Image recognition-based paint anti-aging performance detection method

Abstract

The present application relates to the technical field of image information processing, and especially relates to a paint anti-aging performance detection method based on image recognition, which determines a first comparison sample and a plurality of second comparison samples corresponding to the current paint according to the comparison of component information, identifies and obtains the aging trend characteristic quantity of each sample through historical test images, draws an aging trend fluctuation curve, determines a plurality of feature analysis time windows for the aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample respectively, determines the quantity proportion of each comparison sample screening by calculating the anti-aging performance difference of each type of feature analysis time window, and determines the anti-aging life prediction interval of the current paint according to the anti-aging historical life of the first comparison sample and the second comparison sample meeting the quantity proportion. The present application avoids the limitation that the paint anti-aging performance detection relies on long-term test, and improves the detection efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image information processing, and in particular to a paint anti-aging performance detection method based on image recognition. BACKGROUND

[0002] In the quality evaluation of paint products, the anti-aging performance is a key indicator to measure its durability, which is directly related to the appearance maintenance and protection effect of the coated object. Artificial observation and evaluation rely on subjective experience, and the consistency of the results is poor. With the application of image recognition technology in the field of material detection, a new technical means is provided for paint aging evaluation. Aging features such as cracks can be extracted by analyzing surface images, but the existing methods lack pertinence in sample selection, resulting in insufficient representativeness of the reference samples, which makes the prediction of the paint anti-aging performance inaccurate. Therefore, it is an urgent problem for current technical personnel to efficiently and accurately evaluate the anti-aging performance of paint.

[0003] For example, Chinese Patent Publication No. CN118735915A discloses a paint detection method and system based on multi-dimensional visual analysis, which belongs to the technical field of visual detection. A machine learning model is used to detect the target component of the image to be detected, and output the detection result of whether the target component has paint. The construction method of the machine learning model includes: S1, obtaining a basic image set; S2, performing target detection and labeling on each basic image in the basic image set; S3, processing the identified image; establishing an array with paint color and without paint color; S4, dividing the array into a training set and a test set, and using a machine learning model to learn the training set and the test set to obtain the machine learning model. The machine learning model in the present application is trained by an array including coordinate points, HSV values, and paint color identification. The trained machine learning model is used for intelligent detection of the image to be detected.

[0004] The existing technology also has the following problems:

[0005] The existing technology requires a long test period for paint anti-aging performance detection, which cannot meet the demand for rapid performance verification, affecting the efficiency of paint anti-aging performance detection. SUMMARY

[0006] Therefore, the present application provides a paint anti-aging performance detection method based on image recognition to overcome the problem that the existing technology requires a long test period for paint anti-aging performance detection, which cannot meet the demand for rapid performance verification.

[0007] To achieve the above purpose, the present application provides a paint anti-aging performance detection method based on image recognition, which comprises:

[0008] Obtain component information of a plurality of paint samples, and determine a plurality of first comparison samples and a plurality of second comparison samples corresponding to the current paint according to the comparison of the component information;

[0009] Identify and obtain the aging trend characteristic quantity of the historical test images of the first comparison samples and the second comparison samples, draw the aging trend fluctuation curve, and screen the aging sensitive nodes from the aging trend fluctuation curve of the first comparison samples and the aging trend fluctuation curve of the second comparison samples respectively according to the curve characteristics of the aging trend fluctuation curve;

[0010] Sort the aging sensitive nodes in time sequence, determine a plurality of feature analysis time windows based on the preset node distribution position constraint condition and the time interval constraint condition, and determine the category of each feature analysis time window according to the time starting point position distribution of the feature analysis time window;

[0011] Calculate the anti-aging performance difference of each category of feature analysis time window according to the difference of the aging trend characteristic quantity of the first comparison samples and the second comparison samples at a plurality of times and the duration of the feature analysis time window;

[0012] Screen a plurality of first comparison samples and a plurality of second comparison samples that meet the quantity proportion, and determine the anti-aging life prediction interval of the current paint according to the anti-aging historical life of the first comparison samples and the second comparison samples;

[0013] The quantity proportion is determined according to the anti-aging performance difference of each category of feature analysis time window.

[0014] Further, the process of determining the first comparison samples and the plurality of second comparison samples corresponding to the current paint according to the comparison of the component information comprises:

[0015] Obtain the anti-aging component proportion in the component information of the current paint;

[0016] Determine the paint with an anti-aging component proportion greater than or equal to the anti-aging component proportion of the current paint as the first comparison sample corresponding to the current paint;

[0017] Determine the paint with an anti-aging component proportion less than the anti-aging component proportion of the current paint as the second comparison sample corresponding to the current paint.

[0018] Further, the process of identifying and obtaining the aging trend characteristic quantity of the historical test images of the first comparison samples and the second comparison samples comprises:

[0019] Obtain the historical test images of the first comparison samples and the second comparison samples, wherein the historical test images are surface images of the first comparison samples and the second comparison samples in historical tests;

[0020] obtaining crack area proportions of the paint surface at a plurality of test time nodes in the historical test images of the first comparison samples, and crack area proportions of the paint surface at a plurality of test time nodes in the historical test images of the second comparison samples;

[0021] determining an average value of the crack area proportions in the historical test images of the first comparison samples as the first aging trend characteristic quantity, and determining an average value of the crack area proportions in the historical test images of the second comparison samples as the second aging trend characteristic quantity.

[0022] Further, a first aging trend fluctuation curve of the first aging trend characteristic quantity changing with time and a second aging trend fluctuation curve of the second aging trend characteristic quantity changing with time are drawn under the same coordinate system;

[0023] The horizontal axis of the coordinate system is time, and the vertical axis is the aging trend characteristic quantity.

[0024] Further, the process of screening the aging sensitive nodes for the aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample includes:

[0025] The curve slopes of a plurality of data points are determined on the first aging trend fluctuation curve and the second aging trend fluctuation curve respectively, and the data points with the curve slopes exceeding a preset slope screening value are determined as the aging sensitive nodes.

[0026] Further, the process of determining the feature analysis time window includes:

[0027] Two aging sensitive nodes meeting the node distribution position constraint condition and the time length interval constraint condition are respectively taken as a time starting point and a time ending point of the feature analysis time window to determine a plurality of feature analysis time windows.

[0028] The node distribution position constraint condition is that any one of the two aging sensitive nodes is on the first aging trend fluctuation curve, and the other aging sensitive node is on the second aging trend fluctuation curve.

[0029] The time length interval constraint condition is that the interval time length of the two aging sensitive nodes does not exceed a preset interval time length threshold.

[0030] Further, the process of determining the category of each feature analysis time window according to the time starting point position distribution of the feature analysis time window includes:

[0031] If the time starting point is on the first aging trend fluctuation curve, the feature analysis time window is determined as the first dominant category.

[0032] If the time starting point is located on the second aging trend fluctuation curve, the feature analysis time window is determined as the second dominant category.

[0033] Further, the process of calculating the anti-aging performance difference of each feature analysis time window includes:

[0034] The aging trend characteristic difference of the first aging trend fluctuation curve and the aging trend characteristic difference of the second aging trend fluctuation curve in each feature analysis time window are determined respectively.

[0035] The aging trend characteristic difference maximum value and the duration of the feature analysis time window are weighted to obtain the anti-aging performance difference of each feature analysis time window.

[0036] The average of the anti-aging performance differences of the first dominant category is determined as the first anti-aging performance difference, and the average of the anti-aging performance differences of the second dominant category is determined as the second anti-aging performance difference.

[0037] Further, the process of determining the anti-aging life prediction interval of the current paint includes:

[0038] The anti-aging historical life composition values corresponding to the first comparison sample and the second comparison sample in the quantity proportion are formed into a numerical set.

[0039] The anti-aging life prediction interval of the current paint is determined according to the average of the numerical values in each numerical set.

[0040] The interval starting point of the anti-aging life prediction interval is the minimum value of the average of the numerical values, and the interval ending point is the maximum value of the average of the numerical values.

[0041] Further, the process of determining the quantity proportion includes:

[0042] The ratio of the first anti-aging performance difference and the second anti-aging performance difference is calculated, and the reciprocal of the ratio is determined as the quantity proportion of the first comparison sample and the second comparison sample.

[0043] Compared with the prior art, the beneficial effects of the present application are that the component information of the paint sample of the present application determines the first comparison sample and several second comparison samples corresponding to the current paint, identifies and obtains the aging trend characteristic quantity of each sample through the historical test image, draws the aging trend fluctuation curve, screens the aging sensitive nodes of the aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample respectively, determines the number of characteristic analysis time windows and the category of each characteristic analysis time window, determines the number ratio of the first comparison sample and the second comparison sample screened through the calculation of the anti-aging performance difference quantity of each category of characteristic analysis time window, and determines the anti-aging life prediction interval of the current paint according to the anti-aging historical life of the first comparison sample and the second comparison sample meeting the number ratio, thereby avoiding the limitation that the anti-aging performance detection of the paint depends on long-term test, and improving the detection efficiency.

[0044] Further, the present application converts the abstract aging phenomenon into specific numerical indicators by obtaining the surface image of the sample in the historical test and extracting the crack area proportion at different test time nodes, realizes the accurate quantification of the aging degree of the paint, and uses the average value of the crack area proportions of the several first comparison samples and second comparison samples as the first and second aging trend characteristic quantities, which can effectively reduce the contingency and deviation of the test data of a single sample and more accurately reflect the overall aging law of the same type of sample.

[0045] Further, the present application screens the aging sensitive nodes of the aging trend fluctuation curve of the first comparison sample and the second comparison sample, determines the data points exceeding the preset slope screening value as the aging sensitive nodes by calculating the slope of several data points of the curve, and the greater the slope, the faster the change of the crack area proportion at this time node, so as to sharply lock the key time point of the sudden change of the aging speed of the paint and focus on the stage of significant change of the aging characteristics, thereby realizing the accurate identification of the key turning point in the aging process.

[0046] Further, the present application determines the prediction interval according to the average value of the numerical values in each numerical set, which can smooth the accidental fluctuation of the historical life of a single sample, so that the obtained interval can more reflect the overall life level of the same type of sample, and the setting of the starting point and the ending point of the interval not only covers the anti-aging life range that the current paint can reach, but also avoids the distortion of the interval caused by individual extreme sample data through the constraint of the average value.

[0047] Further, the present application does not need to perform long-term aging test on the current paint, but quickly determines the anti-aging life prediction interval by means of historical sample data and scientific analysis method, greatly shortens the detection period, avoids the limitation that the anti-aging performance detection of the paint depends on long-term test, and improves the detection efficiency.

[0048] Further, the present application adopts the inverse of the ratio as the quantity ratio, because the performance difference quantity is inversely related to the sample reference value. When the first anti-aging performance difference quantity is larger, it indicates that the aging performance fluctuation of the first comparison sample is more intense, and the stability of its life data is poorer. At this time, the proportion of the first comparison sample in the prediction sample needs to be reduced. Conversely, if the second anti-aging performance difference quantity is larger, the proportion of the second comparison sample needs to be reduced. This inverse conversion mechanism can automatically balance the weights of the two types of samples, so that the composition of the reference sample matches the actual performance fluctuation characteristics, and finally makes the anti-aging life prediction interval determined based on these samples more consistent with the real performance level of the current paint, avoiding the prediction result being biased towards the extreme value due to the excessive participation of a certain type of sample, and improving the detection accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 A step diagram of the paint anti-aging performance detection method based on image recognition of an embodiment of the present application;

[0050] Figure 2 A step diagram of determining the first comparison sample and the second comparison sample of an embodiment of the present application;

[0051] Figure 3 A step diagram of obtaining the aging trend characterization quantity of an embodiment of the present application;

[0052] Figure 4 A logic flow diagram of determining the category of each feature analysis time window of an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to make the purpose and advantages of the present application clearer and more apparent, the present application will be further described below in conjunction with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0054] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0055] It should be noted that in the description of the present application, the terms "upper", "lower", "inner", "outer" and the like indicating the direction or positional relationship are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0056] Please refer to Figure 1 shown, which is a step diagram of the paint anti-aging performance detection method based on image recognition of an embodiment of the present application. The paint anti-aging performance detection method based on image recognition of the present application comprises:

[0057] In step S100, component information of a plurality of paint samples is acquired, and a plurality of first comparison samples and a plurality of second comparison samples corresponding to the current paint are determined according to the comparison condition of the component information.

[0058] Specifically, the component information in the present application includes the component content proportion of the antioxidant. It is well known to those skilled in the art that the antioxidant is added to the paint. In the present application, the specific component of the antioxidant is not limited, which can be any component of hindered phenol or phosphite or a combination of a plurality of components. In the present application, the same type of paint with the same type of antioxidant is selected as the paint sample for comparison analysis.

[0059] In the implementation, the number of the first comparison samples and the second comparison samples needs to ensure that there is sufficient sample data for reference comparison. The number of the first comparison samples and the second comparison samples can be set to 10.

[0060] In step S200, the aging trend characterization quantity of the historical test image of the first comparison sample and the second comparison sample is identified and acquired to draw an aging trend fluctuation curve. The aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample are screened according to the curve characteristics of the aging trend fluctuation curve.

[0061] Specifically, the curve characteristics in the present application can be the curve slope at a plurality of data points on the aging trend fluctuation curve.

[0062] In step S300, each aging sensitive node is sorted in time sequence, and a plurality of feature analysis time windows are determined based on a preset node distribution position constraint condition and a time interval constraint condition. The category of each feature analysis time window is determined according to the time starting position distribution of the feature analysis time window.

[0063] In step S400, the anti-aging performance difference quantity of each category of feature analysis time window is calculated according to the difference of the aging trend characterization quantity of the aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample at a plurality of times and the duration of the feature analysis time window.

[0064] In step S500, a plurality of first comparison samples and a plurality of second comparison samples meeting the number proportion are screened, and the anti-aging life prediction interval of the current paint is determined according to the anti-aging historical life of the first comparison sample and the second comparison sample.

[0065] The number proportion is determined according to the anti-aging performance difference quantity of each category of feature analysis time window.

[0066] The skilled in the art can understand that, from the essence of paint aging, the anti-aging performance of paint depends on the stability and structural integrity of its components. The paint with excellent anti-aging performance can resist the erosion of environmental factors such as light, temperature and humidity for a long time. In the same feature analysis time window, the change range of the aging trend characterization quantity of the crack area ratio of the sample with poor anti-aging performance will be significantly greater than that of the sample with excellent performance still in the slow aging stage, resulting in an increase in the difference in the degree of aging of the two samples. The difference is essentially the difference in performance degradation of the two samples under the same environmental conditions.

[0067] Specifically, in the historical process of the anti-aging test of the paint sample, a plurality of simulation natural aging or accelerated aging links of the same test process are included. First, all paint samples need to be uniformly prepared to form a coating on a substrate of the same material with the same thickness and brushing process to ensure consistency in the initial state. Then, state adjustment is performed by placing the samples in a standard environment, such as a temperature of 23℃±2℃ and a relative humidity of 50%±5%, for a certain period of time to dry the coating completely and reach a stable state. Next, the aging test execution link is performed, which can use a UV aging test box to set a specific wavelength and light intensity to continuously irradiate the sample. Then, wet and hot aging is simulated using a constant temperature and humidity chamber to control high temperature, such as 40-60℃, and high humidity, such as 80%-95% humidity environment. Outdoor comprehensive aging can also be simulated in combination with conditions such as water spraying and temperature changes to be closer to the natural environment. During the test process, a regular monitoring link needs to be set to detect the performance of the sample, including the crack area ratio. When the crack area ratio of the sample reaches a predetermined test endpoint, such as a crack area ratio of 20%, the test is stopped, and the duration from the start of the test to the predetermined test endpoint of all samples is collected as the anti-aging historical life.

[0068] Specifically, please refer to Figure 2 As shown in the figure, it is a step diagram for determining the first comparison sample and the second comparison sample according to the component information comparison of the current paint. The process of determining the first comparison sample and the second comparison sample corresponding to the current paint according to the component information comparison includes:

[0069] Step S101, obtaining the anti-aging component ratio in the component information of the current paint;

[0070] Step S102, determining the paint with an anti-aging component ratio greater than or equal to the anti-aging component ratio of the current paint as the first comparison sample corresponding to the current paint;

[0071] Step S103, determining the paint with an anti-aging component ratio less than the anti-aging component ratio of the current paint as the second comparison sample corresponding to the current paint.

[0072] Specifically, the anti-aging component content in the present application can be the content of the antioxidant, which will not be repeated here.

[0073] Specifically, the anti-aging component content in the present application can be the content of the antioxidant, which will not be repeated here.

[0074] Specifically, please refer to Figure 3 As shown in the figure, it is a step diagram for obtaining the aging trend characterization quantity of the embodiment of the present application. The process of identifying and obtaining the aging trend characterization quantity of the historical test image of the first comparison sample and the second comparison sample includes:

[0075] Step S201, obtaining the historical test image of the first comparison sample and the second comparison sample, the historical test image being the surface image of the first comparison sample and the second comparison sample in historical testing;

[0076] Step S202, obtaining the crack area ratio of the paint surface at a plurality of test time nodes in the historical test image of each first comparison sample, and the crack area ratio of the paint surface at a plurality of test time nodes in the historical test image of each second comparison sample;

[0077] Step S203, determining the average value of the crack area ratio in the historical test image of a plurality of first comparison samples as the first aging trend characterization quantity, and determining the average value of the crack area ratio in the historical test image of a plurality of second comparison samples as the second aging trend characterization quantity.

[0078] Specifically, the process of obtaining the crack area ratio in the historical test image includes denoising and enhancing the collected original image; according to the distribution characteristics of the pixel gray value, the crack area is automatically distinguished, and the crack area is quantified in the binary image by the pixel counting method: the total number of pixels representing the crack in the image is counted, and the total number of pixels in the entire test area is divided to obtain the crack area ratio at the time node, which is the prior art and will not be repeated here.

[0079] It can be understood that the present application converts the abstract aging phenomenon into specific numerical indicators by obtaining the surface image of the sample in the historical test and extracting the crack area ratio at different test time nodes, realizes the accurate quantification of the aging degree of the paint, and uses the average value of the crack area ratio of a plurality of first comparison samples and second comparison samples as the first and second aging trend characterization quantities, which can effectively reduce the contingency and deviation of the test data of a single sample, and more accurately reflect the overall aging law of the same kind of sample.

[0080] Specifically, the first aging trend fluctuation curve of the first aging trend characterization quantity changing over time and the second aging trend fluctuation curve of the second aging trend characterization quantity changing over time are drawn under the same coordinate system.

[0081] The horizontal axis of the coordinate system is time, and the vertical axis is the aging trend characterization quantity.

[0082] Specifically, the process of screening the aging sensitive nodes from the aging trend fluctuation curve of the first comparison sample and the aging trend fluctuation curve of the second comparison sample includes:

[0083] The curve slopes of a plurality of data points on the first aging trend fluctuation curve and the second aging trend fluctuation curve are determined respectively, and the data points with the curve slopes exceeding a preset slope screening value are determined as the aging sensitive nodes.

[0084] Specifically, the preset slope screening value can be in the range of [0.25, 0.4]. In order to avoid missing the screening of the aging sensitive nodes due to the preset slope screening value being too large, and to avoid misjudgment of the screening of the aging sensitive nodes due to the preset slope screening value being too small, preferably, a slope screening value is provided herein, and the slope screening value can be 0.3.

[0085] It can be understood that the aging trend fluctuation curves of the first comparison sample and the second comparison sample are screened for aging sensitive nodes, the curve slopes of a plurality of data points are calculated, the data points with the curve slopes exceeding a preset slope screening value are determined as the aging sensitive nodes, and the greater the slope, the faster the crack area ratio changes at this time node. Therefore, the key time point of the sudden change of the paint aging speed is locked in, the stage of significant change in the aging characteristics is focused on, and the accurate identification of the key turning point in the aging process is achieved.

[0086] Specifically, the process of determining the feature analysis time window includes:

[0087] The two aging sensitive nodes meeting the node distribution position constraint condition and the time length interval constraint condition are respectively taken as the time start point and the time end point of the feature analysis time window to determine a plurality of feature analysis time windows.

[0088] The node distribution position constraint condition is that any one of the two aging sensitive nodes is on the first aging trend fluctuation curve, and the other aging sensitive node is on the second aging trend fluctuation curve.

[0089] The time length interval constraint condition is that the interval time length of the two aging sensitive nodes does not exceed a preset interval time length threshold.

[0090] Specifically, the preset interval duration threshold value can be preset by a person skilled in the art, to ensure that the aging performance occurring in the screened feature analysis time window is generated for the same test condition, the interval duration threshold value can be set to [3, 10], the interval unit is day, preferably, the interval duration threshold value can be 5 days.

[0091] Specifically, the present application limits two aging sensitive nodes to the first and second aging trend fluctuation curves through the node distribution position constraint condition, ensures that each feature analysis time window can simultaneously cover the aging sensitive information of the first comparison sample and the second comparison sample, and focuses the targeted analysis in the window on the correlation of the two types of samples on the key aging nodes. The interval duration constraint condition limits the interval duration of the two sensitive nodes, so that each feature analysis time window can concentrate on reflecting the correlation change of the aging characteristics of the two types of samples in the similar time range, thereby improving the consistency and comparability of the analysis data, and the evaluation of the anti-aging performance of the paint is deepened to the feature analysis in the specific time interval.

[0092] Specifically, please refer to Figure 4 The figure is a logic flow chart for determining the category of each feature analysis time window in the embodiment of the present application, and the process of determining the category of each feature analysis time window according to the time start position distribution of the feature analysis time window includes:

[0093] If the time start point is located on the first aging trend fluctuation curve, the feature analysis time window is determined as the first dominant category;

[0094] If the time start point is located on the second aging trend fluctuation curve, the feature analysis time window is determined as the second dominant category.

[0095] Specifically, the window with the time start point located on the first aging trend fluctuation curve is determined as the first dominant category, and the window with the time start point located on the second aging trend fluctuation curve is determined as the second dominant category. It can be understood that the first dominant category window focuses more on reflecting the aging condition of the paint with the first comparison sample key aging node as the reference, and the second dominant category window focuses on the aging condition of the paint with the second comparison sample key aging node as the reference.

[0096] Specifically, the process of calculating the anti-aging performance difference amount of each category of feature analysis time window includes:

[0097] The aging trend representation difference of the first aging trend fluctuation curve and the aging trend representation difference of the second aging trend fluctuation curve in each feature analysis time window are determined respectively;

[0098] The absolute value of the difference between the aging trend characteristic quantity difference of the first aging trend fluctuation curve and the aging trend characteristic quantity difference of the second aging trend fluctuation curve is combined with the duration of the feature analysis time window to obtain the anti-aging performance difference quantity of each feature analysis time window through weighted calculation.

[0099] The average of the anti-aging performance difference quantities of the several feature analysis time windows of the first dominant category is determined as the first anti-aging performance difference quantity, and the average of the anti-aging performance difference quantities of the several feature analysis time windows of the second dominant category is determined as the second anti-aging performance difference quantity.

[0100] For example, the anti-aging performance difference quantity = aging trend characteristic quantity difference factor × (difference absolute value / difference absolute value reference value) + duration factor × (duration / duration reference value). The aging trend characteristic quantity difference factor can be calculated in advance. The absolute value of the difference between the aging trend characteristic quantity difference of the several first aging trend fluctuation curves and the aging trend characteristic quantity difference of the second aging trend fluctuation curve is averaged, and the average of the calculated difference absolute value is determined as the difference absolute value reference value. The value of the duration reference value can be determined according to the preset interval duration threshold. The duration reference value = interval duration threshold × 0.8. The sum of the aging trend characteristic quantity difference factor and the duration factor is 1. The specific values of the two factors can be selected by a person skilled in the art according to the influence of the aging trend characteristic quantity difference factor and the duration factor on the calculation result in historical data. Here, a value of the aging trend characteristic quantity difference factor and the duration factor is provided. The aging trend characteristic quantity difference factor can be set to 0.5, and the duration factor can be set to 0.5.

[0101] Specifically, the aging trend characteristic quantity difference of the first and second aging trend fluctuation curves in each window is determined respectively, which can directly obtain the aging change degree of the two types of samples in the same test process. When the aging performance of the first aging trend fluctuation curve appears preferentially, that is, the starting point of the window is on the aging trend fluctuation curve, and the aging trend characteristic quantity difference of the first comparison sample is greater than that of the second comparison sample, the greater the absolute value of the difference, the more significantly the performance fluctuation of the first comparison sample exceeds that of the second comparison sample in the window is reflected by the weighting of the duration. The average of the first dominant category corresponds to the performance fluctuation degree in the scenario where the aging performance of the first comparison sample appears preferentially, and the average of the second dominant category corresponds to the performance fluctuation degree in the scenario where the aging performance of the second comparison sample appears preferentially. Therefore, the scientificity and reliability of the prediction are improved.

[0102] It can be understood that if the aging performance of the first aging trend fluctuation curve appears in the test process prior to the aging performance of the second aging trend fluctuation curve, the time starting point of the characteristic analysis time window embodied in the appearance is on the first aging trend fluctuation curve, if the aging performance of the first aging trend fluctuation curve appears in the test process behind the aging performance of the second aging trend fluctuation curve, the time starting point of the characteristic analysis time window embodied in the appearance is on the second aging trend fluctuation curve, and the difference between the aging trend characteristic quantities of the first aging trend fluctuation curve in the characteristic analysis time window represents the aging change degree of the first comparison sample in the characteristic analysis time window, the difference between the aging trend characteristic quantities of the second aging trend fluctuation curve in the characteristic analysis time window represents the aging change degree of the second comparison sample in the characteristic analysis time window, and the absolute value of the difference between the two represents the difference in the aging degree change in the same characteristic analysis time window. In fact, the earlier the aging performance appears, the greater the difference in the aging degree change in the same characteristic analysis time window, and the more unstable the performance of the comparison sample.

[0103] Specifically, the process of determining the anti-aging life prediction interval of the current paint includes:

[0104] The anti-aging historical life values of the first comparison sample and the second comparison sample in accordance with the quantity ratio are combined to form a numerical set;

[0105] The anti-aging life prediction interval of the current paint is determined according to the average value of the numerical set;

[0106] The interval starting point of the anti-aging life prediction interval is the minimum value of the average value, and the interval ending point is the maximum value of the average value.

[0107] Specifically, the anti-aging historical life values of the first comparison sample and the second comparison sample in accordance with the quantity ratio are combined to form a numerical set, the anti-aging ingredient content of the first comparison sample is not less than that of the current paint, and its historical life can be used as a potential upper limit reference for the anti-aging performance of the current paint. The anti-aging ingredient content of the second comparison sample is lower than that of the current paint, and its historical life can be used as a potential lower limit reference. The determination of the prediction interval according to the average value of the numerical set can smooth the accidental fluctuations of the historical life of the individual sample, so that the obtained interval can better reflect the overall life level of the same sample. The setting of the interval starting point and the interval ending point not only covers the anti-aging life range that the current paint can reach, but also avoids the distortion of the interval caused by individual extreme sample data through the constraint of the average value. The present application does not need to perform long-term aging test on the current paint, but uses historical sample data and scientific analysis method to quickly determine the anti-aging life prediction interval, greatly shortens the detection period, avoids the limitation of long-term test for the detection of the anti-aging performance of the paint, and improves the detection efficiency.

[0108] Specifically, the process of determining the quantity ratio includes:

[0109] The ratio of the first anti-aging performance difference and the second anti-aging performance difference is calculated, and the reciprocal of the ratio is determined as the quantity ratio of the first comparison sample and the second comparison sample.

[0110] Specifically, in order to facilitate calculation, the last digit after the decimal point can be selected during the calculation of the quantity ratio, and the final determined quantity of the first comparison sample and the second comparison sample can be ensured to be an integer according to the rounding rule.

[0111] Specifically, the present application dynamically correlates the performance fluctuation characteristics of the two types of comparison samples, accurately controls the sample quantity ratio for life prediction, thereby improving the reliability of the prediction interval. It can be understood that the first anti-aging performance difference reflects the performance fluctuation with the first comparison sample as the reference, and the second anti-aging performance difference corresponds to the fluctuation of the second comparison sample. The ratio of the two essentially reflects the performance differentiation rate of the two types of samples at the key aging stage, and the reciprocal of the ratio is used as the quantity ratio because the performance difference is inversely related to the reference value of the sample. When the first anti-aging performance difference is large, it means that the aging performance of the first comparison sample fluctuates more violently, and its stability is poor. At this time, the proportion of the first comparison sample in the prediction sample needs to be reduced. Conversely, if the second anti-aging performance difference is larger, the quantity proportion of the second comparison sample needs to be reduced. This reciprocal conversion mechanism can automatically balance the weights of the two types of samples, so that the composition of the reference sample matches the actual performance fluctuation characteristics, and ultimately makes the anti-aging life prediction interval determined based on these samples more consistent with the real performance level of the current paint, avoiding the prediction result being biased towards the extreme value due to the excessive participation of a certain type of sample, and improving the detection accuracy.

[0112] For example, if 10 first comparison samples and 10 second comparison samples are selected for the current paint, the first anti-aging performance difference is 3.2, and the second anti-aging performance difference is 2.

[0113] The ratio of the first anti-aging performance difference and the second anti-aging performance difference is 3.2 / 2.0=1.6, and its reciprocal is 0.6, that is, for every 0.6 first comparison samples, 1 second comparison sample needs to be selected. In order to facilitate calculation, the last digit after the decimal point is retained, and the final determined quantity needs to be rounded to ensure that it is an integer. In the case of a quantity ratio of 0.6, in 10 first comparison samples and 10 second comparison samples, the sample quantity combination participating in the anti-aging life prediction can be 5 second comparison samples corresponding to 3 first comparison samples, or 10 second comparison samples corresponding to 6 first comparison samples. In order to reduce sample deviation, the combination with the largest sample quantity is selected, that is, 6 first comparison samples combined with 10 second comparison samples.

[0114] The anti-aging history life (in years) of the 10 first comparison samples and the 10 second comparison samples is as follows:

[0115]

[0116] The average values of the two sets of numerical values are compared, wherein the minimum value of the average value is 96.3 months, and the maximum value is 97.8 months, so the anti-aging life prediction interval of the current paint is [96.3, 97.8] months.

[0117] The technical scheme of the present application has been described in combination with the preferred embodiments shown in the drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical schemes after the changes or replacements will fall within the protection scope of the present application.

[0118] The above only describes the preferred embodiments of the present application and is not used to limit the present application; the present application can have various changes and variations for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An image recognition-based method for detecting the anti-aging performance of paint, characterized in that, The method comprises the following steps: Obtain the component information of a plurality of paint samples, and determine a plurality of first comparison samples and a plurality of second comparison samples corresponding to the current paint according to the comparison of the component information; Identify the historical test images of the first comparison samples and the second comparison samples and obtain the aging trend characteristic quantities to draw the aging trend fluctuation curves, and screen the aging sensitive nodes from the aging trend fluctuation curves of the first comparison samples and the aging trend fluctuation curves of the second comparison samples respectively according to the curve characteristics of the aging trend fluctuation curves; Sort the aging sensitive nodes in time sequence, and determine a plurality of feature analysis time windows based on the preset node distribution position constraint condition and the time interval constraint condition, and determine the categories of the feature analysis time windows according to the time starting position distribution of the feature analysis time windows; Calculate the anti-aging performance difference of each category of the feature analysis time windows according to the difference of the aging trend characteristic quantities of the first comparison samples and the second comparison samples at a plurality of time points and the duration of the feature analysis time windows; Screen a plurality of first comparison samples and a plurality of second comparison samples that meet the quantity proportion, and determine the anti-aging life prediction interval of the current paint according to the anti-aging historical life of the first comparison samples and the second comparison samples; The quantity proportion is determined according to the anti-aging performance difference of each category of the feature analysis time windows.

2. The image recognition-based paint anti-aging performance detection method according to claim 1, characterized in that, The process of determining the first comparison samples and the plurality of second comparison samples corresponding to the current paint according to the comparison of the component information comprises: Obtain the anti-aging component proportion in the component information of the current paint; Determine the paint with an anti-aging component proportion greater than or equal to the anti-aging component proportion of the current paint as the first comparison sample corresponding to the current paint; Determine the paint with an anti-aging component proportion less than the anti-aging component proportion of the current paint as the second comparison sample corresponding to the current paint.

3. The image recognition-based paint anti-aging performance detection method according to claim 2, characterized in that, The process of identifying the historical test images of the first comparison samples and the second comparison samples and obtaining the aging trend characteristic quantities comprises: Obtain the historical test images of the first comparison samples and the second comparison samples, which are the surface images of the first comparison samples and the second comparison samples in the historical test; Obtain the crack area proportion of the paint surface at a plurality of test time nodes in the historical test images of each first comparison sample, and the crack area proportion of the paint surface at a plurality of test time nodes in the historical test images of each second comparison sample; Determine the average value of the crack area proportion in the historical test images of a plurality of first comparison samples as the first aging trend characteristic quantity, and determine the average value of the crack area proportion in the historical test images of a plurality of second comparison samples as the second aging trend characteristic quantity.

4. The image recognition-based paint anti-aging performance detection method according to claim 3, characterized in that, Draw the first aging trend fluctuation curve of the first aging trend characteristic quantity changing with time and the second aging trend fluctuation curve of the second aging trend characteristic quantity changing with time in the same coordinate system; The horizontal axis of the coordinate system is time, and the vertical axis is the aging trend characteristic quantity.

5. The image recognition-based paint anti-aging performance detection method according to claim 4, characterized in that, The process of screening the aging sensitive nodes from the aging trend fluctuation curves of the first comparison samples and the aging trend fluctuation curves of the second comparison samples comprises: The curve slopes of a plurality of data points on the first aging trend fluctuation curve and the second aging trend fluctuation curve are determined, and the data points with the curve slopes exceeding a preset slope screening value are determined as aging sensitive nodes.

6. The image recognition-based paint anti-aging performance detection method according to claim 5, characterized in that, The process of determining the feature analysis time window comprises: Two aging sensitive nodes meeting a node distribution position constraint condition and a time interval constraint condition are respectively taken as a time start point and a time end point of the feature analysis time window to determine a plurality of feature analysis time windows. The node distribution position constraint condition is that any one of the two aging sensitive nodes is on the first aging trend fluctuation curve and the other aging sensitive node is on the second aging trend fluctuation curve. The time interval constraint condition is that the interval time length of the two aging sensitive nodes does not exceed a preset interval time length threshold.

7. The image recognition-based paint anti-aging performance detection method according to claim 6, characterized in that, The process of determining the category of each feature analysis time window according to the time start point position distribution of the feature analysis time window comprises: If the time start point is on the first aging trend fluctuation curve, the feature analysis time window is determined as the first explicit category. If the time start point is on the second aging trend fluctuation curve, the feature analysis time window is determined as the second explicit category.

8. The image recognition-based paint anti-aging performance detection method according to claim 7, characterized in that, The process of calculating the anti-aging performance difference amount of each category of feature analysis time window comprises: The aging trend characteristic value difference of the first aging trend fluctuation curve and the aging trend characteristic value difference of the second aging trend fluctuation curve in each feature analysis time window are respectively determined. The aging trend characteristic value difference maximum value and the duration time length of the feature analysis time window are weighted to obtain the anti-aging performance difference amount of each feature analysis time window. The average value of the anti-aging performance difference amounts of a plurality of feature analysis time windows of the first explicit category is determined as the first anti-aging performance difference amount, and the average value of the anti-aging performance difference amounts of a plurality of feature analysis time windows of the second explicit category is determined as the second anti-aging performance difference amount.

9. The image recognition-based paint anti-aging performance detection method according to claim 8, characterized in that, The process of determining the anti-aging life prediction interval of the current paint comprises: A numerical value set is composed of the anti-aging historical life amounts corresponding to the first comparison sample and the second comparison sample meeting the quantity proportion. The anti-aging life prediction interval of the current paint is determined according to the average value of the numerical values in each numerical value set. The interval start point of the anti-aging life prediction interval is the minimum value of the average value, and the interval end point is the maximum value of the average value.

10. The image recognition-based paint anti-aging performance detection method according to claim 9, characterized in that, The process of determining the quantity proportion comprises: The ratio of the first anti-aging performance difference amount to the second anti-aging performance difference amount is calculated, and the reciprocal of the ratio is determined as the quantity proportion of the first comparison sample and the second comparison sample.

Images