Image recognition-based brick structure wall masonry parameter detection method

Abstract

The invention discloses a brick structure wall brick and stone parameter detection method based on image recognition, belongs to the technical field of wall brick and stone detection, and solves the problems of low efficiency and large error in manual detection of brick structure wall surface brick and stone parameters. According to the method, the brick and the wall are rapidly detected and evaluated by calculating the distribution uniformity, the irregular shape parameter and the divergence of the brick, the current state of the brick and the wall is reflected, the problem of limitation of the brick of the brick structure wall based on manual detection in the prior art is solved, the detection efficiency is improved, and the detection cost is reduced. The feasibility, comprehensiveness and objectivity of brick structure wall masonry detection are analyzed, and the reliability and authenticity of brick structure wall masonry detection analysis results are guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of wall masonry detection, and in particular to a method for detecting masonry parameters of a brick-built wall structure based on image recognition. Background Art

[0002] Brickwork is a traditional form of building structure. Its raw materials are widely available and easy to obtain. The process is simple and does not require formwork or special construction equipment. Brickwork can save a lot of cement, steel, wood and other building materials, thereby reducing the project cost. Therefore, brickwork has been widely used in construction projects.

[0003] The engineering quality inspection of brickwork structures is of great significance to ensure the safety, stability and long-term service life of brickwork buildings. It is very necessary to extract the parameters of masonry materials on the wall surface of brickwork structures. The traditional method of detecting masonry parameters on the wall surface of brickwork structures is generally manual inspection, which extracts masonry parameters through direct visual observation combined with auxiliary inspection tools. This method is not only inefficient, but may also cause errors due to human factors or factors of the inspection tools themselves. The parameter extraction accuracy is poor and the inspection effect is poor. It is difficult to meet the accuracy requirements for wall masonry parameter extraction and the high requirements for brickwork structure engineering quality inspection. Summary of the invention

[0004] In view of the above problems in the prior art, the present invention provides a method for detecting masonry parameters of a brick structure wall based on image recognition, which solves the problems of low efficiency and large errors in manually detecting masonry parameters on the surface of a brick structure wall.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows: A method for detecting masonry parameters of a brick structure wall based on image recognition is provided, comprising the steps of: S1. Obtain a detection image of a brick wall and preprocess the detection image.

[0006] S2. Input the preprocessed detection image into the SAM model to generate a segmentation mask image that identifies the masonry area.

[0007] S3. Binarize, open and close the segmentation mask image in sequence, and perform edge detection on the processed segmentation mask image to obtain the outline of the masonry unit; read the position data of all pixels in the processed segmentation mask image to obtain the number of the masonry unit and the corresponding pixel coordinates.

[0008] S4. According to the outline and pixel coordinates of each masonry unit, the geometric moment, perimeter and minimum circumscribed rectangle of the corresponding masonry unit are calculated, and the area of ​​the corresponding masonry unit is obtained through the geometric moment.

[0009] S5. According to the area of ​​each masonry unit, the area ratio of each masonry unit to the detection image is calculated, each masonry unit is classified according to the area ratio of each masonry unit and the aspect ratio of the minimum circumscribed rectangle, and the masonry distribution uniformity is calculated; the masonry distribution uniformity reflects whether the masonry units are evenly distributed in the entire wall structure. By controlling the masonry distribution uniformity, local over-crowding or over-sparseness can be avoided, thereby ensuring the consistency of the wall.

[0010] S6. Calculate the shape irregularity parameter of each masonry unit based on the perimeter and classification information of each masonry unit, and perform weighted processing on the shape irregularity parameter of each masonry unit to obtain the wall shape irregularity characterization parameter. The shape irregularity parameter reflects the edge complexity and shape regularity of a single masonry unit. It measures the degree to which the masonry unit is close to a circle or square through the relationship between the perimeter and the area, which can help identify those masonry units whose shapes deviate significantly from the expected standard, indicating manufacturing defects or construction problems. The wall shape irregularity characterization parameter can evaluate the complexity of the internal structure of the wall, which is very important for predicting the mechanical behavior and long-term performance of the wall. Complex shapes may lead to stress concentration and affect the stability and durability of the wall.

[0011] S7. Calculate masonry convergence based on the area, perimeter, and pixel coordinates of each masonry unit. Masonry convergence can quantify whether masonry units tend to cluster together or disperse evenly. By identifying whether there are abnormal clustering or dispersion phenomena in certain areas, it indicates potential construction problems or design defects.

[0012] In this solution, each masonry unit is classified by combining the aspect ratio of its minimum circumscribed rectangle and its area ratio, which improves the accuracy of masonry unit classification compared to the method of classifying masonry units only by area ratio. By calculating the uniformity of masonry distribution, shape irregularity parameters and convergence, the masonry and wall are quickly detected and evaluated, reflecting the current status of the masonry and wall, solving the limitations of the current traditional manual detection of brick-built wall masonry, not only improving the detection efficiency, but also realizing the comprehensive and objective analysis of the feasibility of brick-built wall masonry detection, and ensuring the reliability and authenticity of the brick-built wall masonry detection analysis results.

[0013] Furthermore, the preprocessing of the detection images includes: using the PIL image processing library to unify the image resolution of all detection images, and performing denoising on each detection image through Gaussian filtering. This ensures the consistency of all input images, provides clearer images for subsequent analysis, and thus improves the accuracy of the detection results.

[0014] Further, Masonry units are recorded as , The geometric moment of is:

[0015] in, For the Masonry units Order moment, and are all non-negative integers, is a natural number; and are the width and height of the detected image respectively; and Respectively The horizontal coordinate set and vertical coordinate set of the edge pixel points of the masonry unit, and They are the position index variables of the detected image in the width and height directions respectively. The coordinates are Gray value of pixel point. Geometric moment is used to describe the mathematical characteristics of the shape of an object. It can extract the precise position and morphological information of masonry units from the segmentation mask image, and provide basic data for further size measurement and classification. Among them, the 0th order moment is used to obtain the area of ​​the masonry unit, the 1st order moment is used to calculate the coordinates of the centroid of the masonry unit, and the 2nd order moment is used to describe the direction and shape characteristics of the masonry unit, such as the direction of the major axis and the minor axis.

[0016] Further, The area share is:

[0017] in, for The area share of for The area of , ; and are the width and height of the detected image respectively.

[0018] Further, obtain The minimum enclosing rectangle method includes: S4.1. Get the bracket Convex hull of all pixels in the image and obtain all convex hull points of the convex hull, rotate all edges of the convex hull to the coordinate axis in sequence, and obtain the rotation coordinates of all convex hull points on the convex hull during each rotation; S4.2. Obtain the four corner points of the convex hull in each rotation process and obtain a circumscribed rectangle through the coordinate extreme values ​​of the rotation coordinates of all convex hull points in each rotation process; S4.3. Calculate the areas of all bounding rectangles obtained during the rotation process, and take the bounding rectangle with the smallest area as The minimum enclosing rectangle of S4.4, The aspect ratio of the minimum bounding rectangle is:

[0019] in, is the aspect ratio of the minimum enclosing rectangle, which is aspect ratio, is the width of the minimum enclosing rectangle, is the height of the minimum enclosing rectangle.

[0020] Further, obtain The convex hull methods include: S4.11. All the points in the pairwise construction straight line, ,in, is the number of combinations, for the number of midpoints; S4.12, determine the remaining Are all the points on the same side of the line? If so, the line is used as one side of the convex hull, otherwise, no action is taken.

[0021] S4.13, connect all the straight lines that are one side of the convex hull from beginning to end to obtain The convex hull of .

[0022] Furthermore, the calculation method of masonry distribution uniformity is: ,

[0023] in, is the uniformity of masonry distribution; is the total number of masonry categories in all masonry units, For the The ratio of the number of masonry units to the total number of masonry units in the class, is a natural number. When the MUI value is closer to 1, it means that the masonry types in the detected image are evenly distributed, otherwise, the distribution is less uniform; Furthermore, the expression for calculating the shape irregularity parameter of each masonry unit is: ,

[0024] in, For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; For the The number of masonry-like materials; is the total number of masonry units; is the total number of masonry categories in all masonry units; and All are natural numbers.

[0025] Furthermore, the expression for calculating the characterization parameter of the irregular wall shape is:

[0026] in, Characterize the parameters for the irregular shape of the wall; For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​a masonry unit.

[0027] Furthermore, the calculation method for calculating the masonry convergence of masonry is: ，

[0028] in, is the masonry aggregation; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; and are the width and height of the detected image respectively.

[0029] The present invention discloses a method for detecting masonry parameters of a brick structure wall based on image recognition, which has the following beneficial effects: 1. The present invention calculates the uniformity of masonry distribution, shape irregularity parameters and convergence, and quickly detects and evaluates masonry and walls, reflecting the current status of masonry and walls, and solves the limitations of the current traditional brick structure walls based on manual detection. It not only improves the detection efficiency, but also realizes the comprehensive and objective analysis of the feasibility of masonry detection of brick structure walls, and ensures the reliability and authenticity of the masonry detection and analysis results of brick structure walls.

[0030] 2. The present invention has high recognition accuracy and efficiency. Compared with the traditional method of manually detecting brick-built wall structures, the present invention can identify brick-built wall areas under complex backgrounds, and realize accurate extraction of masonry parameters of brick-built structures under images of different unit sizes. It can improve the accuracy of detection data and avoid errors caused by human factors or environmental factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of a method for detecting masonry parameters of a brick structure wall based on image recognition; Figure 2 For multiple detection images; Figure 3 is the segmentation mask image; Figure 4 is the segmentation mask image after binary processing; Figure 5 Schematic diagram of the extracted contour edge; Figure 6 is the minimum enclosing rectangle of multiple masonry units. DETAILED DESCRIPTION

[0032] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0033] refer to Figure 1 This embodiment provides a method for detecting masonry parameters of a brick structure wall based on image recognition, comprising the steps of: S1. Obtain a detection image of a brick wall and pre-process the detection image. In this embodiment, a mobile phone camera is directly used to take a picture of a brick wall under natural light conditions. Some detection images are referenced to Figure 2 In addition, this embodiment uses the PIL image processing library to modify the image resolution of all detection images, obtains a masonry wall image with the same resolution of 1000*1500, and denoises the image through Gaussian filtering to reduce the impact of noise.

[0034] S2. Input the preprocessed detection image into the SAM model to generate a segmentation mask image that identifies the masonry area. Figure 3 .

[0035] S3, the segmentation mask image is processed by binarization, opening operation and closing operation in sequence. Specifically, the foreground and background of the input binary image are made clear. In this embodiment, the foreground of the binary image is recorded as 1, which is bricks and the background is recorded as 0, which is the background. The effect of removing the noise in the background is obtained by first performing an erosion operation through an opening operation and then performing an expansion operation. The effect is shown in the figure. Figure 4 shown.

[0036] Step S2 is to obtain the outline of the masonry unit by performing edge detection on the processed segmentation mask image, such as Figure 5 As shown, the number of the masonry unit and the corresponding pixel coordinates are obtained by reading the position data of all pixels in the processed segmentation mask image.

[0037] S4. According to the outline and pixel coordinates of each masonry unit, the geometric moment, perimeter and minimum enclosing rectangle of the corresponding masonry unit are calculated, and the area of ​​the corresponding masonry unit is obtained through the geometric moment. Figure 6 shown.

[0038] In this embodiment, Masonry units are recorded as , The geometric moment of is:

[0039] in, For the Masonry units Order moment, and are all non-negative integers, is a natural number; and are the width and height of the detected image respectively; and Respectively The horizontal coordinate set and vertical coordinate set of the edge pixel points of the masonry unit, and They are the position index variables of the detected image in the width and height directions respectively. The coordinates are Gray value of pixel point. Geometric moment is used to describe the mathematical characteristics of the shape of an object. It can extract the precise position and morphological information of masonry units from the segmentation mask image, and provide basic data for further size measurement and classification. Among them, the 0th order moment is used to obtain the area of ​​the masonry unit, the 1st order moment is used to calculate the coordinates of the centroid of the masonry unit, and the 2nd order moment is used to describe the direction and shape characteristics of the masonry unit, such as the direction of the major axis and the minor axis.

[0040] S5. Calculate the area ratio of each masonry unit to the detection image according to the area of ​​each masonry unit.

[0041] The area share is:

[0042] in, for The area share of for The area of , ; and are the width and height of the detected image respectively.

[0043] Each masonry unit is classified by its area ratio and the aspect ratio of the minimum circumscribed rectangle, and the masonry distribution uniformity is calculated. The masonry distribution uniformity reflects whether the masonry units are evenly distributed in the entire wall structure. By controlling the masonry distribution uniformity, local over-density or over-sparseness can be avoided, thereby ensuring the consistency of the wall.

[0044] The expression for calculating the uniformity of masonry distribution is: ,

[0045] in, is the uniformity of masonry distribution; is the total number of masonry categories in all masonry units, For the The ratio of the number of masonry units to the total number of masonry units in the class, is a natural number. When the MUI value is closer to 1, it means that the masonry types in the detected image are evenly distributed, otherwise, the distribution is less uniform.

[0046] S6. Calculate the shape irregularity parameter of each masonry unit based on the perimeter and classification information of each masonry unit, and perform weighted processing on the shape irregularity parameter of each masonry unit to obtain the wall shape irregularity characterization parameter. The shape irregularity parameter reflects the edge complexity and shape regularity of a single masonry unit. It measures the degree to which the masonry unit is close to a circle or square through the relationship between the perimeter and the area, which can help identify those masonry units whose shapes deviate significantly from the expected standard, indicating manufacturing defects or construction problems. The wall shape irregularity characterization parameter can evaluate the complexity of the internal structure of the wall, which is very important for predicting the mechanical behavior and long-term performance of the wall. Complex shapes may lead to stress concentration and affect the stability and durability of the wall.

[0047] In this embodiment, the calculation classification is Class Brick The expression for the shape irregularity parameter of a masonry unit is: ,

[0048] in, For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; For the The number of masonry-like materials; is the total number of masonry units; is the total number of masonry categories in all masonry units; and All are natural numbers.

[0049] In this embodiment, the expression for calculating the parameter representing the irregular shape of the wall is:

[0050] in, Characterize the parameters for the irregular shape of the wall; For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​a masonry unit.

[0051] S7. Calculate the masonry convergence based on the area, perimeter, and pixel coordinates of each masonry unit. Masonry convergence can quantify whether the masonry units tend to cluster together or disperse evenly. By identifying whether there are abnormal clustering or dispersion phenomena in certain areas, potential construction problems or design defects are indicated.

[0052] In this embodiment, the expression for calculating the masonry convergence of masonry is: ,

[0053] in, is the masonry aggregation; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; and are the width and height of the detected image respectively. When , the masonry is in a completely dispersed state; When , the masonry is randomly distributed; When , the bricks and stones are in a clustered state.

[0054] As a further solution of this embodiment, obtain The minimum enclosing rectangle method includes: S4.1. Get the bracket The convex hull of all pixel points is obtained, and all the convex hull points are obtained. All the edges of the convex hull are rotated to the coordinate axis in turn, and the rotation coordinates of all the convex hull points on the convex hull are obtained once during each rotation.

[0055] Get The convex hull methods include: S4.11. All the points in the pairwise construction straight line, ,in, is the number of combinations, for The number of midpoints.

[0056] S4.12, determine the remaining Are all the points on the same side of the line? If so, the line is used as one side of the convex hull, otherwise, no action is taken.

[0057] S4.13, connect all the straight lines that are one side of the convex hull from beginning to end to obtain The convex hull of .

[0058] S4.2. Obtain the four corner points of the convex hull in each rotation process and obtain a circumscribed rectangle through the coordinate extreme values ​​of the rotation coordinates of all convex hull points in each rotation process.

[0059] S4.3. Calculate the areas of all bounding rectangles obtained during the rotation process, and take the bounding rectangle with the smallest area as The minimum enclosing rectangle of .

[0060] S4.4. Calculate the width and height of the smallest circumscribed rectangle, which is the width and height of the irregular-shaped masonry, and calculate the aspect ratio:

[0061] in, for The aspect ratio of the minimum bounding rectangle of for The width of the minimum enclosing rectangle of for The height of the minimum bounding rectangle.

[0062] In summary, the beneficial effects of this solution are: This solution classifies each masonry unit by combining the aspect ratio of its minimum circumscribed rectangle and its area ratio, which improves the accuracy of masonry unit classification compared to the method of classifying masonry units only by area ratio. By calculating the uniformity of masonry distribution, shape irregularity parameters and convergence, the masonry and walls are quickly detected and evaluated, reflecting the current status of masonry and walls, solving the limitations of the current traditional manual detection of brick-built wall masonry, not only improving the detection efficiency, but also realizing the comprehensive and objective analysis of the feasibility of brick-built wall masonry detection, and ensuring the reliability and authenticity of the brick-built wall masonry detection analysis results.

[0063] Although the specific implementation of the invention is described in detail in conjunction with the drawings, it should not be understood as limiting the scope of protection of this patent. Within the scope described in the claims, various modifications and variations that can be made by those skilled in the art without creative work still fall within the scope of protection of this patent.

Claims

1. A method for detecting masonry parameters of a brick structure wall based on image recognition, characterized in that: Includes steps: S1, obtaining a detection image of a brick wall and preprocessing the detection image; S2, input the preprocessed detection image into the SAM model to generate a segmentation mask image that identifies the masonry area; S3, binarizing, opening and closing the segmentation mask image in sequence, and performing edge detection on the processed segmentation mask image to obtain the outline of the masonry unit; reading the position data of all pixels in the processed segmentation mask image to obtain the number of the masonry unit and the corresponding pixel coordinates; S4, according to the outline and pixel coordinates of each masonry unit, calculate the geometric moment, perimeter and minimum circumscribed rectangle of the corresponding masonry unit, and obtain the area of ​​the corresponding masonry unit through the geometric moment; S5. According to the area of ​​each masonry unit, the area ratio of each masonry unit to the detection image is calculated, each masonry unit is classified according to the area ratio of each masonry unit and the aspect ratio of the minimum circumscribed rectangle, and the masonry distribution uniformity is calculated; S6. Calculate the shape irregularity parameter of each masonry unit according to the perimeter and classification information of each masonry unit, and perform weighted processing on the shape irregularity parameter of each masonry unit to obtain the wall shape irregularity characterization parameter; S7. Calculate the masonry convergence according to the area, perimeter and pixel coordinates of each masonry unit.

2. The detection method according to claim 1, characterized in that: The preprocessing of the detection images includes: using the PIL image processing library to unify the image resolution of all detection images, and performing denoising on each detection image through Gaussian filtering.

3. The detection method according to claim 1, characterized in that: No. Masonry units are recorded as , The geometric moment of is: in, For the Masonry units Order moment, and are all non-negative integers, is a natural number; and are the width and height of the detected image respectively; and Respectively The horizontal coordinate set and vertical coordinate set of the edge pixel points of the masonry unit, and They are the position index variables of the detected image in the width and height directions respectively. The coordinates are Grayscale value of the pixel.

4. The detection method according to claim 3, characterized in that: The area share is: in, for The area share of for The area of , ; and are the width and height of the detected image respectively.

5. The detection method according to claim 3, characterized in that: Get The minimum enclosing rectangle methods include: S4.

1. Get bracketing Convex hull of all pixels in the image and obtain all convex hull points of the convex hull, rotate all edges of the convex hull to the coordinate axis in sequence, and obtain the rotation coordinates of all convex hull points on the convex hull during each rotation; S4.

2. Obtain the four corner points of the convex hull in each rotation process and obtain a circumscribed rectangle through the coordinate extreme values ​​of the rotation coordinates of all convex hull points in each rotation process; S4.

3. Calculate the areas of all bounding rectangles obtained during the rotation process, and take the bounding rectangle with the smallest area as The minimum enclosing rectangle of S4.4, The aspect ratio of the minimum bounding rectangle is: in, is the aspect ratio of the minimum bounding rectangle, is the width of the minimum enclosing rectangle, is the height of the minimum enclosing rectangle.

6. The detection method according to claim 5, characterized in that: Get The convex hull methods include: S4.

11. All the points in the pairwise construction straight line, ,in, is the number of combinations, for the number of midpoints; S4.12, determine the remaining Are all the points on the same side of the line? If so, the line is used as one side of the convex hull, otherwise, no action is taken. S4.13, connect all the straight lines that are one side of the convex hull from beginning to end to obtain The convex hull of .

7. According to the detection method of claim 1, the calculation method of masonry distribution uniformity is: ， in, is the uniformity of masonry distribution; is the total number of masonry categories in all masonry units, For the The ratio of the number of masonry units to the total number of masonry units in the class, is a natural number.

8. The detection method according to claim 1, characterized in that: The expression for calculating the shape irregularity parameter of each masonry unit is: ， in, For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; For the The number of masonry-like materials; is the total number of masonry units; is the total number of masonry categories in all masonry units; and All are natural numbers.

9. The detection method according to claim 1, characterized in that: The expression for calculating the characterization parameters of irregular wall shape is: in, Characterize the parameters for the irregular shape of the wall; For the Class Brick The shape irregularity parameter of each masonry unit; For the Class Brick The area of ​​masonry units; For the The number of masonry types, is the total number of masonry categories in all masonry units.

10. The detection method according to claim 1, characterized in that: The expression for calculating the masonry convergence of each type of masonry unit is: ， in, is the masonry aggregation; For the Class Brick The area of ​​masonry units; For the Class Brick The perimeter of a masonry unit; and are the width and height of the detected image respectively; For the The number of masonry types, is the total number of masonry categories in all masonry units.

Images