An automatic gluing and intelligent detection system and method for a packaging box

Abstract

The application provides a kind of packing box automatic gluing intelligent detection system and method, it is related to intelligent detection technical field, system includes: gluing adaptation analysis module, for the analysis of packing box parameter data obtains gluing adaptation characteristic data.Glue box adaptation module, for the association of glue performance data and gluing adaptation characteristic data obtains glue box matching data.Glue material adaptation analysis module, for the influence degree of analysis glue material dynamic characteristic data to glue box matching data obtains glue material adaptation influence coefficient, gluing monitoring adjustment module, for based on glue material dynamic characteristic data, glue box matching data and glue material adaptation influence coefficient constructs dynamic monitoring model, carries out real-time detection to gluing process, generates parameter adjustment and defect processing instruction.The application can adapt to a variety of production scenarios and material combination, improve production efficiency and early warning potential failure defect, greatly improve the prevention ability, carry out effective dynamic monitoring and adjustment.

Description

Technical Field

[0001] This invention relates to the field of intelligent detection technology, and in particular to an intelligent detection system and method for automatic glue application on packaging boxes. Background Technology

[0002] In the packaging manufacturing industry, the quality of adhesive coating on packaging boxes directly determines the packaging's sealing performance, structural strength, and lifespan, thus affecting the protective effect during product transportation and the appearance presented to end consumers. It is a core quality control point in the packaging production process. With the continuous increase in demands for packaging precision and production efficiency from industries such as e-commerce logistics, food, and pharmaceuticals, traditional packaging box adhesive coating technology is gradually becoming unable to meet the requirements of modern production.

[0003] Current packaging box adhesive coating technologies only focus on the static properties of the adhesive before coating and the static results after coating, lacking effective perception and analysis of the dynamic characteristics of the adhesive and packaging box throughout the entire curing cycle. For example, during the adhesive curing process, the adhesive undergoes a phase transformation from liquid to semi-solid and then to solid, accompanied by dynamic processes such as glue line shrinkage and rapid viscosity increase; simultaneously, the packaging box material undergoes micro-deformation due to glue wetting and moisture absorption, and these dynamic changes directly affect the final bonding effect. Furthermore, existing technologies lack the ability to monitor and dynamically adjust the coating process in real time, resulting in a "post-event detection, passive handling" management model. On the one hand, adhesive quality inspection is mostly concentrated after the coating process is completed, failing to capture anomalies in real time during coating. By the time defects are discovered, a batch of defective products has already been produced, causing material waste and production stoppages. On the other hand, even if coating anomalies are detected, adjustments rely on manual experience, lacking data-driven quantitative adjustment criteria, which not only prolongs the production cycle but also makes it difficult to guarantee the stability of the adjustment effect.

[0004] Furthermore, existing technologies largely rely on human experience to select adhesive types, failing to establish a quantitative matching relationship between the packaging box's geometry, material characteristics, usage scenarios, and the adhesive's dynamic characteristics. This results in low matching accuracy and further limits the level of intelligent control over the adhesive application process.

[0005] In summary, there is an urgent need for an intelligent detection system and method for automatic gluing of packaging boxes to solve the above-mentioned technical pain points. Summary of the Invention

[0006] This invention provides an intelligent detection system and method for automatic glue application on packaging boxes, which addresses the shortcomings of existing technologies that do not adequately consider the dynamic characteristics of glue and lack monitoring and adjustment of the glue application process.

[0007] On one hand, the present invention provides an intelligent detection system for automatic glue application on packaging boxes, comprising:

[0008] The adhesive compatibility analysis module is used to collect packaging box parameter data and adhesive performance data, and to analyze the packaging box parameter data using a geometric material compatibility method to obtain adhesive compatibility feature data.

[0009] The glue box adaptation module is used to analyze the changes in the shape and state of the packaging box caused by the glue over time to obtain the dynamic characteristic data of the glue material. The glue performance data is then correlated with the glue application adaptation characteristic data to obtain the glue box matching data.

[0010] The adhesive material compatibility analysis module is used to analyze the impact of the dynamic characteristics of the adhesive material on the matching data of the adhesive box to obtain the adhesive material compatibility influence coefficient.

[0011] The glue application monitoring and adjustment module is used to build a dynamic monitoring model based on the dynamic characteristics data of the glue material, the matching data of the glue box, and the glue material compatibility influence coefficient. It performs real-time detection of the glue application process and generates parameter adjustment and defect handling instructions.

[0012] This invention provides an intelligent detection system for automatic glue application on packaging boxes, and the geometric material adaptation method includes:

[0013] The basic dimensions and structural form of the packaging box are extracted from the packaging box parameter data as geometric parameters, and the material type of the packaging box is determined, combined with the surface characteristic parameters of the material as material parameters.

[0014] Based on the usage scenario and expected circulation environment parameters of the packaging box, the usage scenario parameters are obtained, and the glue application location and method are determined by combining geometric parameters and material parameters.

[0015] By correlating usage scenario parameters with adhesive strength and material parameters with adhesive properties, adhesive compatibility characteristic data can be obtained through analysis.

[0016] This invention provides an intelligent detection system for automatic glue application on packaging boxes, the method for precise analysis of the adhesive state includes:

[0017] The stage detection unit is used to divide different key monitoring stages according to the adhesive curing cycle and material response characteristics, and to collect dynamic monitoring parameters of adhesive morphology and material state at the corresponding stages.

[0018] The dynamic parameter processing unit is used to clean, smooth, and align the dynamic detection parameters in time and space, eliminate outliers, and unify the time and space references to obtain dynamic processing parameters.

[0019] The dynamic feature unit is used to calculate the rate of change of adhesive line morphology, curing speed coefficient, material deformation coefficient, and interface adaptation stability as dynamic feature indicators based on dynamic processing parameters.

[0020] The adhesive property correction unit is used to differentiate and correct dynamic characteristic indicators for different packaging box materials and adhesive types to obtain dynamic characteristic data of adhesive materials.

[0021] This invention provides an intelligent detection system for automatic glue application on packaging boxes. The steps for calculating dynamic feature indicators by the dynamic feature unit include:

[0022] The average width of the adhesive line at different key monitoring stages was selected from the dynamic processing parameters. The shrinkage or expansion trend during the curing process of the adhesive was analyzed, and the morphological change rate of the adhesive line was obtained by adjusting the parameters according to the type of packaging box material.

[0023] Viscosity and curing time at different time periods are selected from dynamic processing parameters, the initial curing coefficient is calculated, and the curing speed coefficient is obtained by adjusting the initial curing coefficient according to the type of adhesive.

[0024] The initial material deformation coefficient is calculated by selecting the maximum edge warping height and adhesive edge thickness from the dynamic processing parameters. The initial material deformation coefficient is then adjusted according to the rigidity of the packaging box material to obtain the final material deformation coefficient.

[0025] Peeling force at different time points is selected from dynamic processing parameters, initial stability is calculated, and interface adaptation stability is obtained by adjusting the initial stability according to the type of adhesive.

[0026] This invention provides an intelligent detection system for automatic glue application on packaging boxes. The steps for the glue box matching module to obtain glue box matching data include:

[0027] The adhesive performance data and adhesive compatibility characteristic data are simultaneously standardized, and a mapping relationship between the two is established from four dimensions: surface, penetration, adhesive amount and fixation.

[0028] Set hard matching rules based on the type of packaging box material and the usage scenario, and then filter and retain the material types, scenario tags and passing status as the initial adaptation pool.

[0029] Build a machine learning model and train it by collecting historical adhesion quality data. Input the model into the initial adaptation pool and output the historical adhesion quality score.

[0030] A gradient boosting tree model is used, and parameters are optimized through 5-fold cross-validation. The initial matching score is obtained by predicting each combination in the initial adaptation pool.

[0031] The initial matching score is corrected based on the dynamic characteristic data of the adhesive material to obtain the matching data of the adhesive box.

[0032] This invention provides an intelligent detection system for automatic glue application on packaging boxes. The steps of obtaining the glue compatibility influence coefficient by the glue compatibility analysis module include:

[0033] Collect adhesive application failures caused by abnormal dynamic properties of adhesive materials, establish a mapping relationship between dynamic properties and failure types, and determine the impact path.

[0034] The impact weight of the failure is determined based on the frequency and severity of the failure, and the correlation strength between dynamic characteristic indicators and adhesive application failure is calculated.

[0035] By combining the current deviation value, fault impact weight, and correlation strength of the dynamic characteristic data of the adhesive material, the impact scores of different dynamic characteristic indicators on the matching data of the adhesive box are calculated, and the initial coefficients are obtained by integrating them.

[0036] Collect real-time monitoring data of the current production batch and calibrate the initial coefficients to obtain the adhesive material compatibility influence coefficient.

[0037] This invention provides an intelligent detection system for automatic glue application on packaging boxes. The steps of obtaining a dynamic monitoring model by the glue application monitoring and adjustment module include:

[0038] The data processing unit is used to unify the format and dimensions of the dynamic characteristic data of adhesive materials, the matching data of adhesive boxes, and the influence coefficient of adhesive material compatibility, and to perform time correlation according to the adhesive batch ID and spatial alignment according to the workstation number to obtain the comprehensive processed data.

[0039] The fusion feature extraction unit is used to extract fusion features of static adaptation basis, dynamic risk and weight adjustment from the comprehensive processed data to construct a multi-dimensional monitoring feature set.

[0040] The temporal-static fusion unit is used to collect historical production data as a training set for training, employing a hybrid model architecture that combines temporal prediction and static classification.

[0041] The risk threshold setting unit is used to set the risk threshold of the model output based on a multi-dimensional monitoring feature set and different scenarios to obtain a dynamic monitoring model.

[0042] This invention provides an intelligent detection system for automatic glue application on packaging boxes. The steps of obtaining a multi-dimensional monitoring feature set by integrating a feature extraction unit include:

[0043] The final matching degree in the glue box matching data is converted into a 0-1 range value as the basic score for adaptation, and the historical matching pass rate of different glue and packaging box combinations is calculated as the basis for static adaptation.

[0044] The dynamic risk score is obtained by weighting and summing different dynamic characteristic parameters based on the adhesive material compatibility influence coefficient, and then combining the time-series change rate of the dynamic characteristic parameters to obtain the dynamic risk.

[0045] We construct weighted features based on the adhesive material compatibility influence coefficient, adjust the influence ratio of static compatibility basis and dynamic risk to obtain weight adjustment, and cross-integrate with static compatibility basis and dynamic risk to obtain a multi-dimensional monitoring feature set.

[0046] This invention provides an intelligent automatic glue-applying detection system for packaging boxes. The steps of the glue-applying monitoring and adjustment module receiving parameter adjustment and defect handling instructions include:

[0047] Real-time acquisition of dynamic characteristic data of adhesive materials, equipment operating parameters and image data during the adhesive application process, alignment with multi-dimensional monitoring feature sets and input into the dynamic monitoring model, outputting real-time risk status and anomaly contribution analysis.

[0048] If the condition is determined to be normal, the data will be continuously monitored and recorded. If the condition is determined to be suspicious, an early warning will be triggered, and the fluctuation threshold of the dynamic characteristic data of the adhesive material will be extracted. Combined with the adhesive box matching data, a pre-adjustment suggestion will be generated.

[0049] If the condition is determined to be unqualified, the defect type is located based on the abnormal contribution analysis, and parameter adjustment and defect handling instructions are generated in combination with the adhesive material compatibility influence coefficient.

[0050] On the other hand, the present invention provides an intelligent detection method for automatic glue application on packaging boxes, comprising:

[0051] We collected packaging box parameter data and adhesive performance data, and used a geometric material adaptation method to analyze the packaging box parameter data to obtain adhesive adaptation characteristic data.

[0052] The dynamic properties of the adhesive material were obtained by analyzing the changes in the shape and state of the packaging box over time using the colloidal analysis method. The adhesive performance data was then correlated with the adhesive compatibility characteristic data to obtain the box matching data.

[0053] The influence of the dynamic characteristics of the adhesive material on the matching data of the adhesive box is analyzed to obtain the adhesive material adaptation influence coefficient.

[0054] A dynamic monitoring model is constructed based on the dynamic characteristics data of adhesive materials, the matching data of adhesive boxes, and the influence coefficient of adhesive material compatibility. The model is used to monitor the adhesive application process in real time and generate parameter adjustment and defect handling instructions.

[0055] This invention provides an intelligent automatic glue application detection system and method for packaging boxes. Through multi-dimensional monitoring feature sets and dynamic monitoring models, it monitors the glue application process in real time, generating parameter adjustment and defect handling instructions to ensure the stability and consistency of glue application quality. The automated and intelligent glue application detection system reduces manual intervention, improves production efficiency, and shortens the production cycle. Precise glue application control and fault prevention reduce material waste and rework costs, lowering production costs. Dynamic monitoring models and real-time risk assessments identify potential problems early and allow for adjustments, improving system reliability and stability. By making differentiated modifications for different packaging box materials and glue types, the system can adapt to various production scenarios and material combinations, enhancing its adaptability and flexibility. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0057] Figure 1 This is a flowchart illustrating an intelligent detection system for automatic gluing of packaging boxes provided in an embodiment of the present invention.

[0058] Figure 2 This is a flowchart illustrating an intelligent detection method for automatic glue application on packaging boxes provided in an embodiment of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0060] The following is combined with Figures 1-2 This invention describes an intelligent detection system and method for automatic glue application on packaging boxes.

[0061] like Figure 1 As shown in the figure, an intelligent detection system for automatic gluing of packaging boxes provided by an embodiment of the present invention includes:

[0062] The adhesive compatibility analysis module is used to collect packaging box parameter data and adhesive performance data, and to analyze the packaging box parameter data using a geometric material compatibility method to obtain adhesive compatibility feature data.

[0063] Geometric material adaptation methods include:

[0064] The basic dimensions and structural form of the packaging box are extracted from the packaging box parameter data as geometric parameters, and the material type of the packaging box is determined, combined with the surface characteristic parameters of the material as material parameters.

[0065] The basic dimensions include the length, width, and height of the packaging box, as well as the dimensions of key glue-applying areas (such as the fit clearance between the lid and the body, and the depth and width of the glue tank, which affect the amount of glue stored).

[0066] The structural form can include the shape of the packaging box, the number and angle of the folds, and the complexity of the glue application path.

[0067] Determine the main material of the packaging box and indicate whether any special treatment is required. Surface characteristic parameters: Obtain the surface characteristic parameters of the material through sensors, including surface roughness, 24-hour water absorption rate, surface tension, etc. These parameters directly affect the adhesion and penetration of the adhesive.

[0068] Based on the usage scenario and expected circulation environment parameters of the packaging box, the usage scenario parameters are obtained, and the glue application location and method are determined by combining geometric parameters and material parameters.

[0069] By combining the fit clearance and structural morphology in the geometric parameters, the precise adhesive application location is determined. For irregular structures, the coordinate nodes of the adhesive application path are marked.

[0070] The adhesive application method is selected based on the surface characteristics and structural complexity of the material. For example, a combination of "dot adhesive + line adhesive" is used for smooth plastic boxes. "Wide-width line adhesive" is used for rough corrugated cardboard. "Segmented adhesive application" is used for complex, irregularly shaped boxes.

[0071] By correlating usage scenario parameters with adhesive strength and material parameters with adhesive properties, adhesive compatibility characteristic data can be obtained through analysis.

[0072] Adhesive compatibility data may include: if the water absorption rate of corrugated paper is ≥8%, it is associated with "adhesives with a drying speed ≥5mm / min are required". If the surface tension of the laminated color box is ≤30mN / m, it is associated with "adhesives with added adhesion promoters are required".

[0073] If the application scenario is "a courier box with a load-bearing capacity of 5kg", then the requirement is "peel strength after adhesive application ≥ 8N / 25mm". If it is "food packaging", then the requirement is "the adhesive must comply with the GB4806.10 food safety standard".

[0074] The glue box adaptation module is used to analyze the changes in the shape and state of the packaging box over time using the glue state analysis method to obtain the dynamic characteristic data of the glue material. The glue performance data is then correlated with the glue application adaptation characteristic data to obtain the glue box matching data.

[0075] The steps by which the adhesive box adapter module obtains the dynamic property data of the adhesive material include:

[0076] The stage detection unit is used to divide different key monitoring stages according to the adhesive curing cycle and material response characteristics, and to collect dynamic monitoring parameters of adhesive morphology and material state at the corresponding stages.

[0077] The key monitoring stages can be divided into: liquid stage: the glue has not yet started to solidify and is in a flowing state, and the focus is on monitoring the initial form.

[0078] Curing transition period: The adhesive changes from liquid to semi-solid, accompanied by an increase in viscosity and volume shrinkage, which is the stage with the most dramatic morphological changes.

[0079] Solid-state setting period: The adhesive has basically cured, and its form and properties tend to be stable. The focus is on monitoring the interaction between the final state and the material.

[0080] Methods for collecting dynamic monitoring parameters may include: high-speed industrial cameras: capturing changes in the shape of the adhesive line and recording the width, edge contour, and surface defects of the adhesive line at different time points.

[0081] Infrared thermal imager: monitors temperature changes during the curing process of adhesive, indirectly reflecting the curing progress.

[0082] Laser thickness gauge: measures changes in adhesive layer thickness in real time, capturing thickness fluctuations caused by shrinkage or flow.

[0083] Displacement sensor: Fixed to the edge of the packaging box, it records the micro-deformation of the material after the adhesive is applied.

[0084] Colorimeter: For color-changing adhesives, it collects RGB values ​​at different time points to quantify the range of color change.

[0085] This includes variations in adhesive line width / thickness, color changes, surface smoothness, curing speed, edge lifting height of the adhesive area, expansion rate of the material after moisture absorption, and changes in adhesion between the surface and the adhesive interface.

[0086] The dynamic parameter processing unit is used to clean, smooth, and align the dynamic detection parameters in time and space, eliminate outliers, and unify the time and space references to obtain dynamic processing parameters.

[0087] Outlier removal: Remove "jump data" caused by device jitter and replace them with the average of three adjacent time points.

[0088] Smoothing: Gaussian filtering is used to eliminate image noise in the edge data of the adhesive line captured by the high-speed camera, making the edge contour more continuous.

[0089] Imputing missing values: For missing data caused by brief signal interruptions, linear interpolation is used to impute them.

[0090] Unified time reference: Convert the acquisition time of all devices into "seconds relative to the moment when the glue application is completed" to ensure that the glue and material data at the same point in time can be directly correlated.

[0091] Spatial coordinate calibration: A coordinate system is established with the starting point of the adhesive application on the packaging box as the origin. The position of the adhesive line captured by the camera and the amount of material deformation recorded by the displacement sensor are uniformly mapped to this coordinate system to eliminate spatial misalignment caused by equipment installation position deviation.

[0092] Dynamic processing parameters can include glue line width data: the average glue line width during the liquid phase and the average glue line width during the solid setting phase.

[0093] Adhesive viscosity data: initial viscosity during the liquid phase, final viscosity during the solid phase, and curing time.

[0094] Material deformation data: maximum lifting height of the glued area edge, thickness of the adhesive edge of the packaging box, height of the corrugated paper flute peak, and surface tension deviation of the laminated color box.

[0095] Interface adhesion data: peel force at different time points during the solid setting period, temperature deviation of hot melt adhesive, and environmental humidity deviation of water-based adhesive, etc.

[0096] The dynamic feature unit is used to calculate the rate of change of adhesive line morphology, curing speed coefficient, material deformation coefficient, and interface adaptation stability as dynamic feature indicators based on dynamic processing parameters.

[0097] The steps for obtaining dynamic feature indices from dynamic feature units include:

[0098] The average width of the adhesive line at different key monitoring stages was selected from the dynamic processing parameters. The shrinkage or expansion trend during the curing process of the adhesive was analyzed, and the morphological change rate of the adhesive line was obtained by adjusting the parameters according to the type of packaging box material.

[0099] If the paper is corrugated: measurement deviations caused by corrugation peaks need to be considered.

[0100] For laminated color boxes: the effect of surface tension on shrinkage needs to be considered.

[0101] If it is ordinary cardstock: no additional correction is required.

[0102] Viscosity and curing time at different time periods are selected from the dynamic processing parameters, and the initial curing coefficient is calculated. The formula is as follows:

[0103]

[0104] In the formula, It is the initial viscosity. It is the final viscosity. It is the curing time. It is the initial curing coefficient.

[0105] The curing speed coefficient is obtained by adjusting the initial curing coefficient according to the type of adhesive.

[0106] If it is a hot melt adhesive: the effect of temperature on the curing speed needs to be considered.

[0107] If it is a water-based adhesive / other type of adhesive: no additional correction is required.

[0108] The initial material deformation coefficient is calculated by selecting the maximum edge warping height and the adhesive edge thickness from the dynamic processing parameters. The formula is expressed as follows:

[0109]

[0110] In the formula, It is the maximum height of the edge that is raised. It refers to the thickness of the adhesive edge. It is the initial material deformation coefficient.

[0111] The material deformation coefficient is obtained by adjusting the initial material deformation coefficient based on the rigidity of the packaging box material.

[0112] If it is cardboard: Cardboard has high rigidity and small deformation deviation.

[0113] If it is corrugated paper: the corrugated structure is easily deformed, so the actual deformation needs to be magnified.

[0114] If it is a plastic box: it is highly rigid and has minimal deformation.

[0115] By selecting the peel force at different time points from the dynamic processing parameters, the initial stability is calculated, and the formula is expressed as:

[0116]

[0117] In the formula, It is the peeling force at five seconds. That is the peeling force at thirty seconds. It refers to initial stability.

[0118] The interface adaptation stability is obtained by adjusting the initial stability according to the type of adhesive.

[0119] If it is a water-based adhesive: the effect of ambient humidity on adhesion needs to be considered.

[0120] If it is hot melt adhesive / other glue: no additional correction is required.

[0121] The adhesive property correction unit is used to differentiate and correct dynamic characteristic indicators for different packaging box materials and adhesive types to obtain dynamic characteristic data of adhesive materials.

[0122] The steps by which the glue box adapter module obtains glue box matching data include:

[0123] The adhesive performance data and adhesive compatibility characteristic data are simultaneously standardized, and a mapping relationship between the two is established from four dimensions: surface, penetration, adhesive amount and fixation.

[0124] Set hard matching rules based on the type of packaging box material and the usage scenario, and then filter and retain the material types, scenario tags and passing status as the initial adaptation pool.

[0125] The strict matching rules include: material-specific rules: corrugated paper: the adhesive viscosity and curing time must meet the preset standards, otherwise it will be rejected.

[0126] Low surface tension and low moisture absorption laminated color boxes: must meet the requirements of affinity and the adhesive must contain adhesion promoters; otherwise, they will be rejected.

[0127] Paperboard: Sufficient glue must be applied; otherwise, it will be rejected.

[0128] Scenario adaptation rules: Food packaging scenario: The adhesive must pass "food-grade certification" and have no volatile harmful substances after curing. If it does not meet these requirements, it will be rejected directly.

[0129] Outdoor / cold chain scenarios: The adhesive must meet the requirements of "weather resistance from -30℃ to 60℃" and "bonding strength retention rate ≥90% at 90% humidity for 24 hours". If it does not meet these requirements, the basic compatibility score will be deducted.

[0130] Build a machine learning model and train it by collecting historical adhesion quality data. Input the model into the initial adaptation pool and output the historical adhesion quality score.

[0131] A gradient boosting tree model is used, and parameters are optimized through 5-fold cross-validation. The initial matching score is obtained by predicting each combination in the initial adaptation pool.

[0132] The initial matching score is corrected based on the dynamic characteristic data of the adhesive material to obtain the matching data of the adhesive box.

[0133] For example, the adhesive model is J-005, the corresponding packaging box type is three-layer corrugated cardboard, and the material / scene label is express packaging. This combination passed the surface adaptation and penetration adaptation rules, with an initial score of 83 points. Due to the material deformation coefficient in the historical dynamic characteristic data, a dynamic prediction correction deducted 3 points, resulting in a final matching degree of 80 points and an adaptation level of excellent.

[0134] The adhesive model is J-008, the corresponding packaging box type is a laminated color box, and the material / scene label is food packaging. This combination passed the rules for surface compatibility and penetration compatibility, with an initial score of 79 points. Due to the presence of an adhesion promoter in the adhesive, a dynamic prediction correction increased the score by 2 points, resulting in a final matching score of 81 points and a compatibility level of excellent.

[0135] The adhesive material compatibility analysis module is used to analyze the impact of the dynamic characteristics of the adhesive material on the matching data of the adhesive box to obtain the adhesive material compatibility influence coefficient.

[0136] The steps for obtaining the adhesive compatibility influence coefficient by the adhesive compatibility analysis module include:

[0137] Collect adhesive application failures caused by abnormal dynamic properties of adhesive materials, establish a mapping relationship between dynamic properties and failure types, and determine the impact path.

[0138] Adhesive application defects can be categorized as follows:

[0139] Faults related to the shape of the adhesive thread: shrinkage and breakage of the adhesive thread, and expansion and overflow of adhesive thread.

[0140] Curing efficiency related faults: incomplete curing leading to delamination, and curing too quickly leading to insufficient wetting.

[0141] Material deformation type of failure: the edges of the packaging box are curled up, and the adhesive surface is misaligned.

[0142] Persistent interface failures: delamination within 24 hours, decreased adhesive strength after high and low temperature cycling.

[0143] The mapping relationship is as follows: the fault type is adhesive line shrinkage and breakage, and the associated dynamic characteristic parameter is the adhesive line morphology change rate. The abnormal threshold of the parameter is ≤-15%, and the impact on the adhesive box matching data is a loss of 30-40 points in matching degree.

[0144] The fault type is delamination due to excessively rapid curing. The associated dynamic characteristic parameter is the curing speed coefficient, with an abnormal threshold of ≥250 mPa・s / s. The impact on the glue box matching data is a loss of 20-30 points in matching degree.

[0145] The fault type is edge warping, and its associated dynamic characteristic parameter is the material deformation coefficient. The abnormal threshold of the parameter is ≥0.2, and its impact on the matching data of the glue box is a loss of 15-25 points in matching degree.

[0146] The fault type is delamination within 24 hours. The associated dynamic characteristic parameter is interface adaptation stability. The abnormal threshold of the parameter is ≤70%. The impact on the glue box matching data is a loss of 40-50 points in matching degree.

[0147] The impact weight of the failure is determined based on the frequency and severity of the failure, and the correlation strength between dynamic characteristic indicators and adhesive application failure is calculated.

[0148] By combining the current deviation value, fault impact weight, and correlation strength of the dynamic characteristic data of the adhesive material, the impact scores of different dynamic characteristic indicators on the matching data of the adhesive box are calculated, and the initial coefficients are obtained by integrating them.

[0149] Collect real-time monitoring data of the current production batch and calibrate the initial coefficients to obtain the adhesive material compatibility influence coefficient.

[0150] The glue application monitoring and adjustment module is used to build a dynamic monitoring model based on the dynamic characteristics data of the glue material, the matching data of the glue box, and the glue material compatibility influence coefficient. It performs real-time detection of the glue application process and generates parameter adjustment and defect handling instructions.

[0151] The steps for the glue application monitoring and adjustment module to obtain the dynamic monitoring model include:

[0152] The data processing unit is used to unify the format and dimensions of the dynamic characteristic data of adhesive materials, the matching data of adhesive boxes, and the influence coefficient of adhesive material compatibility, and to perform time correlation according to the adhesive batch ID and spatial alignment according to the workstation number to obtain the comprehensive processed data.

[0153] The fusion feature extraction unit is used to extract fusion features of static adaptation basis, dynamic risk and weight adjustment from the comprehensive processed data to construct a multi-dimensional monitoring feature set.

[0154] The steps for obtaining a multi-dimensional monitoring feature set by fusing feature extraction units include:

[0155] The final matching degree in the glue box matching data is converted into a value in the range of [0-1] as the basic score for adaptation, and the historical matching pass rate of different glue and packaging box combinations is calculated as the static adaptation basis.

[0156] The dynamic risk score is obtained by weighting and summing different dynamic characteristic parameters based on the adhesive material compatibility influence coefficient, and then combining the time-series change rate of the dynamic characteristic parameters to obtain the dynamic risk.

[0157] We construct weighted features based on the adhesive material compatibility influence coefficient, adjust the influence ratio of static compatibility basis and dynamic risk to obtain weight adjustment, and cross-integrate with static compatibility basis and dynamic risk to obtain a multi-dimensional monitoring feature set.

[0158] The temporal-static fusion unit is used to collect historical production data as a training set for training, employing a hybrid model architecture that combines temporal prediction and static classification.

[0159] Production data includes normal glue-coating samples, suspicious samples, and unqualified samples. A lightweight LSTM (Long Short-Term Memory) network is used, taking "weighted dynamic characteristic values ​​of the past 10 seconds" as input and outputting "dynamic risk prediction values ​​for the next 2 seconds" to capture the changing trends of dynamic characteristics. XGBoost (Extreme Gradient Boosting Tree) is used, taking "adaptation base score", "adaptation stability feature" and "scenario label" as input and outputting "static adaptation risk level", matching the "real-time risk score of the glue-coating process" with the corresponding label.

[0160] The risk threshold setting unit is used to set risk thresholds for the model output based on a multi-dimensional monitoring feature set and different scenarios to obtain a dynamic monitoring model. An initial defect judgment baseline is set based on the historical pass rate in the glue box matching data. The prediction model is switched by combining the parameter fluctuation trend in the dynamic characteristic data of the glue material. The glue material compatibility influence coefficient is used as a weighting factor to adjust the fusion ratio of static compatibility features and dynamic risk features.

[0161] The steps by which the glue application monitoring and adjustment module receives parameter adjustment and defect handling instructions include:

[0162] Real-time acquisition of dynamic characteristic data of adhesive materials, equipment operating parameters and image data during the adhesive application process, alignment with multi-dimensional monitoring feature sets and input into the dynamic monitoring model, outputting real-time risk status and anomaly contribution analysis.

[0163] If the condition is determined to be normal, the data will be continuously monitored and recorded. If the condition is determined to be suspicious, an early warning will be triggered, and the fluctuation threshold of the dynamic characteristic data of the adhesive material will be extracted. Combined with the adhesive box matching data, a pre-adjustment suggestion will be generated.

[0164] If the condition is determined to be unqualified, the defect type is located based on the abnormal contribution analysis, and parameter adjustment and defect handling instructions are generated in combination with the adhesive material compatibility influence coefficient.

[0165] For parametric defects, generate equipment parameter adjustment instructions. For structural defects, generate defect handling instructions. After the instructions are executed, real-time dynamic characteristic data after adjustment is collected and fed back to the model to verify the handling effect: if the effect is satisfactory, the adjusted parameters are recorded as the optimization benchmark for this scenario. If the effect is not satisfactory, the instructions are iteratively corrected until the risk is eliminated.

[0166] like Figure 2 As shown, based on the same inventive concept, this invention also protects an intelligent detection method for automatic glue application on packaging boxes. The intelligent detection method includes:

[0167] We collected packaging box parameter data and adhesive performance data, and used a geometric material adaptation method to analyze the packaging box parameter data to obtain adhesive adaptation characteristic data.

[0168] The dynamic properties of the adhesive material were obtained by analyzing the changes in the shape and state of the packaging box over time using the colloidal analysis method. The adhesive performance data was then correlated with the adhesive compatibility characteristic data to obtain the box matching data.

[0169] The influence of the dynamic characteristics of the adhesive material on the matching data of the adhesive box is analyzed to obtain the adhesive material adaptation influence coefficient.

[0170] A dynamic monitoring model is constructed based on the dynamic characteristics data of adhesive materials, the matching data of adhesive boxes, and the influence coefficient of adhesive material compatibility. The model is used to monitor the adhesive application process in real time and generate parameter adjustment and defect handling instructions.

[0171] This embodiment provides an intelligent detection system and method for automatic glue application on packaging boxes. Through multi-dimensional static adaptation analysis and dynamic characteristic quantification, combined with a machine learning model, it significantly improves adaptation accuracy and enhances the stability of glue application quality. By using dynamic feature indicators and the glue material adaptation influence coefficient, it provides early warnings of potential faults and defects, greatly improving prevention capabilities. This improves glue application quality and production efficiency, enabling timely detection of problems and defects in the glue application process. It also generates corresponding parameter adjustment and defect handling instructions based on different real-time risk states, allowing for effective dynamic monitoring and adjustment.

[0172] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A packaging box automatic gluing intelligent detection system, characterized in that, The application relates to a glue application adaptive analysis method and device. The glue application adaptive analysis device comprises a glue application adaptive analysis module, a glue box adaptive module and a glue material adaptive analysis module. The glue application adaptive analysis method comprises the following steps: a stage detection unit is used for dividing different key monitoring stages according to glue curing periods and material response characteristics, and collecting dynamic monitoring parameters of glue forms and material states in corresponding stages; a dynamic parameter processing unit is used for carrying out cleaning, smoothing and space-time alignment processing on the dynamic monitoring parameters, eliminating abnormal values and unifying time and space references to obtain dynamic processing parameters; a dynamic characteristic unit is used for calculating glue line form change rates, curing speed coefficients, material deformation coefficients and interface adaptive stabilities as dynamic characteristic indexes based on the dynamic processing parameters; a glue material characteristic correction unit is used for correcting the dynamic characteristic indexes to obtain the glue material dynamic characteristic data according to different packaging box materials and glue types. The dynamic characteristic unit calculates the dynamic characteristic indexes in the following steps: glue line width averages of different key monitoring stages are selected from the dynamic processing parameters, shrinkage or expansion trends in the glue curing process are analyzed, and the glue line form change rates are obtained by adjusting the glue line width averages according to packaging box material types; viscosities and curing times of different time periods are selected from the dynamic processing parameters, initial curing coefficients are calculated, the initial curing coefficients are adjusted according to glue types to obtain the curing speed coefficients; edge maximum lifting heights and bonding edge thicknesses are selected from the dynamic processing parameters, initial material deformation coefficients are calculated, and the initial material deformation coefficients are adjusted according to packaging box material rigidities to obtain the material deformation coefficients; peeling forces of different time points are selected from the dynamic processing parameters, initial stabilities are calculated, and the initial stabilities are adjusted according to the glue types to obtain the interface adaptive stabilities; a glue material adaptive analysis module is used for analyzing influences of the glue material dynamic characteristic data on the glue box matching data to obtain glue material adaptive influence coefficients; a glue application monitoring and adjustment module is used for constructing a dynamic monitoring model based on the glue material dynamic characteristic data, the glue box matching data and the glue material adaptive influence coefficients, carrying out real-time detection on a glue application process, and generating parameter adjustment and defect processing instructions. The geometric material adaptive method comprises the following steps:

2. The automatic gluing and intelligent detection system for packaging boxes according to claim 1, characterized in that, basic sizes and structural forms of packaging boxes are extracted from packaging box parameter data as geometric parameters, and material types of the packaging boxes are determined, and material surface characteristic parameters are combined as material parameters; use scene parameters are obtained according to use scenes and expected circulation environment parameters of the packaging boxes, and glue application positions and manners are determined according to the geometric parameters and the material parameters; the use scene parameters are associated with glue application strengths, the material parameters are associated with glue performances, and the glue application adaptive characteristic data are analyzed and obtained. ​ 3. The automatic gluing and intelligent detection system for packaging boxes according to claim 1, characterized in that, The step of obtaining the glue box matching data by the glue box adaptation module comprises: synchronizing and standardizing the glue performance data and the glue application adaptation feature data, and establishing a mapping relationship between the two after standardization from four dimensions of surface, penetration, glue amount and fixation; setting a hard matching rule in combination with the packaging box material type and use scenario, and performing screening, and retaining the material type, scenario label and passing condition that pass as an initial adaptation pool; constructing a machine learning model, collecting historical bonding quality data for training, inputting the initial adaptation pool, and outputting a historical bonding quality score; adopting a gradient boosting tree model, optimizing parameters through 5-fold cross-validation, and predicting each combination in the initial adaptation pool to obtain an initial matching degree score; based on the glue material dynamic characteristic data, correcting the initial matching degree score to obtain the glue box matching data.

4. The automatic gluing and intelligent detection system for packaging boxes according to claim 1, characterized in that, The step of obtaining the glue material adaptation influence coefficient by the glue material adaptation analysis module comprises: collecting glue application failures caused by abnormal glue material dynamic characteristics, establishing a mapping relationship between dynamic characteristics and failure types, and determining an influence path; determining a failure influence weight based on failure frequency and severity, and calculating the correlation strength of the dynamic characteristic indicators and the glue application failure; combining the current deviation value of the glue material dynamic characteristic data, the failure influence weight and the correlation strength, calculating the influence score of different dynamic characteristic indicators on the glue box matching data, and integrating to obtain an initial coefficient; collecting real-time monitoring data of the current production batch, calibrating the initial coefficient to obtain the glue material adaptation influence coefficient.

5. The automatic gluing and intelligent detection system for packaging boxes according to claim 1, characterized in that, The step of obtaining the dynamic monitoring model by the glue application monitoring and adjustment module comprises: a data comprehensive processing unit, configured to unify the format and dimension of the glue material dynamic characteristic data, the glue box matching data and the glue material adaptation influence coefficient, and perform time correlation according to the glue application batch ID and spatial alignment according to the station number to obtain comprehensive processing data; a fusion feature extraction unit, configured to extract fusion features of static adaptation basis, dynamic risk and weight adjustment from the comprehensive processing data to construct a multi-dimensional monitoring feature set; a time sequence static fusion unit, configured to collect historical production data as a training set for training by adopting a hybrid model architecture of time sequence prediction and static classification; a risk threshold setting unit, configured to set a risk threshold output by the model based on the multi-dimensional monitoring feature set to obtain the dynamic monitoring model.

6. The automatic gluing and intelligent detection system for packaging boxes according to claim 5, characterized in that, The step of obtaining the multi-dimensional monitoring feature set by the fusion feature extraction unit comprises: converting the final matching degree in the glue box matching data into a 0-1 interval value as an adaptation basis score, and counting the historical matching qualification rate of different glue and packaging box combinations as the static adaptation basis; weighting and summing different dynamic characteristic parameters according to the glue material adaptation influence coefficient to obtain a dynamic risk score, and combining the time sequence change rate of the dynamic characteristic parameters to obtain the dynamic risk; Based on the glue material adaptation influence coefficient, a weight feature is constructed, the weight adjustment is obtained by adjusting the influence proportion of the static adaptation basis and the dynamic risk, and the multi-dimensional monitoring feature set is obtained by cross-fusing the static adaptation basis and the dynamic risk.

7. The automatic gluing and intelligent detection system for packaging boxes according to claim 6, characterized in that, The step of obtaining the parameter adjustment and defect processing instruction by the glue coating monitoring and adjusting module includes: Real-time acquisition of glue material dynamic characteristic data, equipment operating parameters and image data during the glue coating process, alignment with the multi-dimensional monitoring feature set, input into the dynamic monitoring model, output of real-time risk state and abnormal contribution degree analysis; If it is determined to be a normal state, the data is continuously monitored and recorded, if it is determined to be a suspicious state, a warning is triggered, and the fluctuation threshold of the glue material dynamic characteristic data is extracted, and a pre-adjustment suggestion is generated in combination with the glue box matching data; If it is determined to be an unqualified state, the defect type is located based on the abnormal contribution degree analysis, and the parameter adjustment and defect processing instruction are generated in combination with the glue material adaptation influence coefficient.

8. A method for automatically detecting the intelligent detection of a packaging box, which uses the intelligent detection system for automatically detecting the intelligent detection of a packaging box according to any one of claims 1 to 7, characterized in that, The intelligent detection method includes: Collecting packaging box parameter data and glue performance data, using a geometric material adaptation method to analyze the packaging box parameter data to obtain glue coating adaptation feature data; Using a colloidal analysis method to analyze the shape and state change of the glue over time to obtain glue material dynamic characteristic data, correlating the glue performance data and the glue coating adaptation feature data to obtain glue box matching data; Analyzing the influence degree of the glue material dynamic characteristic data on the glue box matching data to obtain a glue material adaptation influence coefficient; Based on the glue material dynamic characteristic data, the glue box matching data and the glue material adaptation influence coefficient, a dynamic monitoring model is constructed, and the glue coating process is detected in real time to generate parameter adjustment and defect processing instructions.

Images