CN118392054B - Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement - Google Patents
Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement Download PDFInfo
- Publication number
- CN118392054B CN118392054B CN202410846294.3A CN202410846294A CN118392054B CN 118392054 B CN118392054 B CN 118392054B CN 202410846294 A CN202410846294 A CN 202410846294A CN 118392054 B CN118392054 B CN 118392054B
- Authority
- CN
- China
- Prior art keywords
- point
- points
- point cloud
- area
- neighborhood
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 84
- 238000005259 measurement Methods 0.000 title claims abstract description 28
- 230000011218 segmentation Effects 0.000 claims abstract description 10
- 238000012216 screening Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims description 30
- 239000013598 vector Substances 0.000 claims description 21
- 238000009826 distribution Methods 0.000 claims description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 5
- 238000003491 array Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000012847 principal component analysis method Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims 1
- 238000012545 processing Methods 0.000 description 9
- 238000000691 measurement method Methods 0.000 description 4
- 238000007792 addition Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000009683 ultrasonic thickness measurement Methods 0.000 description 2
- 238000009827 uniform distribution Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Landscapes
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The invention relates to a method for measuring thickness of a spaceflight wallboard based on point cloud data and Haoskov measurement, which comprises the following steps: the laser scanner scans the front and back sides of the wallboard respectively and performs point cloud pretreatment; calculating the local geometric characteristics of each point in the preprocessed point cloud, and dividing uniform and non-uniform areas of the surface of the point cloud; selecting initial seed points, and extracting a plurality of points in the area to be called growing inner points; checking and screening out points with consistent characteristics with the growing inner points from the rough segmentation area space, and forming an area together with the growing inner points, namely the areaRecorded as point cloud; Method for searching other surface area of wallboard by adopting octreeCalculate the areaArea of contactThe Haoskov distance between the two is obtained, and the minimum value is the thickness of the wallboard. According to the method, the nearest neighborhood point on the other surface is searched for calculating the distance by searching the neighborhood point through the octree, so that the thickness measurement efficiency of the aerospace wallboard is improved, and the accuracy of a measurement result is ensured.
Description
Technical Field
The invention relates to the technical field of computer vision and space wallboard detection, in particular to a space wallboard thickness measurement method based on point cloud data and Haosdorf measurement.
Background
In the aerospace engineering, the wallboard is used as an important component of a spacecraft structure, and accurate measurement of the thickness of the wallboard has important significance for guaranteeing the success of an aerospace task. The thickness of the aerospace panel is directly related to the structural strength, durability, and overall performance of the spacecraft. Therefore, in order to ensure safe operation of the spacecraft in extreme environments, accurate and rapid measurement of the thickness of the wallboard is essential. By accurately measuring the thickness of the wallboard, the abrasion, corrosion or other potential defects of the wallboard can be found in time, so that repair or replacement measures can be taken in time, and the reliability and safety of the spacecraft are ensured. Aiming at the thickness measurement requirement of the space wallboard, the traditional method mainly comprises ultrasonic thickness measurement, X-ray thickness measurement and other technologies. However, these conventional methods have many limitations and challenges. For example, ultrasonic thickness measurement requires direct contact with the surface of the object to be measured, and is not applicable to wall plates with complex structures; while X-ray thickness measurement is limited by radiation safety and cost issues, it is difficult to apply in large scale in practical aerospace engineering.
In recent years, with rapid development of laser scanning technology and three-dimensional point cloud processing algorithm, a thickness measurement method of a space wallboard based on three-dimensional point cloud data is attracting attention. According to the method, three-dimensional point cloud data of the surface of the wallboard are obtained through equipment such as a laser scanner, and then the thickness of the wallboard is rapidly and accurately measured by utilizing a point cloud data processing and analyzing algorithm. Although the point cloud-based method has many advantages such as simple operation, rich data, non-contact, etc., some challenges remain in practical applications. For example, the speed of a point cloud data processing algorithm may be limited by the size of the data volume and the complexity of the algorithm, requiring further optimization to improve measurement efficiency; in addition, because of the variety of shapes and surface characteristics of the aerospace wall panels, the analysis and processing algorithm of the point cloud data needs to have certain adaptability and universality so as to meet measurement requirements under different conditions. Therefore, how to overcome these difficulties and improve the accuracy, efficiency and application range of the space wallboard thickness measurement method based on the point cloud is still one of the important directions of current research.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a method for measuring the thickness of a space wallboard based on point cloud data and Haoskov measurement, which solves the problems of low efficiency and accuracy in measuring the thickness of the space wallboard in the prior art; according to the method, ideal seed points on the surface of the wallboard model are selected to obtain the upper surface and the lower surface of the wallboard, so that the accuracy of surface extraction is improved, and then the nearest neighbor points on the other surface are searched for calculating the distance by using an octree neighbor point searching method, so that the thickness measurement efficiency of the aerospace wallboard is remarkably improved, the accuracy of measurement results is ensured, and meanwhile, the wide applicability of the method in different measurement occasions is also shown.
In order to solve the technical problems, the invention provides the following technical scheme: a thickness measurement method of a space wallboard based on point cloud data and Haoskov measurement comprises the following steps:
s1, respectively scanning the front surface and the back surface of a wallboard by a laser scanner, performing rough registration operation to obtain complete point cloud data, and performing point cloud pretreatment;
s2, calculating the preprocessed point cloud Dividing uniform and non-uniform areas of the point cloud surface according to the local geometric features of each point in the cloud surface;
s3, selecting a surface with most uniform and continuous point distribution in a surface area space, selecting initial seed points in the area, and extracting a plurality of points in the surface area, namely growing inner points;
S4, checking and screening points with consistent characteristics with the growing internal points from the rough-divided area space, and forming an area together with the growing internal points, thereby extracting one surface, namely the area, of the wallboard model Recorded as point cloud；
S5, searching the other surface of the wallboard by adopting an octree methodI.e. areaAt the point cloudUpper part of the cylinderEach neighborhood pointCalculate the areaArea of contactThe Haoskov distance between the two is obtained, and the minimum value is the thickness of the wallboard.
Further, the specific process of step S1 includes the following steps:
S11, firstly manually removing block noise near the target point cloud, and then removing outlier data points by adopting a filtering algorithm;
S12, in the downsampling process, optimizing the point cloud data by adopting a voxel filter method, and creating a plurality of edge lengths on the whole point cloud data Inside each voxel grid we merge the originally dispersed points and use the centroid position of these points to represent the center point of the whole grid.
Further, the specific process of step S2 includes the following steps:
s21: point-to-point cloud Each point of (3)Find itSets of neighborhood pointsThe 3 angular coordinate vectors of the neighborhood points form a covariance matrixThen calculateMaximum characteristic value of (2)Covariance matrixThe following formula is shown:
；
In the method, in the process of the invention, Respectively represent a single pointA kind of electronic deviceRectangular coordinates of each neighborhood point in three-dimensional space, the coordinates are given in the form of one-dimensional arrays, and the length of each array isCorresponding to a pointA kind of electronic deviceCoordinates, covariance of nearest neighbor pointsRepresenting vector variablesAndCovariance between, which measuresAndThe degree of linear correlation between, and likewise,AndRespectively representAnd、AndCovariance between;
S22: numbering each area, performing rough segmentation to roughly determine the positions of the areas in a three-dimensional space, and reducing the search range of subsequent fine segmentation and feature analysis; subdividing each coarsely segmented face space into a plurality of three-dimensional cube grids, the grids having edges of all ；
S23: setting a proper threshold according to the maximum characteristic value calculation result of each pointIf the threshold value is exceeded, the point is considered to belong to a point with less uniform distribution;
s24: dividing the grids into uniform point grids and non-uniform point grids, and counting that the characteristic value in each grid exceeds a set threshold value If the ratio exceeds 80%, then defining the grid as a non-uniform grid of points; otherwise, the point-sharing grid is adopted, so that the area of the point cloud distribution is divided into uniform and non-uniform areas.
Further, in step S3, the specific process includes the following steps:
S31, calculating point cloud by adopting principal component analysis method Normal vector and curvature of each point in (a);
s32, respectively extracting Points in the uniform point grid in the individual area space constituteIndividual point of uniformity datasetAnd calculateCentroid point of (2)Coordinates, finding the homopoint datasetMid to centroid pointPoints of minimum euclidean distanceDefined asInitial seed points of the number area, so that uniform and continuous distribution of point clouds near the initial seed points is ensured;
s33, if the included angle between the neighborhood point of each seed point and the normal vector of the seed point is smaller than a certain angle threshold value The neighborhood point is considered as the point in the target area, if the curvature of the neighborhood point is smaller than a certain curvature threshold valueThen the neighborhood point is added to the set of seed points,Deleting the current seed point after the neighborhood point is judged, and entering the growth process of the next seed point;
S34, repeating the step S303 until the seed point set is empty.
Further, the specific process of step S4 includes the following steps:
s41, adding the growing inner points to the inner point set as initial points In (a) and (b);
S42, appointing Points other than the growing interior points in the individual area space constitute an alternative datasetSequentially pop upIn (2) point verification consistency conditions forFitting a plane by adopting a least square method;
S43, the distance between the current popping candidate point and the fitting plane is smaller than a certain distance threshold value The included angle between the normal vector of the current pop-up candidate point and the normal vector of the fitting plane is smaller than a certain angle value；
S44, regarding the candidate points passing the consistency test as the points in the target surface and dynamically adding the candidate points to the inner point setIn (a) and (b);
S45, repeating the checking step until all points in the alternative data set are popped up, and finally obtaining an inner point set I.e. the points in the target area form a point cloud。
Further, the specific process of step S5 includes the following steps:
s51, recursively dividing the point cloud space into 8 equal cubes, which are divided into 8 subcubes, i.e. cells, when a predefined maximum level is reached When there are no more points in the cell, the recursive subdivision process is stopped;
S52, representing the membership of the points in the unit cell in a digital coding mode, wherein each point is allocated with a unique digital code, and the code is formed by 3 Bit groupsComposition by linking them to a sequence from 1 toForming global coordinates of the cell, each point being associated with a unique numerical index indicating the location of the cell in which it is located;
S53, determining which neighborhood points Point and query PointIn the same cell, the distances from all these points to the query point are calculated, and they are sorted according to this distance and kept closestA plurality of neighborhood points;
s54, three kinds of Code for interleaving groups of bits together to reconstruct adjacent cells, set by 3 additionsIs thatIs used to determine the coordinates of the adjacent cells,And performing a reverse process to find codes of adjacent cells to find the wall plateOf face, i.e. face BEach neighborhood pointThe construction principle of octree can be expressed as:
；
Wherein, As the coordinates of the current cell,For the coordinate offset of the adjacent cells,For the highest weight bit it is possible to use,Is bit 1 of the code from the left;
s55, calculating the area In (a) and (b)Point to pointAll distances of the points take their minimum value.
Further, the area is calculated in the step S55In (a) and (b)Point to pointTaking the minimum value of all the distances of the points, specifically comprising: in the area ofSelecting a plurality ofPoints, respectively calculating their arrival point cloudsIs the most recent of (a)Each neighborhood pointThe distance between these two points, called peers, the set of distances is:
；
Known as point cloud Opposite surfaceThe hausdorff distance of the nearest neighbor of (1), namely:
。
By means of the technical scheme, the invention provides a method for measuring the thickness of the aerospace wallboard based on point cloud data and Haoskov measurement, which has the following beneficial effects:
The method adopts the region growing method to extract the upper and lower surfaces of the numbered area point cloud space wall plate, wherein the point on one surface searches the nearest neighbor point on the other surface by using the octree to search the neighbor point to calculate the Hausdorff distance, thereby not only remarkably improving the thickness measurement efficiency of the aerospace wall plate and ensuring the accuracy of the measurement result, but also showing the wide applicability of the thickness measurement wall plate in different measurement occasions.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:
FIG. 1 is a flow chart of a method for measuring thickness of a space wallboard based on point cloud data and Haoskov metrics;
FIG. 2 is a point cloud image of a panel of the present invention after a panel scan pretreatment;
FIG. 3 shows the upper surface of a wall panel extracted by the area growing method of the present invention;
fig. 4 is a cross-sectional view of a wall panel of the present invention and an extracted upper surface of the wall panel.
Detailed Description
In order that the above-recited objects, features and advantages of the present application will become more readily apparent, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. Therefore, the realization process of how to apply the technical means to solve the technical problems and achieve the technical effects can be fully understood and implemented.
Those of ordinary skill in the art will appreciate that all or a portion of the steps in a method of implementing an embodiment described above may be implemented by a program to instruct related hardware, and thus, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Referring to fig. 1-4, a specific implementation manner of the present embodiment is shown, in the present embodiment, the upper and lower surfaces of the wall plate are extracted from the numbered area point cloud space by adopting a region growing method, wherein the point on one surface searches the nearest neighbor point on the other surface by searching the neighbor point through an octree to calculate the hausdorff distance, so that the thickness measurement efficiency of the space wall plate is significantly improved, the accuracy of the measurement result is ensured, and the wide applicability of the space wall plate in different measurement occasions is also shown.
Referring to fig. 1, the embodiment provides a method for measuring thickness of a space wallboard based on point cloud data and haustorium measurement, which includes the following steps:
s1, respectively scanning the front surface and the back surface of a wallboard by a laser scanner, performing rough registration operation to obtain complete point cloud data, and performing point cloud pretreatment;
as a preferred embodiment of step S1, the specific procedure comprises the steps of:
S11, firstly manually removing block noise near the target point cloud, and then removing outlier data points by adopting a filtering algorithm;
S12, in the downsampling process, optimizing the point cloud data by adopting a voxel filter method, and creating a plurality of edge lengths on the whole point cloud data Inside each voxel grid we merge the originally dispersed points and use the centroid position of these points to represent the center point of the whole grid.
In this embodiment, the laser scanner scans the front and back surfaces of the wall plate respectively, and selects corresponding points to perform coarse registration operation to obtain a complete point cloud, and since a large number of noise and outlier points exist around the original point cloud and contain a large number of points, the distance between two points at the most dense position is less than 1mm, downsampling is required, and fig. 2 is a complete point cloud diagram of the wall plate after preprocessing; in the downsampling process, a voxel filter method is adopted, the technology can realize the optimization processing of point cloud data, and specifically, a series of fixed edge lengths are constructed on the whole point cloud dataWhich constitute a voxel structure in three-dimensional space. Inside each voxel grid we merge the originally dispersed points and use the centroid position of these points to represent the center point of the whole grid. By the method, the number of the data can be effectively reduced while the basic form and the characteristics of the point cloud data are maintained, and the speed and the efficiency of subsequent processing are improved.
S2, calculating the preprocessed point cloudDividing uniform and non-uniform areas of the point cloud surface according to the local geometric features of each point in the cloud surface;
As a preferred embodiment of step S2, the specific procedure comprises the steps of:
s21: point-to-point cloud Each point of (3)Find itSets of neighborhood pointsThe 3 angular coordinate vectors of the neighborhood points form a covariance matrixThen calculateMaximum characteristic value of (2). Covariance matrixThe following formula is shown:
；
In the method, in the process of the invention, Respectively represent a single pointA kind of electronic deviceRectangular coordinates of each neighborhood point in three-dimensional space, the coordinates are given in the form of one-dimensional arrays, and the length of each array isCorresponding to a pointA kind of electronic deviceCoordinates, covariance of nearest neighbor pointsRepresenting vector variablesAndCovariance between, which measuresAndThe degree of linear correlation between, and likewise,AndRespectively representAnd、AndCovariance between;
at the calculation point Selecting an appropriate normal vectorIt is important.The value is too small, meaning that the neighborhood points are insufficient to adequately describe the pointsThe characteristics of the local plane, thereby causing inaccurate normal vector calculation; whileIf the value is too large, the point may be causedThe neighborhood point set of (c) contains too many points from adjacent surfaces, which again leads to normal vector distortion and increases the computational time cost.
S22: numbering each area, performing rough segmentation to roughly determine the positions of the areas in a three-dimensional space, and reducing the search range of subsequent fine segmentation and feature analysis; subdividing each coarsely segmented face space into a plurality of three-dimensional cube grids, the grids having edges of all；
Referring to fig. 2, each area is first numbered for subsequent identification and processing. Next, rough segmentation is performed, which is mainly to roughly determine the coordinate range of each area in the three-dimensional space. Because of the coarse segmentation, these ranges may contain all points of the corresponding area, possibly even some noise points. Thereafter, for further refinement, each of the coarsely segmented face spaces is subdivided into a plurality of three-dimensional cube meshes, the meshes having edges of all. The purpose of this step is to provide a basis for subsequent fine segmentation and feature analysis.
S23: setting a proper threshold according to the maximum characteristic value calculation result of each pointIf the threshold value is exceeded, the point is considered to belong to a point with less uniform distribution;
Specifically, a maximum eigenvalue of each point is calculated, and this eigenvalue can reflect the distribution characteristics of the points in its neighborhood. Then, a proper threshold value is determined according to the distribution condition of the characteristic values . If the maximum eigenvalue of a point exceeds this threshold, then the neighborhood of that point can be considered to be relatively distributed in the direction of the corresponding eigenvector, i.e., the point is unevenly distributed within its neighborhood.
S24: dividing the grids into uniform point grids and non-uniform point grids, and counting that the characteristic value in each grid exceeds a set threshold valueIf the ratio exceeds 80%, then defining the grid as a non-uniform grid of points; otherwise, the grid of equal points means that the points in the grid are approximately equal in distance in each arrangement direction, and the characteristic of array distribution is presented; while non-uniform dot grids represent non-uniform spacing of dots in all directions, typically exhibiting a banded distribution, such non-uniformity often being accompanied by discontinuities. By the method, the area of the point cloud distribution is divided into uniform and non-uniform areas, and important basis is provided for subsequent processing and analysis.
S3, selecting a surface with most uniform and continuous point distribution in a surface area space, selecting initial seed points in the area, and extracting a plurality of points in the surface area, namely growing inner points;
As a preferred embodiment of step S3, the specific procedure comprises the steps of:
S31, calculating point cloud by adopting principal component analysis method Normal vector and curvature of each point in (a);
s32, respectively extracting Points in the uniform point grid in the individual area space constituteIndividual point of uniformity datasetAnd calculateCentroid point of (2)Coordinates, finding the homopoint datasetMid to centroid pointPoints of minimum euclidean distanceDefined asInitial seed points of the number area, so that uniform and continuous distribution of point clouds near the initial seed points is ensured;
s33, if the included angle between the neighborhood point of each seed point and the normal vector of the seed point is smaller than a certain angle threshold value The neighborhood point is considered as the point in the target area, if the curvature of the neighborhood point is smaller than a certain curvature threshold valueThen the neighborhood point is added to the set of seed points,Deleting the current seed point after the neighborhood point is judged, and entering the growth process of the next seed point;
S34, repeating the step S303 until the seed point set is empty.
S4, checking and screening points with consistent characteristics with the growing internal points from the rough-divided area space, and forming an area together with the growing internal points, thereby extracting one surface, namely the area, of the wallboard modelRecorded as point cloud；
As a preferred embodiment of step S4, the specific process comprises the steps of:
s41, adding the growing inner points to the inner point set as initial points In (a) and (b);
S42, appointing Points other than the growing interior points in the individual area space constitute an alternative datasetSequentially pop upIn (2) point verification consistency conditions forFitting a plane by adopting a least square method;
S43, the distance between the current popping candidate point and the fitting plane is smaller than a certain distance threshold value The included angle between the normal vector of the current pop-up candidate point and the normal vector of the fitting plane is smaller than a certain angle value；
S44, regarding the candidate points passing the consistency test as the points in the target surface and dynamically adding the candidate points to the inner point setIn (a) and (b);
S45, repeating the checking step until all points in the alternative data set are popped up, and finally obtaining an inner point set I.e. the points in the target area form a point cloud。
In this embodiment, please refer to fig. 3, in the areaIn the process, the characteristic values of each point are countedIs 0.0006. Test multipleValues (0.0003-0.0007) and different grid sizes(0.05 M-0.15 m). Discovery ofAt <0.0004, the uniform grid is often misjudged as non-uniform; at >0.0004, non-uniform region identification is insufficient. The too small blur limit is set to be too small,Too much masks the non-uniformity. Final selection=0.0004，=0.10M to accurately divide the region and reduce erroneous judgment. The region growing method can only play a role in the region with better point continuity, and the rest of internal points are extracted through consistency test.
When calculating curvature, the curvature is greatly influenced by the continuity of the point cloud distribution and the neighborhood pointsThe influence of numerical values, in terms of areaFor example, multiple locations have poor continuity and tend to block the growth process of the seed points, so that more desirable initial seed points occur in a relatively continuously distributed region. Finally, we setWhen =10, the point of minimum curvature is used as seed point, fig. 3 is the upper surface of the wall plate extracted by regional growth, i.e. point cloud。
S5, searching the other surface of the wallboard by adopting an octree methodI.e. areaAt the point cloudUpper part of the cylinderEach neighborhood pointCalculate the areaArea of contactThe Haoskov distance between the two is obtained, and the minimum value is the thickness of the wallboard.
As a preferred embodiment of step S5, the specific process comprises the steps of:
s51, recursively dividing the point cloud space into 8 equal cubes, which are divided into 8 subcubes, i.e. cells, when a predefined maximum level is reached When there are no more points in the cell, the recursive subdivision process is stopped;
S52, representing the membership of the points in the unit cell in a digital coding mode, wherein each point is allocated with a unique digital code, and the code is formed by 3 Bit groupsComposition by linking them to a sequence from 1 toForming global coordinates of the cell, each point being associated with a unique numerical index indicating the location of the cell in which it is located;
S53, determining which neighborhood points Point and query PointIn the same cell, the distances from all these points to the query point are calculated, and they are sorted according to this distance and kept closestA plurality of neighborhood points; the distance being measured essentially by finding the surroundings of the query pointThe method comprises the steps of calculating the distance between a query point and each neighborhood point;
s54, each point is allocated a unique digital code, and the code is formed by 3 Bit group%) Composition is prepared. The 3-bit group can accurately identify the position of the subcube containing the point in a higher level. Since we have 8 subcubes, 3 bits groups (each bit may be 0 or 1) are sufficient to cover all possible subcube positions. Each bit corresponds to a dimension of the three-dimensional space, encoding the position of a point relative to a plane orthogonal to that dimension, thus dividing the cube into two equal-sized portions.
These 3-bit groups not only identify local positions, but also by concatenating them from 1 to 1Forms the global coordinates of the cells. This means that either the coarse-grained location of the highest level or the finer-level precise location can be uniquely determined by this numerical code.
The digital codes and the point list are stored in parallel, and are sequenced, so that strict consistency between the codes and the points is maintained, and an additional index table is not required to be maintained in a memory, thereby saving space. The ordered list enables us to efficiently search for a given code using dichotomy, thereby quickly locating the corresponding cell.
Is provided withTo perform the octree level of an operation, the data of the cells of the group consisting of the immediately adjacent cells (i.e., one or more cells of the current cell) may be considered onlyThe code portion consisting of the individual significant bits applies binary addition and subtraction. Is provided withAs the coordinates of the current cell,For the coordinate shift of adjacent cells, setIs thatCoordinates of adjacent cells of (and hence we also have a condition). According to each dimension(Three dimensions in space are to be taken into accountRespectively mapped to numbers (0, 1, 2) to extract the corresponding numbersThe highest weight bits，Is bit 1 of the code, starting from the left). For example, if the current dimension isThen we will choose bits 1, 4, 7 from the left; if it isDimension, we will choose bits 2, 5, 8; if it isDimension, we will choose bits 3, 6, 9. At these pointsOn the bit group, corresponding to the position of the cell in the dimension, if、Or (b)Is strictly negative, then the application decrements if、Or (b)Is strictly positive, then the application increments, if they are empty, then no action is taken. In practice, decrements and increments correspond to just the offset (left or right) of one cell relative to the current position, and depend on the dimension under consideration. After processing each dimension we have done by combining threeThe groups of bits are interleaved together to reconstruct the codes of the neighboring cells. The construction principle of octree can be expressed as:
；
Wherein, As the coordinates of the current cell,For the coordinate offset of the adjacent cells,For the highest weight bit it is possible to use,Is bit 1 of the code from the left;
Thus, by 3 additions Deriving the coordinates of the adjacent cells and performing a reverse process to find the codes of the adjacent cells, thereby finding the wall plateFace (i.e. face B)Each neighborhood point。
S55, calculating the areaIn (a) and (b)Point to pointAll distances of the points take their minimum value.
More specifically, the area is calculated in step S55In (a) and (b)Point to pointTaking the minimum value of all the distances of the points, specifically comprising: referring to FIG. 4, in the areaSelecting a plurality ofPoints, respectively calculating their arrival point cloudsIs the most recent of (a)Each neighborhood pointThe distance between these two points, called peers, the set of distances is:
；
Known as point cloud Opposite surfaceThe hausdorff distance of the nearest neighbor of (1), namely:
。
In summary, the method of region growth is adopted to extract the upper and lower surfaces of the numbered area point cloud space wallboard, the point on one surface searches the nearest neighbor point on the other surface by using the method of searching neighbor points by using an octree to calculate the Hausdorff distance, the efficiency of measuring the thickness of the aerospace wallboard is remarkably improved, the accuracy of measuring results is ensured, and meanwhile, the wide applicability of the method in different measuring occasions is also shown.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
Logic and/or steps represented in the flowcharts or otherwise described herein, e.g., a ordered listing of executable instructions for implementing logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions.
The foregoing embodiments have been presented in a detail description of the invention, and are presented herein with a particular application to the understanding of the principles and embodiments of the invention, the foregoing embodiments being merely intended to facilitate an understanding of the method of the invention and its core concepts; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
Claims (6)
1. The method for measuring the thickness of the aerospace wallboard based on the point cloud data and the Haosdorf measurement is characterized by comprising the following steps of:
s1, respectively scanning the front surface and the back surface of a wallboard by a laser scanner, performing rough registration operation to obtain complete point cloud data, and performing point cloud pretreatment;
S2, calculating the local geometric characteristics of each point in the preprocessed point cloud Q, and dividing uniform and non-uniform areas of the point cloud surface according to the local geometric characteristics;
s3, selecting a surface with most uniform and continuous point distribution in a surface area space, selecting initial seed points in the area, and extracting a plurality of points in the surface area, namely growing inner points;
S4, checking and screening points with the same characteristics as the growing internal points from the rough-division area space, and forming an area together with the growing internal points, so that one surface of the wallboard model, namely an area A, is extracted and marked as point cloud S;
s5, searching the other surface S' of the wallboard by adopting an octree method, namely, calculating the Hausdorff distance between the surface A and the surface B by using k neighborhood points p of the surface B on the point cloud S, and obtaining the thickness of which the minimum value is the wallboard;
The specific process comprises the following steps:
S51, recursively dividing the point cloud space into 8 equal cubes, wherein the cubes are divided into 8 subcubes, namely cells, and when a predefined maximum level N is reached and no more points exist in the cells, the recursion subdivision process is stopped;
S52, representing membership of points in the cell in a digital coding mode, wherein each point is allocated a unique digital code, the code consists of N3 bit groups b xbybz, global coordinates of the cell are formed by connecting the N3 bit groups b xbybz to all subdivision levels from 1 to N, and each point is associated with a unique digital index which indicates the position of the cell where the point is located;
s53, determining which neighborhood points p and query points p' are in the same cell, calculating the distances from all the points to the query points, sequencing the points according to the distances, and keeping the nearest k neighborhood points;
S54, interweaving three n-bit groups together to reconstruct codes of adjacent cells, setting (x 0+Δx,y0+Δy,z0 +Deltaz) as { Deltax, deltay, deltaz }. Epsilon [ -1,1] 3 coordinates of the adjacent cells, and carrying out a reverse process to find codes of the adjacent cells, thereby finding k neighborhood points p of a wallboard S' face, namely a face B, so that the construction principle of the octree can be expressed as follows:
Where (x 0,y0,z0) is the coordinates of the current cell, (Δx, Δy, Δz) is the coordinate offset of the neighboring cell, b 3I+dim is the highest weighted bit, I ε [0, n-1] }, b I is bit 1 from the left of the code;
S55, calculating all distances from the p' point to the p point in the area B, and taking the minimum value.
2. The method for measuring the thickness of the aerospace panel based on the point cloud data and the hausdorff metric according to claim 1, wherein the method comprises the following steps of: the specific process of the step S1 comprises the following steps:
S11, firstly manually removing block noise near the target point cloud, and then removing outlier data points by adopting a filtering algorithm;
And S12, in the downsampling process, a voxel filter method is adopted to optimize point cloud data, a plurality of cube grids with the edge length of a are created on the whole point cloud data, namely three-dimensional voxel grids, and in each voxel grid, original scattered points are combined and processed, and the center point of the whole grid is represented by the centroid positions of the points.
3. The method for measuring the thickness of the aerospace panel based on the point cloud data and the hausdorff metric according to claim 1, wherein the method comprises the following steps of: the specific process of the step S2 comprises the following steps:
S21: each point Q i in the point cloud Q obtains k neighborhood point sets N i, 3 angular coordinate vectors of the neighborhood points form a covariance matrix C, and then the maximum eigenvalue E of the covariance matrix C is obtained, where the covariance matrix C is represented by the following formula:
Wherein X, Y, Z respectively represent rectangular coordinates of k neighborhood points of a single point q i in three-dimensional space, the coordinate values are given in the form of one-dimensional arrays, the length of each array is l, the coordinates of k nearest neighborhood points corresponding to the point q i, covariance cov (X, Y) represents covariance between vector variables X and Y, which measures the degree of linear correlation between X and Y, and likewise cov (X, Z) and cov (Y, Z) respectively represent covariance between X and Z, Y and Z;
s22: numbering each area, performing rough segmentation to determine the positions of the areas in a three-dimensional space, and reducing the search range of subsequent fine segmentation and feature analysis; subdividing each coarsely divided face space into a plurality of three-dimensional cube grids, the edges of which are b;
s23: setting a proper threshold E th according to the maximum characteristic value calculation result of each point, and considering that the point belongs to a point with uneven distribution if the threshold is exceeded;
S24: dividing the grids into uniform point grids and non-uniform point grids, counting the proportion of points with characteristic values exceeding a set threshold E th in each grid, and defining the grid as the non-uniform point grid if the proportion exceeds 80%; otherwise, the point-sharing grid is adopted, so that the area of the point cloud distribution is divided into uniform and non-uniform areas.
4. The method for measuring the thickness of the aerospace panel based on the point cloud data and the hausdorff metric according to claim 1, wherein the method comprises the following steps of: in step S3, the specific process includes the following steps:
s31, calculating the normal vector and curvature of each point in the point cloud Q by adopting a principal component analysis method;
S32, respectively extracting points in the uniform point grids in j area spaces, forming j uniform point data sets D j, calculating the coordinates of a centroid point M j of D j, searching a point S j with the minimum Euclidean distance from the centroid point M j in the uniform point data set D j, and defining the point as an initial seed point of a j area, thereby ensuring that the point cloud distribution near the initial seed point is uniform and continuous;
s33, if the included angle between the neighborhood point of each seed point and the normal vector of the seed point is smaller than a certain angle threshold Ar th, the neighborhood point is considered as a target in-plane point, if the curvature of the neighborhood point is smaller than a certain curvature threshold C th, the neighborhood point is added into the seed point set, the current seed point is deleted after the judgment of the k neighborhood point is finished, and the next seed point growth process is started;
S34, repeating the step S303 until the seed point set is empty.
5. The method for measuring the thickness of the aerospace panel based on the point cloud data and the hausdorff metric according to claim 1, wherein the method comprises the following steps of: the specific process of the step S4 comprises the following steps:
S41, adding the growing inner points serving as initial points into an inner point set I j;
S42, designating points except growing inner points in j area spaces to form an alternative data set E j, sequentially popping up point verification consistency conditions in E j, and fitting the points in I j to a plane by adopting a least square method;
S43, the distance between the current popup candidate point and the fitting plane is smaller than a certain distance threshold D th, and the included angle between the normal vector of the current popup candidate point and the normal vector of the fitting plane is smaller than a certain angle value Ac th;
S44, the candidate points passing the consistency test are regarded as the points in the target surface and are dynamically added into an interior point set I j;
S45, repeating the checking step until all points in the alternative data set are popped up, and finally obtaining an inner point set I j which is the inner point of the target area to form a point cloud S.
6. The method for measuring the thickness of the aerospace panel based on the point cloud data and the hausdorff metric according to claim 1, wherein the method comprises the following steps of: in the step S55, all distances from the p' point to the p point in the area B are calculated, and the minimum value is taken, which specifically includes: selecting a plurality of p 'points in the area B, and respectively calculating the distances between the p' points and the nearest k neighborhood points p in the point cloud s, wherein the two points are called as peer points, and the set of the distances is as follows:
{dp′＝d(p′,s),p∈s}
The Hastedorff distance, called the nearest neighborhood of the point cloud s relative to the surface s', is:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410846294.3A CN118392054B (en) | 2024-06-27 | 2024-06-27 | Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410846294.3A CN118392054B (en) | 2024-06-27 | 2024-06-27 | Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118392054A CN118392054A (en) | 2024-07-26 |
| CN118392054B true CN118392054B (en) | 2024-08-30 |
Family
ID=91985414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410846294.3A Active CN118392054B (en) | 2024-06-27 | 2024-06-27 | Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118392054B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111735384A (en) * | 2020-04-28 | 2020-10-02 | 南京工程学院 | Engine installation measurement and installation method and device based on dynamic interference analysis |
| CN112634457A (en) * | 2021-01-06 | 2021-04-09 | 广西科技大学 | Point cloud simplification method based on local entropy of Hausdorff distance and average projection distance |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11430224B2 (en) * | 2020-10-23 | 2022-08-30 | Argo AI, LLC | Systems and methods for camera-LiDAR fused object detection with segment filtering |
| CN113902688A (en) * | 2021-09-23 | 2022-01-07 | 无锡市城市重点建设项目管理中心 | Non-uniform point cloud surface area segmentation method based on improved region growing method |
| CN116817771B (en) * | 2023-08-28 | 2023-11-17 | 南京航空航天大学 | Aerospace part coating thickness measurement method based on cylindrical voxel characteristics |
| CN117808703B (en) * | 2024-02-29 | 2024-05-10 | 南京航空航天大学 | Multi-scale large-scale component assembly gap point cloud filtering method |
-
2024
- 2024-06-27 CN CN202410846294.3A patent/CN118392054B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111735384A (en) * | 2020-04-28 | 2020-10-02 | 南京工程学院 | Engine installation measurement and installation method and device based on dynamic interference analysis |
| CN112634457A (en) * | 2021-01-06 | 2021-04-09 | 广西科技大学 | Point cloud simplification method based on local entropy of Hausdorff distance and average projection distance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118392054A (en) | 2024-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | Adaptive simplification of point cloud using k-means clustering | |
| CN109859226B (en) | Detection method of checkerboard corner sub-pixels for graph segmentation | |
| CN111831660B (en) | Method and device for evaluating metric space division mode, computer equipment and storage medium | |
| CN109887015A (en) | A kind of point cloud autoegistration method based on local surface feature histogram | |
| CN107038717A (en) | A kind of method that 3D point cloud registration error is automatically analyzed based on three-dimensional grid | |
| CN111275724B (en) | Airborne point cloud roof plane segmentation method based on octree and boundary optimization | |
| CN105654483A (en) | Three-dimensional point cloud full-automatic registration method | |
| CN112819962B (en) | Non-uniform grid division and local grid density method in digital image correlation | |
| JP2011238204A (en) | Method for recognition and position attitude determination of three-dimensional object at three-dimensional scene | |
| CN111028327A (en) | Three-dimensional point cloud processing method, device and equipment | |
| CN110533774B (en) | Three-dimensional model reconstruction method based on smart phone | |
| CN110232742A (en) | 3D printing hierarchical control algorithm | |
| CN114663373A (en) | Point cloud registration method and device for detecting surface quality of part | |
| CN109345571B (en) | Point cloud registration method based on extended Gaussian image | |
| CN117710603B (en) | Unmanned aerial vehicle image three-dimensional building modeling method under constraint of linear geometry | |
| CN114358166A (en) | Multi-target positioning method based on self-adaptive k-means clustering | |
| CN118392054B (en) | Method for measuring thickness of aerospace wallboard based on point cloud data and Haosdorf measurement | |
| CN111597367B (en) | Three-dimensional model retrieval method based on view and hash algorithm | |
| CN112767429A (en) | Ground-snow surface point cloud rapid segmentation method | |
| CN111539966A (en) | Colorimetric sensor array image segmentation method based on fuzzy c-means clustering | |
| Chappuis et al. | Improving surface meshing from discrete data by feature recognition | |
| CN114998397B (en) | Multi-view satellite image stereopair optimization selection method | |
| CN108734180A (en) | A kind of SIFT feature gradient generation method based on calculation optimization | |
| CN116091562A (en) | Building point cloud automatic registration method based on two-dimensional projection line segments | |
| CN111009004A (en) | Hardware optimization method for accelerating image matching |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |