A waterborne polyurethane coating detection method and system

Abstract

The invention relates to the technical field of intelligent detection, and discloses a waterborne polyurethane coating detection method and a system thereof. The waterborne polyurethane coating detection method acquires sample coating data after infrared light irradiation to obtain spectral information data; then pre-processes the spectral information data and performs layered marking to obtain marking data; then identifies the marking data to obtain real spectral information data; then inputs the real spectral information data into an intelligent detection model for training to output detection data; finally, generates a coating detection report according to the output detection data, and displays the report in a user interface with visual graphics and text, so that the spectral waveform diagram of infrared light irradiation can be processed by the intelligent detection model, so that the waterborne polyurethane coating can be accurately detected by analyzing the infrared spectrum, so that the coating data is more accurate and reliable, and the user can more clearly understand the detection data of various properties of the waterborne polyurethane coating.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to a waterborne polyurethane coating detection method and a system thereof. Background Art

[0002] In daily life, paint factories need to test waterborne polyurethane coatings before packing them into barrels. At present, the traditional detection method of using infrared light to detect coatings still has some defects. When the traditional detection method detects coatings through infrared light irradiation, it will be interfered by temperature and humidity factors. At the same time, due to problems with the coating technique, the coating will have uneven thickness, resulting in overlapping or blocked spectra. Therefore, the traditional detection method of using infrared light irradiation to detect the quality of waterborne polyurethane coatings is difficult to obtain more accurate and reliable detection data.

[0003] With respect to the traditional detection method, the present invention collects data of the waterborne polyurethane coating after infrared light irradiation, marks the collected data, and analyzes the marked data to determine the direction of the next experimental operation.

[0004] The analysis data is input into the intelligent detection model for training, so that the intelligent detection model can process the spectrum that is interfered by temperature and humidity or overlapped and separated, and further output the detection results.

[0005] In summary, how to use infrared light to detect the quality of waterborne polyurethane coatings and make the detection data more accurate and reliable is an urgent problem to be solved and optimized in the waterborne polyurethane coating detection system. Summary of the invention

[0006] The present invention provides a waterborne polyurethane coating detection method and system thereof, which solve the technical problem in the related art of detecting the quality of waterborne polyurethane coating by infrared light irradiation and can obtain more accurate and reliable detection data.

[0007] In order to solve the above technical problems, the present invention provides a waterborne polyurethane coating detection method and system thereof, and the specific technical scheme is as follows:

[0008] In a first aspect, a method for detecting a waterborne polyurethane coating comprises the following steps:

[0009] Collect the sample coating data after infrared light irradiation to obtain spectrum information data;

[0010] Preprocessing the spectral information data and performing hierarchical labeling to obtain labeled data;

[0011] Identify the marked data to obtain real spectral information data;

[0012] Input the real spectral information data into the intelligent detection model for training to output the detection data;

[0013] Generate a coating inspection report based on the output inspection data, and display the report in the user interface with visual graphics and text.

[0014] Furthermore, the sample coating data after infrared light irradiation is collected to obtain spectral information data, including:

[0015] Parallel collection of coating data of multiple samples after infrared light irradiation;

[0016] Acquire spectrum information data based on the absorption light wavelength data of the coating data after infrared light irradiation;

[0017] The coating photometric distribution is obtained through spectral information data to obtain photometric distribution data.

[0018] Further, the spectral information data is preprocessed and hierarchically labeled to obtain labeled data, including:

[0019] Preprocessing is performed based on the photometric distribution data of the coatings of multiple sample panels to obtain preprocessed data;

[0020] The pre-processed photometric distribution data are marked in sequence;

[0021] Adding timestamp information to the marked preprocessed data and performing layered processing to obtain layered marked data;

[0022] Performing sub-tag processing on the layered tag data to obtain sub-tag information data;

[0023] Capturing the sub-mark information data at a fixed point in the spectral region to obtain explicit spectral data;

[0024] Draw a spectral waveform graph based on the explicit spectral data.

[0025] Further, the sub-mark information data is captured at a fixed point in the spectral region to obtain explicit spectral data, including:

[0026] The luminosity distribution data is obtained by: ; In the formula, y i is the photometric data point after baseline correction, x i is the wavelength position of the original luminosity distribution, a0, a1, ..., a n The series coefficients are used to perform baseline correction to obtain clear data;

[0027] Determine the photometric characteristic absorption peaks at 350nm and 550nm in the photometric distribution data based on the clear data;

[0028] pass: ; In the formula, G (x i ) is the derivative value of the luminosity distribution data, yi+1 and i-1 is the luminosity value of the adjacent data point; x i+1 and x i-1 is the wavelength position of adjacent data points, from which characteristic peaks are identified;

[0029] Perform fixed-point capture according to the characteristic peak position to extract the characteristic luminosity distribution intensity value;

[0030] The characteristic photometric distribution data of multiple sub-markers are summarized and statistically analyzed to obtain a variety of dominant photometric characteristic data.

[0031] Furthermore, the marked data is identified to obtain the real spectral information data, including:

[0032] Identify the marked data and remove interference features caused by temperature and humidity changes to obtain valid marked data;

[0033] Based on the effective marked data, the overlapping and separated spectra are further identified according to the spectral waveform diagram to obtain the real spectral information data.

[0034] Furthermore, the marked data is identified and interference features caused by temperature and humidity changes are eliminated to obtain valid marked data, including:

[0035] pass: ; In the formula, r is the correlation coefficient, x is the marked spectral characteristic data, y is the temperature and humidity data, and n is the number of samples, thereby obtaining the temperature and humidity correlation spectral characteristic data;

[0036] The temperature and humidity correlation spectral feature data is separated from the labeled data and passed through: ; In the formula, Y is the response matrix of temperature and humidity, A is the mixing matrix, s is the spectral data, B is the regression coefficient matrix, and E is the error matrix, which are eliminated to obtain valid labeled data.

[0037] Furthermore, overlapping and cut-off spectra are further identified based on the effective marking data according to the spectral waveform diagram to obtain real spectral information data, including:

[0038] Visualize the effective marked data in the form of a waveform graph, and preliminarily identify the peak position, bandwidth and morphological characteristics of the waveform graph;

[0039] Identify overlapping characteristic peak areas through the spectral waveform, and use: Where A i is the amplitude, μ i is the center position, σ i is the standard deviation, n is the number of characteristic peaks, e is the overlap threshold, and x is the position of the spectrum waveform, thereby separating the overlapping peaks;

[0040] The separated spectral waveform is further used to identify the breakpoints of the spectrum, and untrustworthy bands are removed or the bands are repaired to the broken areas to obtain real spectral information data.

[0041] Furthermore, the real spectral information data is input into the intelligent detection model for training to output detection data, including:

[0042] Acquire coating performance data based on real spectral information data, and obtain performance processing data through data processing;

[0043] The performance processing data is compared with the coating performance data entered into the model database to output the coating inspection data.

[0044] Furthermore, a coating inspection report is generated based on the output inspection data, and the report is displayed on the user interface in a visual graphic format, including:

[0045] Design the test report structure based on the test results output by the intelligent test model;

[0046] Edit the text contents of two test reports based on the coating test data;

[0047] And generate two sets of chart information data based on the test data;

[0048] Embed a chart information data into a test report text content, so that the text content describes the chart information in detail;

[0049] Design a personalized user interface and provide filtering and sorting functions so that users can view specific content based on their needs.

[0050] In a second aspect, the system provides an electronic device including a memory, a processor, and a waterborne polyurethane coating detection method program stored in the memory and executable on the processor, wherein the waterborne polyurethane coating detection method program implements the steps of the waterborne polyurethane coating detection method when executed by the processor. The system includes:

[0051] Coating acquisition module: It is used to collect the coating data of the sample after infrared light irradiation to obtain spectral information data;

[0052] Data analysis module: It is used to pre-process the spectral information data and perform hierarchical labeling to obtain labeled data;

[0053] Intelligent detection module: It is used to input real spectral information data into the intelligent detection model for training to output detection data;

[0054] User interface module: It is used to generate a coating inspection report based on the output inspection data, and display the report in the user interface with visual graphics and text.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The above scheme of the present invention acquires spectral information data by collecting sample coating data after infrared light irradiation; facilitates subsequent detection of coating data of each layer after infrared light irradiation, and then pre-processes the spectral information data and performs layered labeling to obtain labeling data; then inputs the real spectral information data into the intelligent detection model for training to output detection data; finally, generates a coating detection report according to the output detection data, and displays the report in the user interface with visual graphics, so that the spectral waveform diagram of infrared light irradiation can be processed by the intelligent detection model, so that the water-based polyurethane coating can be accurately detected by analyzing the infrared spectrum, and the detection results are output, so that the coating data is more accurate and reliable, and the user can have a clearer understanding of the detection data of various properties of the water-based polyurethane coating. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The accompanying drawings are used for further understanding of the present invention and are used to explain the present invention together with the embodiments of the present invention, but do not constitute a limitation of the present invention.

[0058] Figure 1 It is a schematic flow chart of a waterborne polyurethane coating detection method provided by an embodiment of the present invention;

[0059] Figure 2 is a schematic diagram of a waterborne polyurethane coating detection system provided by an embodiment of the present invention;

[0060] Figure 3 This is a first waterborne polyurethane coating diagram provided by an embodiment of the present invention;

[0061] Figure 4 is a second waterborne polyurethane coating diagram provided by an embodiment of the present invention;

[0062] Figure 5 This is an infrared spectrum of two waterborne polyurethane coatings provided by an embodiment of the present invention;

[0063] Figure 6 A waterborne polyurethane coating halo pattern provided by an embodiment of the present invention

[0064] Figure 7 It is a waterborne polyurethane coating partition diagram provided by an embodiment of the present invention;

[0065] Figure 8 The present invention provides a waterborne polyurethane coating partition diagram in an embodiment. DETAILED DESCRIPTION

[0066] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the present specification. Various examples may omit, replace, or add various processes or components as needed. In addition, the features described in some examples may also be combined in other examples.

[0067] In order to better understand the above technical scheme, the technical scheme of the present invention will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification. It should be pointed out that the above steps are only a preferred implementation order. In the specific implementation process, some steps can be replaced without affecting the overall implementation effect. In order to more clearly explain the technical scheme of this application, the following content explains this scheme in a preferred manner.

[0068] like Figures 1 to 8 As shown, the present invention provides a method for detecting a waterborne polyurethane coating, comprising the following steps:

[0069] Step 1, collecting sample coating data after infrared light irradiation to obtain spectrum information data;

[0070] Step 2, preprocessing the spectral information data and performing hierarchical labeling to obtain labeled data;

[0071] Step 3, identifying the marked data to obtain real spectral information data;

[0072] Step 4, inputting the real spectral information data into the intelligent detection model for training to output detection data;

[0073] Step 5: Generate a coating inspection report based on the output inspection data, and display the report in a user interface in the form of visual graphics and text.

[0074] In a preferred embodiment of the present invention, two identical sample plates of 2×3 decimeter specifications are selected, and two coats of waterborne polyurethane coating are applied to the surfaces of the two identical sample plates in turn. After the two sample plates are air-dried for 40 minutes, the two sample plates coated with waterborne polyurethane coating are tested in turn.

[0075] Adjust the wavenumber range of the infrared spectrometer to 14000-4000cm -1 ; Resolution is 1-2cm -1 ; Scanning speed is 10sps; Sampling is by transmission method.

[0076] Specifically, according to Figure 3 and Figure 4The two waterborne polyurethane coating images shown in the figure show that when the coating is applied to the sample, the white irregular bright spots in the image indicate the bubbles generated when the coating is applied. Figure 6 The uneven halo shown in the image indicates that the coating thickness of a sample is not uniform. Figure 7 and Figure 8 There are dark mottled marks of varying degrees in the image, which indicates that depending on the flatness of the sample surface, different degrees of separation and breakage will occur when the coating is applied, which will affect the subsequent infrared spectrum detection of the coating.

[0077] Through step 1, the coating sample is scanned using a high-precision infrared spectral imaging device, and the collected data may include spectral data, image data, and thickness data. The collected data is then marked, and two sample coatings of the same material are marked. Coatings with different formulations and configuration processes are applied to each sample, so that the characteristics or properties of the coating are also different, such as adhesion, wear resistance, etc. High-quality original data is obtained through precise collection equipment, providing a solid foundation for subsequent data analysis and modeling. At the same time, the marked data can directly reflect the true characteristics of the coating, which is convenient for identifying important performance-related features in subsequent model training. Finally, since the marked data is the basis of supervised learning, it can be used to train the intelligent detection model to improve the detection ability and prediction accuracy of the model; according to step 1, step 2 cleans, denoises, and standardizes the spectral information data for subsequent analysis. Commonly used preprocessing techniques include noise removal, missing value filling, baseline drift correction and normalization. The preprocessed data is analyzed to identify patterns and important features in the data for labeling. The preprocessing step removes unnecessary noise and deviations, improves the quality and availability of the data, and labels the data based on the preprocessed data. It can help identify spectral features or other data features that have a significant impact on coating performance, and provide effective input for model training. By analyzing the data, important information and trends in the coating data can be extracted, providing data support for optimizing production and quality control.

[0078] Then, the marked data is identified through step 3 to obtain the real spectral information data. In spectral analysis, changes in temperature and humidity will cause noise and interference. To solve this problem, it is necessary to obtain real-time temperature and humidity data from environmental sensors and record spectral data synchronously. Through analysis, the features highly related to temperature and humidity are identified, the interference signals are separated, and the valid data is retained. After eliminating the interference caused by temperature and humidity, purer and more stable spectral data are obtained, thereby improving the accuracy of subsequent model training and prediction, avoiding the misleading of environmental factors, and using effective marked data and spectral waveforms to preliminarily identify overlapping areas and breakpoints, and perform peak detection, separate overlapping peaks, and extract independent spectral signals. Subsequently, the broken parts in the spectrum are identified, and these areas are eliminated or repaired to ensure the continuity of the spectrum. In this way, overlapping and blocked spectra are effectively separated to obtain more real data, provide a reliable basis for subsequent analysis, and improve the accuracy of spectral analysis. Based on the real spectral data after de-overlapping and blocking, key performance features such as spectral peak position, intensity, integrated area, etc. are extracted to reflect the chemical and physical properties of the coating;

[0079] Then input the real spectral information data of step 3 into the intelligent detection model of step 4. According to the data characteristics and task requirements, the model is trained using the preprocessed and labeled data, and the model parameters are adjusted to optimize the model's performance. During the training process, the model's ability to recognize and predict coating characteristics is improved through continuous iteration. At the same time, the coating data after infrared light irradiation is labeled to further identify and verify the authenticity and effectiveness of the infrared spectrum, thereby improving the accuracy, stability and generalization ability of the model. Therefore, the trained intelligent model can quickly and accurately detect new samples and predict the performance data of the coating, thereby reducing the time and errors of manual detection and realizing an efficient automated detection process. The model can continue to learn and optimize based on new data to improve its long-term adaptability and accuracy. Finally, step 5 is used to integrate the detection data output by the model with other relevant data for comprehensive analysis to obtain the coating's Performance evaluation and improvement suggestions are provided to facilitate further quality control of waterborne polyurethane coating testing. First, the analysis results are compiled into a detailed test report. The report structure includes summary, test methods, results, conclusions and suggestions. Data visualization tools are used to convert the test results into charts (such as line charts, bar charts, heat maps, etc.), which are intuitively displayed in the user interface. The test results are then presented in a graphic and textual manner to make complex data easier to understand and analyze. The report provides comprehensive test information and suggestions to provide support for production adjustments and quality control decisions. The intuitive interface design and interactive functions enhance the user experience and make it easier for users to obtain and analyze test results.

[0080] In a preferred embodiment of the present invention, the above step 1 further includes:

[0081] Step 11, collecting coating data of multiple samples after being irradiated with infrared light in parallel;

[0082] Step 12, obtaining spectrum information data based on the absorption light wavelength data of the coating data after infrared light irradiation;

[0083] Step 13, obtaining the coating luminosity distribution through the spectral information data to obtain luminosity distribution data.

[0084] In the embodiment of the present invention, two sample coatings of the same material are prepared through step 11 to ensure that they have different coating formulations or treatment processes, so as to comprehensively evaluate the coating performance. The two sample coatings are respectively collected using a multi-channel infrared spectroscopy device. The spectral data, thickness data, etc. of multiple coatings can be obtained simultaneously through parallel collection. The parallel collection greatly shortens the data collection time and improves the overall detection efficiency. At the same time, the collected sample data has higher consistency and comparability, reduces the impact of external environmental changes on the data, and can cover different coating types, providing more samples for subsequent analysis and model training.

[0085] Based on step 11, step 12 is used to collect the absorption spectrum data of the coating under infrared light irradiation using a spectrometer, and then a suitable wavelength range (such as 400nm to 700nm) is selected to collect the spectrum data, and the absorption characteristics of the coating under infrared light irradiation are captured to provide a detailed data basis for subsequent analysis. Coatings with different characteristics have different absorption characteristics at different wavelengths, and the spectrum data can reflect the composition and structural characteristics of the coating;

[0086] Then, according to step 13, the photometric distribution of each wavelength, that is, the light intensity value corresponding to each wavelength, is calculated from the spectral data collected in step 12, and then the drawn photometric distribution diagram is displayed through a visual interface to show the light intensity distribution at different wavelengths. The photometric distribution diagram can intuitively display the absorption characteristics of the coating at different wavelengths, which is convenient for identifying characteristic peaks and characteristic areas. The photometric distribution data can further analyze the optical properties of the coating, and the spectral information data can be obtained again to provide a basis for subsequent feature extraction.

[0087] In a preferred embodiment of the present invention, the above step 2 may include:

[0088] Step 21, preprocessing based on the photometric distribution data of the coatings of the multiple sample plates to obtain preprocessed data;

[0089] Step 22, marking the pre-processed luminosity distribution data in sequence;

[0090] Step 23, adding timestamp information to the marked pre-processed data, and performing layered processing to obtain layered marked data;

[0091] Step 24, performing sub-tag processing on the layered tag data to obtain sub-tag information data;

[0092] Step 25, performing spectral region fixed-point capture on the sub-mark information data to obtain explicit spectral data;

[0093] Step 26, drawing a spectrum waveform diagram according to the explicit spectrum data.

[0094] In the embodiment of the invention, according to step 11, step 21 is first used to denoise the photometric distribution data to remove random noise, and then baseline correction is performed by correcting the baseline drift to ensure the stability of the data. Finally, the photometric distribution data is normalized to standardize the data, remove noise and baseline drift, and obtain purer and more stable spectral data. The normalization process makes the data of different samples comparable, making the photometric distribution data clearer for subsequent feature marking. Based on the pre-processed data of step 21, step 22 is used to feature mark the spectral information data of each layer of coating, and the marking content includes spectral characteristics (such as peak value, integral value), physical properties (such as thickness, adhesion), time information, etc., and the data of each coating is processed in turn. Ensure the accuracy and completeness of the marking, mark the coating data one by one, ensure that all important features and attributes are marked in detail, and lay the foundation for subsequent comparative analysis. According to the marked data obtained in step 22, add timestamp information to each marked data through step 23, record the time point of data acquisition and marking, and deeply understand the performance differences of different coatings. Add timestamp information to two different marked data, and perform layered processing according to the timestamp information and other key features to obtain layered marked data; the layering can be based on the acquisition time, coating type, detection conditions, etc., and the timestamp enables the data to be analyzed in time series, which is helpful to study the law of coating performance changes over time. The layered processing makes data management more systematic and convenient for subsequent sub-marking and in-depth analysis;

[0095] After the specific layered labeling processing according to step 23, the layered labeling data is further processed by sub-labeling through step 24. The sub-labeling can be a more refined feature, such as performance changes in a specific time period and spectral features in a specific area. By extracting the sub-labeling information data and appending it to the original labeling data for further analysis, and obtaining the subtle features of the single-layer coating of each coating, the sub-labeling can make the data analysis more refined and accurate, which can reveal the subtle differences in coating performance. Sub-labeling is performed according to specific needs to provide more targeted data support; according to the sub-labeling information data obtained in step 24, a spectral region with significant features, such as absorption peaks and fluorescence peaks, is selected according to the spectral characteristics through step 25, and then the sub-labeling information data obtained in step 24 is captured at a fixed point, and the spectral data in the characteristic region is extracted. The explicit spectral features of the coating are obtained through fixed-point capture, which provides important data for subsequent analysis and modeling. By capturing the data in the characteristic region, irrelevant data is reduced and the analysis efficiency is improved; finally, based on step 25 and step 26, the explicit spectral data is plotted into a spectral waveform diagram (such as Figure 5 shown), according to Figure 7 The transmittance shows a waveform trend under different wave numbers, indicating that different coating thicknesses, bubbles generated, and coating partition phenomena will affect the transmittance of the infrared spectrum, so a clear and intuitive spectral waveform diagram is designed to show the spectral transmittance of different wave numbers. The spectral waveform diagram can intuitively display the spectral characteristics of the coating, facilitate the identification and analysis of characteristics, and assist researchers and engineers in decision-making and analysis through the visual information provided by the chart.

[0096] In a preferred embodiment of the present invention, the above step 25 specifically includes:

[0097] Step 251, the luminosity distribution data is processed by: ; In the formula, y i is the photometric data point after baseline correction, x i is the wavelength position of the original luminosity distribution, a0, a1, ..., a n The series coefficients are used to perform baseline correction to obtain clear data;

[0098] Step 252, determining the photometric characteristic absorption peaks at 350 nm and 550 nm in the photometric distribution data based on the clear data;

[0099] Step 253. Pass: ; In the formula, G (x i ) is the derivative value of the luminosity distribution data, y i+1 and i-1 is the luminosity value of the adjacent data point; x i+1 and x i-1is the wavelength position of adjacent data points, from which characteristic peaks are identified;

[0100] Step 254, performing fixed-point capture according to the characteristic peak position to extract the characteristic luminosity distribution intensity value;

[0101] Step 255 , summarize and statistically analyze the characteristic photometric distribution data of the multiple sub-marks to obtain a variety of explicit photometric characteristic data.

[0102] In the embodiment of the invention, based on step 24, step 251 is further used to further describe the specific implementation means of correcting the baseline of the photometric distribution data to remove the baseline drift in the spectral data, and further introduces the implementation means of removing noise and fluctuation to ensure the stability and clarity of the spectral data. The baseline correction can effectively remove the baseline drift and improve the accuracy and reliability of the spectral data, thereby obtaining clear data to make the characteristic peaks more obvious, which is convenient for subsequent feature extraction and analysis. Then, according to the characteristics of the coating and preliminary research, 350nm and 550nm are determined as the main characteristic absorption peak positions through step 252. Subsequently, in the corrected clear spectral data, the photometric distribution at 350nm and 550nm is analyzed to confirm the position and intensity of the characteristic absorption peak, and according to the predetermined characteristic wavelength, the efficiency and accuracy of the feature extraction are improved to ensure that the extracted spectral data is concentrated in the wavelength range with the most photometric characteristics.

[0103] Step 253 further describes the means of identifying the characteristic peak position in the spectrum, and at the same time confirms the accuracy of the characteristic peak position and ensures the accuracy of the identified spectrum. By automatically identifying the characteristic peak, efficiency is improved and human intervention is reduced. After the characteristic peak is determined according to step 253, step 254 is used to extract the intensity value of the characteristic photometric distribution at the identified characteristic peak position using a fixed-point capture algorithm, and the characteristic peak area is calculated to obtain the intensity value of the characteristic photometric distribution. The fixed-point capture algorithm is used to ensure that only valid data in the characteristic area is extracted to avoid data redundancy, thereby obtaining the intensity value of the key characteristic peak, providing an important basis for performance analysis. Finally, step 255 is used to summarize the characteristic photometric distribution data of multiple sub-marks (different coatings), and statistical analysis is used to summarize the data to extract a variety of explicit photometric feature data.

[0104] In a preferred embodiment of the present invention, the above step 3 may include:

[0105] Step 31, identifying the marked data and eliminating interference features caused by temperature and humidity changes to obtain valid marked data;

[0106] Step 32: further identifying overlapping and cut-off spectra according to the spectrum waveform diagram based on the effective marking data to obtain real spectrum information data.

[0107] In an embodiment of the present invention, when performing spectral analysis through step 31, changes in temperature and humidity may cause noise and interference features in the data. In order to solve this problem, it is first necessary to obtain real-time temperature and humidity data from the environmental sensor, and record these data synchronously with the spectral data. Through spectral analysis, the features in the spectral data that are highly correlated with temperature and humidity are identified, and the interference signals related to temperature and humidity are separated from the true spectral features, thereby eliminating the spectral features that are highly correlated with temperature and humidity, retaining valid data that is not affected by environmental factors, eliminating interference features caused by temperature and humidity, and further obtaining purer and more stable spectral data, so that the valid data can improve the accuracy of subsequent model training and prediction, and avoid misleading the results due to environmental factors; based on the valid labeled data of step 31 according to the spectral waveform diagram of step 26. By observing the waveform graph, the overlapping areas and breakpoints in the spectrum can be preliminarily identified, and the peak value detection of the spectral data can be further performed to determine the overlapping peak areas. The overlapping spectral peaks are then separated to extract independent spectral signals. The separated parts in the spectrum are then identified and these areas are removed or repaired to ensure the continuity of the spectrum, thereby effectively separating the overlapping and separated spectra and obtaining more realistic spectral data, providing a reliable basis for subsequent analysis. By processing the overlapping and separated spectra, errors can be avoided and the accuracy of spectral analysis can be improved. Based on the real spectral data after de-overlapping and separation in step 32, key performance features are extracted. These features may include spectral peak position, intensity, integrated area, etc., reflecting the chemical and physical properties of the coating.

[0108] In another preferred embodiment of the present invention, the above step 31 may include:

[0109] Step 311, by: ; In the formula, r is the correlation coefficient, x is the marked spectral characteristic data, y is the temperature and humidity data, and n is the number of samples, thereby obtaining the temperature and humidity correlation spectral characteristic data;

[0110] Step 312, separating the temperature and humidity correlation spectral feature data from the label data, and performing: ; In the formula, Y is the response matrix of temperature and humidity, A is the mixing matrix, s is the spectral data, B is the regression coefficient matrix, and E is the error matrix, which are eliminated to obtain valid labeled data.

[0111] In an embodiment of the present invention, while collecting spectral data in step 311, a temperature and humidity sensor is used to monitor and record the temperature and humidity data in the experimental environment in real time. Usually, the temperature and humidity data and the spectral data are collected synchronously to ensure the consistency and correlation of the data. By recording the temperature and humidity data at the same time, it is ensured that the influence of temperature and humidity on the spectrum can be accurately identified, thereby ensuring that the collected data is complete and convenient for subsequent correlation analysis; then extract key features from the spectral data, such as the intensity, position, and bandwidth of the characteristic peak, and then use the correlation coefficient to calculate the relevant spectral feature data between the spectral feature and the temperature and humidity, so that those spectral features that are greatly affected by temperature and humidity can be accurately identified, providing a basis for subsequent interference elimination, and through correlation analysis, quantitatively evaluate the interference degree of temperature and humidity on the spectrum, ensure the scientificity and reliability of data processing, and finally mark the temperature and humidity correlation spectral feature data, and mark the spectral features that are significantly affected by temperature and humidity according to the results of the correlation analysis. These marked data will be used to distinguish effective features from interference features in subsequent steps, and the marking enables the disturbed spectral features to be clearly distinguished, laying the foundation for the subsequent elimination of interference features, and early marking can speed up the recognition and processing in the subsequent processing process.

[0112] According to the temperature and humidity correlation labeling data in step 311, the spectral features affected by temperature and humidity are separated from the total spectral data to form two groups of data: one group is the disturbed spectral features, and the other group is the remaining spectral feature data. After separating the interference features, the remaining spectral data is purer, which helps to improve the accuracy of the analysis. Through data separation, the effective data can be processed and analyzed more centrally, reducing data redundancy, and then the temperature and humidity correlation spectral feature data is eliminated, and the identified temperature and humidity correlation spectral feature data is eliminated from the total spectral data, and only the spectral features that are not affected by temperature and humidity or are less affected are retained. This process can be achieved by writing an algorithm or manual screening. After eliminating the interference features that are highly correlated with temperature and humidity, the remaining data can better reflect the essential characteristics of the sample, and the data quality is significantly improved, so that the purity of the effective labeling data improves the stability and reliability of subsequent model training. At the same time, the effective labeling data after elimination is cross-validated to ensure that the elimination process does not cause the loss of key feature information. The accuracy of the elimination process can be verified by using methods such as residual analysis. Cross-validation can ensure that after eliminating interfering features, the remaining data is still complete and reliable. The verified data can better adapt to different environmental conditions and improve the wide adaptability of the model.

[0113] In a preferred embodiment of the present invention, the above step 32 may include:

[0114] Step 321 , visualizing the valid marked data in the form of a waveform, and preliminarily identifying the peak position, bandwidth and morphological characteristics of the waveform;

[0115] Step 322, identifying overlapping characteristic peak regions through the spectrum waveform diagram, and by: Where A i is the amplitude, μ i is the center position, σ i is the standard deviation, n is the number of characteristic peaks, e is the overlap threshold, and x is the position of the spectrum waveform, thereby separating the overlapping peaks;

[0116] Step 323, further identifying the breakpoints of the spectrum from the separated spectrum waveform graph, and removing untrustworthy bands or repairing the broken areas of the bands to obtain real spectrum information data.

[0117] In an embodiment of the present invention, the effective marking data obtained in step 31 is converted into a spectrum waveform diagram (usually a curve diagram of absorbance or transmittance versus wave number or wavelength) in step 26 through step 321 to intuitively display the overall morphology of the spectrum data. The horizontal axis of the waveform diagram usually represents the wave number or wavelength, and the vertical axis represents the absorbance or transmittance. First, the peak position in the waveform diagram is preliminarily identified. The peak position corresponds to the characteristic absorption peak in the spectrum, which represents the derivative value of the spectrum data, and the peak position corresponds to the point where the derivative changes significantly; the bandwidth of each absorption peak is measured, that is, the full width half height of the peak, and the bandwidth can reflect the intensity of the intermolecular interaction and the microstructure of the sample, and the overall morphology of the waveform diagram is analyzed to identify features such as asymmetric peaks, shoulder peaks, and platforms. These morphological features may indicate different components or structural changes in the sample. Through the visualization of the waveform diagram, the overall characteristics and change trends of the spectrum data can be intuitively understood, which is helpful to discover potential problems or anomalies, and the key characteristic peaks in the spectrum can be quickly identified, providing a basis for subsequent fine analysis. When two or more characteristic peaks are close in position in the spectrum, peak overlap occurs, resulting in changes in the morphology of a single peak. By observing the waveform, we can identify the peak regions with obvious overlapping features. Usually, these regions have large peak widths or abnormal peak shapes. We can perform deconvolution analysis on the overlapping characteristic peak regions and separate the overlapping peaks.

[0118] By adjusting the fitting parameters to minimize the residuals, we ensure that the peak morphology after separation can accurately reflect the independent features in the actual spectrum. After successfully separating the overlapping peaks, we can more clearly identify and quantify each independent characteristic peak, thereby accurately understanding the composition and structural information of the sample. By separating the overlapping peaks, we can reduce the misjudgment of peak area and position, thereby improving the accuracy of spectral analysis.

[0119] The separated spectral waveform is then used to further identify the breakpoints of the spectrum, and untrustworthy bands are removed or the bands are repaired to obtain the real spectral information data, so that the breakpoints of the spectrum may appear when the data acquisition is unstable or the sample state changes. These breakpoints appear as sudden signal interruptions or jumps on the waveform. These discontinuities are identified by edge detection algorithms. For obvious breakpoints, if the bands cannot be naturally connected, these untrustworthy bands are removed to ensure that only complete and continuous spectral data are retained. If the break area is small and can be supplemented by interpolation or other data correction methods (such as polynomial fitting, spline interpolation, etc.), the broken bands are repaired: , where represents the repaired spectral value, a n is the fitting coefficient, x is the wavelength or wavenumber, and the continuity of the spectrum is restored, thereby ensuring that the final spectral data is complete and continuous by eliminating or repairing broken bands, which helps to accurately analyze the characteristic information in the spectrum, eliminate unreliable bands, avoid errors caused by these areas in subsequent analysis, and improve the reliability of the data.

[0120] In a preferred embodiment of the present invention, the above step 4 may include:

[0121] Step 41, obtaining coating performance data according to the real spectrum information data, and obtaining performance processing data through data processing;

[0122] Step 42, comparing the performance processing data with the coating performance data entered into the model database to output coating detection data.

[0123] In an embodiment of the present invention, the extracted performance data is standardized, normalized or otherwise processed in step 41 to ensure data comparability. The data is further processed to obtain coating performance data based on a variety of real spectral information data. The explicit photometric characteristic data obtained in step 2 is required. These data may include the intensity value of the absorption peak, the spectral integral value, and the characteristic wavelength. The photometric characteristic data is associated with the coating performance (such as adhesion, wear resistance, hardness, etc.) through historical data to obtain specific performance data of the coating, such as predicting the wear resistance of the coating based on the intensity value of the absorption peak. The photometric characteristic data is used to accurately predict various properties of the coating, reducing the time and cost of actual testing. At the same time, a variety of photometric characteristic data are used to provide a comprehensive coating performance evaluation and increase the reliability of the prediction. The coating performance data is further removed or corrected for outliers and missing values ​​to ensure data integrity and accuracy. Noise in the data is then removed to improve data quality. Finally, the data is standardized so that data with different characteristics are at the same level to facilitate subsequent analysis, so that the system can automatically predict the performance of new samples and improve detection efficiency, thereby being able to extract data from a large amount of data. The complex features and patterns are learned to improve the accuracy of prediction. Finally, the newly input performance processing data is predicted through step 42, the performance data of the coating is output, and the prediction results are compared with the coating performance data already entered in the model database to verify the accuracy and consistency of the prediction results. According to the comparison results, the final test results are output, and a detailed test report is generated. By comparing with the database, the accuracy of the prediction results is verified to ensure the reliability of the test results. The generated test report contains detailed performance data and analysis results to provide users with a comprehensive coating performance evaluation. By continuously updating the model database, the model performance is continuously optimized to improve the accuracy of long-term predictions. The actual test data in the database is compared with the standard data in the database to obtain the error or deviation between them, determine the qualification of the coating or possible defects, and output the final coating test data based on the comparison results. A test report is generated, which includes detailed analysis results of the coating performance. By comparing with the database, the accuracy of the test results is verified to ensure the reliability of the analysis, so that the system generates a test conclusion of the coating, which is convenient for quickly understanding the quality and performance of the coating, and supports production decisions and quality control.

[0124] In another embodiment of the present invention, step 4 is further described. First, key performance data are selected to determine key performance indicators of the coating, such as adhesion, wear resistance, hardness, and thickness. The coating performance indicators are the targets of subsequent analysis and prediction, specifically including:

[0125] Adhesion refers to the adhesion between the coating and the substrate. By analyzing the spectral characteristics, the adhesion characteristics of the coating can be inferred. The coating with strong adhesion will show changes in certain specific absorption peaks in the spectrum. The spectral absorption peak of the coating with good adhesion is relatively stable, and the morphology will not change due to slight surface deformation. By reusing Raman spectroscopy, the changes and distribution of the molecular bonds of the coating can be observed. The peak value of the intermolecular force in the Raman spectrum of the coating with strong adhesion is higher and the distribution is more uniform.

[0126] Wear resistance refers to the durability of a coating under friction conditions. Spectral characteristics can reflect the changes in the coating before and after wear, thereby evaluating wear resistance. Observing the changes in spectral intensity before and after wear can reflect wear resistance. For coatings with good wear resistance, the intensity changes in the spectrum are small. The changes in surface chemical bonds can be observed through infrared spectroscopy. For coatings with good wear resistance, the infrared spectrum does not change much before and after wear, indicating the stability of the chemical bonds.

[0127] Hardness refers to the ability of a coating to resist deformation. Spectral characteristics can show hardness by reflecting changes in the structure and composition of the material. In the infrared-infrared spectrum, the absorption band of a coating with higher hardness may show specific displacement or morphological changes, reflecting the density and bond energy characteristics of the material. The light scattering spectrum of a coating with higher hardness has higher scattering intensity and more concentrated scattering peak positions.

[0128] Thickness refers to the actual physical thickness of the coating. Spectral characteristics can reflect thickness through effects such as interference and absorption. The thickness of the film can be measured by the spacing of interference fringes in the spectrum. The smaller the spacing of interference fringes, the thicker the coating is, and the position of the absorption edge at a specific wavelength can reflect the thickness of the coating. The displacement and morphology of the absorption edge will change with different thicknesses. Due to different coating thicknesses, the reflectivity peak position and intensity will be different. The thickness of the coating can be inferred from the change in reflectivity.

[0129] By analyzing the photometric distribution characteristics of the above-mentioned performance data and mapping them to the spectral waveform diagram, the spectral characteristic areas related to the key performance indicators are identified. Usually, these areas are the parts of the spectrum with significant characteristics (such as absorption peaks and emission peaks). By selecting key performance data and characteristic areas, the pertinence and efficiency of the analysis are improved, and based on the spectral waveform diagram, the photometric characteristic areas related to the performance are accurately determined, laying the foundation for subsequent data extraction and analysis.

[0130] Then, through a detailed analysis of the performance photometric characteristic areas marked in the spectral waveform diagram, the specific wavelength ranges of these areas are determined, and the starting wavelength and ending wavelength of the characteristic areas are recorded to ensure that the selected wavelength range covers all important spectral features. By determining the wavelength of the characteristic area, it is ensured that all performance-related spectral features are captured to improve the representativeness of the data. After the wavelength range is clarified, data extraction can be carried out in a targeted manner to reduce unnecessary data processing work. Then, within the determined wavelength range of the characteristic area, the photometric distribution characteristic data, including absorption intensity, peak position, etc., is extracted, and the spectral data is processed to accurately extract the characteristic data. The extracted photometric characteristic data contains key information on the coating performance, which provides important data support for subsequent analysis. By using professional algorithms, the extracted data is ensured to be accurate and reliable, and then the extracted photometric characteristic data is used. The data are classified and organized according to each key performance parameter, and the organized data are integrated into a matrix, with each row representing a sample and each column representing a feature. The data matrix form is used for easy management and operation, and a standardized data structure is provided for subsequent analysis and modeling. Moreover, through the data matrix, the data distribution of all samples and features can be comprehensively viewed, which is convenient for discovering potential rules. The data matrix is ​​then used as a coating performance photometric characteristic data set, which contains the characteristics of all samples and the corresponding performance data. At the same time, the data set is preliminarily verified to ensure the integrity and consistency of the data, so that all photometric features and performance data are integrated in the data set for unified analysis and processing, providing a comprehensive data basis for subsequent machine learning model training and performance prediction. Finally, the data set is divided into a training set and a test set in a ratio of 7:3 to ensure the randomness and representativeness of the data distribution.

[0131] Training set: used for model training, accounting for 70% of the data set.

[0132] Test set: used for model validation, accounting for 30% of the dataset.

[0133] The training set provides enough data for model training to improve the generalization ability of the model, while the test set is used to evaluate the performance of the model to ensure that the model performs well on unseen data.

[0134] In another embodiment of the present invention, the training set data obtained by segmentation is input into the intelligent detection model. The training set data contains a variety of explicit photometric features and corresponding coating performance data, so that the model learns the relationship between the features in the training set and the target variable through repeated iterations. Each iteration of the model adjusts the internal parameters to reduce the prediction error, and adjusts the model parameters in real time through cross-validation and hyperparameter optimization methods (such as grid search, random search) to improve the generalization ability and prediction accuracy of the model. Through repeated iterations and parameter adjustment, the model can accurately learn data features and improve prediction ability. Real-time adjustment of parameters ensures that the model performs best on the training data, laying the foundation for subsequent verification and prediction. The test set data obtained by segmentation is input into the trained model for verification, so that the model predicts the performance of the test set data, outputs the prediction results, and calculates the error between the prediction results and the actual values ​​in real time. Common error indicators include mean square error (MSE), mean absolute error (MAE), determination coefficient (R²), etc. The performance of the model on unseen data is evaluated by error indicators to ensure the prediction ability and generalization performance of the model, and error analysis can help identify the deficiencies of the model on specific data and guide further model optimization and improvement.

[0135] Then the new coating photometric distribution feature pre-detection data is input into the trained intelligent detection model. The model predicts the performance based on the input photometric feature data and outputs the corresponding coating performance parameters, such as adhesion, wear resistance, hardness and thickness. Based on the trained model, the new data can be quickly predicted to improve the detection efficiency. At the same time, the predictive ability of the model can be used to detect coating performance on a large scale and efficiently, reducing the workload of laboratory testing. Finally, the coating performance data predicted by the model is summarized, and the performance data is comprehensively analyzed. The coating performance is evaluated based on the analysis results, and a detailed test report is generated, including the distribution of performance parameters, overall evaluation and improvement suggestions, so that the performance of the coating can be comprehensively evaluated and its performance differences under different conditions can be understood. The test report provides detailed analysis results and suggestions to help users make decisions and improve coating processes. The results of the analysis can be used to optimize coating formulations and processes to improve product quality and performance.

[0136] In a preferred embodiment of the present invention, the above step 5 may include:

[0137] Step 51, designing a test report structure according to the test results output by the intelligent test model;

[0138] Step 52, respectively edit the text contents of two test reports based on the coating test data;

[0139] Step 53, generating two sets of chart information data according to the detection data;

[0140] Step 54, embedding a chart information data into a test report text content, so that the text content describes the chart information in detail;

[0141] Step 55, design a personalized user interface and provide filtering and sorting functions so that users can view specific content independently according to their needs.

[0142] In an embodiment of the present invention, based on the coating test results outputted in step 3, the test results outputted by the intelligent test model are analyzed in step 51, including various performance parameters of the coating, such as adhesion, wear resistance, hardness and thickness, etc., and based on the test results, the main content and structure of the test report are determined. The report usually includes the following parts: summary, introduction, test method, test results, analysis and discussion, conclusions and suggestions, appendix, etc. At the same time, the structural framework of the report is designed to ensure that the content of each part is logically clear and well-structured, and can fully display the test results and analysis process. By designing a reasonable report structure, it is helpful to clearly display the content of the report, so that readers can quickly understand the key test results, and ensure that the report content fully covers all important test information and analysis results, providing users with a comprehensive reference, and then use the text editing tool (such as Microsoft Word, LaTeX, etc.) in step 52 to edit the text content of the test report, which specifically includes:

[0143] Report 1: Technical Report: The first test report can be positioned as a technical report, which describes the test process, data analysis methods and results in detail and is suitable for technicians and researchers;

[0144] Report 2: Concise Report: The second test report can be positioned as a concise report, focusing on key results and conclusions, suitable for managers and decision makers.

[0145] Edit the report content according to the needs of different user groups, provide targeted reports, and through professional editing, ensure that the report content is clear and logically rigorous, and improve the readability and professionalism of the report.

[0146] Then, use step 53 to use data visualization tools (such as Python's Matplotlib, Excel, etc.) to generate chart information, including line charts, bar charts, scatter plots, and heat maps, including:

[0147] Chart 1: Detailed chart: The first chart information data contains detailed charts, showing all the data and analysis results during the test, suitable for technical reports;

[0148] Figure 2: Simple Figures: The second set of graphic information data contains simple figures and charts, focusing on key data and results, suitable for concise reports.

[0149] By visually displaying test data and results through charts, the visualization effect of the report is improved, which is easier for readers to understand and analyze. The chart information can enhance the text content and provide richer data information and visual aids.

[0150] Then, in step 54, the first chart information data is embedded into the text content of the technical report to ensure that the position of the chart in the report is reasonable and closely integrated with the relevant text content. In the report text, each chart is described in detail, the data and results displayed in the chart are explained, and their significance and impact are analyzed. The format of the chart and text is adjusted to ensure that the report is beautiful and easy to read as a whole. Through the combination of pictures and text, a more comprehensive display of the test results is provided to facilitate readers to deeply understand and analyze the data. By describing the chart information in detail, the persuasiveness and scientificity of the report are enhanced. Finally, in step 55, a personalized user interface is designed to make the user interface beautiful, concise, and easy to use, and capable of displaying the test report and chart information. A filtering function is provided to allow users to filter the report content and chart information according to specific conditions (such as time, performance parameters, coating type, etc.). A sorting function is provided to allow users to sort the report content and chart information according to specific indicators (such as performance indicator size, date, etc.). An interactive function is also designed to enable users to easily click to view detailed information, thereby enhancing the user experience. Thus, a personalized user interface and filtering and sorting functions are provided to enable users to independently view and analyze specific content according to their own needs, thereby improving the user experience, making the user interface friendly and convenient to operate, and improving the efficiency of users in viewing and analyzing the test report.

[0151] like Figure 2 As shown, a waterborne polyurethane coating detection system comprises:

[0152] Coating acquisition module 201: used to acquire coating data of the sample after infrared light irradiation to obtain spectrum information data;

[0153] Data analysis module 202: used to pre-process the spectral information data and perform hierarchical labeling to obtain label data;

[0154] Intelligent detection module 203: used to identify the marked data to obtain real spectral information data; input the real spectral information data into the intelligent detection model for training to output detection data;

[0155] User interface module 204: It is used to generate a coating inspection report according to the output inspection data, and display the report in the user interface in the form of visual graphics and text.

[0156] An embodiment of the present invention further provides an electronic device according to a water-based polyurethane coating detection method and a system thereof, including a memory, a processor, and a water-based polyurethane coating detection method program stored in the memory and executable on the processor. When the water-based polyurethane coating detection method program is executed by the processor, the steps of the water-based polyurethane coating detection method are implemented. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0157] Those skilled in the art will also appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0158] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0159] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0160] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0161] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0162] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0163] In addition, it should be pointed out that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it can be understood that all or any steps or components of the method and apparatus of the present invention can be implemented in hardware, firmware, software or a combination thereof in any computing device (including processors, storage media, etc.) or a network of computing devices, which can be achieved by those of ordinary skill in the art using basic programming skills after reading the description of the present invention.

[0164] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0165] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A method for detecting a waterborne polyurethane coating, characterized in that: The following steps are involved: Collect the sample coating data after infrared light irradiation to obtain spectrum information data; Parallel collection of coating data of multiple samples after infrared light irradiation; Preprocessing the spectral information data and performing hierarchical labeling to obtain labeled data; Preprocessing is performed based on the photometric distribution data of the coatings of multiple sample panels to obtain preprocessed data; The pre-processed photometric distribution data are marked in sequence; Adding timestamp information to the marked preprocessed data and performing layered processing to obtain layered marked data; Performing sub-tag processing on the layered tag data to obtain sub-tag information data; Capturing the sub-mark information data at a fixed point in the spectral region to obtain explicit spectral data; Draw a spectrum waveform diagram based on explicit spectrum data; The luminosity distribution data is obtained by: ; In the formula, y i is the photometric data point after baseline correction, x i n is the wavelength position number of the original luminosity distribution, a0, a1, ..., a n The series coefficients are used to perform baseline correction to obtain clear data; Determine the photometric characteristic absorption peaks at 350nm and 550nm in the photometric distribution data based on the clear data; pass: ; In the formula, G (x i ) is the derivative value of the luminosity distribution data, y i+1 and i-1 is the luminosity value of the adjacent data point; x i+1 and x i-1 is the wavelength position of adjacent data points, from which characteristic peaks are identified; Perform fixed-point capture according to the characteristic peak position to extract the characteristic luminosity distribution intensity value; Summarize and statistically analyze the characteristic photometric distribution data of multiple sub-markers to obtain multiple dominant photometric characteristic data; Identify the marked data to obtain real spectral information data; Identify the marked data and remove interference features caused by temperature and humidity changes to obtain valid marked data; Based on the effective marking data, the overlapping and cut-off spectra are further identified according to the spectral waveform diagram to obtain the real spectral information data; pass: ; In the formula, r is the correlation coefficient, x is the marked spectral characteristic data, y is the temperature and humidity data, and n is the number of samples, thereby obtaining the temperature and humidity correlation spectral characteristic data; The temperature and humidity correlation spectral feature data is separated from the labeled data and passed through: ; In the formula, Y is the response matrix of temperature and humidity, A is the mixing matrix, s is the spectral data, B is the regression coefficient matrix, and E is the error matrix, which is eliminated to obtain valid labeled data; Visualize the effective marked data in the form of a waveform graph, and preliminarily identify the peak position, bandwidth and morphological characteristics of the waveform graph; Identify overlapping characteristic peak areas through the spectral waveform, and use: Where A i is the amplitude, μ i is the center position, σ i is the standard deviation, n is the number of characteristic peaks, e is the overlap threshold, and x is the position of the spectrum waveform, thereby separating the overlapping peaks; The separated spectrum waveform is further used to identify the breakpoints of the spectrum, and untrustworthy bands are removed or the broken areas are repaired to obtain the real spectrum information data; Input the real spectral information data into the intelligent detection model for training to output the detection data; Generate a coating inspection report based on the output inspection data, and display the report in the user interface with visual graphics and text.

2. A method for detecting a waterborne polyurethane coating according to claim 1, characterized in that: Collect the sample coating data after infrared light irradiation to obtain spectral information data, including: Acquire spectrum information data based on the absorption light wavelength data of the coating data after infrared light irradiation; The luminosity distribution of the coating is obtained through the spectral information data to obtain the luminosity distribution data.

3. A method for detecting a waterborne polyurethane coating according to claim 2, characterized in that: Input real spectral information data into the intelligent detection model for training to output detection data, including: Acquire coating performance data based on real spectral information data, and obtain performance processing data through data processing; The performance processing data is compared with the coating performance data entered into the model database to output the coating inspection data.

4. A method for detecting a waterborne polyurethane coating according to claim 3, characterized in that: Generate a coating inspection report based on the output inspection data, and display the report in a visual graphic on the user interface, including: Design the test report structure based on the test results output by the intelligent detection model; Edit the text contents of two test reports based on the coating test data; And generate two sets of chart information data based on the test data; Embed a chart information data into a test report text content, so that the text content describes the chart information in detail; Design a personalized user interface and provide filtering and sorting functions so that users can view specific content based on their needs.

5. A waterborne polyurethane coating detection system, characterized in that: The system is provided with an electronic device including a memory, a processor, and a waterborne polyurethane coating detection method program stored in the memory and executable on the processor. When the waterborne polyurethane coating detection method program is executed by the processor, the steps of the waterborne polyurethane coating detection method according to any one of claims 1 to 4 are implemented. The system includes: Coating acquisition module: It is used to collect the coating data of the sample after infrared light irradiation to obtain spectral information data; Data analysis module: It is used to pre-process the spectral information data and perform hierarchical labeling to obtain labeled data; Intelligent detection module: It is used to identify the marked data to obtain the real spectral information data; input the real spectral information data into the intelligent detection model for training to output the detection data; User interface module: It is used to generate a coating inspection report based on the output inspection data, and display the report in the user interface with visual graphics and text.

Images