Paperboard flattening control method, system, and paperboard flattening device based on machine vision

Abstract

The application discloses a paperboard flattening control method and system based on machine vision and a paperboard flattening device, and relates to the technical field of machine vision. The paperboard flattening control method based on machine vision acquires image data of paperboard to be flattened, and establishes a three-dimensional model of the paperboard to be flattened based on the image data; surface feature data of the paperboard to be flattened is acquired, and the surface temperature of a flattening plate is adjusted based on the surface feature data so that the surface humidity of the paperboard to be flattened reaches a humidity before flattening; the feature data of the paperboard to be flattened before flattening is acquired based on the three-dimensional model, and the extrusion force of the flattening plate on the paperboard to be flattened after the surface humidity of the paperboard to be flattened reaches the humidity before flattening is determined based on the feature data, thereby solving the problem that the prior art cannot understand the physical state and deformation degree of the paperboard based on machine vision, and achieving more accurate processing.

Description

Technical Field

[0001] This invention relates to the field of machine vision technology, specifically to a machine vision-based method, system, and device for controlling cardboard flattening. Background Technology

[0002] During the manufacturing process of packaging cardboard, the cardboard needs to be coated with glue and laminated. Before this coating and laminating process, a flattening operation is usually performed. Flattening ensures the quality of the glue and lamination, effectively preventing issues caused by cardboard warping that could affect the quality of the coating and lamination. Currently, cardboard flattening is generally achieved using a cardboard flattening machine.

[0003] Chinese invention patent CN113934190B discloses a machine vision-based method for quality control in corrugated cardboard production. The method includes: acquiring a grayscale image of a target surface; acquiring the grayscale gradient direction of each pixel; calculating the defect probability of each pixel; establishing a grayscale histogram, calculating the background probability value for each grayscale level, and acquiring the background region of the grayscale image; calculating the degree of abnormality of each pixel in the grayscale image, and constructing a sequence of all abnormality degrees in the grayscale image; calculating the influence value of each abnormality degree sequence to obtain the overall influence value of the target surface image; and adjusting production machinery parameters based on the overall influence value of the target image. According to the technical means proposed by this invention, the influence degree of defects is calculated through the grayscale image of the target surface, thereby adjusting the machinery parameters and improving product quality and production efficiency.

[0004] Existing cardboard flattening machines are all based on set parameters for flattening, but it is inconvenient to combine machine vision technology to identify and understand the physical state and degree of deformation of the cardboard, which makes it difficult to achieve more precise processing. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a cardboard flattening control method, system, and device based on machine vision, which solves the problem that existing technologies cannot easily understand the physical state and deformation degree of cardboard based on machine vision, thereby achieving more precise processing.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a machine vision-based cardboard flattening control method, comprising the following steps: acquiring image data of the cardboard to be flattened, and establishing a three-dimensional model of the cardboard to be flattened based on the image data; acquiring surface feature data of the cardboard to be flattened, and adjusting the surface temperature of the pressing plate based on the surface feature data of the cardboard to be flattened, so that the surface humidity of the cardboard to be flattened reaches the pre-flattening humidity; acquiring the pre-flattening feature data of the cardboard to be flattened based on the three-dimensional model of the cardboard to be flattened, and determining the extrusion force applied by the pressing plate to the cardboard to be flattened after the surface humidity of the cardboard to be flattened reaches the pre-flattening humidity based on the pre-flattening feature data.

[0007] Furthermore, the process of establishing a three-dimensional model of the cardboard to be flattened based on the image data of the cardboard to be flattened is as follows: Select a region from the image data of the cardboard to be flattened as a template, and search for the registration region most similar to the template in the images to be registered stored in the database; determine the registration transformation between the template and the images corresponding to the registration region; apply the registration transformation to the images to be registered and adjust the images to be registered; repeatedly select all regions in the image data of the cardboard to be flattened and adjust the images to be registered corresponding to each region; stitch all the adjusted images to be registered together to obtain the three-dimensional model of the cardboard to be flattened.

[0008] Further, the process of adjusting the surface temperature of the pressing plate based on the surface feature data of the cardboard to be flattened, so that the surface humidity of the cardboard to be flattened reaches the humidity before flattening, is as follows: Surface feature data of each detection area on the surface of the cardboard to be flattened are obtained. The surface feature data of each detection area includes the surface humidity and surface temperature of the detection area. The detection areas are divided according to the positions of the heating areas on the pressing plate surface. It is determined whether the surface humidity of each detection area is greater than a set humidity threshold. If so, the pressing plate surface temperature of the corresponding detection area is determined based on the surface feature data of the detection area with a surface humidity greater than the set humidity threshold, the dehumidification time, and the humidity before flattening, so that the surface humidity of the cardboard to be flattened reaches the humidity before flattening. Otherwise, the pressing plate surface temperature of the corresponding detection area is not adjusted.

[0009] Furthermore, if the surface humidity of all detection areas is greater than the set humidity threshold, then the humidity before flattening is the set humidity threshold; if the surface humidity of a detection area is less than or equal to the set humidity threshold, then the humidity before flattening is the average surface humidity of the detection areas of the detection areas whose surface humidity is less than or equal to the set humidity threshold.

[0010] Furthermore, the formula for calculating the surface temperature of the pressure plate is as follows:

[0011]

[0012] In the formula, T h T is the surface temperature of the pressure plate. n To detect the surface temperature of the area, H h To detect the surface humidity of the area, H f Q is the set humidity threshold. h Q is the amount of heat used for heating. e C is the heat used for evaporating water. p denoted as , where m is the specific heat capacity of the cardboard, m is the mass of the cardboard in the detection area, and ΔH is the latent heat of vaporization of water.

[0013] Furthermore, the process of determining the extrusion pressure applied to the cardboard by the pressing plate after the surface humidity of the cardboard to be flattened reaches the pre-flattening humidity, based on the pre-flattening feature data, is as follows: the surface temperature of the pressing plate is raised to the temperature at which the cardboard to be flattened is most prone to deformation; the pre-flattening feature data of the cardboard to be flattened is analyzed to obtain the flatness of the cardboard to be flattened and the height of the highest point of the cardboard to be flattened relative to the flattened plane; the auxiliary feature data of the cardboard to be flattened is obtained; and the extrusion pressure applied to the cardboard to be flattened is determined based on the auxiliary feature data of the cardboard to be flattened and the analyzed pre-flattening feature data.

[0014] Furthermore, the process of obtaining the flatness of the paperboard to be flattened is as follows: Several test points on the three-dimensional model of the paperboard to be flattened are randomly selected, and the coordinates of each test point are obtained; the flatness of the paperboard to be flattened is determined based on the distance between the selected coordinates of each test point and the plane on which the paperboard to be flattened is located after being flattened.

[0015] Furthermore, the process of determining the extrusion pressure applied to the paperboard to be flattened based on the attached feature data of the paperboard to be flattened and the parsed pre-flattening feature data is as follows: the pressure index is determined based on the attached feature data of the paperboard to be flattened and the parsed pre-flattening feature data; the pressure index is compared with the pressure index range stored in the database, and the extrusion pressure corresponding to the successfully matched pressure index range is obtained as the extrusion pressure applied to the paperboard to be flattened.

[0016] The machine vision-based cardboard flattening control system, used in the aforementioned machine vision-based cardboard flattening control method, includes a 3D modeling module, a pressing plate surface temperature adjustment module, and an extrusion pressure determination module. Specifically: the 3D modeling module acquires image data of the cardboard to be flattened and establishes a 3D model of the cardboard based on this image data; the pressing plate surface temperature adjustment module acquires surface feature data of the cardboard to be flattened and adjusts the pressing plate surface temperature based on this data to bring the surface humidity of the cardboard to the level before flattening; the extrusion pressure determination module acquires the pre-flattening feature data of the cardboard based on the 3D model and determines the extrusion pressure applied by the pressing plate to the cardboard after the surface humidity reaches the pre-flattening level.

[0017] The machine vision-based cardboard flattening device includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, cause the processor to perform the machine vision-based cardboard flattening control method described above. It also includes a flattening support and a pressing plate mounted on the flattening support. The pressing plate is fixedly connected to the flattening support via a lifting cylinder. Heating modules are fixedly connected to both the pressing plate and the flattening support, and humidity sensors are installed on the heating modules. A drive roller is mounted on one side of the flattening support via a drive motor. A lifting frame is fixedly connected to the flattening support via an adjusting cylinder, and a driven roller is movably connected inside the lifting frame.

[0018] The present invention has the following beneficial effects:

[0019] (1) The paperboard flattening control method based on machine vision uses machine vision to acquire image data of paperboard and establish a three-dimensional model, which can accurately understand the physical state and deformation degree of paperboard, thereby achieving more precise processing. By adjusting the temperature of the pressing plate surface to match the humidity of the paperboard before flattening, it can ensure that the paperboard maintains ideal humidity during processing and avoid quality problems caused by dryness or excessive moisture.

[0020] (2) The machine vision-based paperboard flattening control method automates the entire flattening process, reducing the need for manual intervention, improving production speed and continuity, and can quickly respond to changes in the paperboard state, adjusting the extrusion pressure and temperature settings in real time to cope with different batches or conditions of paperboard. By precisely controlling the processing, material defects and waste can be minimized, and material utilization can be improved. Precise control of the temperature of the pressing plate and the applied extrusion pressure can reduce energy consumption, especially in heating and pressure applications.

[0021] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0022] Figure 1 This is a flowchart of the cardboard flattening control method based on machine vision according to the present invention.

[0023] Figure 2 This is a function graph of the flatness of the cardboard to be flattened in the cardboard flattening control method based on machine vision of the present invention.

[0024] Figure 3 This is a function graph of the pressure index of the cardboard flattening control method based on machine vision according to the present invention.

[0025] Figure 4 This is a flowchart of the cardboard flattening control system based on machine vision according to the present invention.

[0026] Figure 5 This is a schematic diagram of the overall structure of the cardboard flattening device based on machine vision according to the present invention.

[0027] Figure 6 This is a schematic diagram of the pressing plate of the cardboard flattening device based on machine vision according to the present invention.

[0028] In the diagram, 1. Flattening support; 2. Flat plate; 3. Lifting cylinder; 4. Heating module; 5. Humidity sensor; 6. Drive motor; 7. Active roller; 8. Adjusting cylinder; 9. Lifting frame; 10. Driven roller. Detailed Implementation

[0029] The embodiments of this application, through a machine vision-based cardboard flattening control method, system, and flattening device, address the problem that existing technologies cannot easily understand the physical state and deformation degree of cardboard based on machine vision, thereby achieving more precise processing.

[0030] The problem in this application embodiment, the overall steps are as follows:

[0031] First, image data of the cardboard to be flattened is acquired using a machine vision system, and a 3D model of the cardboard is built based on this data. Simultaneously, surface feature data of the cardboard is acquired. Machine vision is a technology that enables computer systems to acquire, process, and understand image information through cameras, sensors, or other vision devices. It utilizes computer vision algorithms and models to allow computer systems to mimic the functions of the human visual system, such as recognizing objects, detecting motion, and analyzing image content. Based on the surface feature data of the cardboard to be flattened, especially humidity data, the surface temperature of the pressing plate is adjusted to control the surface humidity of the cardboard to an optimal state. Based on the pre-flattening feature data acquired from the 3D model of the cardboard, the changes before and after flattening, as well as the required compression force, are determined. Based on this data, the flattening parameters are optimized to ensure that the ideal flatness of the cardboard is achieved without over-flattening.

[0032] Please see Figure 1 The present invention provides a technical solution: a paperboard flattening control method based on machine vision, comprising the following steps: acquiring image data of the paperboard to be flattened, and establishing a three-dimensional model of the paperboard to be flattened based on the image data; acquiring surface feature data of the paperboard to be flattened, and adjusting the surface temperature of the pressing plate based on the surface feature data of the paperboard to be flattened, so that the surface humidity of the paperboard to be flattened reaches the humidity before flattening; acquiring the pre-flattening feature data of the paperboard to be flattened based on the three-dimensional model of the paperboard to be flattened, and determining the extrusion force applied by the pressing plate to the paperboard to be flattened after the surface humidity of the paperboard to be flattened reaches the pre-flattening humidity based on the pre-flattening feature data.

[0033] By using machine vision systems and 3D modeling, the flatness and warping height of the cardboard to be flattened can be analyzed, enabling precise control of the flattening process and ensuring product quality stability and consistency. By acquiring surface characteristic data of the cardboard and adjusting the surface temperature of the pressing plate based on this data, the surface humidity of the cardboard to be flattened can be optimized, thereby improving the flattening processing parameters and enhancing the processing quality and stability of the cardboard.

[0034] Specifically, the process of establishing a three-dimensional model of the cardboard to be flattened based on the image data of the cardboard to be flattened is as follows: Select a region from the image data of the cardboard to be flattened as a template, and search for the registration region most similar to the template in the images to be registered stored in the database; determine the registration transformation between the template and the images corresponding to the registration region; apply the registration transformation to the images to be registered and adjust the images to be registered; repeatedly select all regions in the image data of the cardboard to be flattened and adjust the images to be registered corresponding to each region; stitch together all the adjusted images to be registered to obtain the three-dimensional model of the cardboard to be flattened.

[0035] In this implementation scheme, firstly, a region is randomly selected from the image data of the cardboard to be flattened as a template. The template contains important features or structures of the cardboard to be flattened, such as curvature, size, and creases, to ensure the accuracy and reliability of subsequent registration. A set of images to be registered is stored in a database. For the selected template region, the most similar registration region is searched. This can be achieved using image feature matching algorithms, such as feature descriptor-based matching algorithms or deep learning-based feature extraction and matching methods.

[0036] The matching process using deep learning-based feature extraction and matching methods is as follows:

[0037] A pre-trained deep learning model, such as a convolutional neural network (CNN), like ResNet or VGG, is selected to extract features from the image. The image to be registered and the template image are input into this model to obtain their feature maps. These feature maps capture key information in the image, such as edges, corners, and other texture features. Feature matching algorithms (such as KD-Tree or FLANN) are used to quickly match the extracted feature points. RANSAC or learning-based matching optimization algorithms can be applied to improve the robustness and accuracy of the matching. The matching results are then examined to eliminate false matches and retain high-confidence matching pairs. Methods such as matching score thresholds can be used to determine which matches are reliable.

[0038] The registration transformation between the template region and the registration region is determined. This registration transformation includes translation, rotation, scaling, etc. The determination of the registration transformation is based on the corresponding points or features between the template and the registration region, and is solved using a registration algorithm. The solution process is as follows:

[0039] Choose an appropriate transformation model based on the application scenario, such as affine transformation or perspective transformation. Affine transformation includes rotation, scaling, translation, and tilt adjustment; perspective transformation considers more complex changes in viewpoint. Using the least squares method or other optimization techniques, solve for the transformation matrix based on matching point pairs. Optimization algorithms such as gradient descent can be applied to accurately solve for the transformation parameters, ensuring optimal registration between images. Apply the solved transformation to the images to be registered, and evaluate the registration quality by visual inspection or calculating the error in the overlapping area.

[0040] The determined registration transform is applied to the image to be registered and adjusted. This step involves interpolating and transforming the pixel values ​​of the image to be registered to maintain the image's geometry and structure. The above steps are repeated for all areas in the image data of the flattened cardboard, applying the registration transform to the corresponding image to be registered and adjusting accordingly.

[0041] All the adjusted images to be registered are stitched together to obtain a complete 3D model of the cardboard to be flattened. This can be achieved through image stitching algorithms, such as feature matching-based stitching algorithms or deep learning-based image stitching methods.

[0042] The process of the deep learning-based image stitching method is as follows:

[0043] Global image registration is performed using features extracted through deep learning to ensure alignment accuracy between images before stitching. A deep learning model is used to predict the least noticeable stitching line positions to reduce visual stitching artifacts. Image segmentation and edge detection techniques can be used to assist this process. Generative Adversarial Networks (GANs) are used for style transfer and color correction of the images, adjusting color and lighting differences within the stitching area to make the stitched image more visually unified.

[0044] By registering and adjusting the various regions in the image data of the cardboard to be flattened with the images to be registered in the database, a three-dimensional model of the entire cardboard was obtained. This ensured the accuracy and completeness of the model, allowing subsequent flattening control to be based on an accurate three-dimensional model. By selecting a template and performing registration transformations, the precision and accuracy of the three-dimensional model were ensured, making the subsequent control process more reliable. Processing all regions in the image data of the cardboard to be flattened ensured that all parts of the entire cardboard were included in the three-dimensional model, without missing any important information.

[0045] Specifically, the process of adjusting the surface temperature of the pressing plate based on the surface characteristic data of the paperboard to be flattened, so that the surface humidity of the paperboard to be flattened reaches the pre-flattening humidity, is as follows: First, acquire the surface characteristic data of each detection area on the surface of the paperboard to be flattened. The surface characteristic data of each detection area includes the surface humidity and surface temperature of the detection area. These detection areas are divided according to the position of each heating area on the surface of the pressing plate to ensure coverage of the entire surface of the paperboard to be flattened. For each detection area, determine whether the surface humidity of each detection area is greater than the set humidity threshold. This humidity threshold can be set according to the requirements of the paperboard to be flattened and the environmental conditions. The dehumidification temperature can be adjusted according to different humidity environments and paperboard requirements to adapt to different production needs and environmental conditions. If the surface humidity of a certain detection area is greater than the set humidity threshold, it indicates that the area needs to be adjusted. Then, based on the surface characteristic data of the detection area where the surface humidity is greater than the set humidity threshold, the dehumidification time, and the pre-flattening humidity, determine the surface temperature of the pressing plate corresponding to the detection area, so that the surface humidity of the paperboard to be flattened reaches the pre-flattening humidity. If the surface humidity of a certain detection area is not greater than the set humidity threshold, the surface temperature of the pressing plate corresponding to the detection area is not adjusted.

[0046] By monitoring and acquiring surface characteristic data of various areas on the surface of the cardboard to be flattened, it is determined whether the humidity has reached the preset threshold. Then, the temperature is adjusted accordingly as needed to ensure that the humidity of the cardboard surface reaches the state before flattening, which helps to maintain the stability and consistency of the cardboard quality.

[0047] If the surface humidity of all detection areas is greater than the set humidity threshold, the humidity before flattening is the set humidity threshold; if the surface humidity of some detection areas is less than or equal to the set humidity threshold, the humidity before flattening is the average surface humidity of the detection areas whose surface humidity is less than or equal to the set humidity threshold.

[0048] If the surface humidity of all detection areas is greater than the set humidity threshold, it indicates that the entire surface of the cardboard to be flattened is relatively damp. Therefore, the humidity before flattening is set to the set humidity threshold. If the surface humidity of some detection areas is less than or equal to the set humidity threshold, it indicates that there are some dry areas on the cardboard surface. In this case, to more accurately determine the humidity before flattening, the average humidity of all detection areas with humidity less than or equal to the set humidity threshold is taken as the humidity before flattening.

[0049] It can flexibly determine the humidity before flattening according to the actual situation, which can meet the needs of both fully humid conditions and relatively dry conditions in some areas, so as to ensure that the humidity of the cardboard surface after flattening reaches the expected state. It determines the humidity before flattening according to the actual humidity of different areas, which more accurately reflects the humidity of the entire cardboard surface. It takes into account the humidity differences in different areas, making the determination of humidity before flattening more flexible and adaptable to different situations.

[0050] The surface temperature of the press plate can be calculated using historical empirical data, such as by calculating the average surface temperature of the press plate under the same conditions. Alternatively, it can be obtained using the following formula: The formula for calculating the surface temperature of the press plate is as follows. Before calculation, each parameter should be scalarized:

[0051]

[0052] In the formula, T h T is the surface temperature of the pressure plate. n To detect the surface temperature of the area, H h To detect the surface humidity of the area, H f Q is the set humidity threshold. h Q is the amount of heat used for heating. e C is the heat used for evaporating water. p ΔH represents the specific heat capacity of the paperboard, which can be obtained experimentally based on its composition and density, typically ranging from 1.3 to 1.4 J / (g·℃). m represents the mass of the paperboard in the detection area, which can be calculated by dividing the total mass by the number of monitoring areas. ΔH represents the latent heat of vaporization of water, approximately 2260 kJ / kg or 540 kcal / kg. In this embodiment, the calculation of the platen surface temperature is based on the principle of energy balance, considering the contributions of two parts of heat: heat used for heating and heat used for evaporating moisture. The sum of the heating heat and the evaporating heat equals the total heat absorbed by the paperboard, i.e., the sum of the heat used to raise the paperboard temperature and the heat used to evaporate moisture. The heating heat is the heat absorbed by the paperboard, which is related to the specific heat capacity of the paperboard, the mass of the paperboard, and the temperature difference between the platen surface temperature and the detection area surface temperature. The evaporating heat is the heat absorbed by the paperboard, which is related to the latent heat of vaporization of water, the mass of the paperboard, and the humidity difference between the surface humidity of the detection area and the set humidity threshold.

[0053] By taking into account the heat contribution from heating and evaporating moisture, the surface temperature of the pressing plate can be calculated more accurately, thereby achieving precise control of the surface humidity of the paperboard to be pressed. The formula comprehensively considers the energy conversion process of heating and evaporating moisture, which helps to optimize energy utilization efficiency and reduce energy waste. Furthermore, by combining thermodynamic principles and paperboard characteristics, the surface temperature of the pressing plate can be accurately calculated, thereby achieving effective control of the surface humidity of the paperboard, improving production efficiency and product quality.

[0054] Specifically, the process of determining the extrusion pressure applied to the cardboard by the pressing plate after the surface humidity of the cardboard to be flattened reaches the pre-flattening humidity, based on the pre-flattening characteristic data, is as follows: First, the surface temperature of the pressing plate is raised to the temperature at which the cardboard is most prone to deformation. This temperature is determined based on experience or experiments, usually according to the characteristics of the cardboard material and processing requirements. Next, the pre-flattening characteristic data of the cardboard is analyzed to obtain the flatness of the cardboard and the height of the highest point of the cardboard relative to the flattened plane. Then, the auxiliary characteristic data of the cardboard is obtained. Based on the auxiliary characteristic data and the analyzed pre-flattening characteristic data, the extrusion pressure applied to the cardboard is determined. This process requires considering factors such as the strength, deformation characteristics, and surface characteristics of the cardboard material to determine the optimal extrusion pressure.

[0055] By raising the surface temperature of the pressing plate to the temperature most prone to deformation and analyzing the characteristic data of the paperboard to be pressed, the required extrusion pressure is determined after the surface humidity of the paperboard to be pressed reaches the humidity before pressing. This logic ensures precise control and adjustment of the paperboard during the pressing process to obtain ideal processing results and product quality, avoiding resource waste caused by under- or over-processing, including paperboard materials and energy, thereby saving production costs.

[0056] The process of obtaining the flatness of the paperboard to be flattened is as follows: First, randomly select several test points on the three-dimensional model of the paperboard to be flattened and obtain the coordinates of each test point. These test points should be distributed throughout the entire surface of the paperboard to fully reflect the overall flatness of the paperboard. For each selected test point, calculate the distance between it and the plane on which the paperboard to be flattened will be located after being flattened to determine the flatness of the paperboard to be flattened. This distance can be determined by the difference between the height coordinate value of the test point and the height value of the plane on which the paperboard to be flattened is located.

[0057] Besides being identified and obtained through existing paperboard flatness models, the flatness of the paperboard to be flattened can also be obtained in the following way. The calculation formula for the flatness of the paperboard to be flattened is as follows, and each parameter is processed before calculation:

[0058]

[0059] In the formula, ZP represents the flatness of the paperboard to be flattened, i is the test point number (i = 1, 2, 3, ..., N), N is the total number of test points, and γ is a weighting factor for the distance between the test point coordinates and the plane where the paperboard will be flattened. Based on historical experiments, the influence of different γ values ​​on flatness is compared. Combined with existing flatness analysis algorithms, the flatness obtained by the algorithm is obtained. The difference between the flatness calculated by different γ values ​​and the flatness calculated by the algorithm is compared, and the γ value with the smallest difference is selected. i Let γ be the height coordinate of the test point at the i-th test point, Z0 be the height of the plane where the paper to be flattened will be located after flattening, and B be the weight adjustment base. Based on the method of determining γ, the flatness of the paperboard to be flattened calculated without weight adjustment base is compared with the flatness calculated by the algorithm, and the B with the smallest difference is selected. The algorithm used for flatness analysis includes, but is not limited to, curvature analysis, Gaussian filtering, and image processing techniques. Z is the adjustment coefficient for the weighting factor γ. i The larger the value, the greater the proportion of γ.

[0060] Based on the calculated distance to each test point, the flatness of the cardboard to be flattened is determined by a weighted average. The deviation of the distance between each test point was taken into account. The weighting was based on the ratio of the height coordinate value of the test point distance to the height value of the plane on which the cardboard to be flattened is located. By selecting multiple test points, the flatness of the cardboard to be flattened can be evaluated more comprehensively. This avoids the problem of inaccurate evaluation results caused by only considering local areas. The flatness is calculated by using a numerical method, which reduces the influence of human subjective factors and improves the objectivity and reliability of the evaluation results.

[0061] The experimental data of the height coordinates of each test point are shown in Table 1:

[0062] Table 1. Experimental data on the height coordinates of each test point.

[0063]

[0064] The flatness ZP of the cardboard to be flattened, calculated from the weighted distance in Table 1, is 8.18. This is a weighted average result, taking into account the height difference between each test point and the plane height, as well as their respective weights. i The larger the absolute value of -Z0|, the stronger the... The larger Z is, then at this time i The greater the weight, the greater the weighted distance, which in turn leads to a greater deviation from Z0.

[0065] Figure 2The graph shows the function of the flatness of the cardboard to be flattened, illustrating the relationship between the height of each test point and the preset plane height Z0 = 5. This visually demonstrates the distribution of test point heights and the deviation from the plane height. |Z i The larger the absolute value of -Z0|, the stronger the... The larger Z is, the greater Z becomes. i The greater the weight, the greater the weighted distance, which in turn leads to a greater deviation from Z0.

[0066] The process of determining the extrusion pressure applied to the cardboard to be flattened based on its associated characteristic data and the analyzed pre-flattening characteristic data is as follows: First, a pressure index is determined based on the associated characteristic data and the analyzed pre-flattening characteristic data of the cardboard. The pressure index comprehensively evaluates the stress on the cardboard based on factors such as the cardboard's stiffness, elastic modulus, and the area under pressure, as well as the analyzed pre-flattening characteristic data. The pressure index is then compared with pressure index ranges stored in a database. These ranges are derived from previous experiments or experience. The extrusion pressure corresponding to the successfully matched pressure index range is taken as the extrusion pressure applied to the cardboard to be flattened. This extrusion pressure is a result obtained beforehand based on experiments or simulations to ensure that appropriate pressure is applied to the cardboard. The associated characteristic data of the cardboard to be flattened includes the cardboard's stiffness, elastic modulus, and the area under pressure.

[0067] By comprehensively considering the auxiliary characteristic data of the paperboard to be flattened and the analyzed pre-flattening characteristic data, a more accurate pressure index is determined, which can better reflect the stress situation of the paperboard. Based on the pressure index range stored in the database, the extrusion pressure can be flexibly adjusted according to different situations and requirements to adapt to different paperboard materials, sizes and processing requirements.

[0068] Besides using computer-aided engineering software (such as finite element analysis software) to simulate the behavior of cardboard under pressure, the pressure index can also be obtained in the following way. The formula for calculating the pressure index is as follows, and each parameter is scalarized before calculation:

[0069]

[0070] In the formula, Yz is the pressure index, St is the stiffness of the cardboard, E is the elastic modulus of the cardboard, t is the pressure plate application time, which is determined based on the time consumed by the pressure plate to flatten the cardboard in the past multiple times, and the average value of historical events is taken as the pressure plate application time, A is the area of ​​the cardboard under pressure, α1 is the weighting factor of the auxiliary characteristic data of the cardboard to be flattened, cZP is the flatness benchmark reference value, which is determined based on the average flatness of the cardboard that has been flattened in the past, HP is the height of the highest point of the cardboard to be flattened relative to the flattened plane, cHP is the height benchmark reference value, which is determined based on the average height of the warped edge of the cardboard in the past, and α2 is the weighting factor of the analyzed pre-flattening characteristic data. α1 and α2 can be obtained by analyzing experimental data, such as multiple linear regression analysis, to determine the degree of influence of the auxiliary characteristic data of the cardboard to be flattened and the analyzed pre-flattening characteristic data on Yz.

[0071] In this implementation scheme, the formula combines two main categories of data: auxiliary characteristic data of the cardboard to be flattened and analyzed pre-flattening characteristic data. This provides a more comprehensive assessment to determine the extrusion pressure applied to the cardboard, thereby optimizing the processing quality and efficiency. The auxiliary characteristic data, based on physical properties, calculates the required base extrusion pressure. Stiffness and modulus of elasticity are indicators describing a material's resistance to deformation, while the pressing time and pressure area directly affect the extrusion process. The analyzed pre-flattening characteristic data helps assess the unevenness of the cardboard surface, influencing the adjustment of the extrusion pressure.

[0072] The values ​​of the pre-flattening feature data and auxiliary feature data after analysis for each group are shown in the table below:

[0073] Table 2 shows the values ​​of pre-flattening characteristic data and auxiliary characteristic data for each group after analysis.

[0074]

[0075] Where α1=0.7, α2=0.3, cZP=10, cHP=10.

[0076] Each ID corresponds to a set of parsed pre-flattening feature data. As the amount of data in the parsed pre-flattening feature data increases, the pressure index also gradually increases. Figure 3 The graph is a function of the stress index, showing the increasing trend of the configuration ID. Each point represents the stress index of a configuration, clearly demonstrating the systematic growth as the configuration changes.

[0077] With other parameters fixed, as ZP and HP increase, it leads to An increase in pressure, which in turn increases Yz, indicates that greater pressure is needed to flatten the cardboard.

[0078] A machine vision-based cardboard flattening control system is used in the aforementioned machine vision-based cardboard flattening control method, such as... Figure 4 As shown, the system includes a 3D modeling module, a pressing plate surface temperature adjustment module, and an extrusion pressure determination module. The 3D modeling module acquires image data of the cardboard to be flattened and establishes a 3D model of the cardboard based on this image data. The pressing plate surface temperature adjustment module acquires surface feature data of the cardboard to be flattened and adjusts the pressing plate surface temperature based on this data to bring the surface humidity of the cardboard to the level before flattening. The extrusion pressure determination module acquires the pre-flattening feature data of the cardboard based on the 3D model and determines the extrusion pressure applied by the pressing plate to the cardboard after the surface humidity reaches the pre-flattening level.

[0079] In this implementation scheme, machine vision is used to acquire image data of the cardboard and build a three-dimensional model, which can accurately understand the physical state and degree of deformation of the cardboard, thereby achieving more precise processing. By adjusting the temperature of the pressing plate surface to match the humidity of the cardboard before pressing, it can be ensured that the cardboard maintains ideal humidity during processing, avoiding quality problems caused by dryness or excessive moisture.

[0080] The automation of the entire flattening process reduces the need for manual intervention, increases production speed and continuity, and can quickly respond to changes in the condition of the cardboard, adjusting the extrusion pressure and temperature settings in real time to cope with different batches or conditions of cardboard.

[0081] By precisely controlling the processing, material defects and waste can be minimized, and material utilization can be improved. Precise control of the temperature of the press plate and the applied extrusion pressure can reduce energy consumption, especially in heating and pressure applications.

[0082] Machine vision-based cardboard flattening devices, such as Figures 5-6 The system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, cause the processor to perform the aforementioned machine vision-based cardboard flattening control method. It also includes a flattening support 1 and a pressing plate 2 mounted on the flattening support 1. The pressing plate 2 is fixedly connected to the flattening support 1 via a lifting cylinder 3. Heating modules 4 are fixedly connected to both the pressing plate 2 and the flattening support 1. A humidity sensor 5 is installed on the heating module 4. An active roller 7 is also mounted on one side of the flattening support 1 via a drive motor 6. A lifting frame 9 is also fixedly connected to the flattening support 1 via an adjusting cylinder 8. A driven roller 10 is movably connected inside the lifting frame 9.

[0083] In this embodiment, the cardboard to be flattened is placed on the flattening support 1. The pressing plate 2 moves downward under the action of the lifting cylinder 3, so that the humidity sensor 5 on the pressing plate 2 comes into contact with the cardboard to be flattened. At the same time, the humidity sensor 5 on the flattening support 1 comes into contact with the cardboard to be flattened. The humidity of each area of ​​the cardboard to be flattened is obtained through the humidity sensor 5, and the heating module is determined to heat according to the humidity.

[0084] Then, the lifting cylinder 3 moves the pressing plate 2 upward. The image data of the cardboard to be pressed can be acquired by the external image acquisition device. The image data of the cardboard to be pressed is processed and modeled to obtain the surface feature data of the cardboard to be pressed and to determine the pressing force of the pressing plate 2.

[0085] The lifting cylinder 3 moves the pressing plate 2 downward again, and the heating module 4 heats the cardboard to be flattened to the temperature at which it is most easily deformed. At the same time, the calculated extrusion force is used to press the cardboard to be flattened.

[0086] After the pressure is applied for a period of time, the lifting cylinder 3 moves the pressing plate 2 upward, and then the cardboard is squeezed a second time through the driving roller 7 and the driven roller 10 to further flatten it.

[0087] An electronic device includes a processor and a memory, in which computer program instructions are stored, which, when executed by the processor, cause the processor to perform the above-described machine vision-based cardboard flattening control method.

[0088] A computer-readable storage medium for storing a program that, when executed by a processor, implements the machine vision-based cardboard flattening control method described above.

[0089] In summary, this application has at least the following effects:

[0090] By using machine vision to acquire image data of cardboard and build a 3D model, the physical state and degree of deformation of the cardboard can be accurately understood, thus enabling more precise processing. By adjusting the temperature of the pressing plate surface to match the humidity of the cardboard before pressing, it can be ensured that the cardboard maintains ideal humidity during processing, avoiding quality problems caused by dryness or excessive moisture.

[0091] The automation of the entire flattening process reduces the need for manual intervention, increases production speed and continuity, and can quickly respond to changes in the condition of the cardboard, adjusting the extrusion pressure and temperature settings in real time to cope with different batches or conditions of cardboard.

[0092] By precisely controlling the processing, material defects and waste can be minimized, and material utilization can be improved. Precise control of the temperature of the press plate and the applied extrusion pressure can reduce energy consumption, especially in heating and pressure applications.

[0093] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0094] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0095] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0096] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0097] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0098] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A machine vision-based method for controlling cardboard flattening, characterized in that, Includes the following steps: Acquire image data of the cardboard to be flattened, and build a three-dimensional model of the cardboard based on the image data; Acquire surface feature data of the paperboard to be flattened, and adjust the surface temperature of the pressing plate based on the surface feature data of the paperboard to be flattened so that the surface humidity of the paperboard to be flattened reaches the humidity before flattening; Based on the three-dimensional model of the paperboard to be flattened, the pre-flattening feature data of the paperboard to be flattened is obtained, and based on the pre-flattening feature data, the extrusion force applied by the pressing plate to the paperboard to be flattened after the surface humidity of the paperboard to be flattened reaches the pre-flattening humidity is determined. The process of determining the compressive force applied to the cardboard by the pressing plate after the surface moisture of the cardboard to be flattened reaches the pre-flattening moisture level, based on the pre-flattening characteristic data, is as follows: Raise the surface temperature of the pressing plate to the temperature at which the cardboard to be pressed is most prone to deformation; The temperature at which the cardboard is most prone to deformation when it is flattened is determined based on experience or experiments, according to the characteristics of the cardboard material and processing requirements. Analyze the pre-flattening feature data of the cardboard to be flattened to obtain the flatness of the cardboard to be flattened and the height of the highest point of the cardboard to be flattened relative to the flattened plane; Obtain the attached feature data of the cardboard to be flattened; The extrusion force applied to the paperboard to be flattened is determined based on the attached feature data of the paperboard to be flattened and the analyzed pre-flattening feature data. The process of determining the extrusion force applied to the paperboard to be flattened based on the attached feature data of the paperboard to be flattened and the parsed pre-flattening feature data is as follows: The pressure index is determined based on the attached feature data of the cardboard to be flattened and the analyzed pre-flattening feature data. The pressure index is compared with the pressure index range stored in the database, and the extrusion pressure corresponding to the successfully matched pressure index range is used as the extrusion pressure applied to the cardboard to be flattened.

2. The machine vision-based cardboard flattening control method according to claim 1, characterized in that, The process of establishing a 3D model of the cardboard to be flattened based on the image data of the cardboard to be flattened is as follows: Select a region from the image data of the cardboard to be flattened as a template, and search for the registration region that is most similar to the template in the images to be registered stored in the database; Determine the registration transformation between the template and the image corresponding to the registration region; The registration transformation is applied to the image to be registered, and the image is adjusted accordingly. Repeatedly select all regions in the image data of the cardboard to be flattened, and adjust the image to be registered for each region; All the adjusted images to be registered are stitched together to obtain the 3D model of the cardboard to be flattened.

3. The machine vision-based cardboard flattening control method according to claim 1, characterized in that, The process of adjusting the surface temperature of the pressing plate based on the surface characteristic data of the cardboard to be flattened, so that the surface humidity of the cardboard to be flattened reaches the humidity before flattening, is as follows: Surface feature data of each detection area on the surface of the cardboard to be flattened are obtained. The surface feature data of each detection area includes the surface humidity and surface temperature of the detection area. The detection areas are divided according to the position of each heating area on the surface of the pressing plate. Determine whether the surface humidity of each detection area is greater than the set humidity threshold; If so, the surface temperature of the pressing plate corresponding to the detection area is determined based on the surface feature data of the detection area where the surface humidity is greater than the set humidity threshold, the dehumidification time, and the humidity before flattening, so that the surface humidity of the cardboard to be flattened reaches the humidity before flattening. Otherwise, do not adjust the surface temperature of the pressure plate corresponding to the detection area.

4. The machine vision-based cardboard flattening control method according to claim 3, characterized in that, If the surface humidity of all detection areas is greater than the set humidity threshold, then the humidity before flattening is the set humidity threshold. If the surface humidity of a detection area is less than or equal to the set humidity threshold, then the humidity before flattening is the average surface humidity of the detection areas where the surface humidity is less than or equal to the set humidity threshold.

5. The machine vision-based cardboard flattening control method according to claim 3, characterized in that, The formula for calculating the surface temperature of the pressure plate is as follows: In the formula, The surface temperature of the pressure plate. To detect the surface temperature of the area, To detect the surface humidity of the area, The set humidity threshold, For the heat used for heating, The heat used for evaporating water. This refers to the specific heat capacity of the cardboard. To test the quality of the cardboard in the inspection area, It is the latent heat of vaporization of water.

6. The machine vision-based cardboard flattening control method according to claim 1, characterized in that, The process for obtaining the flatness of the cardboard to be flattened is as follows: Randomly select several test points on the 3D model of the cardboard to be flattened and obtain the coordinates of each test point; The flatness of the paperboard to be flattened is determined based on the distance between the selected test point coordinates and the plane on which the paperboard will be flattened.

7. A machine vision-based cardboard flattening control system, used in the machine vision-based cardboard flattening control method according to any one of claims 1-6, characterized in that, It includes a 3D modeling module, a platen surface temperature adjustment module, and an extrusion pressure determination module, among which: The 3D modeling module is used to acquire image data of the cardboard to be flattened and to build a 3D model of the cardboard to be flattened based on the image data. The surface temperature adjustment module of the pressing plate is used to acquire the surface characteristic data of the paperboard to be pressed and adjust the surface temperature of the pressing plate based on the surface characteristic data of the paperboard to be pressed so that the surface humidity of the paperboard to be pressed reaches the humidity before pressing. The extrusion pressure determination module is used to obtain the pre-flattening feature data of the cardboard based on the three-dimensional model of the cardboard to be flattened, and to determine the extrusion pressure applied by the pressing plate to the cardboard after the surface humidity of the cardboard to be flattened reaches the pre-flattening humidity based on the pre-flattening feature data.

8. A cardboard flattening device based on machine vision, characterized in that, It includes a processor and a memory, wherein computer program instructions are stored in the memory, and the computer program instructions, when executed by the processor, cause the processor to perform the machine vision-based cardboard flattening control method as described in any one of claims 1-6; It also includes a flattening support base (1) and a flattening plate (2) set on the flattening support base (1). The flattening plate (2) is fixedly connected to the flattening support base (1) by a lifting cylinder (3). A heating module (4) is fixedly connected to both the flattening plate (2) and the flattening support base (1). A humidity sensor (5) is installed on the heating module (4). One side of the flattening support (1) is also equipped with an active roller (7) via a drive motor (6), and a lifting frame (9) is fixedly connected to the flattening support (1) via an adjusting cylinder (8), and a driven roller (10) is movably connected inside the lifting frame (9).

Images