Performance detection method for hygienic product finished product

Abstract

The invention relates to the technical field of hygienic product detection, in particular to a hygienic product finished product performance detection method, which comprises the following steps: S1, obtaining a hygienic product to be detected, and determining a performance index of hygienic product detection; s2, comparing the detection parameters obtained under the performance indexes, generating a detection data set according to the detection parameters, and generating a performance index difference curve according to the difference detected by the detection data set according to the performance indexes; s3, judging index difference distribution of the performance index difference curve, and calibrating an index difference region of the performance index difference curve; s4, historical detection parameters on the index difference area are called, multi-dimensional feature indexes of the detection parameters on the current index difference area are obtained through matching, and then difference degree factors of the multi-dimensional feature indexes are generated; s5, tracing the hygienic product detection process according to the difference degree factor to obtain a decision sequence of hygienic product detection; and the detection efficiency and accuracy of the hygienic products are improved.

Description

Technical Field

[0001] The invention relates to the technical field of sanitary product detection, in particular to a method for detecting the performance of a finished product of a sanitary product. Background Art

[0002] In the production process of sanitary products, it is crucial to ensure that the performance of the finished products meets the established standards. However, traditional sanitary product performance testing methods often rely on manual sampling and laboratory testing, which is not only time-consuming and labor-intensive, but may also lead to inaccurate test results due to human factors. In addition, as consumers' requirements for the quality of sanitary products continue to increase, higher accuracy and efficiency requirements are also put forward for the performance testing of sanitary products.

[0003] For example, Chinese patent publication number CN115266565A discloses a toilet paper softness detection method and system, and a toilet paper softness detection system is an artificial intelligence system in the production field. It includes: obtaining the roughness of toilet paper; obtaining the wrinkle area in the first frame image and the last frame image during the movement of toilet paper; obtaining the grayscale difference of the wrinkle area in the first frame image and the last frame image; obtaining the entropy value difference of the wrinkle area in the first frame image and the last frame image based on the grayscale co-occurrence matrix of the wrinkle area in the first frame image and the last frame image; obtaining the degree of damage of the wrinkle area of ​​toilet paper based on the grayscale difference and entropy value difference of the wrinkle area; obtaining the softness of toilet paper according to the degree of damage of the wrinkle area of ​​toilet paper and the roughness of toilet paper.

[0004] For example, Chinese patent publication number CN115266536A discloses a method for detecting the water absorption performance of diapers. The method includes: obtaining the surface images of the diaper before absorbing water and after the copper sulfate solution has been diffused, and graying them to obtain the first image and the second image; obtaining the water absorption uniformity description factor of the diaper using the mean of the gradient and information entropy of each pixel in the second image; the pixel blocks containing the edge of the detection reference imprint in the first and second images are the target pixel blocks and the pixel blocks to be matched respectively; obtaining the motion vector of each target pixel block based on the similarity between each target pixel block and each pixel block to be matched by a three-step search method; obtaining the water absorption performance evaluation index of the diaper based on the motion vector of each target pixel block and the water absorption uniformity description factor.

[0005] The prior art describes that the softness and uniform water absorption of sanitary products are judged by the grayscale difference of images. However, it is difficult to handle a large number of different types of sanitary products through image detection alone, resulting in reduced efficiency in sanitary product detection. It is necessary to conduct comprehensive processing on the differences of sanitary products under different indicators to identify the differences of sanitary products under corresponding indicators and improve the reliability and accuracy of sanitary product detection. Summary of the invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for testing the performance of finished sanitary products, including: S1, obtaining the sanitary products to be tested, and determining the performance indicators of the sanitary products testing; the performance indicators include absorption amount, absorption rate, re-infiltration amount, surface attachments, pH value and microbial indicators.

[0007] S2, comparing the detection parameters obtained under the performance indicators, and generating a detection data set according to the detection parameters, and detecting the difference of the detection data set according to the performance indicators to generate a performance indicator difference curve.

[0008] S3, determining the indicator difference distribution of the performance indicator difference curve, and calibrating the indicator difference area of ​​the performance indicator difference curve.

[0009] S4, retrieve the historical detection parameters on the indicator difference area, match and obtain the multi-dimensional characteristic indicators of the detection parameters on the current indicator difference area, and then generate the difference degree factor of the multi-dimensional characteristic indicator.

[0010] S5, tracing back the sanitary product testing process according to the differential degree factor to obtain the decision sequence of sanitary product testing.

[0011] The beneficial effects of the present invention are as follows: 1. The present invention automatically performs parameter detection and data analysis by acquiring sanitary products to be tested and determining performance indicators; the acquired detection parameters are screened according to the differences in performance indicators and the difference metrics under different performance indicators to obtain the target performance indicators currently required, which can reduce the amount of data manually processed during detection and reduce errors caused by differences in different performance indicators to improve detection effects.

[0012] 2. The present invention generates a performance index difference curve and calibrates the index difference area. When setting the index difference area, different performance indicators are fitted according to the calculated information entropy, and the edge area that appears when the information entropy is calculated is extracted. It can highlight the differences in the calculation of different data amounts, improve the processing effect for detection scenarios with a large number and a variety of types, and calibrate the existing areas using time intervals and data relative proportional rates. It can highlight the relationship between the data acquisition interval and the corresponding difference, and can intuitively reflect the performance differences of sanitary products in different performance indicators, providing a basis for subsequent decision-making.

[0013] 3. The present invention can more comprehensively reflect the performance characteristics of sanitary products by extracting multidimensional characteristic indicators and calculating differential degree factors. By matching the maximum expansion factor, unit water absorption factor and maximum time factor in the multidimensional performance indicators with relevant data, the data statistics on the indicator difference area can be more inclined to calculate the difference of the data itself relative to the basic water absorption performance, and the indicator difference area is combined with this information, thereby improving the accuracy and reliability of the detection. The differential degree factors are analyzed by using a random forest model, and the key factors affecting the performance of sanitary products can be automatically identified. The data in the sanitary product detection process is output to obtain a decision sequence, and the guiding data for the production and detection of sanitary products is provided, meeting the needs of efficient detection of sanitary products. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0015] Figure 1 The present invention is a flow chart of a method for testing the performance of finished products of sanitary products.

[0016] Figure 2 The present invention is a flow chart of S1 of a method for testing the performance of finished products of sanitary products.

[0017] Figure 3 The present invention is a flow chart of S2 of a method for testing the performance of finished products of sanitary products.

[0018] Figure 4 The present invention is a flow chart of S3 of a method for testing the performance of finished products of sanitary products.

[0019] Figure 5 The present invention is a flow chart of S32 of a method for testing the performance of finished products of sanitary products.

[0020] Figure 6 The present invention is a flow chart of S4 of a method for testing the performance of finished products of sanitary products. DETAILED DESCRIPTION

[0021] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0022] See also Figure 1 A method for testing the performance of finished sanitary products includes: S1, obtaining sanitary products to be tested and determining performance indicators for the sanitary products testing; the performance indicators include absorption capacity, absorption speed, re-infiltration capacity, surface attachments, pH value and microbial indicators.

[0023] S2, comparing the detection parameters obtained under the performance indicators, and generating a detection data set according to the detection parameters, and detecting the difference of the detection data set according to the performance indicators to generate a performance indicator difference curve.

[0024] S3, determining the indicator difference distribution of the performance indicator difference curve, and calibrating the indicator difference area of ​​the performance indicator difference curve.

[0025] S4, retrieve the historical detection parameters on the indicator difference area, match and obtain the multi-dimensional characteristic indicators of the detection parameters on the current indicator difference area, and then generate the difference degree factor of the multi-dimensional characteristic indicator.

[0026] S5, tracing back the sanitary product testing process according to the differential degree factor to obtain the decision sequence of sanitary product testing.

[0027] In one embodiment of the present invention, step S1 mainly obtains performance indicators existing in the sanitary products to be tested, and tests the sanitary products according to the data presented by these performance indicators.

[0028] The performance indicators set here will be classified according to the values ​​required by the data. For example, when the sanitary products tested are sanitary napkins, sanitary napkins, toilet paper and other different products, different performance indicators will be set according to the model and size of the tested products to determine the actual values ​​and types of these products tested.

[0029] The absorption amount and absorption speed will show the performance of sanitary products under normal water absorption, and determine whether the absorption amount and absorption speed meet the standards. For example, the absorption amount must reach seven times the weight of the sanitary product; the pH value indicates that it will not irritate the skin when used, such as the pH value of the sanitary product before and after water absorption meets the normal use standards; the back permeability is a test to determine whether the sanitary products will have reverse osmosis, resulting in poor use effect. At this time, the method of applying pressure after water absorption is used to understand the data of the product under different water absorption amounts and pressures, and to determine whether the overall form can remain dry to improve the user experience; microbial indicators are to determine whether there are corresponding problems during production, that is, the data of bacteria, mold and yeast detection during production is used to determine whether contamination occurs.

[0030] According to these set test items, the corresponding situation during the current test can be discovered. At the same time, according to the different performance indicators of the test, it will be judged whether there is correlation between the data under the performance indicators under the corresponding differences, so as to identify other situations that may exist in the current sanitary products when the corresponding problems are detected, and finally complete the comparison of the data.

[0031] When determining the performance indicators for sanitary product testing, we must first determine the classification of the sanitary products to be tested, as well as the items to be tested in combination during testing, so as to obtain a set of performance indicators.

[0032] like Figure 2 As shown, the implementation method of step S1 also includes: S11, taking the sanitary products to be tested as input, and obtaining the test items of the sanitary products to be tested; the test items are setting items that include a part of performance indicators. For example, during abstract testing, multiple sanitary products are tested together, and the sanitary products will correspond to the items that need to be tested, and then be tested in turn.

[0033] S12, the test items of the sanitary products to be tested, construct a selection matrix corresponding to the sanitary products to be tested, and use the selection matrix to determine the performance indicators of the sanitary products testing; at this time, the selection matrix is ​​composed according to the test items and the test priority to indicate the performance indicators that should be tested first for the current sanitary products, and give priority to testing some performance indicators.

[0034] In one embodiment of the present invention, step S2 mainly compares the detection parameters obtained under the performance index, and generates an index difference curve according to the differences presented by the detection parameters.

[0035] The test parameters will indicate the specific quantitative values ​​of the performance indicators tested at this time. For example, the test parameters may be: the absorption capacity of a sanitary napkin sample is 200ml, the time required for a sanitary napkin sample to completely absorb 50ml of liquid is 30 seconds, the pH value before absorbing water is 6.5, and the pH value after absorbing water is 7.0; under a specific pressure, the reverse osmosis volume of a sanitary napkin sample is 5g, the total bacterial count is 100CFU / g, etc. These data will be expressed in the form of data for performance indicators under different test quantities. Finally, a performance indicator difference curve is obtained based on the difference between this part of the data and the numerical values ​​when the sanitary products are designed.

[0036] The detection data set will include multiple groups of detection parameters detected according to the performance indicator type, generating a data set containing the test data of the current sanitary product. For example, when the detection parameters are pH value and chemical composition, the data displayed by the current sanitary product under the corresponding test will be determined according to the main detected part at this time, such as the numerical change of pH value before and after testing with the test agent, and whether the pH value changes differently due to different absorption amounts when the absorption amount is large.

[0037] At the same time, if the structure of the sanitary product is displayed before and after use, the test data set will contain data in the form of images to determine whether the structure of the current sanitary product is stable during the test, and whether the expected image shape before and after the test can meet the design standards, thereby determining whether the currently produced finished product meets the hygiene and usage standards.

[0038] For the coordinate system of the performance indicator difference curve, the horizontal axis can represent different performance indicator names or test numbers. For example, when comparing the absorption capacity, absorption speed and other performance of multiple products, the X-axis can be labeled with absorption capacity, absorption speed, pH value, etc. in turn; the vertical axis displays the value of the specific parameter. At the same time, according to the different forms of the collected indicator data, the double Y-axis coordinates will be used to display the current collected data.

[0039] For example, in view of the performance of the performance indicators before and after water absorption, when generating the performance indicator difference curve, the absorption amount and absorption speed will be used as the horizontal coordinates of the coordinate system, and other measured and calculated performance indicators will be used as the vertical coordinates of the coordinate system, and then a performance indicator difference curve will be formed to evaluate the differences under the corresponding water absorption capacity. At the same time, when the data collected by the performance indicators change over time, the time will be used as the horizontal coordinate, and the other measured and calculated performance indicators will be used as the vertical coordinate to set the performance indicator difference curve; the double vertical coordinate is to use two indicators with small dimensional difference and close values ​​in the performance indicators as two vertical coordinates, and then use time as the horizontal coordinate for comprehensive display; for example, the performance indicators select the two values ​​of re-infiltration amount and absorption amount, and after converting these two values ​​into values ​​in the range of 0-1, use the performance indicator difference curve to compare the re-infiltration amount and absorption amount at different times at the same time.

[0040] When generating a performance index difference curve based on the difference in performance index detection of the detection data set, it is necessary to consider the setting form of the corresponding performance index when the detection parameter difference is the smallest, such as Figure 3 As shown, the implementation method of step S2 also includes: S21, setting the mapping relationship between the detection parameters and the performance indicators according to the acquisition location of the detection parameters. Since the detection parameters are the actual data taken by the performance indicators for a certain part, and the performance indicators represent the form of the classification in which the detection parameters are located, determining the mapping relationship is to initially define the detection parameters and the performance indicators, and generate a table covering the mapping relationship to determine the relationship between the two parameters when the content of the performance indicator test and the detection parameter value change.

[0041] S22, connect all the detection parameters according to the values ​​of the detection parameters, select the detection parameters with the smallest adjacent difference under a single performance indicator, and take the sum of the distances of the connected detection parameters as the first relationship factor; this factor focuses on the differences between adjacent detection parameters within a single performance indicator. Specifically, it connects by selecting the detection parameters with the smallest adjacent difference under a specific performance indicator, and calculates the sum of the distances after these connections. The distance here can be understood as the absolute difference between the two detection parameter values ​​or the standardized difference measure. Helps identify the consistency and continuity of the data within a certain performance indicator; a smaller first relationship factor means that under this performance indicator, the changes in the values ​​of each detection parameter are relatively smooth and stable, without abrupt fluctuations.

[0042] S23, connect the detection connections with the smallest adjacent comprehensive differences under all performance indicators, and take the sum of the distances of the detection parameters after connection as the second relationship factor; compared with the first relationship factor, the second relationship factor focuses on the comprehensive performance between all performance indicators. It considers the differences between adjacent detection parameters across different performance indicators, selects those detection parameters with the smallest adjacent comprehensive differences for connection, and calculates the sum of the distances after these connections. At this time, when calculating the adjacent comprehensive differences, all performance indicator values ​​will be converted to the range of 0-1 to obtain a dimensionless comprehensive difference. This comprehensive difference can evaluate the overall coordination and consistency of the entire product in multi-dimensional performance indicators. A lower second relationship factor indicates that the various performance indicators are well balanced, without obvious shortcomings or extremes.

[0043] S24, comparing the first relationship factor and the second relationship factor with the set target value, and taking the compared performance index as the target performance index, and generating a performance index difference curve according to the target performance index.

[0044] The purpose of comparing the two relationship factors is to ensure that the performance indicators support each other, avoid excessive emphasis on one aspect at the expense of other aspects of performance, and thus improve the overall user satisfaction; the set target value will select different values ​​according to the different performance indicators corresponding to the first relationship factor and the second relationship factor. For example, the implementation method of the target performance indicator is: when the first relationship factor and the second relationship factor are both greater than the set target value, the performance indicators corresponding to the first relationship factor and the second relationship factor are used as the target performance indicator; when the first relationship factor is greater than the set target value and the second relationship factor is less than the set target value, the performance indicator corresponding to the first relationship factor is used as the target performance indicator; when the first relationship factor is less than the set target value and the second relationship factor is greater than the set target value, the performance indicator corresponding to the second relationship factor is used as the target performance indicator. This is to select performance indicators that may change significantly in the form of deviations. If the data obtained in this way is not ideal compared to the actual detection, the different changes in the values ​​of the first relationship factor and the second relationship factor over time can also be used for setting. For example, the performance indicator when the trends of the first relationship factor and the second relationship factor are the same is set as the target performance indicator. This same trend represents the part where the values ​​of the first relationship factor and the second relationship factor increase and decrease together over time, and the performance indicator corresponding to this part is identified. By knowing the current trend of sanitary product detection, it is convenient to adjust the detection method and quantity in a timely manner.

[0045] In one embodiment of the present invention, step S3 mainly determines the indicator difference distribution of the performance indicator difference curve, and after obtaining multiple areas according to the indicator difference distribution, calibrates the indicator difference area belonging to the current performance indicator difference area; the indicator difference distribution indicates under what circumstances there are large differences and small differences in the performance indicator difference curve when the corresponding numerical value is selected for the current detection parameter, so as to quantify the relationship between the performance indicators, identify those areas with large differences, and mark them at the corresponding sanitary product detection position, and further analyze the sanitary products based on this part of the data.

[0046] At this time, the distribution of identified indicator differences will present various forms. It is necessary to clarify the main performance indicators and detection parameters recorded when there is a performance indicator difference curve, as well as the relative regional distribution of these data, and select the indicator difference situation that you want to deal with.

[0047] like Figure 4As shown, the implementation method of step S3 also includes: S31, using each performance indicator existing in the performance indicator difference curve to calculate the information entropy value of each performance indicator at the corresponding detection position; the information entropy value is calculated according to the probability of occurrence of each point in the performance indicator difference curve, and the information entropy value corresponding to each performance indicator in this curve is calculated. For the detection position, it represents the batch, number and type of sanitary products when the performance indicator is obtained, and represents the actual source of the corresponding data under the current performance indicator; at this time, it is necessary to map the corresponding data of the conventional detection position of the sanitary products and the performance indicator difference curve, and finally find the occurrence probability in these areas to quantify the uncertainty or degree of change of each performance indicator at the corresponding detection position.

[0048] S32, according to the distribution of information entropy values, the detection positions are fitted and divided into multiple edge areas; based on the fitting results, the positions where the information entropy values ​​change significantly can be identified, and these positions are used as boundaries to divide multiple areas with similar characteristics. These areas will indicate similar areas where the information entropy values ​​of sanitary products will change significantly during testing.

[0049] In order to divide into multiple regions with similar edges, such as Figure 5 As shown, step S32 also includes: S321, recording the detection position of the median value of each performance indicator when calculating the information entropy, and connecting the detection position to obtain multiple detection areas; at this time, the median value when calculating the information entropy is taken as a point, and the points where other performance indicators are calculating the information entropy are found, and these points are connected, so that the areas represented by all the data can be divided into multiple detection areas, and when these areas are connected, there will be situations where the values ​​are close.

[0050] S322, adjust the amount of data when calculating the information entropy, and extract the adjacent area difference and the maximum area difference of the detection area when the amount of data calculated by the information entropy changes; this area is to determine the change in the size of the original detection area after the amount of data calculated by the information entropy changes, and extract this part of the difference area that exists under the continuous adjustment, and then obtain the maximum area difference of the detection area after the data amount is adjusted. This can reduce the total amount of data required for information entropy processing to shorten the calculation time of information entropy. At the same time, this will tend to emphasize sparse and distributed data. Using these data as the difference in the detection of sanitary products will help identify mutations or abnormal areas in the data, more intuitively display the characteristics of the data itself, and enhance the visualization of the data.

[0051] S323, setting the area corresponding to the adjacent area difference and the maximum area difference as the edge area. In this way, according to the different areas where the data exists on the performance index difference curve, a region with intuitive data features and small data calculation is selected to identify the prominent feature information when detecting sanitary products.

[0052] Afterwards, the edge area is identified, and the information existing in the edge area is marked to set the indicator difference area that needs to be output. Therefore, the subsequent implementation method of step S3 is: S33, the edge area is identified for difference, and the indicator difference distribution of the performance indicator difference curve is obtained; this part essentially identifies what performance indicators exist in the edge area, so as to know what characteristic information exists in the currently extracted edge area, and mark this part of the existing data.

[0053] S34, according to the time interval and data relative ratio rate of each performance indicator in the indicator difference distribution, calibrate the indicator difference area of ​​the performance indicator difference curve. This part further identifies the features existing in the edge area, and further describes the data in this area according to the marked indicator difference distribution. The time interval is to display the indicator difference distribution in different time intervals as data areas with different colors or marks. The data relative ratio rate is to indicate the change ratio of the average values ​​of these data at different time intervals. After these two values ​​are used to identify the corresponding data, the indicator difference area described is obtained.

[0054] In one embodiment of the present invention, step S4 mainly compares the difference between the historical detection parameters and the current detection parameters in the indicator difference area to generate a difference degree factor describing the similarity between the current data and the historical data. At this time, the historical detection parameters are matched in the same area, and the historical detection parameters are set with the same value as the current one to see whether there is difference information corresponding to the current situation, and a multidimensional feature index is generated according to these existing identical parts. This multidimensional feature index will finally get a difference degree factor, which will indicate the total difference degree represented by the multidimensional feature index.

[0055] In order to make the description of the actual situation of sanitary products better, the performance indicators described in the indicator difference area will be matched according to the area of ​​each detection parameter, and then a multidimensional feature indicator will be generated. This multidimensional feature indicator will be strengthened according to the factors of the detection parameters in water absorption to generate a differential degree factor for the performance invention of sanitary products. This differential degree factor can describe the characteristics corresponding to different differential degrees of basic performance of sanitary products after performance testing, so as to facilitate the subsequent summary of these existing characteristics and find a method suitable for the current production or adjustment of sanitary products to improve the efficiency and accuracy of performance testing.

[0056] like Figure 6 As shown, the implementation method of step S4 includes: S41, performing multidimensional feature extraction on the detection parameters in the current indicator difference area to obtain multidimensional feature indicators; the multidimensional feature indicators include: maximum expansion factor, unit water absorption factor and maximum time factor.

[0057] The maximum expansion factor refers to the increase in volume of sanitary products, such as sanitary napkins or diapers, after they have completely absorbed liquid. It reflects the physical changes that may occur in the product during use, affecting the comfort and fit of the product. The maximum expansion factor uses the ratio of pixels occupied by sanitary products in the image before and after expansion to determine the increase in the proportion of absorbed liquid and whether excessive expansion has occurred; the increase in the proportion after expansion is divided by the normal expansion ratio in the historical data to obtain a ratio. The largest value in this ratio is the maximum expansion factor currently required.

[0058] The specific absorption factor can be used to measure how much liquid a hygiene product can absorb per unit area or unit weight. This helps to evaluate the product's water absorption efficiency and economics. The amount of liquid that can be absorbed is determined by the ratio of the weight after water absorption to the weight before water absorption, and this ratio is divided by the normal absorption ratio to quantify the ratio of each hygiene product relative to normal conditions.

[0059] The maximum time factor refers to the maximum time required for sanitary products to completely absorb liquid from the beginning of contact. This indicator is very important to users because it directly affects the instant drying effect and user experience of the product. The maximum time factor will select the data with the longest liquid absorption time under the same conditions to obtain the maximum time factor. This maximum time factor will be proportional to the normal water absorption time to obtain a maximum time factor. When calculating the three indicators included in the multidimensional feature indicators, the normal data used is the average value of the historical data with the same conditions as the current indicator difference area to quantify the relative outstanding differences of the products at this time.

[0060] S42, calculating the cosine similarity between the detection parameter and the historical difference parameter in the current indicator difference area under the value of the multidimensional feature indicator, and weighted averaging the cosine similarity to obtain the difference degree factor of the multidimensional feature indicator.

[0061] At this time, the actual values ​​of the multidimensional feature indicators are obtained, and the values ​​of the detection parameters in the historical data are selected. The similarity between these values ​​and the current detection parameters is calculated to determine whether the historical data is highly correlated with the current data when the same indicator value as the current detection appears in the historical data, to determine whether the problems existing in the current detection can be adjusted according to the content in the historical data. Since the value of cosine similarity is determined according to the number of multidimensional feature indicators, the cosine similarities of the three indicators included are weighted and averaged to quantify the average value of the similarity between the current data and the historical data, and this average cosine similarity value is considered to be the differential degree factor currently required. This is to identify the pattern similarity of these distribution areas in the basic water absorption performance of sanitary products after the processing of the indicator difference distribution is completed, and the impact of this similarity as a key dimension is used to complete the evaluation of sanitary products. It is mainly to improve the analysis of the product after the different analysis segments are analyzed to see whether there are factors affecting the quality.

[0062] At the same time, under the value of the multidimensional feature index, the detection parameters to be calculated are generally all the parameters related to the detection parameters used to calculate the multidimensional feature index. In addition to the relevant parameters included in the calculation, such as liquid absorption time, water absorption, size, weight, etc., the detection parameters under performance indicators such as wrinkle degree, pH value, water seepage, chemical composition and microbial indicators that may correspond to the multidimensional feature index will also be compared with the historical detection parameters, and then an average value of the cosine similarity is obtained to complete the processing of this multidimensional feature index.

[0063] In one embodiment of the present invention, step S5 mainly traces the sanitary product inspection process according to the differential degree factor, and after processing the data obtained at this time using the random forest model, constructs a decision sequence with part of the output data to obtain data with high accuracy and accurate description of the current inspection situation. The decision sequence represents the root cause of the existence of sanitary products after tracing, and this root cause represents the link where the sanitary products have problems, and the reasons that may cause the current sanitary products to be unqualified. After finding the source of these reasons, it is possible to know the process that needs to be paid attention to and adjusted during the production of the current sanitary products, and speed up the verification of subsequent products, and quickly complete the inspection of sanitary products.

[0064] When generating a decision sequence, the differential degree factor will be used as the basis to generate multiple decision trees. In the form of a random forest model, the possible relevant contents of the differential degree factor will be listed, and the prediction situation of different differential degree factors will be emphasized to complete the adjustment of the overall sanitary product testing process and data processing process, so as to improve the reliability and generalization ability of sanitary product testing.

[0065] The implementation method of step S5 includes: taking the differential degree factor as an input variable, and setting a random forest model corresponding to the differential degree factor; the implementation method of the random forest model is: taking the differential degree factor as the root node of the decision tree in the random forest model, and the data corresponding to the differential degree factor as the intermediate nodes and leaf nodes of the decision tree in the random forest model, and dividing the intermediate nodes into leaf nodes according to the Gini impurity, so as to obtain the random forest model; when calculating the Gini impurity at this time, it is necessary to determine the maximum Gini impurity value of each intermediate node and leaf node; identifying the information in the leaf nodes of the decision tree in the random forest model, obtaining the node information bias represented by the differential degree factor, and taking the part of the data with the largest value in the node information bias as the decision sequence for sanitary product detection.

[0066] Node information bias includes calculating the relative importance and out-of-bag error of the current leaf node; relative importance indicates the contribution of the characteristic value of the data in the current leaf node to the characteristic value of the data in all leaf nodes. For example, the ratio of the amount of data on the current leaf node to the amount of data on all leaf nodes is used, and multiplied by the difference between the Gini impurity of the current leaf node and the Gini impurity of the leaf node to obtain the characteristic importance of the leaf node; for the out-of-bag error, the average error rate of all leaf nodes except the data of the current leaf node is determined. The error rate will use the value of the data in the leaf node to compare with the predicted value in the database to obtain the corresponding error rate. Finally, when the relative importance and out-of-bag error are the largest, the corresponding data will be output, so that we can know the group of data with the largest relative importance in the difference when there are differences in the current sanitary product detection. This group of data will have the greatest impact on the sanitary product detection process and needs to be output as the sequence data needed for subsequent decision-making.

[0067] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. A method for testing the performance of finished products of sanitary products, characterized in that: include: S1, obtaining sanitary products to be tested and determining performance indicators of the sanitary products testing; S2, comparing the detection parameters obtained under the performance index, and generating a detection data set according to the detection parameters, and generating a performance index difference curve by comparing the detection data set with the difference of the performance index detection; S3, determining the indicator difference distribution of the performance indicator difference curve, and calibrating the indicator difference area of ​​the performance indicator difference curve; S4, retrieve the historical detection parameters on the indicator difference area, match and obtain the multi-dimensional characteristic indicators of the detection parameters on the current indicator difference area, and then generate the difference degree factor of the multi-dimensional characteristic indicators; S5, tracing back the sanitary product testing process according to the differential degree factor to obtain the decision sequence of sanitary product testing.

2. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: Performance indicators include absorption capacity, absorption rate, re-infiltration amount, surface attachment, pH value and microbial indicators.

3. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: The implementation of step S1 also includes: S11, taking the sanitary products to be tested as input, and obtaining the test items of the sanitary products to be tested; S12, constructing a selection matrix corresponding to the sanitary products to be tested for the test items of the sanitary products to be tested, and using the selection matrix to determine the performance index of the sanitary products test.

4. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: The implementation of step S2 also includes: S21, setting a mapping relationship between the detection parameters and the performance indicators according to the acquisition location of the detection parameters; S22, connecting all detection parameters according to the values ​​of the detection parameters, selecting the detection parameters with the smallest adjacent difference under a single performance indicator, and taking the sum of the distances of the connected detection parameters as the first relationship factor; S23, connecting the detection connections with the smallest adjacent comprehensive differences under all performance indicators, and taking the sum of the distances of the detection parameters after the connection as the second relationship factor; S24, comparing the first relationship factor and the second relationship factor with the set target value, and taking the compared performance index as the target performance index, and generating a performance index difference curve according to the target performance index.

5. A method for testing the performance of finished products of sanitary products according to claim 4, characterized in that: The target performance indicators are achieved by: When the first relationship factor and the second relationship factor are both greater than the set target values, the performance indicators corresponding to the first relationship factor and the second relationship factor are used as the target performance indicators; When the first relationship factor is greater than the set target value and the second relationship factor is less than the set target value, the performance indicator corresponding to the first relationship factor is used as the target performance indicator; When the first relationship factor is less than the set target value and the second relationship factor is greater than the set target value, the performance indicator corresponding to the second relationship factor is used as the target performance indicator.

6. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: The implementation of step S3 also includes: S31, using each performance indicator existing in the performance indicator difference curve, calculating the information entropy value of each performance indicator at the corresponding detection position; S32, fitting the detection position according to the distribution of the information entropy value, and dividing it into multiple edge areas; S33, performing difference identification on the edge area to obtain an indicator difference distribution of a performance indicator difference curve; S34, calibrating the indicator difference region of the performance indicator difference curve according to the time interval and data relative ratio rate of obtaining each performance indicator in the indicator difference distribution.

7. A method for testing the performance of finished products of sanitary products according to claim 6, characterized in that: The step S32 further includes: S321, recording the detection position of the median value of each performance indicator during information entropy calculation, and connecting the detection positions to obtain multiple detection areas; S322, adjusting the data volume during information entropy calculation, and extracting the adjacent region difference and the maximum region difference of the detection area when the data volume of information entropy calculation changes; S323: Set the area corresponding to the adjacent area difference and the maximum area difference as the edge area.

8. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: The implementation of step S4 includes: S41, performing multidimensional feature extraction on the detection parameters in the current indicator difference area to obtain multidimensional feature indicators; the multidimensional feature indicators include: maximum expansion factor, unit water absorption factor and maximum time factor; S42, calculating the cosine similarity between the detection parameter and the historical difference parameter in the current indicator difference area under the value of the multidimensional feature indicator, and weighted averaging the cosine similarity to obtain the difference degree factor of the multidimensional feature indicator.

9. A method for testing the performance of finished products of sanitary products according to claim 1, characterized in that: The implementation of step S5 includes: The differential degree factor is used as the input variable, and a random forest model corresponding to the differential degree factor is set; the information in the leaf nodes of the decision tree in the random forest model is identified to obtain the node information bias represented by the differential degree factor, and the part of the data with the largest value in the node information bias is used as the decision sequence for sanitary product testing.

10. A method for testing the performance of finished products of sanitary products according to claim 9, characterized in that: The implementation of the random forest model is as follows: The difference degree factor is used as the root node of the decision tree in the random forest model, and the data corresponding to the difference degree factor is used as the intermediate nodes and leaf nodes of the decision tree in the random forest model. The intermediate nodes are divided into leaf nodes according to the Gini impurity, so as to obtain the random forest model.

Images