Machine vision board line distinguishing and edge sealing machine ink-jet printing system

Abstract

The invention discloses a machine vision board line distinguishing and edge banding machine ink-jet printing system. A visual collection unit collects board lines, a data processing unit extracts multi-dimensional texture features and generates dynamic edge banding patterns and two-dimensional codes, an edge banding execution unit carries out edge banding through a white edge banding belt, and an ink-jet printing unit carries out printing after scanning the codes. The system solves the problems that the edge band lines are not unified with the plates and personalized customization is limited, achieves accurate line matching, reduces edge band reserve and is suitable for processing of high-grade plate furniture.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of plate furniture processing, in particular to a machine vision plate pattern recognition and edge banding machine inkjet printing system. BACKGROUND

[0002] In the field of high-grade plate furniture processing, edge banding technology is a key link affecting the overall appearance and quality of furniture. In the prior art, when high-grade plate furniture is edge banded, an edge band is used to wrap the four edges of the plate. However, due to the lack of precise matching mechanism between the pattern design of the edge band and the surface pattern of the plate, the edge band is prone to reveal its own color after processing, resulting in a significant visual difference between the edge banded area and the plate surface, which seriously affects the overall appearance of the furniture and makes the high-grade plate furniture have a grade gap compared to solid wood furniture in terms of visual effect. At the same time, edge band manufacturers usually require a certain minimum order length and provide a limited number of pattern designs, which cannot meet the diversified demand for personalized custom furniture in the market, limiting the further development of the personalized custom furniture market. Currently, there is no technical solution that can achieve automatic recognition of plate patterns, dynamic generation of matching edge patterns, and automatic printing combined with an edge banding machine.

[0003] Based on the above problems, there is an urgent need for a technical solution that can solve the problem of non-uniformity between edge band patterns and plates and the limitation of personalized customization, in order to improve the edge quality and customization flexibility of high-grade plate furniture. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art, and the machine vision plate pattern recognition and edge banding machine inkjet printing system proposed in the present application comprises a visual acquisition unit, a data processing unit, an edge execution unit and an inkjet printing unit, characterized in that the visual acquisition unit acquires the pattern of the plate surface, the data processing unit extracts multi-dimensional texture features from the acquired pattern data, generates a dynamic pattern required for four-edge edge banding of the plate based on the extracted texture features and a two-dimensional code associated with the pattern information, the edge execution unit uses a white edge band to edge band the four edges of the plate, and the inkjet printing unit reads the pattern information in the two-dimensional code through a code reading unit, and performs inkjet printing operation on the four edges of the edge banded plate based on the read pattern information. The multi-dimensional texture feature extraction includes comprehensive extraction of the density, direction and gray scale distribution of the plate pattern, and the dynamic pattern generation forms a precise match with the surface pattern of the plate.

[0005] Preferably, the visual acquisition unit comprises an optical lens assembly and an image preprocessing subunit, the optical lens assembly is used to capture high-definition pattern images of the plate surface, and the image preprocessing subunit performs illumination compensation and anti-reflection suppression processing on the high-definition pattern images to eliminate the interference of environmental light changes on pattern acquisition.

[0006] Further preferably, the data processing unit comprises a texture feature extraction subunit and a pattern generation subunit, the texture feature extraction subunit quantitatively extracts the density value, the strike angle value and the gray distribution variance value of the plate texture through a gray gradient algorithm and a strike vector analysis algorithm, the pattern generation subunit generates a four-side edge sealing pattern consistent with the plate surface texture based on the quantitatively extracted texture feature parameters through a texture mapping algorithm, and encodes the resolution, color parameters and matching coordinate information of the pattern into a two-dimensional code.

[0007] Further preferably, the edge sealing execution unit comprises an edge sealing tape conveying subunit and a pressing subunit, the edge sealing tape conveying subunit accurately conveys the white edge sealing tape to the four-side preset edge sealing position of the plate, and the pressing subunit adopts a constant temperature pressing mode to tightly adhere the white edge sealing tape to the plate edge, so as to ensure that there is no gap between the edge sealing tape and the plate edge.

[0008] Further preferably, the texture feature extraction subunit calculates the effectiveness of the extracted features through a texture feature confidence calculation formula, and the texture feature confidence calculation formula is as follows:

[0009] ;

[0010] wherein is the texture feature confidence, is a texture density weight coefficient, is a texture density quantitative value, is a strike consistency weight coefficient, is an included angle between the texture strike and the plate edge, is a gray distribution weight coefficient, is a gray distribution variance value, and the formula realizes comprehensive effectiveness evaluation of multi-dimensional texture features.

[0011] Further preferably, the inkjet printing unit comprises a printing parameter adaptation subunit, the printing parameter adaptation subunit calculates a printing parameter adjustment coefficient based on the texture feature confidence , and the printing parameter adjustment coefficient calculation formula is as follows:

[0012] ;

[0013] wherein is the printing parameter adjustment coefficient, is a confidence influence coefficient, is an edge sealing tape material influence coefficient, is a material adsorption coefficient of the white edge sealing tape, is a printing speed influence coefficient, is a moving speed of the inkjet printhead, and the formula dynamically adjusts the ink drop size and the printing density parameters.

[0014] Further preferably, the edge sealing execution unit further comprises an edge sealing flatness detection subunit, which obtains flatness data of the edge of the plate after edge sealing by a laser ranging method, and triggers the pressing subunit to perform secondary pressing adjustment when the flatness data exceeds a preset threshold.

[0015] Further preferably, the printing parameter adaptation subunit of the inkjet printing unit adapts the printing parameter based on the printing parameter adjustment coefficient adjusts the ink drop size to:

[0016] ;

[0017] adjusts the printing density to:

[0018] ;

[0019] wherein is a reference ink drop size, is a reference printing density, to achieve precise adaptation of the printing parameter to the texture feature and the edge sealing tape material.

[0020] Further preferably, the inkjet printing unit further comprises a four-side pattern joint calibration subunit, which extracts pattern edge features of adjacent edges, calculates the coincidence degree of the edge features, and performs micro-adjustment on the printing path based on the coincidence degree data to ensure that there is no misalignment at the joint of the four-side pattern.

[0021] Further preferably, the four-side pattern joint calibration subunit calculates the printing path adjustment amount by a joint calibration offset calculation formula, which is:

[0022] ;

[0023] wherein is a joint calibration offset, is a joint adjustment coefficient, is the printing parameter adjustment coefficient, is the pattern edge coincidence degree of the A side and the B side, is the pattern edge coincidence degree of the B side and the C side, is the pattern edge coincidence degree of the C side and the D side, is the pattern edge coincidence degree of the D side and the A side, to achieve dynamic calibration of the four-side pattern joint by the formula.

[0024] The technical effects achieved by the above embodiments include:

[0025] The creative technical point of the present application is that through the cooperation of the visual acquisition unit, the data processing unit, the edge sealing execution unit and the inkjet printing unit, the multi-dimensional feature extraction of the plate pattern, the dynamic edge sealing pattern generation and the code scanning printing integrated process are realized. The technical scheme solves the main problems of the background art that the high-grade plate furniture edge sealing strip pattern is not unified with the plate, the natural color is exposed after processing, and the personalized customization is limited, realizes the accurate matching of the edge sealing pattern and the plate, improves the appearance of the furniture, and at the same time, without the need to reserve a plurality of pattern edge sealing strips, the flexibility of personalized customization is enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 The connection block diagram of the machine vision plate pattern recognition and edge sealing machine inkjet printing system of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0028] In the existing high-grade plate furniture edge sealing process, the edge sealing strip pattern is difficult to unify with the plate surface pattern, and the edge sealing strip is easy to expose the natural color after processing, which affects the overall appearance. At the same time, the length requirement of the edge sealing strip and the limited number of manufacturer patterns seriously limit the development of the personalized customization furniture market.

[0029] Based on this, please refer to Figure 1 The present embodiment provides a machine vision plate pattern recognition and edge sealing machine inkjet printing system, which comprises: a visual acquisition unit for acquiring the pattern of the plate surface, a data processing unit for extracting multi-dimensional texture features from the acquired pattern data, a dynamic pattern generation based on the extracted texture features and a two-dimensional code associated with the pattern information required for four-edge sealing of the plate, an edge sealing execution unit for four-edge sealing of the plate using a white edge sealing strip, and an inkjet printing unit for reading the pattern information in the two-dimensional code through a code scanning unit and performing inkjet printing operation on the four edges of the sealed plate based on the read pattern information. The multi-dimensional texture feature extraction includes comprehensive extraction of the density, direction and gray scale distribution of the plate pattern, and the dynamic pattern generation forms an accurate match with the plate surface pattern.

[0030] It is worth mentioning that the visual acquisition unit selects a 200 million pixel industrial optical lens, the lens focal length is set to 16 mm, the working distance is controlled at 300 mm, and the captured board surface texture image resolution reaches 1920x1080 pixels, which can clearly present the details of fine lines such as wood grain and stone grain. During the acquisition process, the board is fixed on the conveying platform, the conveying speed is synchronized with the lens acquisition frame rate, the frame rate is set to 15 fps, and image blur is avoided. The data processing unit uses an Intel Core i7-12700K processor with 32 GB of DDR4 memory to ensure real-time multi-dimensional texture feature extraction. When extracting multi-dimensional texture features, the density is extracted by calculating the number of lines in a unit area using the gray scale gradient algorithm to obtain the texture density quantization value; the strike is extracted by using the strike vector analysis algorithm to map the direction of the lines to an angle value of 0-π, which is the angle between the line direction and the edge of the board; the gray scale distribution is extracted by calculating the variance of the image pixel gray scale value to reflect the uniformity of the light and dark changes of the lines. Dynamic pattern generation uses texture mapping algorithm, taking the board surface texture as the benchmark, and mapping the extracted density, strike and gray scale features to the four edge sealing areas of the board to ensure that the edge sealing pattern is visually consistent with the board texture. For example, when the board surface is a longitudinal wood grain, the edge sealing pattern is also generated with longitudinal wood grain, and the wood grain density and gray scale change are consistent with the board. The two-dimensional code generation uses the QR code encoding standard, and the encoding content includes the resolution, color parameter and matching coordinates of the dynamic pattern to ensure that the inkjet printing unit can accurately obtain the pattern information. The edge sealing execution unit selects a full-automatic edge sealing machine, the white edge sealing tape is made of PVC material with a thickness of 1.5 mm, and the width is adapted according to the thickness of the board. The conveying platform speed is set to 5 m / min during the edge sealing process to ensure that the edge sealing tape is aligned with the edge of the board. The inkjet printing unit selects an industrial-grade UV inkjet printer, the code scanning unit is integrated at the feeding end of the printer, and a CMOS code scanning module is used. The scanning distance is set to 100 mm, the scanning success rate is more than 99.9%, and the dynamic pattern is accurately printed on the surface of the white edge sealing tape according to the pattern information in the two-dimensional code during printing. The printing speed matches the conveying speed of the edge sealing machine to ensure that the printed pattern is not stretched or misaligned.

[0031] This embodiment realizes accurate matching of edge sealing lines and boards through multi-unit cooperation, solves the aesthetic problem of traditional edge sealing, and dynamically generates patterns without the need to store multiple edge sealing tapes, meeting the personalized customization demand.

[0032] In the existing visual acquisition process, changes in environmental light can easily cause uneven brightness and overexposure in the board texture image, affecting the accuracy of subsequent texture feature extraction.

[0033] Therefore, the visual acquisition unit comprises an optical lens assembly and an image preprocessing subunit, the optical lens assembly is used to capture a high-definition texture image of the surface of the plate, and the image preprocessing subunit performs illumination compensation and reflection suppression processing on the high-definition texture image to eliminate the interference of environmental illumination changes on texture acquisition.

[0034] It is worth mentioning that, in addition to the 200 million pixel industrial lens, the optical lens assembly is also equipped with a ring-shaped LED light source, the color of the light source is white, the color temperature is set to 5500K, the illumination intensity can be adjusted through a PWM signal, and the adjustment range is 100-1000 lux, which ensures uniform illumination of the surface of the plate. The image preprocessing subunit realizes the algorithm based on the OpenCV open source library, the illumination compensation adopts an adaptive histogram equalization algorithm, the brightness of the area with too low brightness in the image is enhanced, and the over-bright area is suppressed, so that the overall gray scale distribution of the image tends to be uniform; the reflection suppression processing adopts a combination of Gaussian filtering and threshold segmentation, first, the image is smoothed through a 5*5 Gaussian filtering kernel to eliminate high-frequency noise, and then the gray threshold is set, and the bottom color of the plate is adjusted, for example, the threshold of the light-colored plate is set to 230, the reflection area with a gray value higher than the threshold is replaced with the average gray value of the adjacent non-reflection area, to avoid that the reflection area covers the texture details. The preprocessed image is transmitted in real time to the data processing unit through an HDMI interface, the transmission delay is controlled within 100 ms, and the real-time performance of subsequent processing is ensured.

[0035] The embodiment eliminates the environmental illumination interference through illumination compensation and reflection suppression, and ensures stable texture image quality.

[0036] In the existing data processing process, the extraction of plate texture features is mostly single-dimensional, which cannot fully reflect the texture characteristics, resulting in low matching degree of the generated edge sealing pattern and the plate texture, and the pattern information transmission lacks a standardized way, which is easy to cause the loss or error of printing parameters.

[0037] Therefore, the data processing unit comprises a texture feature extraction subunit and a pattern generation subunit, the texture feature extraction subunit quantitatively extracts the density value, the direction angle value and the gray scale distribution variance value of the plate texture through a gray scale gradient algorithm and a direction vector analysis algorithm, the pattern generation subunit generates a four-side edge sealing pattern coherent with the surface texture of the plate based on the quantitatively extracted texture feature parameters through a texture mapping algorithm, and encodes the resolution, color parameters and matching coordinate information of the pattern into a two-dimensional code.

[0038] It is worth mentioning that the gray scale gradient algorithm of the texture feature extraction subunit adopts the Sobel operator to calculate the gradient values of the image in the x-axis and y-axis directions respectively, judges the texture edge through the gradient amplitude value, and then counts the number of edges in the unit pixel area to obtain the density value in the unit of strip per square centimeter. Through the conversion relationship between the pixel size and the actual size, 100x100 pixels correspond to an actual 2x2 square centimeter area; the strike vector analysis algorithm finds the direction corresponding to the histogram peak value by calculating the histogram of the gradient direction, that is, the texture strike angle value, which is in the unit of radian. In the calculation process, the non-maximum suppression algorithm is adopted to avoid the interference of false edges on the strike judgment; the gray scale distribution variance value is obtained by calculating the average value of the square difference between all pixel gray scale values and the average gray scale value. The gray scale value range is normalized to 0-255, and the variance value range is 0-65025. The larger the variance is, the more obvious the light and dark changes of the texture are. The texture mapping algorithm of the pattern generation subunit adopts a triangular subdivision mapping method to divide the plate surface texture image into multiple triangular grids. According to the extracted feature parameters, the grid vertex coordinates are adjusted so that after the grid is mapped to the edge sealing area, the density and strike of the texture are consistent with the plate. For example, if the plate surface texture density is 5 strips per square centimeter and the strike angle is π / 2 (vertical), the grid of the edge sealing area is adjusted to be arranged vertically, and the texture density in the grid is also 5 strips per square centimeter. The two-dimensional code encoding adopts the QRCodeVersion10 standard, and the encoding capacity is 1264 characters. The encoding content includes the pattern resolution (such as 300 dpi), the color parameters such as the red channel 200, the green channel 150, the blue channel 100, the matching coordinates such as the left upper corner of the plate as the origin, the A edge printing starting coordinate (0, 0), and the B edge (180, 0) in the unit of millimeter. The resolution of the generated two-dimensional code image after encoding is 200x200 pixels, which is convenient for the scanning unit to identify.

[0039] This embodiment improves the matching degree of the edge sealing pattern and the accuracy of information transmission through multi-dimensional feature extraction and standardized two-dimensional code encoding.

[0040] In the existing edge sealing execution process, the edge sealing tape conveying is prone to deviation, which leads to poor alignment of the edge sealing tape and the plate edge, and the compression method is mostly normal temperature compression. The edge sealing tape and the plate are not tightly attached, which is prone to gaps or falling off, affecting the edge sealing quality.

[0041] Therefore, the edge sealing execution unit includes an edge sealing tape conveying subunit and a compression subunit. The edge sealing tape conveying subunit accurately conveys the white edge sealing tape to the plate four edge sealing positions, and the compression subunit adopts a constant temperature compression method to tightly attach the white edge sealing tape and the plate edge, ensuring that there is no gap between the edge sealing tape and the plate edge.

[0042] It is worth mentioning that the edge band conveying subunit includes a driving feeding roller and a guide roller. The driving feeding roller is driven by a stepper motor, the motor model is 57 stepper motor, the step angle is 1.8°, the rotating speed is controlled by a pulse signal, and the rotating speed is synchronous with the plate conveying speed, for example, when the plate conveying speed is 5 m / min, the rotating speed of the feeding roller is set to 100 rpm; the guide roller is made of polyurethane material, the surface is smooth without lines, the distance between the guide rollers is adjusted according to the width of the edge band, for example, when the width of the edge band is 19 mm, the distance is set to 20 mm, so that the edge band does not deviate left and right during conveying, and the alignment accuracy is controlled within ±0.1 mm. The pressing subunit includes a heating block and a pressing roller. The heating block is made of aluminum alloy material, and nickel-chromium heating wire is built-in. The heating power is 500 W. The temperature of the heating block is controlled at 180-200℃ by a temperature controller. The pressing pressure is adjusted by a cylinder, and the pressure value is set to 0.5 MPa. The rotating speed of the pressing roller is synchronous with the plate conveying speed, so that the edge band is closely attached to the plate edge after heating, and the edge band is quickly cooled to room temperature after attachment to avoid deformation. During the edge sealing process, the edge sealing sequence of the four edges of the plate is A edge (long edge) → B edge (short edge) → C edge (long edge) → D edge (short edge). After each edge sealing is completed, the edge band is cut off by a cutter. The cutter is made of high-speed steel material, and the cutting accuracy is ±0.5 mm.

[0043] The embodiment ensures that the edge band is aligned with the plate edge and closely attached through accurate conveying and constant temperature pressing.

[0044] In the existing texture feature extraction process, there is a lack of evaluation mechanism for the effectiveness of the extracted features, which is easy to cause feature extraction errors due to image noise and blurred lines, thereby affecting the matching degree of the edge sealing pattern and being unable to determine whether the extracted features can be used for subsequent pattern generation.

[0045] Therefore, the texture feature extraction subunit calculates the effectiveness of the extracted features by a texture feature confidence calculation formula, and the texture feature confidence calculation formula is:

[0046] ;

[0047] wherein is the texture feature confidence, is the texture density weight coefficient, is the texture density quantization value, is the consistency weight coefficient, is the angle between the line direction and the plate edge, is the gray scale distribution weight coefficient, is the gray scale distribution variance value, and the comprehensive effectiveness of the multi-dimensional texture features is evaluated by the formula.

[0048] It is worth mentioning that, firstly, the definitions and dimensions of each parameter in the formula should be clearly defined: This is a dimensionless texture density weighting coefficient, adjusted according to the type of wood grain, such as wood grain boards. Set to 0.4, stone-patterned slabs Set it to 0.3; This is the texture density quantization value, with units of lines per square centimeter and a range of 1-10 lines per square centimeter. It is calculated using a grayscale gradient algorithm. The dimensionless direction consistency weight coefficient takes values ​​that are related to... Complementary, wood grain panels Set to 0.3, stone-patterned slabs Set it to 0.4; The angle between the grain direction and the edge of the board, in radians, with a value range of 0-π; This is a dimensionless grayscale distribution weighting coefficient with a value of 0.3, applicable to both wood grain and stone grain panels; This represents the variance of the grayscale distribution, which is dimensionless. Since the grayscale values ​​have been normalized to 0-255, the variance is normalized after calculation, and its value range is 0-1. The theoretical design basis of the formula is that the effectiveness of texture features needs to comprehensively consider three dimensions: density, direction, and grayscale. Density reflects the density of the texture, direction reflects the directional continuity of the texture, and grayscale distribution reflects the light and dark details of the texture. The influence of each dimension is balanced through weighting coefficients. The introduction of this is because the smaller the angle between the grain direction and the edge of the board, the better. The larger the value, the better the consistency and the higher the feature effectiveness; gray-level distribution variance The larger the value, the richer the details of the texture and the higher the effectiveness of the features. The calculation logic derivation process is as follows: First, calculate the weighted value of each dimension separately ( , , Then sum the three weighted values, and finally divide by the sum of the weight coefficients. ),get value, The value range is 0-1. A value ≥0.7 indicates that the feature extraction is valid and can be used for subsequent pattern generation; If the value is less than 0.7, the feature extraction is deemed invalid, triggering the visual acquisition unit to re-acquire the image. For example, after acquiring wood grain panels, =5 strips / square centimeter =π / 2 (vertical direction) =0.8, =0.4, =0.3, =0.3, then =0.4×5=2, = 0.3 x cos(pi / 2) = 0, = 0.3 x 0.8 = 0.24, summing up to 2.24, the weight sum is 1, = 2.24? Here needs to be corrected, because needs to be normalized, in actual calculation to 0-1 (such as 10 lines / cm2 corresponding to 1, 5 lines / cm2 corresponding to 0.5), after correction = 0.5, = 0.4 x 0.5 = 0.2, = 0.3 x 0 = 0, = 0.3 x 0.8 = 0.24, summing up to 0.44, = 0.44 < 0.7, invalid, trigger reacquisition, until = 0.7.

[0049] This embodiment realizes quantitative evaluation of the effectiveness of feature extraction by texture feature confidence calculation, avoiding the influence of invalid features on pattern generation.

[0050] In the existing inkjet printing process, the printing parameters are mostly fixed values, without considering the influence of texture feature effectiveness, edge sealing tape material and printing speed, which is easy to cause problems such as blurred printing pattern, uneven color or ink penetration, affecting the visual effect of the edge sealing pattern.

[0051] Therefore, the inkjet printing unit comprises a printing parameter adaptation subunit, which calculates a printing parameter adjustment coefficient based on the texture feature confidence The formula for calculating the printing parameter adjustment coefficient is as follows:

[0052] ;

[0053] Wherein is the printing parameter adjustment coefficient, is the confidence influence coefficient, is the edge sealing tape material influence coefficient, is the material adsorption coefficient of white edge sealing tape, is the printing speed influence coefficient, is the moving speed of the inkjet printhead, which dynamically adjusts the ink drop size and printing density parameters through the formula.

[0054] It is worth mentioning that the definition and dimension of each parameter in the formula are as follows: is the dimensionless confidence influence coefficient, taking the value of 0.5, is the texture feature confidence (dimensionless, 0-1); is the dimensionless material influence coefficient, taking the value of 0.3, is the edge band material adsorption coefficient (dimensionless, 0-1, reflecting the material's adsorption ability to ink, PVC material =0.6, ABS material =0.8); is a dimensionless speed influence coefficient, taking a value of -0.05 (because the faster the printing speed, the ink usage needs to be reduced to avoid pattern blur), is the printing head moving speed (unit: mm / s, value range: 50-200 mm / s), is the natural logarithm operation, ensuring that the influence of speed on is nonlinear, avoiding sudden changes in The theoretical design basis of the formula is: printing parameters need to adapt to feature effectiveness, material adsorption ability, and printing speed, The higher the feature, the higher the printing precision needs to be improved (corresponding to a reasonable range); The higher the material's adsorption ink ability, the more ink usage can be increased; The faster, the shorter the ink stays on the edge band surface, the less ink usage needs to be reduced to avoid penetration. The calculation logic derivation process is: first calculate the contribution value of each factor ( , , ), then sum the three to get , The value range is controlled between 0.2-0.8, and when it exceeds the range, the boundary value is taken (such as 0.8 when 0.2 when). For example =0.8, =0.6 (PVC edge band), =100 mm / s, =0.5, =0.3, =-0.05, then =0.5x0.8=0.4, =0.3x0.6=0.18, =ln(100)≈4.605, =-0.05x4.605≈-0.230, summing the three gives 0.4+0.18-0.230=0.35, =0.35, which is within the range of 0.2-0.8, and can be used to adjust the printing parameters. If =200 mm / s, ≈5.298, ≈-0.05x5.298≈-0.265, summing up 0.4+0.18-0.265=0.315, =0.315, because the speed is accelerated, Slightly reduced, in line with the demand for reducing ink consumption.

[0055] This embodiment realizes the adaptation of printing parameters to multiple factors by dynamically calculating the printing parameter adjustment coefficient, thereby improving the quality of printed patterns. After the existing edge sealing is performed, there is a lack of detection mechanism for the flatness of the sealed edge. After the sealed edge is attached to the edge of the plate, there may be flatness problems such as protrusions and depressions, which can easily cause pattern deformation during subsequent inkjet printing, affecting the final aesthetic appearance, and the flatness problem cannot be corrected in time, resulting in an increase in the scrap rate.

[0056] Therefore, the edge sealing execution unit further comprises an edge sealing flatness detection subunit. The edge sealing flatness detection subunit obtains the flatness data of the edge of the plate after edge sealing by a laser ranging method. When the flatness data exceeds a preset threshold, the pressing subunit is triggered to perform secondary pressing adjustment.

[0057] It is worth mentioning that the edge sealing flatness detection subunit uses a laser displacement sensor, model Keyence IL-600, measurement range 0-10 mm, measurement accuracy ±0.001 mm, and sampling frequency set to 1000 Hz, ensuring rapid capture of flatness changes. The sensor is installed behind the pressing subunit of the edge sealer, with a distance of 5 mm from the edge of the plate. The measurement direction is perpendicular to the edge sealing surface and moves at a uniform speed along the edge of the plate. The moving speed is synchronized with the conveying speed of the plate, and the height data of the edge sealing surface is collected in real time. The flatness data is calculated by the absolute value of the height difference between adjacent sampling points, and the preset threshold is set to 0.05 mm. That is, when the absolute value of the height difference between adjacent sampling points exceeds 0.05 mm, it is determined that the flatness is not up to standard. When a non-compliant area is detected, the sensor transmits a position signal to the controller of the edge sealing execution unit, which triggers the secondary pressing mechanism of the pressing subunit. During secondary pressing, the pressing temperature is maintained at 180-200°C, the pressing pressure is increased to 0.6 MPa, and the pressing roller speed is reduced to 3 m / min. The non-compliant area is slowly pressed, and after pressing, the laser displacement sensor is used for detection again until the flatness meets the standard. For example, the distance between the edge of plate A and the starting end is 100 mm, and the height difference between adjacent sampling points is 0.06 mm, which exceeds the threshold. The controller triggers the secondary pressing, and after pressing the area, the height difference is detected again to be 0.03 mm, which meets the requirements.

[0058] This embodiment ensures that the flatness of the sealed edge meets the standard by laser ranging detection and secondary pressing adjustment, avoiding subsequent pattern deformation.

[0059] In the existing printing parameter adjustment process, the ink drop size and printing density are only set by empirical values, without precise adaptation based on quantitative adjustment coefficients, resulting in the adjusted parameters still unable to match the comprehensive needs of texture characteristics, edge sealing tape material and printing speed, affecting the clarity and color consistency of the printed pattern.

[0060] Based on this, the printing parameter adaptation subunit of the inkjet printing unit adjusts the ink drop size and the printing density based on the printing parameter adjustment coefficients to

[0061] ;

[0062] to

[0063] ;

[0064] wherein is the reference ink drop size, is the reference printing density, realizing precise adaptation of the printing parameters and texture characteristics and edge sealing tape material.

[0065] It is worth mentioning that the reference ink drop size According to the printer model setting, the reference ink drop size of the selected UV inkjet printer is 10 pl, and the reference printing density is set to 50%, i.e. the ink ejection amount is 50% of the maximum ejection amount. The ink drop size adjustment formula wherein is the adjusted ink drop size, is the printing parameter adjustment coefficient, and the adjusted range is 10×(1+0.2)=12pl to 10×(1+0.8)=18pl, which can ensure that the ink drop forms clear dots on the edge sealing tape surface without excessive diffusion or incoherent dots. The printing density adjustment formula wherein is the adjusted printing density, and the setting of the 0.5 coefficient is because the printing density is more sensitive to the visual effect of the pattern than the ink drop size, and needs to be adjusted moderately to avoid color overflow due to too high density or dull pattern due to too low density. The adjusted range is 50%×(1+0.5×0.2)=55% to 50%×(1+0.5×0.8)=70%. For example =0.35, =10×(1+0.35)=13.5pl, =50%×(1+0.5×0.35)=50%×1.175=58.75%, at this time the ink drop size and density adapt to the adsorption capacity of the PVC edge sealing tape and the printing speed of 100mm / s, and the printed wood grain pattern is clear and the color is consistent with the surface of the board. If =0.8( High, High, Slow), =18pl, =70%, which can enhance the color saturation of the pattern and is suitable for stone patterns and other color-rich patterns.

[0066] This embodiment realizes precise adaptation of printing parameters by adjusting the ink drop size and printing density through a quantitative formula.

[0067] In the existing inkjet printing process, the pattern connection of the four edges of the board is prone to misalignment, such as the pattern of edge A not being aligned with that of edge B, which leads to poor overall pattern continuity and affects the visual integrity of the furniture edge banding. Moreover, there is a lack of calibration mechanism for the connection, which cannot correct the misalignment problem.

[0068] Therefore, the inkjet printing unit further comprises a four-edge pattern connection calibration subunit. The four-edge pattern connection calibration subunit extracts the edge features of the patterns of adjacent edges, calculates the coincidence degree of the edge features, and makes a micro adjustment to the printing path based on the coincidence degree data to ensure that there is no misalignment at the connection of the four-edge patterns.

[0069] It is worth mentioning that the four-edge pattern connection calibration subunit is integrated into the image processing module of the inkjet printing unit and uses an FPGA chip to realize real-time edge feature extraction and coincidence degree calculation. The edge feature extraction uses the Canny edge detection algorithm to detect the edge of the pattern of the adjacent edge and obtain the coordinate information of the edge pixels, i.e., the x and y coordinates based on the edge of the board. The coincidence degree calculation uses the normalized cross-correlation algorithm to match the edge features of the end of edge A with those of the starting end of edge B and calculate the cross-correlation coefficient of the two, which is the coincidence degree of the edge features, with a value range of 0-1, where 1 represents complete coincidence and 0 represents complete non-coincidence. The preset coincidence degree threshold is 0.9. When the coincidence degree is ≥0.9, it is determined that the connection is qualified and there is no need to adjust the printing path. When the coincidence degree is <0.9, the adjustment amount of the printing path is calculated according to the coincidence degree difference. The direction of the adjustment amount is the deviation direction, such as the edge of edge A being on the left side of the edge of edge B, which means that the printing path of edge B needs to be adjusted to the left. The size of the adjustment amount is proportional to the coincidence degree difference, i.e., the greater the coincidence degree difference, the greater the adjustment amount. For example, the coincidence degree of the edge features of edge A and edge B is 0.85, which is lower than the threshold of 0.9, and the coincidence degree difference is 0.05. Assuming that the adjustment coefficient is 0.1 mm / 0.01 difference, the adjustment amount is 0.05 x 10 x 0.1 = 0.05 mm, i.e., the printing path of edge B is shifted 0.05 mm towards edge A, and the coincidence degree is calculated again after the adjustment, until it is ≥0.9. During the calibration process, the adjustment of the printing path is realized by controlling the servo motor of the printhead. The model of the servo motor is Panasonic A6 series, and the positioning accuracy is ±0.001 mm, which ensures the accuracy of the adjusted printing path.

[0070] The embodiment ensures the continuity of the four-side pattern by edge feature extraction and coincidence calibration. The technical effect is that the coincidence of the four-side pattern is up to 99%, the misalignment is controlled within ±0.05 mm, and there is no visual trace of misalignment.

[0071] In the existing four-side pattern connection calibration process, the calculation of the adjustment amount does not combine the printing parameter adjustment coefficient and the relative relationship of the edge coincidence, which is easy to cause over-adjustment or insufficient adjustment, resulting in misalignment at the connection position, and does not consider the priority of different edge connections, which cannot ensure the continuity of the overall pattern. Therefore, the four-side pattern connection calibration subunit calculates the printing path adjustment amount by the connection calibration offset calculation formula, which is:

[0072]

[0073]

[0074] Among them, is the connection calibration offset, is the connection adjustment coefficient, is the printing parameter adjustment coefficient, is the pattern edge coincidence of A side and B side, is the pattern edge coincidence of B side and C side, is the pattern edge coincidence of C side and D side, is the pattern edge coincidence of D side and A side, and the formula realizes the dynamic calibration of the four-side pattern connection.

[0075] It is worth mentioning that the definition and dimension of each parameter in the formula are as follows: The connection calibration offset is mm, and the value range is 0-0.1 mm; The connection adjustment coefficient is dimensionless, and the value is 0.1 mm, which is the basic adjustment amplitude of the offset; The printing parameter adjustment coefficient (0.2-0.8) reflects the influence of the printing parameter on the connection, The larger the printing precision is, the smaller the adjustment amount can be; All of them are dimensionless edge coincidences (0-1); The maximum value of the four coincidences is used for normalization processing to ensure that ​​​​​The value range is 0-1, avoiding that a certain side coincidence degree is too low to cause the adjustment amount to be too large. The theoretical design basis of the formula is: the connection calibration offset needs to be combined with the printing accuracy and the relative level of each side coincidence degree. The higher the coincidence degree, the smaller the adjustment amount, avoiding excessive adjustment. The lower the coincidence degree, the larger the adjustment amount, ensuring that the connection meets the standard. The calculation logic derivation process is: first, calculate to obtain the relative value of the coincidence degree of side A and side B; then multiply and to obtain . For example, =0.1mm, =0.35, =0.85, =0.92, =0.88, =0.90, =0.92, then , =0.1x0.35x0.924≈0.032mm, that is, the printing path adjustment amount of side A and side B is 0.032mm, which combines the printing accuracy and the relative level of the coincidence degree, avoiding excessive adjustment. If =0.7, =0.95, then , =0.1x0.35x0.737≈0.026mm, which needs to be corrected. Because the lower the coincidence degree, the larger the adjustment amount should be, the actual formula should be , that is, the relative value of the difference in coincidence degree. After correction, , =0.1x0.35x1=0.035mm, meeting the requirement of large adjustment amount for low coincidence degree.

[0076] This embodiment realizes dynamic and accurate calculation of the adjustment amount through the connection calibration offset formula, ensuring that the four-side pattern connection meets the standard.

[0077] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any person skilled in the art can modify or change the above disclosed technical content to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made on the basis of the technical essence of the present application to the above embodiments still belongs to the protection scope of the technical solution of the present application.

Claims

1. A machine vision-based sheet metal texture recognition and edge banding inkjet printing system, comprising a vision acquisition unit, a data processing unit, an edge banding execution unit, and an inkjet printing unit, characterized in that, The visual acquisition unit collects the texture of the board surface, and the data processing unit extracts multi-dimensional texture features from the collected texture data. Based on the extracted texture features, a dynamic pattern required for sealing the four sides of the board and a QR code associated with the pattern information are generated. The sealing execution unit seals the four sides of the board with white sealing tape. The inkjet printing unit reads the pattern information in the QR code through the scanning unit and performs inkjet printing on the four sides of the sealed board based on the read pattern information. The multi-dimensional texture feature extraction includes the comprehensive extraction of the density, direction and grayscale distribution of the board texture. The dynamic pattern generation forms a precise match with the texture of the board surface.

2. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 1, characterized in that, The visual acquisition unit includes an optical lens assembly and an image preprocessing subunit. The optical lens assembly is used to capture high-definition texture images of the board surface, and the image preprocessing subunit performs illumination compensation and reflection suppression processing on the high-definition texture images to eliminate the interference of ambient light changes on texture acquisition.

3. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 1, characterized in that, The data processing unit includes a texture feature extraction subunit and a pattern generation subunit. The texture feature extraction subunit uses gray-level gradient algorithm and direction vector analysis algorithm to quantize and extract the density value, direction angle value and gray-level distribution variance value of the board texture. Based on the quantized texture feature parameters, the pattern generation subunit uses texture mapping algorithm to generate a four-sided edge-sealing pattern that is consistent with the texture of the board surface, and encodes the resolution, color parameters and matching coordinate information of the pattern into a QR code.

4. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 1, characterized in that, The edge banding execution unit includes an edge banding tape conveying subunit and a pressing subunit. The edge banding tape conveying subunit accurately conveys the white edge banding tape to the preset edge banding positions on the four sides of the board. The pressing subunit uses a constant temperature pressing method to tightly bond the white edge banding tape to the edge of the board, ensuring that there are no gaps between the edge banding tape and the edge of the board.

5. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 3, characterized in that, The texture feature extraction subunit calculates the effectiveness of the extracted features using a texture feature confidence calculation formula, which is: ； in For texture feature confidence, This is the texture density weighting coefficient. For texture density quantization values, To achieve consistent weighting coefficients, The angle between the grain direction and the edge of the board. The grayscale distribution weighting coefficient is... The variance of the grayscale distribution is used to evaluate the comprehensive effectiveness of multi-dimensional texture features using this formula.

6. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 5, characterized in that, The inkjet printing unit includes a printing parameter adaptation subunit, which adapts the printing parameters based on the confidence level of the texture features. The printing parameter adjustment factor is calculated using the following formula: ； in Adjust the coefficients for printing parameters. The confidence level influence coefficient. The influence coefficient of the edge banding material is... The material adsorption coefficient of the white edge banding tape. The coefficient representing the impact of printing speed. The formula represents the moving speed of the inkjet printhead, and it dynamically adjusts the droplet size and print density parameters.

7. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 4, characterized in that, The edge banding execution unit also includes an edge banding flatness detection subunit. The edge banding flatness detection subunit obtains the flatness data of the edge of the board after edge banding by laser ranging. When the flatness data exceeds the preset threshold, the pressing subunit is triggered to perform secondary pressing adjustment.

8. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 6, characterized in that, The printing parameter adaptation subunit of the inkjet printing unit adjusts the printing parameters based on the printing parameter adjustment coefficient. Adjust the ink droplet size to: ； Adjust the print density to: ； in Based on the ink droplet size, Based on the baseline printing density, it achieves precise matching of printing parameters with texture features and edge banding material.

9. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 1, characterized in that, The inkjet printing unit also includes a four-sided pattern alignment calibration subunit. This subunit extracts the pattern edge features of adjacent sealing edges, calculates the overlap of the edge features, and makes fine adjustments to the printing path based on the overlap data to ensure that there is no misalignment at the four-sided pattern alignment.

10. The machine vision-based sheet metal texture recognition and edge-sealing inkjet printing system according to claim 8, characterized in that, The four-sided pattern connection calibration subunit calculates the print path adjustment amount using the connection calibration offset calculation formula, which is as follows: ； in To connect the calibration offset, To adjust the coefficients, Adjust the coefficients for the printing parameters. Let A and B be the overlap of the pattern edges. Let B and C be the overlap of the pattern edges. Let C be the overlap degree of the pattern edges between sides D. The formula represents the overlap between the pattern edges of side D and side A, and is used to achieve dynamic calibration of the four-sided pattern connection.

Images