Massive ground penetrating radar disease data three-dimensional synthesis method

By expanding the ground penetrating radar disease data into a cube and computing it, the problems of low efficiency of three-dimensional synthesis and loss of depth information in the prior art are solved, and efficient and accurate disease data synthesis is achieved.

CN118962668APending Publication Date: 2024-11-15ANHUI GUIMU ROBOT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410960349.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

When processing massive ground-penetrating radar disease data, the three-dimensional synthesis efficiency is low, the DBSCAN clustering speed is slow, and the disease depth information cannot be retained, resulting in inaccurate analysis of the results.

Method used

By expanding the disease results detected on the profile into cubes, the intersection ratio between each pair of cubes is calculated, and the diseases with the same index are extracted for projection, the synthetic data is obtained, which improves the calculation efficiency and the accuracy of the results.

Benefits of technology

Compared with DBSCAN clustering, the computing efficiency is improved, and the depth information of the disease is retained, which improves the position accuracy of the synthesis results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118962668A_ABST
    Figure CN118962668A_ABST
Patent Text Reader

Abstract

The invention discloses a mass ground penetrating radar disease data three-dimensional synthesis method, which comprises the following steps: S1, acquiring all diseases needing to be synthesized and corresponding position information, and traversing all the diseases to generate corresponding cubes; s2, the intersection-to-union ratio of different diseases in the three-dimensional space and the minimum vertex coordinate and the maximum vertex coordinate of each cube are calculated; s3, calculating the intersection of the bounding boxes of the two groups of cubes and the size of the bounding box of the intersection; s4, calculating the volumes of the two groups of cubes, and calculating the intersection-to-union ratio ioU between each pair of cubes; and S5, removing ioU diagonal elements, initializing an array kep, extracting diseases with the same index in the kep, and calculating the projection of the diseases on the map to obtain synthesized data. A disease result detected on a section is expanded into a cube, then spatial intersection calculation is carried out, compared with DBSCAN clustering, the calculation efficiency is improved, and compared with a geographic space calculation method, the method has depth information of the disease, and the accuracy of a synthesis result position is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of road damage detection, and in particular to a three-dimensional synthesis method for massive ground penetrating radar damage data. Background Art

[0002] Road safety has always been a top priority for social and economic development and has promoted economic development. However, due to the fact that roads are easily worn out and maintenance and management technology is limited, safety accidents caused by road diseases are emerging in an endless stream, seriously affecting the safety of people's lives and property.

[0003] 3D ground penetrating radar technology uses array antenna multi-channel high-frequency antenna scanning. It is convenient for data collection and can realize efficient collection of the three-dimensional space of the road. It can also obtain massive three-dimensional data, realize high-precision positioning of abnormal disease bodies, accurately restore the location of underground diseases, and has extremely high detection accuracy and spatial resolution. However, the amount of data collected by 3D ground penetrating radar is large. In essence, it is manual detection and annotation of diseases, or through the analysis of the slice data of the 3D radar, and then locate the profile of a certain channel for verification, which takes a lot of time; intelligent detection can also be introduced into disease detection to greatly improve the efficiency of data analysis, but intelligent detection only detects radar profiles and cannot well show the precise location of the disease and the size in three-dimensional space.

[0004] The existing method for intelligent detection of ground penetrating radar data is to use DBSCAN clustering to perform three-dimensional synthesis. Due to the huge number of diseases after detection, the DBSCAN clustering speed is extremely slow, and the disease synthesis cannot be quickly analyzed after detection. In addition, the disease results are unevenly distributed. DBSCAN clustering needs to adjust parameters to achieve a certain effect, and different parameters need to be adjusted for different disease distributions, and it is impossible to achieve intelligent synthesis in one go. Another method is to use geographic information for spatial calculation, because in the process of ground penetrating radar data collection, each data has its own spatial position, and the projection of the detection results on the map is used for spatial intersection calculation. This method is more efficient and is not affected by parameters, but the depth information of the disease itself will be lost after projection, and the results of disease synthesis cannot be located in space. After long-term research by the inventor, a three-dimensional synthesis method for massive ground penetrating radar disease data is proposed. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a three-dimensional synthesis method for massive ground penetrating radar disease data.

[0006] The object of the present invention is achieved by the following technical solution: A method for three-dimensional synthesis of massive ground penetrating radar disease data, comprising the following steps:

[0007] S1: Obtain all diseases to be synthesized and their corresponding location information, and traverse all diseases to generate their corresponding cubes;

[0008] S2: Calculate the intersection and union ratios between different diseases in three-dimensional space and the minimum and maximum vertex coordinates of each cube;

[0009] S3: Calculate the intersection of the two sets of cube bounding boxes and the size of the intersection bounding box;

[0010] S4: Calculate the volumes of the two groups of cubes and calculate the intersection-and-union ratio ioU between each pair of cubes;

[0011] S5: Remove the diagonal elements of ioU and initialize the array keep, extract the diseases with the same index in keep and calculate their projections on the map to obtain the synthesized data.

[0012] Preferably, step S1 further includes the following steps:

[0013] S11: The obtained single disease is a rectangular frame, and the height of the disease frame is set to h, the coordinates of the upper right corner are (x1, y1), and the coordinates of the upper left corner are (x2, y2);

[0014] S12: Calculate the disease direction of the disease frame,

[0015] angle=arctan((y2-y1) / (x2-x1));

[0016] S13: traverse all diseases and generate corresponding cubes. Each disease cube is represented by an array of 1 row and 6 columns, recorded as [x, y, z, l, w, h], where x, y, z respectively represent the position information of the center of the cube in space, and l, w, h respectively represent the length, width and height of the cube;

[0017] S14: Obtain a two-dimensional matrix with n rows and 6 columns, where n is the number of diseases. The two-dimensional matrix represents information of all diseases to be synthesized, which is recorded as cubes.

[0018] Preferably, step S2 further includes the following steps:

[0019] S21: Extract the center coordinates of the cube from cubes, and take the first three columns of the matrix cubes to obtain the center coordinates of all cubes center = cubes[:,:,3];

[0020] S22: Extract the size of the cube from cubes, and take the last three columns of the matrix cubes to obtain the length, width and height matrix dim=cubes[:,3,:] of all cubes;

[0021] S23: Calculate the minimum vertex coordinates and maximum vertex coordinates of each cube,

[0022] min = center-dim / 2;

[0023] max=center+dim / 2.

[0024] Preferably, step S3 further includes the following steps:

[0025] S31: Calculate the intersection of the two sets of cube bounding boxes, change the shape of min from (N,3) to (N,1,3), calculate the maximum value of the minimum coordinate point of each pair of cubes in the two 3D cube sets, and change the shape of max from (N,3) to (N,1,3), calculate the minimum value of the maximum coordinate point of each pair of cubes in the two 3D cube sets,

[0026] min=maximum(min[:,None,:],min[None,:,:]);

[0027] max=minimum(max[:,None,:],max[None,:,:]);

[0028] Among them, min is the minimum vertex of each cube bounding box, max is the maximum vertex of each cube bounding box, maximum means comparing the sizes of each element one by one and retaining the larger one, and minimum means comparing the sizes of each element one by one and retaining the smaller one;

[0029] S32: Calculate the size of the intersection bounding box, calculate the dimensions of the intersection area of ​​the two 3D cubes,

[0030] dim=max-min;

[0031] Among them, maximum∈(dim,0),

[0032] Calculate the product of the dimensions of each intersection bounding box to get the intersection volume,

[0033] dim=maximum(dim,0)[:,0]*maximum(dim,1)[:,0]*maximum(dim,0)[:,2].

[0034] Preferably, step S4 further includes the following steps:

[0035] S41: Calculate the volumes of two sets of cubes,

[0036] vol=dim[:,0]*dim[:,1]*dim[:,2];

[0037] Among them, dim[:,0] is the length of the rectangular parallelepiped obtained by taking the first column of dim, dim[:,1] is the width of the rectangular parallelepiped obtained by taking the first column of dim, and dim[:,2] is the height of the rectangular parallelepiped obtained by taking the first column of dim;

[0038] S42: Calculate the intersection-over-union ratio ioU,

[0039] iou=dim / (vol[:,None]+vol[None,:]-dim);

[0040] Among them, iou is the ratio of the intersection volume divided by the union volume, dim is the intersection volume, and vol[:,None]+vol[None,:] is the union volume.

[0041] Preferably, step S5 further includes the following steps:

[0042] S51: In the iou matrix, the value on the diagonal represents the iou between itself and itself and is 1. All iou minus the unit matrix of the same size as itself,

[0043] iou=iou-I n ;

[0044] Elements with an intersection-union ratio greater than 0 are marked as 1, indicating intersection;

[0045] S52: Initialize keep,

[0046] keep = {0, 1, 2, ..., n-1};

[0047] Get the disease index set I that intersects with di i ,

[0048] I i ={j|iou ij =1};

[0049] When I i When it is not empty, add di to the set.

[0050] I i =I i ∪{i};

[0051] Set the group numbers of these line segments to the smallest index value, and update the group numbers of all segments intersecting with di to k.

[0052] k=min(keep[I i ]);

[0053]

[0054] S53: Initialize the diseased area set,

[0055] DiseaseAreas = φ;

[0056] For each unique group number k,

[0057] I k ={i|keep[i]==k};

[0058] Among them, I k It is the set of disease segment indexes of the same group k.

[0059] The present invention has the following advantages: the present invention expands the disease results detected on the section into cubes, performs spatial intersection calculations, obtains the intersection-and-union ratio (ioU) between each pair of cubes, extracts the diseases with the same index in the array and calculates their projections on the map to obtain synthesized data. Compared with DBSCAN clustering, the present invention improves calculation efficiency and engineering practicability. Compared with the geographic space calculation method, the present method has the depth information of the disease and improves the accuracy of the location of the synthesis result. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a structural diagram of the process of the three-dimensional synthesis method for massive ground penetrating radar disease data. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0062] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0064] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0065] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use, or the positions or positional relationships commonly understood by those skilled in the art, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0066] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0067] In this embodiment, if Figure 1 As shown, a method for 3D synthesis of massive ground penetrating radar disease data includes the following steps:

[0068] S1: Obtain all diseases to be synthesized and their corresponding location information, and traverse all diseases to generate their corresponding cubes;

[0069] S2: Calculate the intersection and union ratios between different diseases in three-dimensional space and the minimum and maximum vertex coordinates of each cube;

[0070] S3: Calculate the intersection of the two sets of cube bounding boxes and the size of the intersection bounding box;

[0071] S4: Calculate the volumes of the two groups of cubes and calculate the intersection-and-union ratio ioU between each pair of cubes;

[0072] S5: Remove the diagonal elements of ioU and initialize the array keep, extract the diseases with the same index in keep and calculate their projections on the map to obtain the synthesized data. By expanding the disease results detected on the profile into cubes, and then performing spatial intersection calculations, the intersection ratio ioU between each pair of cubes is obtained, and the diseases with the same index in the array are extracted and their projections on the map are calculated to obtain the synthesized data. Compared with DBSCAN clustering, this method improves the calculation efficiency and engineering practicability. Compared with the geospatial calculation method, this method has the depth information of the disease and improves the accuracy of the location of the synthetic results.

[0073] Furthermore, step S1 further includes the following steps:

[0074] S11: The obtained single disease is a rectangular frame, and the height of the disease frame is set to h, the coordinates of the upper right corner are (x1, y1), and the coordinates of the upper left corner are (x2, y2);

[0075] S12: Calculate the disease direction of the disease frame,

[0076] angle=arctan((y2-y1) / (x2-x1));

[0077] S13: traverse all diseases and generate corresponding cubes. Each disease cube is represented by an array of 1 row and 6 columns, recorded as [x, y, z, l, w, h], where x, y, z respectively represent the position information of the center of the cube in space, and l, w, h respectively represent the length, width and height of the cube;

[0078] S14: Obtain a two-dimensional matrix of n rows and 6 columns, where n is the number of defects. The two-dimensional matrix represents information of all defects to be synthesized, which is recorded as cubes. Specifically, since there is a fixed spacing between the three-dimensional ground penetrating radar channels, when the defects are synthesized in three dimensions, it is necessary to increase a certain thickness of the rectangular frame in the direction of the ground penetrating radar channel so that the rectangular frame becomes a cube, that is, each cube represents a defect.

[0079] In this embodiment, step S2 further includes the following steps:

[0080] S21: Extract the center coordinates of the cube from cubes, and take the first three columns of the matrix cubes to obtain the center coordinates of all cubes center = cubes[:,:,3];

[0081] S22: Extract the size of the cube from cubes, and take the last three columns of the matrix cubes to obtain the length, width and height matrix dim=cubes[:,3,:] of all cubes;

[0082] S23: Calculate the minimum vertex coordinates and maximum vertex coordinates of each cube,

[0083] min = center-dim / 2;

[0084] max=center+dim / 2.

[0085] Furthermore, step S3 further includes the following steps:

[0086] S31: Calculate the intersection of the two sets of cube bounding boxes, change the shape of min from (N,3) to (N,1,3), calculate the maximum value of the minimum coordinate point of each pair of cubes in the two 3D cube sets, and change the shape of max from (N,3) to (N,1,3), calculate the minimum value of the maximum coordinate point of each pair of cubes in the two 3D cube sets,

[0087] min=maximum(min[:,None,:],min[None,:,:]);

[0088] max=minimum(max[:,None,:],max[None,:,:]);

[0089] Among them, min is the minimum vertex of each cube bounding box, max is the maximum vertex of each cube bounding box, maximum means comparing the sizes of each element one by one and retaining the larger one, and minimum means comparing the sizes of each element one by one and retaining the smaller one;

[0090] S32: Calculate the size of the intersection bounding box, calculate the dimensions of the intersection area of ​​the two 3D cubes,

[0091] dim=max-min;

[0092] Among them, maximum∈(dim,0) ensures that the size of the bounding box will not be negative. When two cubes do not overlap, the size will be truncated to zero;

[0093] Calculate the product of the dimensions of each intersection bounding box to get the intersection volume,

[0094] dim=maximum(dim,0)[:,0]*maximum(dim,1)[:,0]*aximum(dim,0)[:,2].

[0095] In this embodiment, step S4 further includes the following steps:

[0096] S41: Calculate the volumes of two sets of cubes,

[0097] vol=dim[:,0]*dim[:,1]*dim[:,2];

[0098] Among them, dim[:,0] is the length of the rectangular parallelepiped obtained by taking the first column of dim, dim[:,1] is the width of the rectangular parallelepiped obtained by taking the first column of dim, and dim[:,2] is the height of the rectangular parallelepiped obtained by taking the first column of dim;

[0099] S42: Calculate the intersection-over-union ratio ioU,

[0100] iou=dim / (vol[:,None]+vol[None,:]-dim);

[0101] Among them, iou is the ratio of the intersection volume divided by the union volume, dim is the intersection volume, and vol[:,None]+vol[None,:] is the union volume. Specifically, the ioU between each pair of cubes is calculated, and the result shape is (N,N), where N is the number of cubes.

[0102] In this embodiment, step S5 further includes the following steps:

[0103] S51: In the iou matrix, the value on the diagonal represents the iou between itself and itself and is 1. All iou minus the unit matrix of the same size as itself,

[0104] iou=iou-I n ;

[0105] Elements with an intersection-union ratio greater than 0 are marked as 1, indicating intersection, that is, these diseases are intersecting, ious = (ious>0);

[0106] S52: Initialize keep with a length of N to record the merging of diseases. Traverse each disease and find the disease index that intersects with itself through the place where ious is 1. If it exists, merge their indexes, update the keep array, and mark the intersecting diseases as the same group (using the smallest index). Specifically:

[0107] keep = {0, 1, 2, ..., n-1};

[0108] Get the disease index set I that intersects with di i ,

[0109] I i ={j|iou ij =1};

[0110] When I i When it is not empty, add di to the set.

[0111] I i =I i ∪{i};

[0112] Set the group numbers of these line segments to the smallest index value, and update the group numbers of all segments intersecting with di to k.

[0113] k=min(keep[I i ]);

[0114]

[0115] S53: Initialize the diseased area set,

[0116] DiseasAreas = φ;

[0117] For each unique group number k,

[0118] I k ={i|keep[i]==k};

[0119] Among them, I k It is the set of disease segment indexes of the same group k.

[0120] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for 3D synthesis of massive ground penetrating radar disease data, characterized by: The following steps are involved: S1: Obtain all diseases to be synthesized and their corresponding location information, and traverse all diseases to generate their corresponding cubes; S2: Calculate the intersection and union ratios between different diseases in three-dimensional space and the minimum and maximum vertex coordinates of each cube; S3: Calculate the intersection of the two sets of cube bounding boxes and the size of the intersection bounding box; S4: Calculate the volumes of the two groups of cubes and calculate the intersection-and-union ratio ioU between each pair of cubes; S5: Remove the diagonal elements of ioU and initialize the array keep, extract the diseases with the same index in keep and calculate their projections on the map to obtain the synthesized data.

2. The method for 3D synthesis of massive ground penetrating radar damage data according to claim 1, characterized in that: The step S1 further includes the following steps: S11: The obtained single disease is a rectangular frame, and the height of the disease frame is set to h, the coordinates of the upper right corner are (x1, y1), and the coordinates of the upper left corner are (x2, y2); S12: Calculate the disease direction of the disease frame, angle=arctan((y2-y1) / (x2-x1)); S13: traverse all diseases and generate corresponding cubes. Each disease cube is represented by an array of 1 row and 6 columns, recorded as [x, y, z, l, w, h], where x, y, z respectively represent the position information of the center of the cube in space, and l, w, h respectively represent the length, width and height of the cube; S14: Obtain a two-dimensional matrix with n rows and 6 columns, where n is the number of diseases. The two-dimensional matrix represents information of all diseases to be synthesized, which is recorded as cubes.

3. The method for 3D synthesis of massive ground penetrating radar damage data according to claim 2, characterized in that: The step S2 further includes the following steps: S21: Extract the center coordinates of the cube from cubes, and take the first three columns of the matrix cubes to obtain the center coordinates of all cubes center = cubes[:,:,3]; S22: Extract the size of the cube from cubes, and take the last three columns of the matrix cubes to obtain the length, width and height matrix dim=cubes[:,3,:] of all cubes; S23: Calculate the minimum vertex coordinates and maximum vertex coordinates of each cube, min = center-dim / 2; max=center+dim / 2.

4. The method for 3D synthesis of massive ground penetrating radar damage data according to claim 3, characterized in that: The step S3 further includes the following steps: S31: Calculate the intersection of the two sets of cube bounding boxes, change the shape of mmin from (N, 3) to (N, 1, 3), calculate the maximum value of the minimum coordinate point of each pair of cubes in the two 3D cube sets, and change the shape of max from (N, 3) to (N, 1, 3), calculate the minimum value of the maximum coordinate point of each pair of cubes in the two 3D cube sets, mmin=maximum(mmin[:,None,:],mmin[None,:,:]); max=minimum(max[:,None,:],max[None,:,:]); Among them, mmin is the minimum vertex of each cube bounding box, max is the maximum vertex of each cube bounding box, maximum means comparing the sizes of each element one by one and retaining the larger one, and minimum means comparing the sizes of each element one by one and retaining the smaller one; S32: Calculate the size of the intersection bounding box, calculate the dimensions of the intersection area of ​​the two 3D cubes, dim=max--mmin; Among them, maximum∈(dimm,0), Calculate the product of the dimensions of each intersection bounding box to get the intersection volume, dim=maximum(dimm,0)[:,0]*maximum(dim,1)[:,0]*maximum(dimm,0)[:,2].

5. The method for 3D synthesis of massive ground penetrating radar damage data according to claim 4, characterized in that: The step S4 further includes the following steps: S41: Calculate the volumes of two sets of cubes, vol=dim[:,0]*dim[:,1]*dim[:,2]; Among them, dim[:,0] is the length of the rectangular parallelepiped obtained by taking the first column of dim, dim[:,1] is the width of the rectangular parallelepiped obtained by taking the first column of dim, and dim[:,2] is the height of the rectangular parallelepiped obtained by taking the first column of dim; S42: Calculate the intersection-over-union ratio ioU, iou=dim / (vol[:,None]+vol[None,:]-dim); Among them, iou is the ratio of the intersection volume divided by the union volume, dim is the intersection volume, and vol[:,None]+vol[None,:] is the union volume.

6. The method for 3D synthesis of massive ground penetrating radar damage data according to claim 5, characterized in that: The step S5 further includes the following steps: S51: In the iou matrix, the value on the diagonal represents the iou between itself and itself and is 1. All iou minus the unit matrix of the same size as itself, iou=iou-I n ; Elements with an intersection-union ratio greater than 0 are marked as 1, indicating intersection; S52: Initialize keep, keep={0,1,2,…,n-1}; Get the disease index set I that intersects with di i , I i ={j|iou ij =1}; When I i When it is not empty, add di to the set. I i =I i ∪{i}; Set the group numbers of these line segments to the minimum index value, and update the group numbers of all segments intersecting with di to k, k = min (keep [I i ]); S53: Initialize the diseased area set, DiseaseAreas = φ; For each unique group number k, I k ={i|keep[i]==k}; Among them, I k It is the set of disease segment indexes of the same group k.