Environmentally friendly coating detection method and system

Abstract

The present invention relates to the field of electronic digital data processing, and in particular to an environmentally friendly paint detection method and system. An environmentally friendly paint detection method includes the following steps: S1: acquiring an environmentally friendly paint image and a scanned image; S2: acquiring water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image; S3: acquiring three-dimensional information of water droplets and bubble areas of the environmentally friendly paint according to the scanned image of the environmentally friendly paint; S4: acquiring supplementary illumination information according to the water droplets and bubble areas of the environmentally friendly paint and the three-dimensional information of the water droplets and bubble areas; S5: acquiring illumination intensity information of water droplets and bubble areas of the environmentally friendly paint according to the supplementary illumination information; S6: acquiring environmentally friendly paint detection information according to the illumination intensity information of water droplets and bubbles. The present invention realizes the detection of environmentally friendly paint by analyzing the water droplets or bubbles appearing on the environmentally friendly paint, considering the principle that light is scattered into seven primary colors when passing through water droplets or bubbles, and analyzing the changes in the illumination intensity of the seven primary colors.

Description

Technical Field

[0001] The present invention relates to the field of electrical digital data processing, and in particular to an environmentally friendly coating detection method and system. Background Art

[0002] In recent years, with the continuous development of my country's industrialization level, the production scale of coatings has gradually expanded, the demand for environmentally friendly coatings has also gradually expanded, and the demand for water resistance and air permeability of environmentally friendly coatings has also gradually increased. High-performance waterproof coatings in environmentally friendly coatings play an important role in building exterior walls, roof waterproofing, basements and underground structures, bridges and tunnels due to their excellent water resistance and good air permeability. As an environmentally friendly coating with a wide range of applications and excellent performance, high-performance waterproof coatings in environmentally friendly coatings will cause significant losses regardless of any unexpected circumstances due to the complexity of the working environment of environmentally friendly coatings and the particularity of their own materials.

[0003] From the perspective of environmentally friendly coatings testing methods, today's environmentally friendly coatings and high-performance waterproof coatings are usually tested based on quantitative simulations in laboratory environments. Although laboratory environmental simulations accelerate the speed of testing, they lack simulation verification for all environmental conditions, especially for environmentally friendly coatings testing methods in areas with frequent freeze-thaw cycles. Testing in actual outdoor application scenarios usually uses manual visual or manual tapping methods to test the performance of coatings. Under special working conditions, professionals are required to carry professional equipment, which causes great inconvenience to the testing of coatings. In the context of the rapid development of intelligence, the testing of environmentally friendly coatings is gradually developing in the direction of complexity and intelligence. Summary of the invention

[0004] In order to overcome the shortcomings of environmentally friendly coating detection in areas with frequent freezing and thawing, such as high dependence on equipment and high requirements for manual experience, the present invention provides an environmentally friendly coating detection method and system.

[0005] The technical implementation scheme of the present invention is: a method for detecting environmentally friendly coatings, comprising the following steps:

[0006] S1: Acquire environmentally friendly paint images and scanned images;

[0007] S2: Obtaining water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image;

[0008] S3: Acquire the three-dimensional information of water droplets and bubble areas of the environmentally friendly coating according to the scanning image of the environmentally friendly coating;

[0009] S4: Obtaining supplementary illumination information based on the water droplet and bubble area of ​​the environmentally friendly coating and the three-dimensional information of the water droplet and bubble area;

[0010] S5: Obtaining the light intensity information of the water droplets and bubble areas of the environmentally friendly coating according to the supplementary light information;

[0011] S6: Obtain environmentally friendly coating detection information based on water droplets and bubble light intensity information;

[0012] S7: Implement environmentally friendly paint detection based on environmentally friendly paint detection information.

[0013] Preferably, the step of obtaining the water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image comprises: using a camera to obtain the positions of water droplets and bubbles in the environmentally friendly paint under good lighting conditions and calculating the areas and volumes of the water droplets and bubbles; specifically: using a high-resolution camera to take photos of the paint surface under good lighting conditions to ensure that the outlines of the water droplets and bubbles are clearly visible; using image denoising, contrast enhancement and brightness adjustment, and histogram equalization technology to improve image quality; applying an edge detection algorithm to identify the outlines of water droplets and bubbles, and using a watershed algorithm in image segmentation technology to accurately segment each independent water droplet or bubble; for each segmented water droplet or bubble area, directly calculating the number of pixels in each water droplet or bubble area, and then converting it into the actual area according to the resolution of the image, the water droplet or bubble area calculation formula is: area = (number of pixels) × (unit pixel area); the bubble volume calculation formula is: ,in, is the radius.

[0014] Preferably, the three-dimensional information of water droplets and bubbles in the environmentally friendly paint is obtained according to the scanned image of the environmentally friendly paint, including: if the light condition is poor, a laser detector is used to scan the environmentally friendly paint, and a camera is used to analyze the scanning situation, to obtain the position of water droplets and bubbles in the environmentally friendly paint and calculate the area, volume and height of the water droplets and bubbles; specifically: one or more laser lines or points are generated by a laser source, irradiated on the environmentally friendly paint, and a camera is used to capture the projection image formed by the laser on the surface of the object; a baseline is determined and the relative position between the laser and the camera is fixed, the baseline is the distance between the laser source and the camera, and the focal length and principal point position information of the camera are obtained; the laser is projected onto the surface of the environmentally friendly paint and the camera is used to capture the pattern formed by the laser spot on the surface of the object, as well as the deformation of water droplets and bubbles; the image is processed and the area, volume and height of the bubble are calculated, wherein the area calculation formula is: , the volume calculation formula is: ,in, is the radius, and the height calculation formula is: ,in, is the horizontal pixel position of the laser spot in the image captured by the camera, is the focal length of the camera, is the angle between the laser source and the camera.

[0015] Preferably, the method of obtaining supplementary illumination information according to the water droplet and bubble area and the three-dimensional information of the water droplet and bubble area of ​​the environmentally friendly coating comprises: determining the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural illumination conditions in the frequent freezing and thawing area according to the three-dimensional information of the water droplet and bubble area in the environmentally friendly coating, combined with the natural illumination information in the frequent freezing and thawing area; if the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural illumination conditions in the frequent freezing and thawing area is good, then no artificial light is supplemented; otherwise, artificial light is supplemented when the natural illumination conditions in the frequent freezing and thawing area are poor; specifically, the method comprises: collecting historical meteorological data in the area, recording the morphology, distribution, and density change information of water droplets and bubbles on the surface of the environmentally friendly coating, and determining the illumination intensity and distribution characteristics required for the optimal light scattering effect on the surface of the environmentally friendly coating; using a ray tracing algorithm to simulate the scattering, reflection, and projection process of natural light and artificial light elements on the coating surface; integrating environmental data, surface state, and illumination requirements, constructing a feature set, and using a random forest model to predict the light scattering effect under different weather conditions, and verifying the prediction accuracy of the model through historical data;

[0016] Among them, the illumination compensation algorithm is used to automatically adjust the intensity, angle and color temperature of the artificial light source according to the current natural lighting conditions and the predicted light scattering effect to supplement the insufficient natural light and ensure that the overall lighting is uniform and meets the scattering requirements.

[0017] Preferably, the method of obtaining the illumination intensity information of the water droplets and bubbles in the environmentally friendly paint area according to the illumination supplementary information includes: dividing the environmentally friendly paint area into N areas, calculating the illumination intensity of seven colors scattered by the water droplets and bubbles in the environmentally friendly paint by natural light in each area; and calculating the interaction between the colors of the environmentally friendly paint and the combination effect of specific colors according to the distribution of the illumination intensity. If the environmentally friendly paint is qualified in the detection, a specific painting method or pigment ratio is selected according to the interaction between the colors and the combination effect of specific colors; specifically, dividing the N areas into 7 color channels, calculating the relative intensity of the seven color channels in each area, and then performing normalization processing, assuming C={R, O, Y, G, B, I, V}, representing the set of seven colors, for any area i (i=1, 2, 3, ..., N), N is the total number of areas, and the relative intensity of each color is expressed by express, ,satisfy , that is, the sum of all color frequencies in the region is 1, ensuring normalization and calculating the color comprehensive index in the region , the calculation formula is: ,in, It is a weight that reflects the importance and is determined according to actual needs, taking into account the interaction between colors and the effect of specific color combinations. The calculation formula is: , It's color pair , The combination weight of represents two colors.

[0018] Preferably, the obtaining of environmental protection coating detection information according to the light intensity information of water droplets and bubbles includes: according to the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles appearing in the frequent freeze-thaw area of ​​the environmental protection coating, combined with the information of micro cracks, mildew spots, and fading of the environmental protection coating, using a model to analyze the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles appearing in the frequent freeze-thaw area of ​​the environmental protection coating, and the information of micro cracks, mildew spots, and fading of the environmental protection coating, so as to obtain the detection information of the environmental protection coating in the frequent freeze-thaw area; specifically: collecting a large amount of information including micro cracks, mildew spots, fading, The high-quality images of the bubble and water drop scattering characteristics are annotated with the specific location and type of defects in each image, and the images are normalized; the CNN model is selected, and the edge, texture, and color distribution features in the image are automatically extracted using the convolutional layer of the CNN. The model is trained using the annotated data set, and the weights are adjusted through back propagation and gradient descent methods so that the model can learn to identify different types of defects; the model is tuned using the validation set, and the best hyperparameter combination is determined using grid search; the model performance is evaluated using the F1 score on an unseen test set, and a color change tree and a color network diagram are constructed. The F1 formula is: .

[0019] Preferably, the construction of the color change tree includes: constructing a color change binary tree to visualize the seven color intensity measurement results reflected by water droplets and bubbles of the same batch of environmentally friendly paints in frequent freeze-thaw areas at different time points; specifically: the root node represents the initial color measurement, the child node represents the subsequent color measurement that changes over time, the edge represents the amount of color change, measured by color difference, and the trend of the paint color change over time is visualized using a tree structure, one child node represents that the color becomes lighter, and the other child node represents that the color becomes darker, and each color change is quantified by calculating the color difference and stored in the corresponding edge;

[0020] The color difference calculation formula is: ,in, : Indicates the difference in brightness (L*) between two colors; : Indicates the difference between two colors in chroma (C*), which reflects the purity or saturation of the color; : Indicates the difference between two colors in hue angle (H*), that is, the hue change of the color; , , : is an empirical constant used to adjust the contribution of brightness, chroma, and hue to the total color difference; , , : is a weighting factor based on brightness and chroma; : is a rotation term related to hue angle and chroma difference.

[0021] Preferably, the construction of the color network diagram comprises: constructing a color network diagram according to the measurement results and color similarities of seven color intensities scattered by water droplets and bubbles of environmentally friendly coatings in frequent freeze-thaw areas of different samples at different time points; specifically: nodes represent color measurements of different samples at different time points, edges represent connections between nodes with similar color features, and the weight of the edge represents the color similarity between two nodes. A K-nearest neighbor algorithm is used to construct a graph, in which each node in the graph is connected to the K nearest nodes, and the weight of the edge is the inverse of the color difference. The higher the weight, the more similar the color is. Color measurements are performed on multiple samples at different times, and a color network diagram is constructed for each sample. Different sample network structures are compared to find common patterns or abnormal patterns.

[0022] Preferably, the color network graph comprises: constructing a graph, for each sample, taking each measurement as a node, using color difference to calculate the similarity between nodes, and constructing weighted edges; analyzing the graph, using a community discovery algorithm to identify sample groups with similar color change trends; and detecting anomalies, finding abnormal samples with color changes through connectivity and density analysis of the graph structure.

[0023] An environmentally friendly coating detection system includes:

[0024] Environmentally friendly paint image acquisition module: used to collect images of environmentally friendly paint taken by the camera;

[0025] Environmentally friendly paint scanning image acquisition module: used to collect the environmentally friendly paint scanning image detected by the laser detector;

[0026] Bubble and water drop area calculation module: used to calculate the area of ​​bubbles and water drops of environmentally friendly coatings;

[0027] Bubble and water drop volume calculation module: used to calculate the volume of bubbles and water droplets of environmentally friendly coatings;

[0028] Bubble and water drop height calculation module: used to calculate the height of bubbles and water drops of environmentally friendly coatings;

[0029] Random forest model building module: used to predict the light scattering effect under different weather conditions;

[0030] CNN model building module: used to obtain environmentally friendly coating detection information;

[0031] Color change tree building module: used to visualize the trend of paint color changes over time based on a tree structure;

[0032] Color network graph construction module: used for connectivity and density analysis of graph structure and finding abnormal samples of color changes.

[0033] Beneficial effects: The present invention starts from the actual natural conditions of the frequent freeze-thaw area. First, when the light conditions in the frequent freeze-thaw area are good, the water droplets and bubbles appearing on the surface of the environmentally friendly paint are photographed by a camera device and the area and volume of the water droplets and bubbles are calculated; when the light conditions in the frequent freeze-thaw area are poor, the paint is scanned by a laser device, and the scanned image is analyzed by a camera device to calculate the area, volume and height of the bubbles and water droplets; when the light conditions in the frequent freeze-thaw area are poor, the paint is supplemented with an artificial light source; by analyzing the principle of the seven primary colors of light formed after the light passes through bubbles and water droplets and is scattered, the paint area is divided into N areas, the light intensity of the seven primary colors of the bubbles and water droplets appearing in each area after the light is scattered is calculated, and the environmentally friendly paint detection result is obtained by analyzing the light intensity; the color change trend of each batch of samples is visualized by using a color binary tree; the abnormal sample set is found by using a color network diagram; and the accuracy and efficiency of paint performance detection are improved by using multiple algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The present invention is a flow chart of an environmentally friendly coating detection method and system. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Example 1: A method for detecting environmentally friendly coatings, such as Figure 1 As shown, the following steps are included:

[0037] S1: Acquire environmentally friendly paint images and scanned images;

[0038] S2: Obtaining water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image;

[0039] S3: Acquire the three-dimensional information of water droplets and bubble areas of the environmentally friendly coating according to the scanning image of the environmentally friendly coating;

[0040] S4: Obtaining supplementary illumination information based on the water droplet and bubble area of ​​the environmentally friendly coating and the three-dimensional information of the water droplet and bubble area;

[0041] S5: Obtaining the light intensity information of the water droplets and bubble areas of the environmentally friendly coating according to the supplementary light information;

[0042] S6: Obtain environmentally friendly coating detection information based on water droplets and bubble light intensity information;

[0043] S7: Implement environmentally friendly paint detection based on environmentally friendly paint detection information.

[0044] Obtaining water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image includes: using a camera to obtain the positions of water droplets and bubbles in the environmentally friendly paint under good lighting conditions and calculating the area and volume of the water droplets and bubbles; specifically: using a high-resolution camera to take photos of the paint surface under good lighting conditions to ensure that the outlines of water droplets and bubbles are clearly visible; using image denoising, contrast enhancement and brightness adjustment, and histogram equalization technology to improve image quality; applying an edge detection algorithm to identify the outlines of water droplets and bubbles, and using a watershed algorithm in image segmentation technology to accurately segment each independent water droplet or bubble; for each segmented water droplet or bubble area, directly calculating the number of pixels in each water droplet or bubble area, and then converting it into the actual area according to the resolution of the image, the water droplet or bubble area calculation formula is: area = (number of pixels) × (unit pixel area); the bubble volume calculation formula is: ,in, is the radius.

[0045] Among them, for the bubble volume, it is assumed that the bubble is spherical, and the bubble diameter is calculated based on the bubble area calculated by the water droplet or bubble area calculation formula, and then the bubble radius is calculated based on the bubble diameter, and finally the bubble volume is calculated.

[0046] The three-dimensional information of water droplets and bubbles in the environmentally friendly paint is obtained according to the scanned image of the environmentally friendly paint, including: if the light condition is poor, the environmentally friendly paint is scanned by a laser detector, and the scanning situation is analyzed by a camera to obtain the position of water droplets and bubbles in the environmentally friendly paint and calculate the area, volume and height of the water droplets and bubbles; specifically: one or more laser lines or points are generated by a laser source, irradiated on the environmentally friendly paint, and the projection image formed by the laser on the surface of the object is captured by a camera; the baseline is determined and the relative position between the laser and the camera is fixed, and the baseline is the distance between the laser source and the camera, and the focal length and principal point position information of the camera are obtained; the laser is projected onto the surface of the environmentally friendly paint and the pattern formed by the laser spot on the surface of the object, as well as the deformation of water droplets and bubbles, is captured by a camera; the image is processed and the area, volume and height of the bubble are calculated, wherein the area calculation formula is: , the volume calculation formula is: ,in, is the radius, and the height calculation formula is: ,in, is the horizontal pixel position of the laser spot in the image captured by the camera, is the focal length of the camera, is the angle between the laser source and the camera.

[0047] Obtaining supplementary illumination information based on the water droplet and bubble area of ​​the environmentally friendly coating and the three-dimensional information of the water droplet and bubble area, including: determining the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural lighting conditions in the frequent freezing and thawing area based on the three-dimensional information of the water droplet and bubble area in the environmentally friendly coating, combined with the natural lighting information in the frequent freezing and thawing area; if the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural lighting conditions in the frequent freezing and thawing area is good, no artificial light will be supplemented; otherwise, artificial light will be supplemented when the natural lighting conditions in the frequent freezing and thawing area are poor; specifically: collecting historical meteorological data in the area, recording the morphology, distribution, and density change information of water droplets and bubbles on the surface of the environmentally friendly coating, and determining the light intensity and distribution characteristics required for the best light scattering effect on the surface of the environmentally friendly coating; using a ray tracing algorithm to simulate the scattering, reflection, and projection process of natural light and artificial light elements on the coating surface; integrating environmental data, surface state, and lighting requirements, constructing a feature set, using a random forest model to predict the light scattering effect under different weather conditions, and verifying the prediction accuracy of the model through historical data;

[0048] Among them, the illumination compensation algorithm is used to automatically adjust the intensity, angle and color temperature of the artificial light source according to the current natural lighting conditions and the predicted light scattering effects to supplement the insufficient natural light and ensure that the overall lighting is uniform and meets the scattering requirements; the temperature difference and condensation conditions are obtained using historical meteorological data, and the first water droplet formation time is obtained based on the temperature difference and condensation conditions; at the same time, the second water droplet formation time formed by tiny cracks or pores in the paint during the physical effect of the freeze-thaw cycle is obtained; at the same time, the third water droplet formation time caused by changes in interlayer or surface tension of the paint is obtained.

[0049] Among them, the first water drop formation time, the second water drop formation time, and the third water drop formation time include: the first water drop formation time (when the temperature drops rapidly, especially at night or when seasons change, the surface temperature is lower than the dew point temperature in the air, causing water vapor in the air to condense on the surface of the coating to form water droplets); the second water drop formation time (during the freeze-thaw cycle, water will penetrate into the tiny or pores of the coating, expand in volume when frozen, and cause pressure on the coating structure. When thawed, this part of water becomes liquid again, forming water droplets on the surface or pushing the water under the original coating to migrate to the surface); the third water drop formation time (rapid temperature changes affect the physical properties of the coating, including surface tension, causing the coating surface to be more susceptible to water droplets or the original tiny bubbles to become larger due to temperature differences, thus becoming visible on the surface).

[0050] The illumination intensity information of water droplets and bubbles in the environmentally friendly paint is obtained according to the supplementary illumination information, including: dividing the environmentally friendly paint area into N areas, calculating the illumination intensity of seven colors scattered by the water droplets and bubbles in the environmentally friendly paint by natural light in each area; and calculating the interaction between the colors of the environmentally friendly paint and the combination effect of specific colors according to the distribution of the illumination intensity. If the environmentally friendly paint is qualified in the inspection, a specific painting method or pigment ratio is selected according to the interaction between the colors and the combination effect of specific colors; specifically, dividing the N areas into 7 color channels, calculating the relative intensity of the seven color channels in each area, and then normalizing them. Let C = {R, O, Y, G, B, I, V}, which represents the set of seven colors. For any area i (i = 1, 2, 3, ..., N), N is the total number of areas, and the relative intensity of each color is expressed by express, ,satisfy , that is, the sum of all color frequencies in the region is 1, ensuring normalization and calculating the color comprehensive index in the region , the calculation formula is: The calculation formula is: ,in, It is a weight that reflects the importance and is determined according to actual needs, taking into account the interaction between colors and the effect of specific color combinations. The calculation formula is: , It's color pair , The combination weight of represents two colors;

[0051] The survival time of the bubbles or water drops is obtained according to the existence time of the bubbles or water drops and the reflection color intensity of the bubbles or water drops.

[0052] The survival time of the bubbles or water drops is the time from the appearance to the disappearance of the water drops or bubbles, including the size, position and density of the water drops or bubbles in this process.

[0053] The environmental protection coating detection information is obtained according to the light intensity information of water droplets and bubbles, including: according to the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles that appear in the frequent freezing and thawing area of ​​the environmental protection coating, combined with the micro cracks, mildew and fading information of the environmental protection coating, the model is used to analyze the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles that appear in the frequent freezing and thawing area of ​​the environmental protection coating, and the micro cracks, mildew and fading information of the environmental protection coating are obtained; specifically: collect a large number of coating micro cracks, mildew, fading, bubbles and water drop scattered The method is to obtain high-quality images with projection features and mark the specific location and type of defects in each image, and normalize the images; select a CNN model, use the convolutional layer of CNN to automatically extract the edge, texture, and color distribution features in the image, train the model using the labeled data set, adjust the weights through back propagation and gradient descent methods, and enable the model to learn to identify different types of defects; use the validation set to tune the model, and use grid search to determine the best hyperparameter combination; use the F1 score to evaluate the model performance on an unseen test set, and construct a color change tree and a color network diagram. The F1 score formula is: ;

[0054] Constructing a color change tree, including: constructing a color change binary tree, visually displaying the seven color intensity measurement results reflected by water droplets and bubbles of the same batch of environmentally friendly paints in frequent freeze-thaw areas at different time points; specifically: the root node represents the initial color measurement, the child node represents the subsequent color measurement that changes over time, the edge represents the amount of color change, measured by color difference, and using a tree structure to visualize the trend of paint color change over time, one child node represents that the color becomes lighter, and another child node represents that the color becomes darker, and each color change is quantified by calculating the color difference and stored in the corresponding edge;

[0055] The color difference calculation formula is: ,in, : Indicates the difference in brightness (L*) between two colors; : Indicates the difference between two colors in chroma (C*), which reflects the purity or saturation of the color; : Indicates the difference between two colors in hue angle (H*), that is, the hue change of the color; , , : is an empirical constant used to adjust the contribution of brightness, chroma, and hue to the total color difference; , , : is a weighting factor based on brightness and chroma; : is a rotation term related to hue angle and chroma difference.

[0056] Constructing a color network diagram, including: constructing a color network diagram according to the measurement results and color similarities of seven color intensities scattered by water droplets and bubbles of environmentally friendly coatings in frequent freeze-thaw areas of different samples at different time points; specifically: nodes represent color measurements of different samples at different time points, edges represent connections between nodes with similar color features, and the weight of the edge represents the color similarity between two nodes. A K-nearest neighbor algorithm is used to construct a diagram, in which each node is connected to the nearest K nodes, and the weight of the edge is the inverse of the color difference. The higher the weight, the more similar the color is. Color measurements are performed on multiple samples at different times, and a color network diagram is constructed for each sample. Different sample network structures are compared to find common patterns or abnormal patterns.

[0057] The color network graph includes: constructing the graph, for each sample, taking each measurement as a node, using color difference to calculate the similarity between nodes, and constructing weighted edges; analyzing the graph, using community discovery algorithm to identify sample groups with similar color change trends; anomaly detection, finding abnormal color change samples through connectivity and density analysis of the graph structure.

[0058] Embodiment 2: Based on Embodiment 1, an environmentally friendly coating detection system includes:

[0059] An environmentally friendly paint image acquisition module is used to collect images of environmentally friendly paint taken by a camera;

[0060] An environmentally friendly paint scanning image acquisition module is used to collect the environmentally friendly paint scanning image detected by the laser detector;

[0061] The bubble and water drop area calculation module is used to calculate the area of ​​bubbles and water drops of environmentally friendly coatings;

[0062] Bubble and water drop volume calculation module, used to calculate the volume of bubbles and water droplets of environmentally friendly coatings;

[0063] Bubble and water drop height calculation module, used to calculate the height of bubbles and water drops of environmentally friendly coatings;

[0064] A random forest model building module to predict the light scattering effects under different weather conditions;

[0065] CNN model building module to obtain environmentally friendly coating detection information;

[0066] A color change tree building module is used to visualize the trend of paint color changes over time based on a tree structure;

[0067] The color network graph construction module is used to analyze the connectivity and density of the graph structure and find abnormal samples of color changes.

[0068] The above is a detailed introduction to the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for detecting environmentally friendly coatings, characterized in that: The following steps are involved: S1: Acquire environmentally friendly paint images and scanned images; S2: Obtaining water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image; S3: Acquire the three-dimensional information of water droplets and bubble areas of the environmentally friendly coating according to the scanning image of the environmentally friendly coating; S4: Obtaining supplementary illumination information based on the water droplet and bubble area of ​​the environmentally friendly coating and the three-dimensional information of the water droplet and bubble area; S5: Obtaining the light intensity information of the water droplets and bubble areas of the environmentally friendly coating according to the supplementary light information; S6: Obtain environmentally friendly coating detection information based on water droplets and bubble light intensity information; S7: Implement environmentally friendly coating detection based on environmentally friendly coating detection information; The method of obtaining supplementary illumination information according to the water droplet and bubble area and the three-dimensional information of the water droplet and bubble area of ​​the environmentally friendly coating includes: determining the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural illumination conditions in the frequent freezing and thawing area according to the three-dimensional information of the water droplet and bubble area in the environmentally friendly coating, combined with the natural illumination information in the frequent freezing and thawing area; if the light scattering effect of the water droplets and bubbles in the environmentally friendly coating under the natural illumination conditions in the frequent freezing and thawing area is good, then no artificial light is supplemented; otherwise, artificial light is supplemented when the natural illumination conditions in the frequent freezing and thawing area are poor; specifically, collecting historical meteorological data in the frequent freezing and thawing area, recording the morphology, distribution, and density change information of water droplets and bubbles on the surface of the environmentally friendly coating, and determining the illumination intensity and distribution characteristics required for the best light scattering effect on the surface of the environmentally friendly coating; using a ray tracing algorithm to simulate the scattering, reflection, and projection process of natural light and artificial light elements on the coating surface; integrating environmental data, surface state, and illumination requirements, constructing a feature set, using a random forest model to predict the light scattering effect under different weather conditions, and verifying the prediction accuracy of the model through historical data; Among them, the illumination compensation algorithm is used to automatically adjust the intensity, angle and color temperature of the artificial light source according to the current natural lighting conditions and the predicted light scattering effect to supplement the insufficient natural light and ensure that the overall lighting is uniform and meets the scattering requirements.

2. The environmentally friendly coating detection method according to claim 1, characterized in that: The method of obtaining water droplets and bubble areas of the environmentally friendly paint according to the environmentally friendly paint image includes: using a camera to obtain the positions of water droplets and bubbles in the environmentally friendly paint under good lighting conditions and calculating the areas and volumes of the water droplets and bubbles; specifically: using a high-resolution camera to take photos of the paint surface under good lighting conditions to ensure that the outlines of water droplets and bubbles are clearly visible; using image denoising, contrast enhancement and brightness adjustment, and histogram equalization technology to improve image quality; applying an edge detection algorithm to identify the outlines of water droplets and bubbles, and using a watershed algorithm in image segmentation technology to accurately segment each independent water droplet or bubble; for each segmented water droplet or bubble area, directly calculating the number of pixels in each water droplet or bubble area, and then converting it into an actual area according to the resolution of the image, the water droplet or bubble area calculation formula is: area = (number of pixels) × (unit pixel area); the bubble volume calculation formula is: ,in, is the radius.

3. The environmentally friendly coating detection method according to claim 1, characterized in that: The method of obtaining three-dimensional information of water droplets and bubbles in the environmentally friendly paint according to the scanned image of the environmentally friendly paint includes: if the light condition is poor, using a laser detector to scan the environmentally friendly paint, and using a camera to analyze the scanning situation, obtaining the positions of water droplets and bubbles in the environmentally friendly paint and calculating the area, volume and height of the water droplets and bubbles; specifically: using a laser source to generate one or more laser lines or points, irradiating the environmentally friendly paint, and using a camera to capture the projection image formed by the laser on the surface of the object; determining a baseline and ensuring that the relative position between the laser and the camera is fixed, the baseline is the distance between the laser source and the camera, and obtaining the focal length and principal point position information of the camera; projecting the laser onto the surface of the environmentally friendly paint and using a camera to capture the pattern formed by the laser spot on the surface of the object, as well as the deformation of water droplets and bubbles; processing the image and calculating the area, volume and height of the bubble, wherein the area calculation formula is: , the volume calculation formula is: ,in, is the radius, and the height calculation formula is: ,in, is the horizontal pixel position of the laser spot in the image captured by the camera, is the focal length of the camera, is the angle between the laser source and the camera.

4. The environmentally friendly coating detection method according to claim 1, characterized in that: The method of obtaining the illumination intensity information of the water droplets and bubbles in the environmentally friendly paint according to the illumination supplementary information includes: dividing the environmentally friendly paint area into N areas, calculating the illumination intensity of seven colors scattered by the water droplets and bubbles in the environmentally friendly paint by natural light in each area; and calculating the interaction between the colors of the environmentally friendly paint and the combination effect of specific colors according to the distribution of the illumination intensity. If the environmentally friendly paint is qualified in the inspection, a specific painting method or pigment ratio is selected according to the interaction between the colors and the combination effect of specific colors; specifically, dividing the N areas into 7 color channels, calculating the relative intensity of the seven color channels in each area, and then performing normalization processing. Let C={R, O, Y, G, B, I, V}, which represents the set of seven colors, for any area ( =1, 2, 3, ..., N), N is the total number of regions, and the relative intensity of each color is expressed in express, ,satisfy , that is, the sum of all color frequencies in the region is 1, ensuring normalization and calculating the color comprehensive index in the region , the calculation formula is: ,in, is the weight, reflecting the importance, determined according to actual needs, while taking into account the interaction between colors and the effect of certain color combinations. The calculation formula is: , It's color pair , The combination weight of represents two colors.

5. The environmentally friendly coating detection method according to claim 1, characterized in that: The environmental protection coating detection information is obtained according to the light intensity information of water droplets and bubbles, including: according to the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles appearing in the frequent freezing and thawing area of ​​the environmental protection coating, combined with the micro cracks, mildew spots, and fading information of the environmental protection coating, a model is used to analyze the existence / disappearance time, three-dimensional information, and light intensity information of water droplets and bubbles appearing in the frequent freezing and thawing area of ​​the environmental protection coating, and the micro cracks, mildew spots, and fading information of the environmental protection coating, so as to obtain the detection information of the environmental protection coating in the frequent freezing and thawing area; specifically: collect a large amount of information including micro cracks, mildew spots, fading, bubbles and water droplets of the coating. The high-quality images of the scattering features are annotated with the specific location and type of defects in each image, and the images are normalized; the CNN model is selected, and the edge, texture, and color distribution features in the image are automatically extracted using the convolutional layer of the CNN. The model is trained using the annotated data set, and the weights are adjusted through back propagation and gradient descent methods so that the model can learn to identify different types of defects; the model is tuned using the validation set, and the best hyperparameter combination is determined using grid search; the model performance is evaluated using the F1 score on an unseen test set, and a color change tree and a color network diagram are constructed. The F1 score formula is: .

6. The environmentally friendly coating detection method according to claim 5, characterized in that: The color change tree is constructed, including: constructing a color change binary tree, visually displaying seven color intensity measurement results reflected by water droplets and bubbles of the same batch of environmentally friendly paints in a frequent freeze-thaw area at different time points; specifically: the root node represents the initial color measurement, the child node represents the subsequent color measurement that changes over time, the edge represents the amount of color change, measured by color difference, and the trend of the color change of the paint over time is visualized by a tree structure, one child node represents that the color becomes lighter, and the other child node represents that the color becomes darker, and each color change is quantified by calculating the color difference and stored in the corresponding edge; The color difference calculation formula is: ,in, : Indicates the difference in brightness (L*) between two colors; : Indicates the difference between two colors in chroma (C*), which reflects the purity or saturation of the color; : Indicates the difference between two colors in hue angle (H*), that is, the hue change of the color; , , : is an empirical constant used to adjust the contribution of brightness, chroma, and hue to the total color difference; , , : is a weighting factor based on brightness and chroma; : is a rotation term related to hue angle and chroma difference.

7. The environmentally friendly coating detection method according to claim 6, characterized in that: Constructing a color network diagram, including: constructing a color network diagram according to the measurement results and color similarities of seven color intensities scattered by water droplets and bubbles of environmentally friendly coatings in frequent freeze-thaw areas of different samples at different time points; specifically: nodes represent color measurements of different samples at different time points, edges represent connections between nodes with similar color features, and the weight of the edge represents the color similarity between two nodes. A K-nearest neighbor algorithm is used to construct a diagram, in which each node is connected to the nearest K nodes, and the weight of the edge is the inverse of the color difference. The higher the weight, the more similar the color is. Color measurements are performed on multiple samples at different times, and a color network diagram is constructed for each sample. The network structures of different samples are compared to find common patterns or abnormal patterns.

8. The environmentally friendly coating detection method according to claim 7, characterized in that: The color network graph includes: constructing a graph, for each sample, taking each measurement as a node, using color difference to calculate the similarity between nodes, and constructing weighted edges; analyzing the graph, using a community discovery algorithm to identify sample groups with similar color change trends; and detecting anomalies, finding samples with abnormal color changes through connectivity and density analysis of the graph structure.

9. An environmentally friendly coating detection system, characterized in that: Included are: Environmentally friendly paint image acquisition module: used to collect images of environmentally friendly paint taken by the camera; Environmentally friendly paint scanning image acquisition module: used to collect the environmentally friendly paint scanning image detected by the laser detector; Bubble and water drop area calculation module: used to calculate the area of ​​bubbles and water drops of environmentally friendly coatings; Bubble and water drop volume calculation module: used to calculate the volume of bubbles and water droplets of environmentally friendly coatings; Bubble and water drop height calculation module: used to calculate the height of bubbles and water drops of environmentally friendly coatings; Random forest model building module: used to predict the light scattering effect under different weather conditions; CNN model building module: used to obtain environmentally friendly coating detection information; Color change tree building module: used to visualize the trend of paint color changes over time based on a tree structure; Color network graph construction module: used for connectivity and density analysis of graph structure and finding abnormal samples of color changes.

Images