Decoration quality intelligent detection method and system

Abstract

The invention relates to the technical field of image processing, in particular to an intelligent decoration quality detection method and system, and the method comprises the steps: carrying out the preprocessing of a historical image of a wall surface, and labeling the historical image, and enabling different labels to represent different defects; setting a scale, taking any pixel point as a central pixel point, taking the scale as a radius to obtain a circular trajectory, and sampling the pixel points on the circular trajectory to obtain neighborhood sampling points; under any scale, calculating the overall consistency of central pixel points in an area where any type of defects are located, and obtaining a corresponding direction set under the scale; according to the direction set corresponding to each type of defects under each scale, calculating to obtain a set difference; training the neural network model to obtain a detection model; and inputting the real-time picture into the detection model, and outputting the probability that the real-time picture belongs to different labels. The method has the effect of improving the wall surface detection precision.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an intelligent decoration quality detection method and system. Background Art

[0002] During the indoor decoration process, factors such as wall paint (for example, poor adhesion of inferior putty), temperature and humidity changes in the external environment, etc., can cause surface defects of the decorated wall. Specific defect types can be cracks, bulges, scratches, etc. The above defects will affect the flatness and aesthetics of the wall on the one hand, and on the other hand, there will be a risk of wall paint falling off, increasing subsequent maintenance costs. Therefore, the detection of the wall after or during the decoration process is an important part of the decoration process. Traditional wall defect detection mostly relies on manual observation, which is easily affected by subjective factors on the one hand, and on the other hand, it will increase the labor burden of the staff, which is not conducive to improving efficiency. In order to reduce the labor burden of the staff, the related technology can detect the decorated wall through image recognition and deep learning algorithms. For example, a wall crack diagnosis method, device, computer equipment and storage medium disclosed in the patent application document with publication number CN118154604A. It first obtains the image information of the wall, and then recognizes the wall image to identify the cracks on the wall surface.

[0003] LBP (Local Binary Patterns) is an algorithm used for image texture analysis and feature extraction. It can be used to extract wall defect information, but the features of the wall have strong directionality. Traditional LBP extracts features through the grayscale value relationship between each pixel and its neighboring pixels. In the extraction process, the importance of neighboring pixels is the same and the scale is single, resulting in the single-scale LBP feature ignoring the direction of the feature, and the effective extraction range has a large deviation, resulting in low accuracy in the recognition of the decorated wall. Summary of the invention

[0004] In order to improve the accuracy of identifying defects on decorated walls, the present application provides an intelligent decoration quality detection method and system.

[0005] In the first aspect, the present application provides an intelligent decoration quality detection method, which adopts the following technical solutions: An intelligent decoration quality detection method includes the steps of: preprocessing and labeling historical images of a wall, wherein different labels represent different defects; Set the scale, take any pixel as the center pixel, take the scale as the radius, get the circular trajectory, sample the pixel points on the circular trajectory to get the neighborhood sampling points; At any scale, calculate the overall consistency of the central pixel points in the area where any type of defect is located to obtain the corresponding direction set at the scale; calculate the set difference based on the direction sets corresponding to each type of defect at each scale; Train the neural network model to obtain the detection model; input the real-time image into the detection model and output the probability that the real-time image belongs to different labels; Among them, the calculation formula for the probability of real-time images belonging to different labels is: ; represents the probability that the real-time image belongs to this label, is the number of selected scales, Indicated on the scale Obtained The set difference between the direction sets corresponding to different defects under the feature map, It indicates the probability that the historical images output during the training of the neural network model belong to different labels. When the probability meets the preset threshold corresponding to any label, it indicates that there is a defect corresponding to the label on the wall.

[0006] In this application, the directional consistency of the central pixel and the neighboring pixels is calculated, and the directional consistency is used as the weight to determine the central pixel. The eigenvalues ​​make the image Feature extraction is more accurate and comprehensive, reducing background noise interference. At the same time, a direction set is constructed based on the overall consistency of the central sampling point in the defect area, and the set difference between different direction sets is calculated; the greater the set difference between the direction sets corresponding to different defects at the same scale, the stronger the ability to distinguish defect types at this scale. Using this feature, weights are assigned to different scales. Combined with the weights of different scales, the probability of the real-time image of the wall belonging to different defects is calculated to improve the accuracy of decoration quality detection.

[0007] Optionally, the set difference is calculated as: ; represents the set difference of different defects at a scale, Indicates Defects and The chi-square test value between the direction sets of the defects; is a linear normalization function; is the number of defect types.

[0008] In the formula, the chi-square test value indicates the difference between different direction sets. Different sets correspond to different defects. The greater the difference, the stronger the ability to distinguish defects. The chi-square test values ​​between different direction sets are summed. The larger the sum, the greater the difference between the sets. Therefore, the extracted Features can better distinguish between multiple defect types, thereby being able to assign weights to different scales.

[0009] Optionally, the method for constructing the direction set includes: selecting multiple scales for the defect area, calculating the overall consistency of each pixel in the defect area at the multiple scales; the overall consistency of multiple pixels in the defect area constitutes the direction set.

[0010] Optionally, the steps of calculating the overall consistency are: calculating the directional consistency between the central sampling point and the corresponding different neighborhood sampling points; calculating the overall consistency of the central pixel point according to the directional consistency between the central pixel point and multiple neighborhood sampling points; The overall consistency is calculated as: ; In the formula, Indicates the overall consistency of the central pixel, represents the number of neighborhood sampling points, Represents a pixel and its The direction consistency of the neighborhood sampling points.

[0011] There are multiple neighborhood sampling points corresponding to a central pixel point, and there is a directional consistency between a central pixel point and a neighborhood sampling point. The directional consistency between the central sampling point and different neighborhood sampling points is calculated, and then the directional consistency between multiple neighborhood sampling points and the central sampling point is processed to obtain the overall trend of the central pixel point and multiple neighborhood sampling points, thereby improving the robustness of detection.

[0012] Optionally, the step of calculating the directional consistency includes: calculating the gradient directions of the neighborhood sampling points and the central pixel point respectively; calculating the directional consistency according to the gradient directions of the neighborhood sampling points and the central pixel point; The calculation formula for directional consistency is: ; Indicates the center pixel and the The direction consistency of the neighborhood sampling points, Indicates The exponential function with base , Represents the gradient direction of the central pixel, Indicates The gradient direction of the neighborhood sampling points is Represents the standard deviation of the gradient direction.

[0013] Calculate the consistency of the gradient direction of the central pixel point and the neighborhood sampling point. The more consistent the gradient directions of the two are, the more consistent the directions of the textures they represent are, which enhances the description of the directional texture of the wall and improves the accuracy of the detection of the decorated wall.

[0014] Optionally, the step of calculating the gradient directions of the neighborhood sampling points and the central pixel point includes: According to the gradient of the neighboring pixels of the neighborhood sampling point, the horizontal gradient and vertical gradient of each neighborhood sampling point are calculated. The calculation method of the vertical gradient is the same as that of the horizontal gradient. The gradient direction of the neighborhood sampling point is calculated by the horizontal gradient and vertical gradient of the neighborhood sampling point. The gradient direction of the central pixel point is obtained in the same way as that of the gradient direction of the neighborhood sampling point. The calculation formula of the horizontal gradient of the neighborhood sampling point is: ; Where: is the horizontal gradient of the neighborhood sampling point, is the first sampling point in the neighborhood The horizontal gradient of neighboring pixels, is the neighborhood sampling point and its The difference between adjacent pixels in horizontal coordinates; The number of neighboring pixels of the selected neighborhood sampling point; is the neighborhood sampling point and its The pixel value difference between adjacent pixels.

[0015] The horizontal gradient and vertical gradient are obtained by interpolating the pixels around the neighborhood sampling point. The change of pixel value in an image is usually the result of the joint effect of multiple neighboring pixels, and As a weight coefficient, it measures the contribution of different neighboring pixels to the gradient of the neighborhood sampling point, ensuring that the neighboring pixels that are closer to the neighborhood sampling point and have larger pixel value differences contribute more to the gradient, thereby improving the accuracy of gradient calculation and thus improving the robustness of subsequent feature extraction.

[0016] Optionally, the pixel values ​​of the neighborhood sampling points are calculated using a bilinear interpolation method.

[0017] Optionally, during the training of the neural network model, historical images and corresponding images at different scales are input. Feature map, output the probability that the historical image belongs to different labels; the loss function of training is Huber loss, and the condition for stopping training is: reaching the preset number of training times or the loss function value is less than the preset threshold.

[0018] Optional, historical images corresponding to The method for obtaining the feature map includes: calculating the The feature value is obtained Feature map; The eigenvalue is calculated as: ; in, is the center pixel Eigenvalues, is the number of neighborhood sampling points, is the center pixel and the The direction consistency of the neighborhood sampling points, It is The pixels of the neighboring sampling points, is the pixel at the center pixel, .

[0019] The directional consistency between the central pixel and the neighborhood sampling points is used as a condition for extraction Eigenvalues, according to different neighborhood points The contribution made by the eigenvalues ​​is weighted so as to better capture the directional texture features in the image.

[0020] In the second aspect, the present application provides an intelligent decoration quality detection system, which adopts the following technical solutions: A decoration quality intelligent detection system comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned decoration quality intelligent detection method is adopted.

[0021] The above-mentioned intelligent decoration quality detection method is generated into a computer program and stored in a memory so as to be loaded and executed by a processor. Thus, a system is made based on the memory and the processor for easy use.

[0022] This application has the following technical effects: Extraction using different neighborhood radii at different scales Features can reduce the omission of texture details in the image. According to the set differences between the direction sets of different defects at different scales, weights are assigned to different scales, and the probability of belonging to multiple defects in the real-time image is calculated based on the weights. In this application, the application of weights fully considers the extracted The difference between features determines the extracted The importance of features is improved, which reduces the error with the true value and improves the detection accuracy of decorated walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a method flow chart of an intelligent decoration quality detection method of the present application.

[0024] Figure 2 This is a method flow chart of step S3 in an intelligent decoration quality detection method of the present application.

[0025] Figure 3 This is a method flow chart of step S4 in an intelligent decoration quality detection method of the present application.

[0026] Figure 4 This is a method flow chart of step S5 in an intelligent decoration quality detection method of the present application. DETAILED DESCRIPTION

[0027] The embodiment of the present application discloses an intelligent detection method for decoration quality, which extracts LBP sign values ​​of images at multiple scales. The direction consistency of the central pixel point and its own neighborhood pixel points in the defect area corresponding to different defects at the same scale constitutes a direction set. Weights are set for different scales according to the set differences between the direction sets corresponding to different defects at the same scale. After the real-time image is input into the neural network model, the probability that the real-time image belongs to different defects is calculated in combination with the scale weight, thereby improving the accuracy of intelligent detection of decoration quality. Figure 1 The intelligent detection method for decoration quality mainly includes steps S1 to S6.

[0028] S1: Preprocess and label the historical images of the wall, and different labels represent different defects.

[0029] In step S1, a high-resolution camera is used to collect historical images of different wall decorations, a Gaussian filter is used to remove noise in the image, the image is grayed, and a label (cracks, paint scratches, stains, bulges, normal labels, all labels except normal are defects) is given to each historical image by technicians in this field.

[0030] S2: Set the scale, take any pixel as the center pixel, and the scale as the radius to obtain a circular trajectory line, and sample the pixel points on the circular trajectory line to obtain neighborhood sampling points.

[0031] The number of neighborhood sampling points is In this embodiment, bilinear interpolation (a method for interpolating discrete data in two-dimensional space. It estimates the value of any point by taking a weighted average of the values ​​of the four nearest neighboring points to estimate the pixel value of the neighboring sampling point) is used to calculate the value of the neighboring sampling point.

[0032] S3: At any scale, calculate the overall consistency of the central pixel points in the area where any type of defect is located, and obtain the corresponding direction set at the scale; The gradient direction of the central pixel is calculated by the horizontal gradient and vertical gradient of the central pixel, and the gradient direction of the neighboring sampling points is calculated by calculating the horizontal gradient and vertical gradient of the neighboring sampling points. The consistency of the gradient directions of the neighboring sampling points and the central pixel sampling point is calculated by the gradient directions of the neighboring sampling points and the central pixel sampling point. The overall consistency of the central pixel is calculated by the direction consistency between the central sampling point and multiple neighboring sampling points. The overall consistency of the central pixel points in the defect area corresponding to different defects at different scales forms a direction set, and the difference of the direction sets corresponding to different defects at different scales is calculated to obtain the set difference.

[0033] In one embodiment, referring to Figure 2 , step 3 includes steps S30 to S36.

[0034] S30: Grayscale image edge filling.

[0035] Because the edge pixels of an image are located on the boundary of the image, their neighboring pixels are not complete, so the neighborhood of the edge pixels of an image can usually only consider one quadrant. For example, if a central pixel is located in the upper left corner of the image, then its neighboring pixels only exist in the lower right corner of the central pixel. In this application, the edge pixels are obtained by filling the edge of the image. Features (in extracting Avoid extracting filled pixels when using features Features); thereby improving the accuracy of intelligent detection of decoration quality.

[0036] The edge filling method is copy filling. The main principle of copy filling is to expand the neighborhood of edge pixels beyond the image boundary and fill the extended area by copying the values ​​of edge pixels. Specifically, if the pixel is located at the upper edge of the image, the pixel value of the upper edge is copied to the extended area for filling. If the pixel is located at the lower edge of the image, the pixel value of the lower edge is copied to the extended area for filling. This is done by analogy to complete the filling of the left and right edges of the image.

[0037] S31: Calculate the horizontal gradient and vertical gradient of each neighborhood sampling point according to the gradients of the neighboring pixels of the neighborhood sampling point; The horizontal gradient of the neighborhood sampling point is calculated based on the gradient of the neighboring pixels of the neighborhood sampling point. The calculation method is: ; Where: is the horizontal gradient of the neighborhood sampling point, is the first sampling point in the neighborhood The horizontal gradient of neighboring pixels, is the neighborhood sampling point and its The difference between the horizontal coordinates of adjacent pixels (vertical gradient The calculation time is vertical coordinate difference); The number of neighboring pixels of the selected neighborhood sampling point; is the neighborhood sampling point and its The pixel value difference of neighboring pixels. The closer the horizontal gradient of the neighboring pixels is to the neighborhood sampling point, the greater the influence on the gradient of the neighborhood sampling point.

[0038] In an image, the change of pixel value is usually the result of the interaction of neighboring pixels. By weighted summation, the horizontal gradient calculation formula can comprehensively consider the influence of multiple neighboring pixels, thereby improving the accuracy of horizontal gradient calculation.

[0039] The vertical gradient is calculated in the same way as the horizontal gradient.

[0040] S32: Calculate the gradient direction of the neighborhood sampling point through the horizontal gradient and the vertical gradient of the neighborhood sampling point.

[0041] The gradient direction calculation method of the neighborhood sampling point is: ; In the formula, Indicates The gradient direction of the neighborhood sampling points; is the inverse tangent function, which is used to calculate the angle between the gradient vector and the horizontal axis; is the horizontal gradient of the neighborhood sampling point, is the vertical gradient of the neighborhood sampling point.

[0042] S33: The gradient direction of the central pixel point is obtained in the same way as the calculation method of the gradient direction of the neighborhood sampling points; the specific calculation is not repeated here.

[0043] S34: Calculate the directional consistency of the gradient direction of the neighborhood sampling point and the gradient direction of the central pixel point.

[0044] In one embodiment, the calculation formula for directional consistency is: ; In the formula, Indicates the center pixel and the The direction consistency of the neighborhood sampling points, is an exponential function, Represents the gradient direction of the central pixel, Indicates The gradient direction of the neighborhood sampling points is Represents the standard deviation of the gradient direction.

[0045] In the formula, Indicates the center pixel and the The absolute value of the difference between the gradient directions of the neighboring sampling points. This difference value is the core part of the extreme consistency of the direction. The smaller the difference, the closer the gradient direction of the central pixel is to that of the neighboring sampling points. It reflects the degree of change of the gradient direction in the entire image or a certain area. If the gradient direction changes greatly, The value of will be larger; on the contrary, if the gradient direction changes less, The value of will be smaller. Divide by the standard deviation , we can normalize the difference so that it is not affected by the overall change in the gradient direction in the image. The minus sign transforms the difference into a decreasing function. The larger the difference, the smaller the value of the expression; the smaller the difference, the larger the value of the expression. Finally, the exponential function Map the difference value to a weight value between 0 and 1 to adjust the neighborhood sampling points to the central pixel point Contribution of eigenvalues.

[0046] in the formula The standard deviation of the gradient direction refers to the standard deviation of the gradient direction between the central pixel and the neighborhood sampling points.

[0047] In another embodiment, the calculation formula for directional consistency is: ; In the formula, Indicates the center pixel and the The direction consistency of the neighborhood sampling points, is an exponential function, Represents the gradient direction of the central pixel, Indicates The gradient direction of the neighborhood sampling points is Represents the standard deviation of the gradient direction.

[0048] When calculating directional consistency in this formula, the square root of the difference between the central pixel and the neighborhood sampling points is used. This calculation method is smoother and can better handle small changes in pixel gradients.

[0049] S35: Calculation of overall consistency; The overall consistency of each pixel in the defect area under the same defect at the same scale can be calculated by the following formula: ; In the formula Indicates the overall consistency of the central pixel, represents the number of neighborhood sampling points, Represents a pixel and its The direction consistency of the neighborhood sampling points.

[0050] In another embodiment, the overall consistency formula for each pixel in the defect area under the same defect at the same scale is: ; In the formula, Indicates the overall consistency of a pixel. represents the number of neighborhood sampling points, Represents a pixel and its The direction consistency of the neighborhood sampling points. represents the minimum function, represents the maximum value function; Indicates the directional consistency between the central pixel and the neighboring pixels.

[0051] S36: constructing a direction set; at one scale, the overall consistency of each central pixel point in the defect area corresponding to one defect forms a direction set.

[0052] The defect area includes multiple pixels, and the overall consistency of each pixel in the defect area constitutes a direction set. Each defect includes a direction set at each scale, so multiple defects form multiple direction sets at different scales.

[0053] S4: Based on the direction sets corresponding to each type of defect at each scale, the set difference is calculated.

[0054] Reference Figure 3 , step S4 includes step S41-step S42.

[0055] S41: Calculate the chi-square test value between the direction sets corresponding to different defects at the same scale.

[0056] The chi-square test value represents the difference between different direction sets, and in other embodiments, it can also be represented by KL divergence (an asymmetric measure used in Kullback-Leibler Divergence information theory to measure the difference between two probability distributions), intra-class variance (referring to the degree of difference between pixel values ​​in the same class), and the like.

[0057] S42: Calculate set differences; The set differences of different defects under this scale are calculated based on the chi-square test values ​​between the sets in each direction. The set differences can be calculated by the following formula: ; It indicates the set difference in the direction set corresponding to different defect types at a scale. The larger the set difference, the larger the scale. The more important the feature map is, Indicates Defects and The chi-square test value between the direction sets of the defects; is the number of defect types; is a linear normalization function.

[0058] The sum of the Karma test values ​​between all direction sets is calculated. For example, when there are three defects (cracks, scratches, and bulges), , then three group differences need to be calculated, namely cracks and bulges, cracks and scratches, and scratches and bulges.

[0059] If a certain scale If the feature can distinguish all defects well, the sum of the chi-square test values ​​between all defect pairs will be large, indicating the importance of this scale for defect detection, thus fully reflecting the The ability of a feature to distinguish between different defects.

[0060] In another embodiment, the calculation formula of the set difference is: ; It indicates the set difference in the direction set corresponding to different defect types at a scale. The larger the set difference, the larger the scale. The more important the feature map is, Indicates Defects and The chi-square test value between the direction sets of the defects; is the number of defect types; is an exponential function; is a linear normalization function.

[0061] S5: Train the neural network model to obtain a detection model.

[0062] Reference Figure 4 , step S5 includes step S51-step S52.

[0063] S51: Calculate the pixel points in the historical image LPB Eigenvalues, get the corresponding historical images Feature map; The calculation method is: ; In the formula is the center pixel Eigenvalues, is the number of neighborhood sampling points, is the center pixel and the The direction consistency of the neighborhood sampling points, It is The pixels of the neighboring sampling points, is the pixel at the center pixel, .

[0064] It reflects the consistency of the gradient direction between the neighborhood pixels and the neighborhood sampling points. The higher the direction consistency, The larger the value. Represents the neighborhood sampling point pair Relative contribution of eigenvalues. Based on the similarity of the gradients of the neighborhood sampling points and the central pixel, the contribution of each neighborhood sampling point is dynamically adjusted to better capture directional texture features in the image, such as wall cracks, paint scratches, etc.

[0065] S52: Neural network model training.

[0066] The input of the neural network model in the training process is the historical image and the corresponding Feature map, the output is the probability that the historical image belongs to different labels; The feature map corresponds to a set of label probabilities. F1-score is used to evaluate the model performance during training, and Huber loss (Huber Loss is a loss function that combines the advantages of mean square error and mean absolute error, aiming to improve the robustness of the model to outliers) is selected as the model loss function. Training is stopped when the preset number of training times is reached or the loss function value is less than the preset threshold. In this embodiment, the preset number of training times is 1000 times, and the preset threshold of the loss function is 0.01; the detection model is obtained after the training is completed.

[0067] S6: Input the real-time image into the detection model and output the probability that the real-time image belongs to different labels.

[0068] During the decoration detection process, take photos of the wall in real time, input the previous real-time pictures into the detection model, and output the probability that the real-time pictures belong to different labels; the probability of the labels can be calculated by the following formula: ; In the formula, represents the probability that the real-time image belongs to this label, is the number of scales, Indicated on the scale Obtained The difference in directional consistency of different defects under the feature map, Representation based on scale Obtained The probability of the model outputting the label when the feature map is used as input.

[0069] In scale The greater the difference in the sets of directions corresponding to different defects, the greater the difference in the sets of directions corresponding to different defects. The more important the feature map is for detecting defects, the more Get the scale The weight of The larger the value, the greater the weight of the scale. When the feature map is used as input, the model outputs the probability of belonging to the label, and the total probability is calculated by weight, and the probability of the real-time image belonging to different labels is obtained, thereby considering features of different scales and improving the accuracy of detection.

[0070] An embodiment of the present application also discloses a decoration quality intelligent detection method system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an decoration quality intelligent detection method according to the present application is implemented.

[0071] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0072] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A decoration quality intelligent detection method, characterized in that: The method comprises the following steps: preprocessing and labeling the historical images of the wall, where different labels represent different defects; Set the scale, take any pixel as the center pixel, take the scale as the radius, get the circular trajectory, sample the pixel points on the circular trajectory to get the neighborhood sampling points; At any scale, calculate the overall consistency of the central pixel points in the area where any type of defect is located to obtain the corresponding direction set at the scale; calculate the set difference based on the direction sets corresponding to each type of defect at each scale; Train the neural network model to obtain the detection model; input the real-time image into the detection model and output the probability that the real-time image belongs to different labels; Among them, the calculation formula for the probability of real-time images belonging to different labels is: ; represents the probability that the real-time image belongs to this label, is the number of selected scales, Indicated on the scale Obtained The set difference between the direction sets corresponding to different defects under the feature map, It indicates the probability that the historical images output during the training of the neural network model belong to different labels. When the probability meets the preset threshold corresponding to any label, it indicates that there is a defect corresponding to the label on the wall.

2. The intelligent decoration quality detection method according to claim 1 is characterized in that: in, The formula for calculating the set difference is: ; represents the set difference of different defects at a scale, Indicates Defects and The chi-square test value between the direction sets of the defects; is a linear normalization function; is the number of defect types.

3. The intelligent decoration quality detection method according to claim 2 is characterized in that: The method for constructing a direction set includes: selecting multiple scales for the defect area, calculating the overall consistency of each pixel in the defect area at the multiple scales; the overall consistency of multiple pixels in the defect area constitutes a direction set.

4. The intelligent decoration quality detection method according to claim 1 is characterized in that: The calculation steps of the overall consistency are: calculating the directional consistency between the central sampling point and the corresponding different neighborhood sampling points; calculating the overall consistency of the central pixel point according to the directional consistency between the central pixel point and multiple neighborhood sampling points; The overall consistency is calculated as: ; In the formula, Indicates the overall consistency of the central pixel, represents the number of neighborhood sampling points, Represents a pixel and its The direction consistency of the neighborhood sampling points.

5. The intelligent decoration quality detection method according to claim 4 is characterized in that: The step of calculating the directional consistency includes: calculating the gradient directions of the neighborhood sampling points and the central pixel point; calculating the directional consistency according to the gradient directions of the neighborhood sampling points and the central pixel point; The calculation formula for directional consistency is: ; Indicates the center pixel and the The direction consistency of the neighborhood sampling points, Indicates The exponential function with base , Represents the gradient direction of the central pixel, Indicates The gradient direction of the neighborhood sampling points is Represents the standard deviation of the gradient direction.

6. The intelligent decoration quality detection method according to claim 5 is characterized in that: The steps of calculating the gradient directions of the neighborhood sampling points and the central pixel point include: According to the gradient of the neighboring pixels of the neighborhood sampling point, the horizontal gradient and vertical gradient of each neighborhood sampling point are calculated. The calculation method of the vertical gradient is the same as that of the horizontal gradient. The gradient direction of the neighborhood sampling point is calculated by the horizontal gradient and vertical gradient of the neighborhood sampling point. The gradient direction of the central pixel point is obtained in the same way as that of the gradient direction of the neighborhood sampling point. The calculation formula of the horizontal gradient of the neighborhood sampling point is: ; Where: is the horizontal gradient of the neighborhood sampling point, is the first sampling point in the neighborhood The horizontal gradient of neighboring pixels, is the neighborhood sampling point and its The difference between adjacent pixels in horizontal coordinates; The number of neighboring pixels of the selected neighborhood sampling point; is the neighborhood sampling point and its The pixel value difference between adjacent pixels.

7. The intelligent decoration quality detection method according to claim 1 is characterized in that: The pixel values ​​of the neighborhood sampling points are calculated using the bilinear interpolation method.

8. The intelligent decoration quality detection method according to claim 1 is characterized in that: During the training of the neural network model, historical images and corresponding images at different scales are input. Feature map, output the probability that the historical image belongs to different labels; the loss function of training is Huber loss, and the condition for stopping training is: reaching the preset number of training times or the loss function value is less than the preset threshold.

9. The intelligent decoration quality detection method according to claim 7 is characterized in that: Historical images corresponding to The method for obtaining the feature map includes: calculating the The feature value is obtained Feature map; The eigenvalue is calculated as: ； in, is the center pixel Eigenvalues, is the number of neighborhood sampling points, is the center pixel and the The direction consistency of the neighborhood sampling points, It is The pixels of the neighboring sampling points, is the pixel at the center pixel, .

10. An intelligent decoration quality detection system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent detection method for decoration quality according to any one of claims 1 to 9 is implemented.

Images