UV hot melt adhesive transfer coating quality monitoring system

Abstract

The invention relates to the technical field of hot melt adhesives, in particular to a UV hot melt adhesive transfer coating quality monitoring system. Firstly, a base material image and a coating image collected by a camera are obtained; then, respectively inputting the base material image and the coating image into a base material feature recognition model and a coating feature recognition model for recognition, and calculating to obtain a first thickness uniformity monitoring value and a second defect feature monitoring value; the thickness uniformity monitoring value and the defect characteristic monitoring value are used for evaluating the thickness uniformity degree and the defect degree, so that the accuracy and the reliability of UV hot melt adhesive transfer coating quality monitoring are improved; thirdly, comprehensively evaluating the first thickness uniformity monitoring value and the second defect characteristic monitoring value to obtain a coating quality monitoring value; and finally, comparing the coating quality monitoring value with a coating quality monitoring threshold value, if the coating quality monitoring value exceeds the threshold value, judging that the coating quality monitoring value is unqualified, otherwise, generating colloid temperature monitoring information according to a data result of the colloid temperature prediction model.

Description

Technical Field

[0001] The invention relates to the technical field of hot melt adhesives, in particular to a UV hot melt adhesive transfer coating quality monitoring system. Background Art

[0002] Ordinary hot melt adhesives have poor adhesion, insufficient heat resistance, limited bonding strength, and poor chemical resistance. These defects have restricted the performance of ordinary hot melt adhesive coating methods. Therefore, UV hot melt adhesive transfer coating technology has been used instead. This technology combines UV curing technology and hot melt adhesive technology to quickly cure hot melt adhesive through ultraviolet (UV) radiation, thereby achieving an efficient and environmentally friendly coating process. Compared with ordinary hot melt adhesives, UV hot melt adhesives have excellent bonding strength, optical transparency, easy flowability, high temperature resistance and aging resistance, so UV hot melt adhesive coating technology is widely used in various fields.

[0003] However, the existing technology still has some shortcomings in UV hot melt adhesive transfer coating technology. On the one hand, when using UV light-curing adhesives in the coating process, the factors that lead to uneven coating thickness are not considered enough, which will reduce the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring; on the other hand, the existing technology does not consider the defect characteristics of coating, and does not adjust the temperature of the colloid according to the actual situation, which will reduce the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0004] Therefore, a UV hot melt adhesive transfer coating quality monitoring system is proposed. Summary of the invention

[0005] The purpose of the present invention is to provide a UV hot melt adhesive transfer coating quality monitoring system. First, a substrate image and a coating image captured by a camera are obtained; then, the substrate image and the coating image are respectively input into a substrate feature recognition model and a coating feature recognition model for recognition, and a first thickness uniformity monitoring value and a second defect feature monitoring value are calculated; the thickness uniformity monitoring value and the defect feature monitoring value are used to evaluate the thickness uniformity and defect degree, so as to improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring; then, the first thickness uniformity monitoring value and the second defect feature monitoring value are comprehensively evaluated to obtain a coating quality monitoring value; finally, the coating quality monitoring value is compared with a coating quality monitoring threshold value, and if it exceeds the threshold value, it is judged as unqualified, otherwise, the colloid temperature monitoring information is generated according to the data results of the colloid temperature prediction model.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A UV hot melt adhesive transfer coating quality monitoring system, comprising: an equipment control module, a data acquisition module, a data monitoring module and an output module;

[0008] The device control module is used to control the start, pause and stop of the device;

[0009] The data acquisition module is used to obtain image data collected by the camera; wherein the image data includes: substrate image and coating image;

[0010] The data monitoring module is used to identify and process the image data to obtain monitoring data; wherein the data monitoring module includes: a substrate roughness index calculation unit, a coating thickness uniformity monitoring unit and a coating defect feature monitoring unit;

[0011] The substrate roughness index calculation unit processes the substrate image using a substrate feature recognition model to obtain a substrate roughness index; the coating thickness uniformity monitoring unit evaluates the substrate roughness index, thickness feature, coating speed, coating pressure and coating amount to obtain a first thickness uniformity monitoring value; the coating defect feature monitoring unit evaluates bubble feature, crack feature and concave feature to obtain a second defect feature monitoring value;

[0012] The output module is used to evaluate the monitoring data, output monitoring results and provide monitoring feedback; wherein, the monitoring output module includes: a judgment unit and a colloid temperature monitoring unit; wherein, the colloid temperature monitoring unit inputs the substrate monitoring data and the coating monitoring data into the colloid temperature prediction model, and generates colloid temperature monitoring information according to the data results.

[0013] Furthermore, the substrate roughness index calculation unit processes the substrate image using the substrate feature recognition model to obtain the substrate roughness index, and the specific implementation process includes:

[0014] Inputting the substrate image into the substrate feature recognition model for recognition to obtain substrate features;

[0015] According to the substrate characteristics, a gray value of the substrate texture characteristics and an area of ​​the substrate flattened region are obtained;

[0016] Compare the characteristic grayscale value of the substrate texture and the area of ​​the flat region of the substrate with the total grayscale value and the total area, respectively, to calculate the substrate roughness index;

[0017] The calculation formula of the substrate roughness index is:

[0018]

[0019] Wherein, cci represents the substrate roughness index; n represents the number of regions divided into substrate images; arctan represents the hyperbolic tangent function; α1 represents the weight factor of the grayscale value ratio of the substrate image texture feature; G (w,i)Represented as the texture feature grayscale value of the i-th region; G i It is represented as the total gray value of the ith region; α2 is represented as the weight factor of the rough area ratio of the substrate image; A i A is the total area of ​​the ith region; (p,i) Expressed as the flattened area of ​​the ith region.

[0020] Furthermore, the coating thickness uniformity monitoring unit evaluates the substrate roughness index, thickness characteristics, coating speed, coating pressure and coating amount to obtain a first thickness uniformity monitoring value, and the specific implementation process includes:

[0021] Obtain coating thickness features output by a coating feature recognition model;

[0022] Obtaining a substrate roughness index calculated according to an output result of a substrate feature recognition model;

[0023] The coating thickness feature, the substrate roughness index and the external monitoring value are weighted with a coating thickness feature weight factor, a substrate roughness index weight factor and an external monitoring value weight factor, respectively, to obtain a plurality of evaluation scores;

[0024] Performing a comprehensive evaluation based on the plurality of evaluation scores to calculate a first thickness uniformity monitoring value;

[0025] The calculation formula of the first thickness uniformity monitoring value is:

[0026]

[0027] Where TUJC is the first thickness uniformity monitoring value; M is the number of monitoring data; β1 is the weight factor of the external monitoring value; WBJC is k It is represented by the external monitoring value of the kth monitoring data; β2 is represented by the weight factor of the substrate roughness index; cci k Expressed as the substrate roughness index of the kth monitoring data; cci th It is represented as the substrate roughness index threshold; β3 is represented as the weight factor of thickness feature; hd k Represents the thickness information of the kth monitoring data; hd th It is expressed as the thickness standard value; wherein the calculation formula of the external monitoring value is:

[0028]

[0029] Wherein, WBJC represents the external monitoring value; N represents the number of regions divided into the substrate image; Expressed as a weighting factor for coating speed; tbv j Expressed as the coating speed of the jth area; tbvth Expressed as the standard value of the coating speed; Expressed as a weighting factor for coating pressure; tbv j Expressed as the coating pressure of the jth region; tbv th Expressed as the standard value of the coating pressure; Expressed as the weight factor of the glue coating amount; tbv j Expressed as the glue amount of the jth area; tbv th It is indicated as the standard value of the glue coating amount.

[0030] Furthermore, the coating defect feature monitoring unit evaluates the bubble feature, the crack feature and the depression feature to obtain the second defect feature monitoring value, and the specific implementation process includes:

[0031] Inputting the coating image into the coating feature recognition model for recognition processing;

[0032] Acquire coating defect features output by the coating feature recognition model; wherein the coating defect features include bubble features, crack features and depression features;

[0033] Acquire bubble information, crack information and depression information from the bubble feature, the crack feature and the depression feature respectively;

[0034] The bubble information, the crack information and the depression information are weighted with the bubble characteristic coefficient, the crack characteristic coefficient and the depression characteristic coefficient respectively to calculate and obtain two defect characteristic monitoring values;

[0035] The calculation formula of the second defect characteristic monitoring value is:

[0036]

[0037] Where, QXJC represents the second defect characteristic monitoring value; D represents the number of defect characteristic samples; γ1 represents the bubble characteristic coefficient; qp m It is expressed as the maximum number of bubbles in the mth sample; It is represented by the diameter of the j1th bubble of the mth sample; γ2 is represented by the crack characteristic coefficient; lh m It is expressed as the maximum number of cracks in the mth sample; It is expressed as the length of the j2th crack of the mth sample; It is represented by the width of the j2nd crack of the mth sample; γ3 is represented by the concave characteristic coefficient; ax m It is expressed as the maximum number of depressions of the mth sample; It is expressed as the area of ​​the j3rd concavity of the mth sample; It is expressed as the depth of the j3th depression of the mth sample.

[0038] Furthermore, the colloid temperature monitoring unit inputs the substrate monitoring data and the coating monitoring data into the colloid temperature prediction model, and the specific process of generating colloid temperature monitoring information according to the data results includes:

[0039] Obtain real-time substrate information from substrate monitoring data;

[0040] Obtain real-time coating information and real-time glue information from coating monitoring data;

[0041] Inputting historical substrate information, historical coating information and historical glue coating information in the historical monitoring database into the colloid temperature prediction model for training to obtain temperature prediction parameters;

[0042] Input the temperature prediction parameters, the real-time substrate information, the real-time coating information and the real-time glue coating information into the colloid temperature prediction model, update the model parameters, and output the real-time temperature parameters; wherein the real-time temperature parameters include: real-time substrate temperature parameters, real-time coating temperature parameters and real-time glue coating temperature parameters;

[0043] Comprehensively evaluate the real-time information and the real-time temperature parameters to obtain colloid temperature monitoring information;

[0044] The calculation formula of the colloid temperature monitoring information is:

[0045]

[0046] Wherein, JTJC represents the colloid temperature monitoring information; P represents the number of data samples; It is represented as the real-time substrate temperature parameter of the i-th sample; INF (i,jc) Represented as the real-time substrate information of the i-th sample; It is represented as the real-time coating temperature parameter of the i-th sample; INF (i,tb) Represented as the real-time coating information of the i-th sample; It is represented as the real-time glue coating temperature parameter of the i-th sample; INF (i,tj) Represented as the real-time gluing information of the i-th sample.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. The present invention proposes a thickness uniformity monitoring function for monitoring the thickness variation of coating; the function evaluates the substrate roughness index, thickness characteristics, coating speed, coating pressure and glue coating amount, and uses the calculated first thickness uniformity monitoring value to record the thickness uniformity of coating; wherein, the substrate roughness index is obtained by processing the substrate image using a substrate feature recognition model, and this parameter reflects the roughness of the substrate surface; the thickness characteristics are obtained by processing the coating image using a coating feature recognition model; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0049] 2. The present invention proposes a defect feature monitoring function for feature evaluation of coating defects; this function compares the defect feature monitoring data output by the coating feature recognition model with the defect feature monitoring threshold, and uses the calculated second defect feature monitoring value to record the degree of coating defects; wherein, the defect feature monitoring data includes bubble features, crack features and depression features; the third defect feature monitoring value is calculated by evaluating the defect features output by the coating feature recognition model; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0050] 3. The present invention proposes a colloid temperature monitoring function for monitoring the influence of the real-time monitoring temperature on the colloid temperature; the function uses historical monitoring data to train the colloid temperature prediction model to obtain temperature prediction parameters; the model parameters are updated using the temperature prediction parameters and real-time monitoring data to obtain real-time temperature parameters; the function evaluates the real-time monitoring information and the real-time temperature parameters to obtain colloid temperature monitoring information; the colloid temperature monitoring information reflects the influence of the real-time monitoring temperature on the colloid temperature; the function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the process of a UV hot melt adhesive transfer coating quality monitoring system of the present invention;

[0052] Figure 2 It is a schematic diagram of the coating environment of the present invention;

[0053] Figure 3 It is a structural schematic diagram of the coating feature recognition model of the present invention;

[0054] Figure 4 The present invention is a schematic diagram of a method for implementing a UV hot melt adhesive transfer coating quality monitoring system. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Ordinary hot melt adhesives have poor adhesion, insufficient heat resistance, limited bonding strength, and poor chemical resistance. These defects restrict the performance of ordinary hot melt adhesive coating methods. Therefore, UV hot melt adhesive transfer coating technology has been used instead. This technology combines UV curing technology and hot melt adhesive technology to quickly cure hot melt adhesive through ultraviolet (UV) radiation, thereby achieving an efficient and environmentally friendly coating process. Compared with ordinary hot melt adhesives, UV hot melt adhesives have excellent bonding strength, optical transparency, easy flowability, high temperature resistance, aging resistance and other characteristics. Therefore, UV hot melt adhesive coating technology is widely used in various fields.

[0057] However, the existing technology still has some shortcomings in UV hot melt adhesive transfer coating technology. On the one hand, when using UV light-curing adhesives in the coating process, the factors that lead to uneven coating thickness are not considered enough, which will reduce the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring; on the other hand, the existing technology does not consider the defect characteristics of coating, and does not adjust the temperature of the colloid according to the actual situation, which will reduce the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0058] Embodiment 1

[0059] The specific implementation process in the embodiment of the present application will be implemented by a UV hot melt adhesive transfer coating quality monitoring system of the present invention, see Figure 1 The process description of the system proposed in the present invention is as follows; A UV hot melt adhesive transfer coating quality monitoring system includes: an equipment control module, a data acquisition module, a data monitoring module and an output module;

[0060] The device control module is used to control the start, pause and stop of the device; the device includes a shooting device, a coating device, a data processing device and a data storage device.

[0061] The data acquisition module is used to obtain image data collected by a camera, external data captured by a sensor, and historical monitoring data in a historical monitoring database; wherein the image data includes an image of a substrate before coating and an image of a coating after coating.

[0062] The data monitoring module is used to identify and process the image data to obtain monitoring data; wherein the data monitoring module includes: a substrate roughness index calculation unit, a coating thickness uniformity monitoring unit and a coating defect feature monitoring unit;

[0063] The substrate roughness index calculation unit is used to monitor the roughness of the substrate; the unit processes the substrate image using the substrate feature recognition model to obtain the substrate roughness index; the specific implementation process of the substrate roughness index includes:

[0064] Inputting the substrate image into the substrate feature recognition model for recognition to obtain substrate features;

[0065] The substrate feature recognition model in this embodiment is a deep learning model based on a convolutional neural network, and mainly includes: an input layer, a feature extraction layer, a substrate feature marking layer, a substrate feature recognition layer and an output layer; the input layer converts the substrate image from the image space to the feature space to facilitate subsequent feature processing; the feature extraction layer uses convolution operations of different sizes to extract multi-scale features and fuses the multi-scale features through channel merging operations; the substrate feature marking layer is used to obtain an anchor frame of substrate features; the substrate feature recognition layer recognizes target features according to the anchor frame; the output layer is used to output substrate features; wherein the substrate features include texture features and flat area features.

[0066] Further, the gray value of the substrate texture feature and the area of ​​the substrate flattened region are obtained according to the substrate feature;

[0067] Further, the characteristic gray value of the substrate texture and the area of ​​the flat region of the substrate are compared with the total gray value and the total area respectively, and the substrate roughness index is calculated;

[0068] Furthermore, the calculation formula of the substrate roughness index is:

[0069]

[0070] α1+α2=1.0;

[0071] Wherein, cci represents the substrate roughness index; n represents the number of regions divided into substrate images; arctan represents the hyperbolic tangent function; α1 represents the weight factor of the grayscale value ratio of the substrate image texture feature; G (w,i) Represented as the texture feature grayscale value of the i-th region; G i It is represented as the total gray value of the ith region; α2 is represented as the weight factor of the rough area ratio of the substrate image; A i A is the total area of ​​the ith region; (p,i) It is expressed as the flattened area of ​​the ith region.

[0072] In this embodiment, the number of divided substrate image regions is set to 4, and the weight values ​​of α1 and α2 are set to 0.4 and 0.6 respectively; wherein, the number of divided regions and the weight factor settings can be adjusted according to actual conditions.

[0073] In this embodiment, a substrate roughness index is proposed for monitoring the uniformity of coating thickness; the index uses the grayscale value of the substrate surface texture and the proportion of the rough area to reflect the roughness of the substrate surface; by combining the roughness of the substrate before coating, the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring can be improved.

[0074] The coating thickness uniformity monitoring unit is used to monitor the coating thickness uniformity change;

[0075] The specific implementation process of the coating thickness uniformity monitoring unit in this embodiment includes: first, obtaining the external data, substrate roughness index data and coating thickness data; wherein the coating thickness data is obtained by identifying the coating image through a coating feature recognition model, wherein the structure of the coating feature recognition model is as follows: Figure 2 The external data includes: coating speed, coating pressure and glue coating amount; then, the substrate roughness index information, thickness information, coating speed information, coating pressure information and glue coating amount information are evaluated to obtain a first thickness uniformity monitoring value.

[0076] Wherein, the calculation formula of the first thickness uniformity monitoring value is:

[0077]

[0078] Where TUJC is the first thickness uniformity monitoring value; M is the number of monitoring data; β1 is the weight factor of the external monitoring value; WBJC is k It is represented as the external monitoring value of the kth monitoring data; β2 is represented as the weight factor of the substrate roughness index; cci k Expressed as the substrate roughness index of the kth monitoring data; cci th It is represented as the substrate roughness index threshold; β3 is represented as the weight factor of thickness feature; hd k Represents the thickness information of the kth monitoring data; hd th Expressed as the thickness standard value; wherein the calculation formula of the external monitoring value is:

[0079]

[0080] Wherein, WBJC represents the external monitoring value; N represents the number of regions divided into the substrate image; Expressed as a weighting factor for coating speed; tbv j Expressed as the coating speed of the jth area; tbv th Expressed as the standard value of the coating speed; Expressed as a weighting factor for coating pressure; tbv j Expressed as the coating pressure of the jth region; tbv th Expressed as the standard value of the coating pressure; Expressed as the weight factor of the glue coating amount; tbv j Expressed as the glue amount of the jth area; tbv th It is expressed as the standard value of the glue coating amount.

[0081] The coating defect feature monitoring unit is used to monitor the degree of coating defects; the unit evaluates the coating defect feature output by the coating feature recognition model to obtain a second defect feature monitoring value; the specific implementation process of the second defect feature monitoring value includes:

[0082] Acquire coating defect features output by a coating feature recognition model; wherein the coating defect features include bubble features, crack features, and depression features;

[0083] Further, bubble information, crack information and depression information are obtained from the bubble feature, the crack feature and the depression feature respectively;

[0084] Further, the bubble information, the crack information and the depression information are weighted with the bubble characteristic coefficient, the crack characteristic coefficient and the depression characteristic coefficient respectively to calculate and obtain two defect characteristic monitoring values;

[0085] Furthermore, the calculation formula of the second defect characteristic monitoring value is:

[0086]

[0087] γ1+γ2+γ3=1.0;

[0088] Where, QXJC represents the second defect characteristic monitoring value; D represents the number of defect characteristic samples; γ1 represents the bubble characteristic coefficient; qp m It is expressed as the maximum number of bubbles in the mth sample; It is represented by the diameter of the j1th bubble of the mth sample; γ2 is represented by the crack characteristic coefficient; lh m It is expressed as the maximum number of cracks in the mth sample; It is expressed as the length of the j2th crack of the mth sample; It is represented by the width of the j2nd crack of the mth sample; γ3 is represented by the concave characteristic coefficient; ax mIt is expressed as the maximum number of depressions of the mth sample; It is expressed as the area of ​​the j3rd concavity of the mth sample; It is expressed as the depth of the j3th depression of the mth sample.

[0089] In this embodiment, the bubble characteristic coefficient, crack characteristic coefficient and depression characteristic coefficient are set according to the different degrees of influence on the coating monitoring quality, and are set to 0.35, 0.35 and 0.3 respectively; the settings of different characteristic coefficients can be adjusted according to actual conditions.

[0090] In this embodiment, a defect feature monitoring function is proposed for feature evaluation of coating defects; this function compares the defect feature monitoring data output by the coating feature recognition model with the defect feature monitoring threshold, and uses the calculated second defect feature monitoring value to record the degree of coating defects; wherein, the defect feature monitoring data includes bubble features, crack features and depression features; the third defect feature monitoring value is calculated by evaluating the defect features output by the coating feature recognition model; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0091] The output module is used to evaluate the monitoring data, output monitoring results and provide monitoring feedback; wherein the monitoring output module includes: a judgment unit and a colloid temperature monitoring unit.

[0092] In this embodiment, the judgment unit first performs a weighted evaluation on the first thickness uniformity monitoring value and the second defect characteristic monitoring value to obtain a coating quality monitoring value; then, the coating quality monitoring value is compared with the coating quality threshold. If the threshold is exceeded, the result is directly output as an unqualified product; otherwise, the colloid temperature monitoring unit continues to execute.

[0093] In this embodiment, the coating quality threshold is set to 0.95. The threshold may be related to factors such as material, working environment, and equipment conditions. Therefore, the threshold setting allows relevant personnel to adjust it according to actual conditions.

[0094] The colloid temperature monitoring unit is used to monitor the influence of real-time temperature information on the colloid temperature; the specific implementation process of the unit includes:

[0095] Obtain real-time substrate information from substrate monitoring data;

[0096] Obtain real-time coating information and real-time glue information from coating monitoring data;

[0097] Inputting historical substrate information, historical coating information and historical glue coating information in the historical monitoring database into the colloid temperature prediction model for training to obtain temperature prediction parameters;

[0098] Input the temperature prediction parameters, the real-time substrate information, the real-time coating information and the real-time glue coating information into the colloid temperature prediction model, update the model parameters, and output the real-time temperature parameters; wherein the real-time temperature parameters include: real-time substrate temperature parameters, real-time coating temperature parameters and real-time glue coating temperature parameters;

[0099] The real-time information and the real-time temperature parameters are comprehensively evaluated to obtain colloid temperature monitoring information; the calculation formula of the colloid temperature monitoring information is:

[0100]

[0101] Wherein, JTJC represents the colloid temperature monitoring information; P represents the number of data samples; INF is the real-time substrate temperature parameter of the i-th sample; (i,jc) The real-time substrate information represented as the i-th sample; INF is the real-time coating temperature parameter of the i-th sample; (i,tb) The real-time coating information represented as the i-th sample; INF is the real-time gluing temperature parameter of the i-th sample; (i,tj) Represented as the real-time gluing information of the i-th sample.

[0102] In this embodiment, a UV hot melt adhesive transfer coating quality monitoring system is proposed to monitor the coating quality. The entire coating environment can be referred to Figure 3 ; The system analyzes the state images before and after coating, and uses the feature recognition model to identify and process the substrate image and the coating image to obtain the substrate feature and the coating feature; the system monitors the coating quality from two perspectives: thickness uniformity and coating defect degree, and obtains the coating quality monitoring value by comprehensively evaluating the calculated first thickness uniformity monitoring value and the second coating defect feature monitoring value; then, the coating quality monitoring value is compared with the coating quality threshold, and the monitoring result is output; the system can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0103] Embodiment 2

[0104] As an embodiment of the present invention, refer to Figure 4 , the implementation method of the UV hot melt adhesive transfer coating quality monitoring system includes:

[0105] S10. Acquire the substrate image and coating image captured by the camera;

[0106] S20. Obtaining historical monitoring data from the historical coating database;

[0107] S30. Inputting the substrate image into a substrate feature recognition model to calculate a substrate roughness index;

[0108] S40. Inputting the coating image into a coating feature recognition model to obtain coating thickness features and coating defect features;

[0109] S50. Evaluate the substrate roughness index, the coating thickness characteristic, and the coating defect characteristic to obtain a first thickness uniformity monitoring value and a second defect characteristic monitoring value, respectively;

[0110] S60. Comprehensively evaluate the first thickness uniformity monitoring value and the second defect characteristic monitoring value to obtain a coating quality monitoring value;

[0111] S70. If the coating quality monitoring value exceeds a preset threshold, it is judged as unqualified; otherwise, the substrate monitoring data and the coating monitoring data are input into the colloid temperature prediction model, and the colloid temperature monitoring information is generated according to the data results.

[0112] Furthermore, the specific implementation process of a UV hot melt adhesive transfer coating quality monitoring system is as follows:

[0113] Acquire the substrate image and coating image captured by the camera;

[0114] Processing the substrate image using a substrate feature recognition model to obtain a substrate roughness index;

[0115] The coating image is processed using a coating feature recognition model to obtain thickness features, bubble features, crack features, and depression features;

[0116] Evaluating the substrate roughness index, the thickness characteristics, coating speed, coating pressure and coating amount to obtain a first thickness uniformity monitoring value;

[0117] Evaluating the bubble feature, the crack feature, and the depression feature to obtain a second defect feature monitoring value;

[0118] The first thickness uniformity monitoring value and the second defect characteristic monitoring value are weightedly evaluated to obtain a coating quality monitoring value; the calculation formula of the coating quality monitoring value is:

[0119] ZLJC=ω*HDJC+(1-ω)*QXJC;

[0120] Wherein, ZLJC represents the coating quality monitoring value; ω represents the assigned weight value; HDJC represents the first thickness uniformity monitoring value; QXJC represents the second defect characteristic monitoring value;

[0121] The coating quality monitoring value is compared with the coating quality threshold. If it exceeds the threshold, it is judged as unqualified; otherwise, the substrate monitoring data and the coating monitoring data are input into the colloid temperature prediction model, and the colloid temperature monitoring information is generated according to the data results.

[0122] For specific explanation, the present invention is described in conjunction with the following embodiments, as follows:

[0123] Acquire the substrate image and coating image captured by the camera;

[0124] In this embodiment, the substrate image captured by the camera is represented as a state image before coating, and the coating image is represented as a state image after coating.

[0125] Furthermore, the substrate image is processed using a substrate feature recognition model to obtain a substrate roughness index; wherein the specific implementation process of the substrate roughness index includes:

[0126] Inputting the substrate image into the substrate feature recognition model for recognition to obtain substrate features;

[0127] The substrate feature recognition model in this embodiment is a deep learning model based on a convolutional neural network, and mainly includes: an input layer, a feature extraction layer, a substrate feature marking layer, a substrate feature recognition layer and an output layer; the input layer converts the substrate image from the image space to the feature space to facilitate subsequent feature processing; the feature extraction layer uses convolution operations of different sizes to extract multi-scale features and fuses the multi-scale features through channel merging operations; the substrate feature marking layer is used to obtain an anchor frame of substrate features; the substrate feature recognition layer recognizes target features according to the anchor frame; the output layer is used to output substrate features; wherein the substrate features include texture features and flat area features.

[0128] Further, the gray value of the substrate texture feature and the area of ​​the substrate flattened region are obtained according to the substrate feature;

[0129] Further, the characteristic gray value of the substrate texture and the area of ​​the flat region of the substrate are compared with the total gray value and the total area respectively, and the substrate roughness index is calculated;

[0130] Furthermore, the calculation formula of the substrate roughness index is:

[0131]

[0132] α1+α2=1.0;

[0133] Wherein, cci represents the substrate roughness index; n represents the number of regions divided into substrate images; arctan represents the hyperbolic tangent function; α1 represents the weight factor of the grayscale value ratio of the substrate image texture feature; G (w,i) Represented as the texture feature grayscale value of the i-th region; G i It is represented as the total gray value of the ith region; α2 is represented as the weight factor of the rough area ratio of the substrate image; A i A is the total area of ​​the ith region; (p,i) It is expressed as the flattened area of ​​the ith region.

[0134] In this embodiment, the number of divided substrate image regions is set to 4, and the weight values ​​of α1 and α2 are set to 0.4 and 0.6 respectively; wherein, the number of divided regions and the weight factor settings can be adjusted according to actual conditions.

[0135] In this embodiment, a substrate roughness index is proposed for monitoring the uniformity of coating thickness; the index uses the grayscale value of the substrate surface texture and the proportion of the rough area to reflect the roughness of the substrate surface; by combining the roughness of the substrate before coating, the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring can be improved.

[0136] Further, the coating image is processed using a coating feature recognition model to obtain thickness features, bubble features, crack features and depression features;

[0137] The coating feature recognition model in this embodiment is a deep learning model based on a convolutional neural network, including: an input layer, a feature extraction layer, a coating feature marking layer, a coating feature recognition layer and an output layer; the input layer converts the coating image from the image space to the feature space to facilitate subsequent feature processing; the feature extraction layer uses convolution operations of different sizes to extract multi-scale features and fuses the multi-scale features through channel merging operations; the coating feature marking layer is used to obtain an anchor frame of the coating feature; the coating feature recognition layer recognizes the target feature according to the anchor frame; the output layer is used to output the coating feature.

[0138] Further, the substrate roughness index, thickness characteristics, coating speed, coating pressure and coating amount are evaluated to obtain a first thickness uniformity monitoring value; the specific implementation process of the first thickness uniformity monitoring value includes:

[0139] Obtain external monitoring data from sensors;

[0140] Furthermore, the external monitoring data is evaluated to obtain an external monitoring value; wherein the external monitoring data includes coating speed, coating pressure and glue coating amount; the calculation formula of the external monitoring value is:

[0141]

[0142] Wherein, WBJC represents the external monitoring value; N represents the number of regions divided into the substrate image; Expressed as the weighting factor of the coating speed; tbv j Denoted as the coating speed of the jth zone; tbv th Expressed as the standard value of the coating speed; Expressed as the weight factor of the coating pressure; tbv j Denoted as the coating pressure of the jth region; tbv th Expressed as the standard value of the coating pressure; Expressed as the weight factor of the glue coating amount; tbv j It is expressed as the glue amount of the jth area; tbv th Expressed as the standard value of the glue coating amount;

[0143] Further, the substrate roughness index, thickness characteristics and the external monitoring value are evaluated to obtain a plurality of evaluation scores;

[0144] Further, a comprehensive evaluation is performed based on the multiple evaluation scores to obtain a first thickness uniformity monitoring value; wherein the calculation formula of the first thickness uniformity monitoring value is:

[0145]

[0146] β1+β2+β3=1.0;

[0147] Wherein, TUJC represents the first thickness uniformity monitoring value; M represents the number of monitoring data; β1 represents the weight factor of the external monitoring value; WBJC k represents the external monitoring value of the kth monitoring data; β2 represents the weight factor of the substrate roughness index; cci k cci is the substrate roughness index of the kth monitoring data; th It is represented as the substrate roughness index threshold; β3 is represented as the weight factor of the thickness feature; hd k Represents the thickness information of the kth monitoring data; hd th Expressed as standard thickness value.

[0148] In this embodiment, the weight factors assigned to the external monitoring value, substrate roughness index and thickness information are set according to the different degrees of influence on the coating thickness uniformity, and are set to 0.3, 0.3 and 0.4 respectively; the allocation of weight factors can be adjusted according to actual conditions.

[0149] In this embodiment, a thickness uniformity monitoring function is proposed to monitor the thickness change of the coating; this function evaluates the substrate roughness index, thickness characteristics, coating speed, coating pressure and glue coating amount, and uses the calculated first thickness uniformity monitoring value to record the thickness uniformity of the coating; wherein, the substrate roughness index is obtained by processing the substrate image using a substrate feature recognition model, and this parameter reflects the roughness of the substrate surface; the thickness characteristics are obtained by processing the coating image using a coating feature recognition model; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0150] Further, the bubble feature, the crack feature and the depression feature are evaluated to obtain a second defect feature monitoring value; wherein the specific implementation process of the second defect feature monitoring value includes:

[0151] Acquire coating defect features output by a coating feature recognition model; wherein the coating defect features include bubble features, crack features, and depression features;

[0152] Further, bubble information, crack information and depression information are obtained from the bubble feature, the crack feature and the depression feature respectively;

[0153] Further, the bubble information, the crack information and the depression information are weighted with the bubble characteristic coefficient, the crack characteristic coefficient and the depression characteristic coefficient respectively to calculate and obtain two defect characteristic monitoring values;

[0154] Furthermore, the calculation formula of the second defect characteristic monitoring value is:

[0155]

[0156] γ1+γ2+γ3=1.0;

[0157] Where, QXJC represents the second defect characteristic monitoring value; D represents the number of defect characteristic samples; γ1 represents the bubble characteristic coefficient; qp m It is expressed as the maximum number of bubbles in the mth sample; It is represented by the diameter of the j1th bubble of the mth sample; γ2 is represented by the crack characteristic coefficient; lh m It is expressed as the maximum number of cracks in the mth sample; It is expressed as the length of the j2th crack of the mth sample; It is represented by the width of the j2nd crack of the mth sample; γ3 is represented by the concave characteristic coefficient; ax m It is expressed as the maximum number of depressions of the mth sample; It is expressed as the area of ​​the j3rd concavity of the mth sample; It is expressed as the depth of the j3th depression of the mth sample.

[0158] In this embodiment, the bubble characteristic coefficient, crack characteristic coefficient and depression characteristic coefficient are set according to the different degrees of influence on the coating monitoring quality, and are set to 0.35, 0.35 and 0.3 respectively; the settings of different characteristic coefficients can be adjusted according to actual conditions.

[0159] In this embodiment, a defect feature monitoring function is proposed for feature evaluation of coating defects; this function compares the defect feature monitoring data output by the coating feature recognition model with the defect feature monitoring threshold, and uses the calculated second defect feature monitoring value to record the degree of coating defects; wherein, the defect feature monitoring data includes bubble features, crack features and depression features; the third defect feature monitoring value is calculated by evaluating the defect features output by the coating feature recognition model; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0160] Furthermore, the substrate monitoring data and the coating monitoring data are input into the colloid temperature prediction model, and the specific process of generating the colloid temperature monitoring information according to the data results includes:

[0161] Obtain real-time substrate information from substrate monitoring data;

[0162] Obtain real-time coating information and real-time glue information from coating monitoring data;

[0163] Inputting historical substrate information, historical coating information and historical glue coating information in the historical monitoring database into the colloid temperature prediction model for training to obtain temperature prediction parameters;

[0164] Input the temperature prediction parameters, the real-time substrate information, the real-time coating information and the real-time glue coating information into the colloid temperature prediction model, update the model parameters, and output the real-time temperature parameters; wherein the real-time temperature parameters include: real-time substrate temperature parameters, real-time coating temperature parameters and real-time glue coating temperature parameters;

[0165] The real-time information and the real-time temperature parameters are comprehensively evaluated to obtain colloid temperature monitoring information; the calculation formula of the colloid temperature monitoring information is:

[0166]

[0167] Wherein, JTJC represents the colloid temperature monitoring information; P represents the number of data samples; INF is the real-time substrate temperature parameter of the i-th sample; (i,jc) The real-time substrate information represented as the i-th sample; INF is the real-time coating temperature parameter of the i-th sample; (i,tb) The real-time coating information represented as the i-th sample; INF is the real-time gluing temperature parameter of the i-th sample; (i,tj) Represented as the real-time gluing information of the i-th sample.

[0168] In this embodiment, a colloid temperature monitoring function is proposed to monitor the influence of the real-time monitoring temperature on the colloid temperature; this function uses historical monitoring data to train the colloid temperature prediction model to obtain temperature prediction parameters; the model parameters are updated using the temperature prediction parameters and real-time monitoring data to obtain real-time temperature parameters; this function evaluates the real-time monitoring information and the real-time temperature parameters to obtain colloid temperature monitoring information; the colloid temperature monitoring information reflects the degree of influence of the real-time monitoring temperature on the colloid temperature; this function can effectively improve the accuracy and reliability of UV hot melt adhesive transfer coating quality monitoring.

[0169] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A UV hot melt adhesive transfer coating quality monitoring system, characterized in that: include: Equipment control module, data acquisition module, data monitoring module and output module; The device control module is used to control the start, pause and stop of the device; The data acquisition module is used to obtain image data collected by the camera; wherein the image data includes: substrate image and coating image; The data monitoring module is used to identify and process the image data to obtain monitoring data; wherein the data monitoring module includes: a substrate roughness index calculation unit, a coating thickness uniformity monitoring unit and a coating defect feature monitoring unit; The substrate roughness index calculation unit processes the substrate image using a substrate feature recognition model to obtain a substrate roughness index; the coating thickness uniformity monitoring unit evaluates the substrate roughness index, thickness feature, coating speed, coating pressure and coating amount to obtain a first thickness uniformity monitoring value; the coating defect feature monitoring unit evaluates bubble feature, crack feature and concave feature to obtain a second defect feature monitoring value; The output module is used to evaluate the monitoring data, output monitoring results and provide monitoring feedback; wherein, the monitoring output module includes: a judgment unit and a colloid temperature monitoring unit; wherein, the colloid temperature monitoring unit inputs the substrate monitoring data and the coating monitoring data into the colloid temperature prediction model, and generates colloid temperature monitoring information according to the data results.

2. A UV hot melt adhesive transfer coating quality monitoring system according to claim 1, characterized in that: The substrate roughness index calculation unit processes the substrate image using the substrate feature recognition model to obtain the substrate roughness index. The specific implementation process includes: Inputting the substrate image into the substrate feature recognition model for recognition to obtain substrate features; According to the substrate characteristics, a gray value of the substrate texture characteristics and an area of ​​the substrate flattened region are obtained; The characteristic grayscale value of the substrate texture and the area of ​​the flat region of the substrate are compared with the total grayscale value and the total area respectively, and the substrate roughness index is calculated.

3. A UV hot melt adhesive transfer coating quality monitoring system according to claim 2, characterized in that: The substrate feature recognition model comprises: an input layer, a feature extraction layer, a substrate feature marking layer, a substrate feature recognition layer and an output layer; Wherein, the input layer converts the substrate image from the image space to the feature space to facilitate subsequent feature processing; The feature extraction layer extracts multi-scale features using convolution operations of different sizes and fuses the multi-scale features through a channel merging operation; The substrate feature marking layer is used to obtain an anchor mark frame of substrate features; The substrate feature recognition layer recognizes target features according to the anchor frame; The output layer is used to output the target features; wherein the target features include texture features and flat area features.

4. A UV hot melt adhesive transfer coating quality monitoring system according to claim 1, characterized in that: The coating thickness uniformity monitoring unit evaluates the substrate roughness index, thickness characteristics, coating speed, coating pressure and coating amount to obtain a first thickness uniformity monitoring value. The specific implementation process includes: Obtain coating thickness features output by a coating feature recognition model; Obtaining a substrate roughness index calculated according to an output result of a substrate feature recognition model; The coating thickness feature, the substrate roughness index and the external monitoring value are weighted with a coating thickness feature weight factor, a substrate roughness index weight factor and an external monitoring value weight factor, respectively, to obtain a plurality of evaluation scores; A comprehensive evaluation is performed based on the multiple evaluation scores to calculate and obtain a first thickness uniformity monitoring value.

5. A UV hot melt adhesive transfer coating quality monitoring system according to claim 4, characterized in that: The specific implementation process of the external monitoring value includes: Acquire external monitoring data from the sensor; wherein the external monitoring data includes coating speed, coating pressure and coating amount; The coating speed, the coating pressure and the coating amount are respectively compared with a coating speed standard value, a coating pressure standard value and a coating amount standard value, and an external monitoring value is calculated.

6. A UV hot melt adhesive transfer coating quality monitoring system according to claim 1, characterized in that: The coating defect feature monitoring unit evaluates the bubble feature, the crack feature and the depression feature to obtain the second defect feature monitoring value, and the specific implementation process includes: Inputting the coating image into the coating feature recognition model for recognition processing; Acquire coating defect features output by the coating feature recognition model; wherein the coating defect features include bubble features, crack features and depression features; Acquire bubble information, crack information and depression information from the bubble feature, the crack feature and the depression feature respectively; The bubble information, the crack information and the depression information are weighted with the bubble characteristic coefficient, the crack characteristic coefficient and the depression characteristic coefficient respectively to calculate and obtain two defect characteristic monitoring values.

7. A UV hot melt adhesive transfer coating quality monitoring system according to claim 6, characterized in that: The coating feature recognition model includes: an input layer, a feature extraction layer, a coating feature marking layer, a coating feature recognition layer and an output layer; the input layer converts the coating image from the image space to the feature space to facilitate subsequent feature processing; the feature extraction layer uses convolution operations of different sizes to extract multi-scale features and fuses the multi-scale features through channel merging operations; the coating feature marking layer is used to obtain the anchor frame of the coating feature; the coating feature recognition layer recognizes the target feature according to the anchor frame; the output layer is used to output the coating feature.

8. The UV hot melt adhesive transfer coating quality monitoring system according to claim 1, characterized in that: The colloid temperature monitoring unit inputs the substrate monitoring data and the coating monitoring data into the colloid temperature prediction model, and the specific process of generating colloid temperature monitoring information according to the data results includes: Obtain real-time substrate information from substrate monitoring data; Obtain real-time coating information and real-time glue information from coating monitoring data; Inputting historical substrate information, historical coating information and historical glue coating information in the historical monitoring database into the colloid temperature prediction model for training to obtain temperature prediction parameters; Input the temperature prediction parameters, the real-time substrate information, the real-time coating information and the real-time glue coating information into the colloid temperature prediction model, update the model parameters, and output the real-time temperature parameters; wherein the real-time temperature parameters include: real-time substrate temperature parameters, real-time coating temperature parameters and real-time glue coating temperature parameters; The real-time information and the real-time temperature parameters are comprehensively evaluated to obtain colloid temperature monitoring information.

9. A UV hot melt adhesive transfer coating quality monitoring system according to claim 8, characterized in that: The colloid temperature prediction model includes: convolutional layers, downsampling and upsampling, skip connections and residual blocks.

10. The UV hot melt adhesive transfer coating quality monitoring system according to claim 1, characterized in that: The weight values ​​required for calculating the first thickness uniformity monitoring value and the second defect characteristic monitoring value satisfy the condition that the weighted sum is 1.

Images