Pipeline deformation identification method, terminal device and storage medium

Abstract

This invention discloses a pipeline deformation identification method, terminal device, and storage medium. Based on a reconstructed three-dimensional pipeline model, this method accurately calculates the major and minor diameters of the pipeline cross section by fitting a cross-sectional function model to a voxel coordinate system. This method then identifies pipeline deformation, significantly improving the accuracy of pipeline deformation identification. This invention addresses the problem of deep learning algorithms being unable to identify the overall deformation state of a pipeline. By analyzing multiple cross sections of the three-dimensional pipeline model, it achieves accurate quantitative analysis of the overall deformation state of the pipeline.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a pipeline deformation recognition method, terminal equipment and storage medium. Background Art

[0002] The urban drainage network is an important infrastructure and lifeline to ensure the operation of the city. Due to the influence of external force extrusion during the production process and operation of the pipeline, the drainage pipeline is prone to overall or regional deformation. If not handled in time, it will further cause pipeline rupture, water leakage, and reduced pipeline drainage capacity, thereby causing accidents such as poor urban drainage, urban waterlogging, environmental pollution, and ground collapse. The identification of drainage pipes requires a comprehensive quantitative analysis of multiple aspects such as the deformation rate of the pipe, the length of the pipe deformation, and the position of the pipe deformation. However, the detection video is composed of multiple frames of detection images, and it is difficult for the image to represent the three-dimensional spatial properties of the pipe. Therefore, at this stage, it is difficult to perform high-precision quantitative analysis of pipe deformation based on intelligent means such as machine vision and artificial intelligence. There are currently two main methods for the identification and rating of drainage pipe deformation:

[0003] 1) Manual interpretation. With extensive experience interpreting video data, urban drainage pipeline inspection professionals can estimate the deformation rate and length of the pipeline based on the shape, location, and size of the image defects, thereby deriving a manually assessed defect grade. This method is influenced by factors such as the professional's spatial imagination, experience, and pipeline environment, and is highly subjective and uncertain.

[0004] 2) Deep learning-based defect interpretation technology. Current deep learning-based drainage pipe defect rating mainly uses image recognition technology and image edge extraction technology to identify pipe deformation. This has low recognition accuracy, requires a large amount of data, and cannot determine the level of pipe deformation.

[0005] Invention patent application CN113763363A discloses a method for detecting and grading drainage pipe rupture defects. This method constructs a structural defect Mask R-CNN model to detect and segment defects in captured video or images, identify structural defects in drainage pipes, and output a rupture defect confidence level, a rupture defect bounding box, and a rupture defect mask. The method then calculates the rupture defect's characteristic parameters, rupture defect rating parameters, radial coverage ratio, and circumferential coverage range to grade the rupture defect. This method can only identify and grade pipeline rupture defects, but cannot accurately identify and grade pipeline deformation. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a pipeline deformation identification method, terminal equipment and storage medium to accurately identify pipeline deformation in view of the shortcomings of the existing technology.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is: a pipeline deformation identification method, comprising the following steps:

[0008] Reconstruct the pipeline 3D model:

[0009] Establish a voxel coordinate system, where the origin of the voxel coordinate system is the center point of the starting cross section of the pipeline, the positive direction of the y-axis is the extension direction of the pipeline perpendicular to the cross section, the positive direction of the x-axis is perpendicular to the y-axis and points to the right of the y-axis, and the positive direction of the z-axis is perpendicular to the xy plane and points upward; convert the points in the three-dimensional pipeline model into a voxel model in the voxel coordinate system; the voxel model includes multiple voxel grids, the pipeline cross section where the voxel coordinate system origin is located is the starting end of the voxel model, and the tail end of the voxel model is the end of the pipeline; each voxel grid has a unique and fixed coordinate value, and the voxel grid stores the point data in the reconstructed three-dimensional pipeline model;

[0010] At both ends of the voxel model, a plurality of pipeline sections are randomly intercepted, and the center points of all the intercepted pipeline sections are fitted to obtain the pipeline centerline;

[0011] Get the point on the pipeline centerline and the normal direction corresponding to the point, and obtain the voxel grid set corresponding to the point based on the normal direction, which is the voxel grid set of the pipeline section where the point is located;

[0012] Projecting the voxel set onto the xz plane, calculating the circumscribed polygon of the voxel set, and extracting the inflection points of the circumscribed polygon;

[0013] Use the inflection points of all circumscribed polygons to fit a closed curve, that is, the curve equation of the pipe section corresponding to the circumscribed polygon, and obtain the maximum and minimum values ​​of sqrt((x-x0)*(x-x0)+(z-z0)*(z-z0)), where x and z are the horizontal and vertical coordinates of the points on the closed curve, and x0 and z0 are the horizontal and vertical coordinates of the center point of the pipe section; the maximum and minimum values ​​represent the longest and shortest radii of the pipe section corresponding to the curve F(x,z)=0, respectively.

[0014] Based on the longest radius and the shortest radius, the ratio of the longest diameter and the shortest diameter of the pipeline section to the designed section diameter is calculated respectively, and the maximum value of the ratio is taken as the pipeline deformation rate.

[0015] Based on the reconstruction of a three-dimensional pipeline model, this invention accurately calculates the major and minor diameters of the pipeline cross section by fitting a cross-sectional function model to a voxel coordinate system. This allows for the identification of pipeline deformation, significantly improving the accuracy of deformation identification. The method of this invention eliminates the need for extensive data, significantly improving the efficiency of pipeline deformation identification. This invention addresses the problem of deep learning algorithms being unable to identify the overall deformation state of a pipeline. By analyzing multiple cross sections of the three-dimensional pipeline model, it achieves accurate quantitative analysis of the overall deformation state of the pipeline.

[0016] In the present invention, after obtaining the pipeline deformation rate, a pipeline deformation rating is also performed. The specific implementation process includes: if the pipeline deformation rate is less than 5%, it is classified as Level 1 deformation; if the pipeline deformation rate is greater than 5% but less than 15%, it is classified as Level 2 deformation; if the pipeline deformation rate is greater than 15% but less than 25%, it is classified as Level 3 deformation; and if the pipeline deformation rate is greater than 25%, it is classified as Level 4 deformation. The present invention quantitatively evaluates the pipeline deformation rate. By analyzing the pipeline deformation state, it facilitates the analysis of subsequent pipeline operation status and pipeline water flow capacity, optimizes pipeline usage, and improves urban flood control decision-making capabilities.

[0017] In the present invention, the formula for converting the point P (x, y, z) in the pipeline three-dimensional model to the voxel coordinate system is:

[0018]

[0019] Where I(x',y',z') is the coordinate of the point P(x,y,z) in the pipeline 3D model converted to the voxel coordinate system, c is the size of the voxel grid, and P represents the point coordinate matrix [P x ,P y ,P z ], P x 、P y 、P z is the actual coordinate value of P(x,y,z), A is the transformation matrix, x min 、y min 、z min It is the minimum value of the set of points in the three-dimensional pipeline model on the x, y, and z axes.

[0020] In the present invention, in order to simplify the calculation process and ensure the calculation accuracy, the longest diameter is twice the longest radius; the shortest diameter is twice the shortest radius.

[0021] In the present invention, the calculation formula of the pipeline deformation ratio is: l is the longest diameter or shortest diameter, and r is the designed cross-sectional diameter.

[0022] In the present invention, the calculation process of the longest radius and the shortest radius of the pipeline section includes: calculating the point set on the cross-sectional curve F(x,z)=0 corresponding to the horizontal coordinate x0 of the cross-sectional center point; when the horizontal coordinate x0 of the cross-sectional center point increases or decreases by an integer multiple K, obtaining a point on the cross-sectional curve F(x,z)=0 once to obtain all point sets that meet the conditions; substituting the point (x,z) in the point set into sqrt((x-x0)*(x-x0)+(z-z0)*(z-z0)), calculating the distance between the point in the point set and the cross-sectional center point, taking the maximum and minimum values ​​of the distances, and obtaining the longest radius and the shortest radius of the pipeline section.

[0023] As an inventive concept, the present invention also provides a terminal device, which includes a processor and a memory; the memory stores a computer program / instructions; the processor executes the computer program / instructions stored in the memory; the computer program / instructions are configured to implement the steps of the above-mentioned method of the present invention.

[0024] As an inventive concept, the present invention also provides a computer storage medium having a computer program / instruction stored thereon; characterized in that the computer program / instruction implements the steps of the above-mentioned method of the present invention when executed by a processor.

[0025] As an inventive concept, the present invention also provides a computer program product, including a computer program / instruction; characterized in that the computer program / instruction implements the steps of the above-mentioned method of the present invention when executed by a processor.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1) The present invention calculates the major and minor diameters of the pipeline cross section by fitting the cross-sectional function model, thereby realizing the identification and rating of pipeline deformation;

[0028] 2) This invention solves the problem that deep learning algorithms cannot identify the overall deformation state of the pipeline. By analyzing multiple sections of the pipeline 3D model, it achieves quantitative analysis of the overall deformation state of the pipeline;

[0029] 3) The present invention significantly improves the accuracy of pipeline deformation analysis and reduces the subjectivity and uncertainty of identification results. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is the deformation level identification and interpretation process of an embodiment of the present invention;

[0031] Figure 2 This is a pipeline cross-section extraction diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The specific implementation steps of the embodiment of the present invention are as follows:

[0033] 1. Reconstruct the 3D pipeline model based on the structure-in-motion (SFM) algorithm. The main steps are as follows:

[0034] 1) Use the OpenCV library to extract frames from the video to obtain discrete pipeline images;

[0035] 2) The SFM algorithm (Structure from Motion) is used to analyze and calculate the discrete pipeline image to achieve 3D point cloud reconstruction of the pipeline;

[0036] 2. Couple voxels and Graham's 3D pipeline boundary fitting. The main steps are as follows:

[0037] 1) Establish a voxel model I, which is a set of voxels, including voxel coordinates and point cloud data within the voxel, so that the point P (x, y, z) meets the following conditions:

[0038] P(x,y,z)∈I(x',y',z')(1)

[0039] Where x', y', z' are the relative coordinates (integers) of the actual coordinates of the point cloud transformed into the coordinate system of the voxel model I, where the center point of the starting cross section of the pipeline is the origin, the positive direction of the y-axis is the extension direction of the pipeline perpendicular to the cross section, the positive direction of the x-axis is perpendicular to the y-axis and points to the right direction of the y-axis, and the positive direction of the z-axis is perpendicular to the xy plane; I(x', y', z') is a single voxel object used to store point cloud data that conforms to formula (1), and its transformation formula is shown in formula (2). Where c is the size of the voxel grid (i.e., voxel), and P is the point cloud point coordinate matrix [P x ,P y ,P z ], P x 、P y 、P z is the actual coordinate of the point cloud; A is the transformation matrix, x min 、y min 、z min The minimum value of the current point cloud set on the x, y, and z axes. If decimals appear during calculation, the decimal place is rounded off to get integer coordinates:

[0040]

[0041] In an embodiment of the present invention, the origin of the voxel coordinate system is the center point of the starting cross section of the pipeline, the positive direction of the y-axis is the extension direction of the pipeline perpendicular to the cross section, the positive direction of the x-axis is perpendicular to the y-axis and points to the right of the y-axis, and the positive direction of the z-axis is perpendicular to the xy plane and points upward. Points in the three-dimensional pipeline model are converted to a voxel model in the voxel coordinate system to form a voxel model. The voxel model includes multiple voxel grids. The pipeline section where the voxel coordinate system origin is located is the starting end of the voxel model, and the end of the voxel model is the end of the pipeline. Each voxel grid has a unique, fixed coordinate value, and the voxel grid stores the point data in the reconstructed three-dimensional pipeline model.

[0042] 2) Voxel pipeline centerline extraction: Multiple sections are randomly cut at both ends of the voxel model (for example, a cross section is cut every 10 meters from one side of the voxel model), and the center points of the sections are calculated as points on the centerline. Based on the multiple center points of the sections, support vector machine regression (SVR) is used to fit the pipeline centerline to achieve pipeline centerline extraction.

[0043] 3) 3D voxel model cross-section extraction: Select a point on the central axis, calculate the normal direction of the point on the central axis, and calculate the cross-section voxel set corresponding to the point based on the normal direction. For example, from the central axis from left to right, take a center point every 20 cm, calculate the intersection of the normal plane of the point and the voxel model, and obtain the voxel point set on the cross-section.

[0044] The following is an example of the deformation rate analysis of one section:

[0045] 4) Graham's 3D pipeline boundary fitting: Using the xz plane perpendicular projection algorithm, the cross-sectional voxel set is projected onto the xz plane; the convex hull of the voxel set is calculated using the Graham algorithm, and the inflection points of the convex hull (circumscribed polygon) are extracted. In this embodiment, the Graham algorithm (i.e., the convex hull algorithm) is used to calculate the minimum circumscribed polygon of the voxel set, and the vertices of the circumscribed polygon are extracted as inflection points (i.e., convex hull inflection points);

[0046] 5) Cross-section curve fitting: Based on the extracted convex hull inflection points, SVR is used to fit the cross-section curve polynomial equation F(x,z)=0 to achieve mathematical modeling of the cross section, such as Figure 2 As shown;

[0047] 6) Calculate the longest and shortest diameters of the cross section: F(x,z)=0 is the curve equation of the pipe cross section corresponding to the circumscribed polygon, and solve the maximum and minimum values ​​of sqrt((x-x0)*(x-x0)+(z-z0)*(z-z0)) under the condition that the closed curve F(x,z)=0 is satisfied. In the formula, sprt is the square root, x and z are points on the curve, and x0 and z0 are the center points of the cross section. Use the equal sampling method to solve the curve equation: first calculate the point set on the cross section curve F(x,z)=0 corresponding to the cross section center point x0; solve the point on the cross section curve F(x,z)=0 every time the x0 coordinate increases or decreases by an integer multiple of 0.1, and finally obtain all the point sets that meet the conditions; finally calculate the distance between the points in the point set and the center point of the cross section, and take the maximum and minimum values ​​of the distance to represent the longest radius and shortest radius of the cross section respectively.

[0048] 3. Deformation rate calculation, pipeline deformation identification and grade judgment:

[0049] 1) Based on the longest and shortest radii obtained in step 6) of step 2, calculate the ratio of the longest diameter (twice the longest radius) and the shortest diameter (twice the shortest radius) of the pipe section to the design section diameter, and take the maximum ratio as the pipe deformation rate. Where ratio is the deformation rate, l is the longest or shortest diameter, and r is the diameter:

[0050]

[0051] 2) Determine whether the pipeline has deformed and the degree of deformation based on the specifications. If the cross-sectional deformation rate is less than 5%, it is classified as Level 1 deformation; if the deformation rate is greater than 5% but less than 15%, it is classified as Level 2 deformation; if the deformation rate is greater than 15% but less than 25%, it is classified as Level 3 deformation; if the deformation rate is greater than 25%, it is classified as Level 4 deformation.

[0052] Generally, Level 1 deformation requires no treatment, Level 2 indicates the need for frequent inspection, and Levels 3 and 4 require repair. Analysis of deformation status facilitates subsequent analysis of pipeline operation and water flow capacity, optimizing pipeline usage and improving urban flood control decision-making.

[0053] Another embodiment of the present invention further provides a terminal device, comprising a processor and a memory; the memory storing a computer program / instructions. The processor executes the computer program / instructions stored in the memory; the computer program / instructions are configured to implement the steps of the method of the above embodiment of the present invention.

[0054] In the embodiment of the present invention, the processor may be a microcontroller or the like.

[0055] Another embodiment of the present invention further provides a computer storage medium having a computer program / instruction stored thereon; when the computer program / instruction is executed by a processor, the steps of the method of the above embodiment of the present invention are implemented.

[0056] In the embodiment of the present invention, the storage medium may be an optical disc or the like.

[0057] Another embodiment of the present invention further provides a computer program product, including a computer program / instruction; when the computer program / instruction is executed by a processor, the steps of the method of the above embodiment of the present invention are implemented.

Claims

1. A pipeline deformation identification method, characterized in that: The following steps are involved: Reconstructing the 3D model of the pipeline based on the structure-in-motion algorithm: Establish a voxel coordinate system, the origin of the voxel coordinate system is the center point of the starting cross section of the pipeline, the positive direction of the y-axis is the extension direction of the pipeline perpendicular to the cross section, the positive direction of the x-axis is perpendicular to the y-axis and points to the right direction of the y-axis, and the positive direction of the z-axis is perpendicular to the xy plane and points upward; convert the points in the three-dimensional pipeline model to the voxel model under the voxel coordinate system; the voxel model includes multiple voxel grids, the pipeline section where the voxel coordinate system origin is located is the starting end of the voxel model, and the end of the voxel model is the end of the pipeline; each voxel grid has a unique and fixed coordinate value, and the voxel grid stores the point data in the reconstructed three-dimensional pipeline model; the points in the three-dimensional pipeline model are converted into voxel grids. The formula for conversion to voxel coordinate system is: ； in, is a point in the 3D model of the pipeline Convert to the coordinates in the voxel coordinate system, c is the size of the voxel grid, P Represents the point coordinate matrix [ P x , P y , P z ], P x 、 P y 、 P z for The actual coordinate value of , A is the transformation matrix, x min 、 y min 、 z min It is the minimum value of the set of points in the three-dimensional pipeline model on the x, y, and z axes; At both ends of the voxel model, multiple pipeline sections are randomly intercepted, and the center points of all the intercepted pipeline sections are fitted using a support vector machine regression to obtain the pipeline centerline; Get the point on the pipeline centerline and the normal direction corresponding to the point, and obtain the voxel grid set corresponding to the point based on the normal direction, which is the voxel grid set of the pipeline section where the point is located; Projecting the voxel set onto the xz plane, calculating the circumscribed polygon of the voxel set, and extracting the inflection points of the circumscribed polygon; Use the inflection points of all circumscribed polygons to fit a closed curve, i.e., the curve equation of the pipe section corresponding to the circumscribed polygon, and obtain the maximum and minimum values ​​of sqrt((x-x0)* (x-x0) + (z-z0)* (z-z0)), where x and z are the abscissa and ordinate of the points on the closed curve, and x0 and z0 are the abscissa and ordinate of the center point of the pipe section. The maximum and minimum values ​​represent the longest and shortest radii of the pipe section corresponding to the curve F(x, z) = 0, respectively. Based on the longest radius and the shortest radius, the ratio of the longest diameter and the shortest diameter of the pipeline section to the designed section diameter is calculated respectively, and the maximum value of the ratio is taken as the pipeline deformation rate.

2. The pipeline deformation identification method according to claim 1, characterized in that: After obtaining the pipeline deformation rate, a pipeline deformation rating is also performed. The specific implementation process includes: if the pipeline deformation rate is less than 5%, it is level 1 deformation; if the pipeline deformation rate is greater than 5% and less than 15%, it is level 2 deformation; if the pipeline deformation rate is greater than 15% and less than 25%, it is level 3 deformation; if the pipeline deformation rate is greater than 25%, it is level 4 deformation.

3. The pipeline deformation identification method according to claim 1 or 2, characterized in that: The longest diameter is twice the longest radius; the shortest diameter is twice the shortest radius.

4. The pipeline deformation identification method according to claim 1 or 2, characterized in that: The pipeline deformation rate ratio The calculation formula is: , l is the longest or shortest diameter, r is the designed cross-sectional diameter.

5. The pipeline deformation identification method according to claim 1 or 2, characterized in that: The calculation process of the longest radius and the shortest radius of the pipeline section includes: calculating the point set on the cross-sectional curve F(x, z)=0 corresponding to the horizontal coordinate x0 of the cross-sectional center point; when the horizontal coordinate x0 of the cross-sectional center point increases or decreases by an integer multiple K, obtaining a point on the cross-sectional curve F(x, z)=0 once, and obtaining all point sets that meet the conditions; substituting the point (x, z) in the point set into sqrt((x-x0)* (x-x0) + (z-z0)* (z-z0)), calculating the distance between the point in the point set and the cross-sectional center point, taking the maximum and minimum values ​​of the distance, and obtaining the longest radius and the shortest radius of the pipeline section.

6. A terminal device, characterized in that: The invention comprises a processor and a memory; the memory stores a computer program / instruction; the processor executes the computer program / instruction stored in the memory; the computer program / instruction is configured to implement the steps of the method according to any one of claims 1 to 5.

7. A computer storage medium having a computer program / instruction stored thereon; characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer program product comprising a computer program / instructions; characterized in that When the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images