNN, Transformer-based text recognition models (e.g., TRBA, ABINet), or open-source systems like PaddleOCR and PP-OCRv3. These networks typically include character detection (e.g., CTPN, EAST) and recognition modules (e.g., LSTM+CTC, Attention decoder), which can directly extract printed characters or handwritten text from the image and output complete string information. The end-to-end architecture eliminates the multi-stage error accumulation problem in traditional OCR processes, improving recognition accuracy and robustness, and is particularly suitable for handling labeled image content with font distortion, tilted angles, or complex backgrounds. Simultaneously, this textual content information forms the basis for subsequent semantic understanding and task-driven tasks, such as automatic classification, path planning, and instruction recognition.
[0178] The text content information is input into the text semantic encoder to extract the text vector of the text modality. The text semantic encoder is a pre-trained language model.
[0179] Specifically, the identified text strings are input into pre-trained language models (such as BERT, RoBERTa, ERNIE, DistilBERT, etc.) for encoding, extracting text vectors of the text modality. These models, pre-trained on large-scale corpora, can understand the grammatical structures and semantic relationships inherent in the text, outputting high-dimensional text feature vectors with context awareness. These vectors not only represent the literal content of the text but also contain its latent semantics, category attributes, or instructional meanings. After structuring the text content into semantic vectors, they can be aligned and fused with image modality vectors, further supporting cross-modal retrieval, printing task classification, personalized pattern recommendation, and other functions, improving the intelligent recognition system's ability to understand complex printed content and its task adaptability.
[0180] The structural feature vector and the text vector are subjected to multimodal fusion processing to obtain a fused feature vector. The fusion processing includes a cross-modal attention mechanism or a feature splicing fusion strategy.
[0181] Specifically, after obtaining the structural feature vectors of the image modality and the semantic feature vectors of the text modality, the system employs a multimodal fusion strategy to integrate feature information from these two different sources. One approach is feature concatenation, where image and text features are directly concatenated dimensionally and then fed into the subsequent model. A more efficient approach is to use a cross-modal attention mechanism, such as a Transformer structure, to fuse image and text feature vectors, enabling the model to learn to capture semantic relationships between images and text. The fused feature vectors can more comprehensively represent the multimodal information of the identified region, including its appearance features and semantic content. This fusion method can overcome the ambiguity and information gaps that may arise from a single modality, improving the robustness and accuracy of the model when processing complex or ambiguous identifiers. For example, when the image content is blurry but the text is clear, or the text is irregular but the image is clear, multimodal fusion can make a comprehensive judgment, significantly improving recognition stability and the accuracy of subsequent tasks.
[0182] The printing information is determined based on the fused feature vector and a preset multi-label classifier.
[0183] Specifically, the fused feature vector is input into a pre-trained multi-label classifier. This classifier can be based on a fully connected neural network structure or use a multi-label discriminant head with a sigmoid activation function (rather than a softmax single-label discriminant). The classifier outputs multiple label fields describing the specific information required for the printing task, including but not limited to: the pattern number to be printed, the required color configuration, the printing layer order, special inkjet channel information, and the printing path planning type. This step infers the complete printing parameters or instruction set from the fused semantics. The advantage of using a multi-label classification model is that it can output information from multiple dimensions simultaneously, meeting the multi-parameter configuration requirements of complex printing tasks and greatly improving automation and customization capabilities. This approach is suitable for large-scale customized industrial production scenarios, automatically parsing the printing requirements behind graphic markings, reducing manual intervention, and improving efficiency and accuracy.
[0184] In summary, the embodiments of the present invention provide a printing device and a visual positioning method for a printing substrate.
[0185] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.
[0186] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.
[0187] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant locality, and corresponding operation entry points shall be provided for the user to choose to authorize or refuse.
[0188] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.
[0189] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.
Claims
1. A printing apparatus employing visual positioning of a substrate, comprising a frame and a first moving component and a second moving component disposed on the frame, characterized in that: The first movable component is provided with a printing component that can move relative to the first movable component, and a visual positioning device is also provided on one side of the printing component; The second moving component is provided with a carrier component for carrying the printing substrate, and the carrier component can enter the printing area covered by the printing component through the second moving component; The carrier component is provided with an information identifier and at least one carrier container for accommodating the printing substrate. The information identifier includes printing information corresponding to the printing substrate medium. The visual positioning device is used to obtain the position information of the substrate and, based on the information identifier, obtain the printing information; The print file is transferred to the printing carriage where the printing component is located to realize printing, wherein the print file is generated from the printing information and the location information.
2. The printing equipment using visual positioning of the substrate according to claim 1, characterized in that, The visual positioning device is mounted on one side of the printing carriage via a mounting bracket; The visual positioning device includes a recognition component and a light source component spaced apart on top of the recognition component; The light source assembly includes a filter disposed below the recognition assembly and a light source disposed on one side of the filter; The frame is provided with a first expansion plate and a second expansion plate on both sides; It also includes a printing area disposed between the first expansion plate and the second expansion plate; The first moving component is mounted on the frame via the first expansion plate and the second expansion plate, and the printing component is provided with a printing interval distance from the printing area.
3. The printing equipment using visual positioning of the substrate according to claim 1, characterized in that, The first moving component includes a guide rail frame and a slide groove disposed on the guide rail frame, and a first limiting member is provided on both sides of the guide rail frame; The guide rail frame is provided with a first set of guide rails on both sides of the slide groove; The first set of guide rails is equipped with a movable seat that is connected to the first set of guide rails via a slider; The movable base is equipped with a moving part; The first moving component drives the moving element to move within the slide groove via a moving cylinder, thereby causing the moving seat to move on the first set of guide rails; The printing component is mounted on the movable base.
4. The printing equipment using visual positioning of the substrate according to claim 1, characterized in that, The second moving component includes a second set of guide rails disposed on both sides of the printing area on the frame and a printing platform that moves based on the second set of guide rails; The printing platform is equipped with positioning holes; The carrier component is detachably mounted on the printing platform through the positioning hole.
5. The printing equipment using visual positioning of the substrate according to claim 2, characterized in that, The frame is also equipped with a third moving component; The third moving component includes lifting parts disposed on both sides of the frame; the lifting parts on both sides of the frame are connected to the connecting shaft via couplings.
6. The printing equipment using visual positioning of the substrate according to claim 5, characterized in that, The lifting unit includes a worm gear transmission seat and a screw driven by the worm gear transmission seat; The worm gear transmission base is also provided with a driven wheel; The screw is connected to a lifting frame, which includes a connecting plate and a first guide rod and a second guide rod disposed on both sides of the connecting plate. The first guide rod and the second guide rod move based on the guide sleeve disposed on the frame; The frame is also equipped with a drive motor, and the output shaft of the drive motor is equipped with a drive pulley; the drive pulley is connected to the driven pulley via a belt. The first guide rod, the second guide rod, and the other end of the screw are equipped with lifting plates; The first extension plate and the second extension plate are respectively disposed on the lifting plates of the lifting parts on both sides of the frame, and move up and down based on the lifting plates.
7. A visual positioning method for a printing device, characterized in that, The method for visually positioning a printing device as described in any one of claims 1-6 includes: The substrate is placed inside the carrier container, and the information label is affixed to the carrier component; The second moving component drives the printing platform to the preset printing position; The printing component is driven by the first moving component to move towards the preset printing position; The spatial location information of the substrate and the printing information in the information label are obtained through the visual positioning device. The location information and the printing information are merged to generate a print file, which is then transmitted to the printing carriage for printing.
8. The visual positioning method for a printing device according to claim 7, characterized in that, The step of obtaining the spatial location information of the substrate and the printing information in the information identifier through the visual positioning device includes: Acquire an image of the substrate, and perform illumination compensation processing on the image of the substrate to obtain the target image; Edge detection is performed on the target image to obtain the contour feature information of the substrate; Geometric fitting is performed on the contour feature information to determine the boundary shape of the substrate; Based on the boundary shape, and combined with a preset perspective transformation and posture calculation algorithm, the spatial position information of the substrate is determined, wherein the spatial position information includes position coordinates and posture angles. The target image is input into a pre-trained identifier segmentation model to determine the identifier region; The identification information of the identified area is extracted, and the printing information is determined based on the extracted identification information. The printing information includes the pattern to be printed, the color to be printed, and the path to be printed.
9. The visual positioning method for a printing device according to claim 8, characterized in that, The step of geometrically fitting the contour feature information to determine the boundary shape of the substrate includes: Based on the contour feature information, an initial edge image corresponding to the target image is obtained; The initial edge image is input into a pre-trained semantic segmentation model to obtain the semantic boundary mask of the substrate, wherein the semantic segmentation model is used to distinguish the substrate region and the background region in the image; The initial edge image and the semantic boundary mask are subjected to boundary fusion processing to obtain the target boundary image; Contour points are detected and extracted from the target boundary image, and the extracted contour point set is filtered to obtain a filtered contour point set, wherein the filtering includes removing outliers and repairing break points; The selected set of contour points is subjected to geometric fitting to determine the boundary shape.
10. The visual positioning method for a printing device according to claim 8, characterized in that, The step of extracting identification information from the identified area and determining the printing information based on the extracted identification information includes: A labeled region image is obtained, and the labeled region image is input into a visual coding model to determine the structural feature vector of the image modality. The visual coding model is an image coding network built based on a convolutional neural network. Based on an end-to-end character detection and recognition network, text recognition is performed on the image of the identified area to obtain text content information; The text content information is input into the text semantic encoder to extract the text vector of the text modality. The text semantic encoder is a pre-trained language model. The structural feature vector and the text vector are subjected to multimodal fusion processing to obtain a fused feature vector. The fusion processing includes a cross-modal attention mechanism or a feature splicing fusion strategy. The printing information is determined based on the fused feature vector and a preset multi-label classifier.
Citation Information
Patent Citations
An IoT mobile phone protective case image printing system
CN107020832B
