CN110570511B - Processing method, device and system of point cloud data and storage medium - Google Patents
Processing method, device and system of point cloud data and storage medium Download PDFInfo
- Publication number
- CN110570511B CN110570511B CN201810575099.6A CN201810575099A CN110570511B CN 110570511 B CN110570511 B CN 110570511B CN 201810575099 A CN201810575099 A CN 201810575099A CN 110570511 B CN110570511 B CN 110570511B
- Authority
- CN
- China
- Prior art keywords
- point
- dimensional
- vector
- determining
- points
- 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
- 238000003672 processing method Methods 0.000 title abstract description 11
- 239000013598 vector Substances 0.000 claims abstract description 149
- 238000000034 method Methods 0.000 claims abstract description 56
- 230000001154 acute effect Effects 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims description 36
- 239000011159 matrix material Substances 0.000 claims description 24
- 239000006185 dispersion Substances 0.000 claims description 6
- 230000029777 axis specification Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 abstract description 11
- 238000005070 sampling Methods 0.000 description 40
- 238000010586 diagram Methods 0.000 description 16
- 238000005192 partition Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 14
- 238000000638 solvent extraction Methods 0.000 description 8
- 230000008859 change Effects 0.000 description 7
- 230000005484 gravity Effects 0.000 description 5
- 238000013507 mapping Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000003746 surface roughness Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/005—General purpose rendering architectures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Graphics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses a processing method, a device, a system and a storage medium of point cloud data. The method comprises the following steps: acquiring a three-dimensional point set of the surface of an object, and determining a sub-surface of the object of a first point for the first point of the surface of the object; determining a first axis passing through the first point and perpendicular to the subsurface; acquiring a viewing angle direction corresponding to a first point, wherein the viewing angle direction is the direction in which the projection position of the first point in corresponding image acquisition equipment points to the space position of the first point; a vector on the first axis at an acute angle to the viewing direction of the first point is determined as a normal vector of the first point. According to the method provided by the embodiment of the invention, the orientation information of the point cloud normal vector is determined through the camera view angle direction, so that the calculation complexity is low and the accuracy is high, and the calculation efficiency of the point cloud normal vector is improved.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method, an apparatus, a system, and a storage medium for processing point cloud data.
Background
With the development and maturity of digital image processing, digital projection display and computer processing technologies, three-dimensional modeling technologies are rapidly developed and widely used. The point data of the outer surface of the object in the real scene is obtained by utilizing a three-dimensional modeling technology, the set of the point data with three-dimensional space information is called as three-dimensional point cloud data, and a three-dimensional model of the object can be constructed based on the three-dimensional point cloud data.
The point cloud normal vector can represent the spatial structure of the three-dimensional object, and the accuracy of the calculation structure of the point cloud normal vector directly influences the accuracy of subsequent point cloud data processing and three-dimensional model construction, and the accuracy of the normal vector characteristic of the point cloud data has important significance for measuring the accuracy of the three-dimensional point cloud structure.
In the prior art, because the inherent physical characteristics of the point cloud acquisition equipment are limited, the obtained point cloud data usually contains loopholes, noise and the like, or because the three-dimensional object itself comprises sharp angles, non-convex structures and other structural features, the calculation result of the normal vector of the point cloud is seriously influenced by the noise or the structural features of the object itself, and the settlement result is easy to make mistakes.
Disclosure of Invention
The embodiment of the invention provides a processing method, a device, a system and a storage medium for point cloud data, which can rapidly and accurately calculate the normal vector of the point cloud data.
According to an aspect of the embodiment of the present invention, there is provided a method for processing point cloud data, including:
acquiring a three-dimensional point set of the surface of an object, and determining a subsurface of the object at a first point of the surface of the object;
determining a first axis passing through the first point and perpendicular to the subsurface;
Acquiring a viewing angle direction corresponding to a first point, wherein the viewing angle direction is the direction in which the projection position of the first point in corresponding image acquisition equipment points to the space position of the first point;
a vector on the first axis at an acute angle to the viewing direction of the first point is determined as a normal vector of the first point.
According to another aspect of the embodiment of the present invention, there is provided a processing apparatus for point cloud data, including:
the subsurface determining module is used for acquiring a three-dimensional point set of the surface of the object, and determining a subsurface of the object at a first point of the surface of the object;
an axis determination module for determining a first axis passing through the first point and perpendicular to the subsurface;
the view angle direction acquisition module is used for acquiring a view angle direction corresponding to the first point, wherein the view angle direction is the direction in which the projection position of the first point in the corresponding image acquisition equipment points to the spatial position of the first point;
and the normal vector determining module is used for determining a vector which forms an acute angle with the viewing angle direction of the first point on the first axis as a normal vector of the first point.
According to still another aspect of the embodiment of the present invention, there is provided a processing system for point cloud data, including: a memory and a processor; the memory is used for storing programs; the processor is used for reading executable program codes stored in the memory to execute the processing method of the three-dimensional point cloud data.
According to still another aspect of the embodiments of the present invention, there is provided a computer-readable storage medium having stored therein instructions that, when executed on a computer, cause the computer to perform the method of processing point cloud data of the above aspects.
According to the processing method, the processing device, the processing system and the storage medium of the point cloud data, the normal vector orientation information can be determined through the camera view angle, the calculation complexity is low, the accuracy is high, and therefore the operation efficiency is improved.
Drawings
In order to more clearly illustrate the technical solution of the embodiments of the present invention, the drawings that are needed to be used in the embodiments of the present invention will be briefly described, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram illustrating normal vectors of sampling points of a surface of an object according to an embodiment of the present invention;
FIG. 2a is a schematic diagram of a normal vector configuration of a surface of an object showing smooth features according to an embodiment of the invention;
FIG. 2b is a schematic diagram of a normal vector configuration of the object surface showing sharp features according to an embodiment of the invention;
FIG. 3a is a schematic diagram of a normal vector structure of an object surface showing a convex structure according to an embodiment of the present invention;
FIG. 3b is a schematic diagram illustrating a normal vector configuration of a surface of an object having a non-convex configuration according to an embodiment of the present invention;
FIG. 4a is a schematic diagram showing the structure of a Kdtree constructed from a three-dimensional point set according to an embodiment of the invention;
FIG. 4b is a schematic diagram showing the effect of cutting a plane corresponding to Kdtree according to an embodiment of the invention;
FIG. 5 is a schematic view showing the viewing angle direction of any point on the surface of an object according to an illustrative embodiment of the invention;
FIG. 6 is a flow diagram illustrating a method of processing point cloud data according to one embodiment of the invention;
fig. 7 is a flowchart illustrating a processing method of point cloud data according to another embodiment of the present invention;
fig. 8 is a schematic structural diagram of a processing device for point cloud data according to an embodiment of the present invention;
fig. 9 is a block diagram illustrating an exemplary hardware architecture of a computing device implementing a method and apparatus for processing point cloud data according to an embodiment of the present invention.
Detailed Description
Features and exemplary embodiments of various aspects of the present invention will be described in detail below, and in order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely configured to illustrate the invention and are not configured to limit the invention. It will be apparent to one skilled in the art that the present invention may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the invention by showing examples of the invention.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
In the embodiment of the invention, point Cloud (Point Cloud) may represent a set of discrete points on the surface of an object, point Cloud Data (Point Cloud Data) may represent a set of Data of discrete points on the surface of the object acquired by a measuring instrument, and the Point Cloud Data may include color information, depth information and geometric position information represented by three-dimensional coordinates of the object, and the three-dimensional size of the object may be obtained through the geometric position information of the object.
In one embodiment, an instrument device capable of performing three-dimensional scanning, such as a three-dimensional scanner, a laser or radar scanner, a stereo camera, or the like, may be used to project structured light on an object in a real scene, so as to perform three-dimensional scanning at multiple viewing angles, such as 360 degrees, at different viewing angles, and obtain multiple point data of the object in the real scene, where the point data forms point cloud data corresponding to the different viewing angles.
In one embodiment, the point cloud data of the object may be input into a file for storage to obtain a three-dimensional point cloud model, the three-dimensional point cloud model is utilized to perform surface reconstruction of the three-dimensional model, the three-dimensional surface shape of the object is determined, and the surface model of the object is obtained, for example, the point cloud model is converted into a surface model such as a polygonal mesh model or a curved surface model. Wherein each polygon in the polygonal mesh model can uniquely define a polygon plane; the surfaces of the surface model may have gaps or overlaps between them. In the three-dimensional model processing, each polygon or each curved surface of the surface model of the object may be regarded as a surface piece to be processed of the three-dimensional model.
FIG. 1 illustrates a schematic normal vector diagram of a sample point of the surface of an object according to an embodiment of the present invention. As shown in fig. 1, in the embodiment of the present invention, the Normal (Normal) is a perpendicular line reaching the surface of the object, and the Normal is a vector, simply referred to as Normal vector. For any point in the point cloud data of the object, the normal vector of the surface of the area where the any point is located may be referred to as the normal vector of the any point.
In the embodiment of the invention, the normal vector can be used for describing the space structure of the three-dimensional object, is an important structural feature of the three-dimensional model of the object, and is described below by taking the surface reconstruction and texture mapping process of the three-dimensional model as an example.
With continued reference to fig. 1, the angle between the incident line of the image acquisition device and the normal vector of the patch to be processed may be used to describe the viewing angle of the shot for the three-dimensional model of the object.
In one embodiment, the point cloud data model at different shooting angles sometimes shows a rotational misalignment or a translational misalignment during the surface reconstruction of the three-dimensional model. Therefore, the point cloud registration is required to be performed on the plurality of point clouds, and coordinate transformation is performed through rotation and/or translation operation, so that consistent alignment of the point cloud data under a unified coordinate system is realized, and a complete three-dimensional point cloud data model is obtained.
In the point cloud registration process, the variation degree of the normal vector of the sampling point on the surface of the object can be used for measuring the fluctuation or flatness degree of the area where the sampling point is located, namely, the larger the variation degree of the normal vector of the sampling point is, the larger the fluctuation of the area where the sampling point is located is; when the normal vector of the sampling point obtains an extreme value in the normal vector changing direction, the normal vector of the current sampling point is shown to have mutation.
In this embodiment, feature matching may be performed on different point cloud data based on a point cloud normal vector, feature point pairs between the different point cloud data are determined, and coordinate transformation is performed on one or more feature points in the feature point pairs to register the different point cloud data.
In one embodiment, in order to improve the authenticity of the three-dimensional model, the texture image of the object may be acquired by the image acquisition device, and the texture mapping is performed by using the mapping relationship between the three-dimensional model of the object and the texture image, and the texture image is mapped to the three-dimensional model of the object, so as to obtain the three-dimensional texture model of the object.
In the texture mapping process of the embodiment, for each to-be-processed surface patch of the three-dimensional model of the object, if the normal vector of the midpoint of each to-be-processed surface patch of the surface model of the object is consistent, the three-dimensional model of the object will show the visual effect of the mirror surface, that is, each surface of the three-dimensional model of the object is spliced like the mirror surface, and the distortion degree is higher.
In this embodiment, a more suitable shooting view angle for each to-be-processed patch of the surface model of the object may be selected according to an included angle between an incident ray of the image acquisition device and a normal vector of the to-be-processed patch, and a texture image subset corresponding to the selected shooting view angle may be determined, so as to perform three-dimensional mapping on the to-be-processed patch of the three-dimensional model, thereby obtaining a more accurate three-dimensional texture model.
According to the embodiment, the normal vector of the three-dimensional point cloud is an important index for measuring the space structure of the three-dimensional point cloud, is an indispensable characteristic attribute in the point cloud data, and is beneficial to improving the precision and quality of the three-dimensional model of the object by carrying out characteristic description and characteristic matching of the three-dimensional model by utilizing the normal vector of the point cloud.
In one embodiment, errors such as noise and loopholes exist in the point cloud data obtained through three-dimensional scanning, and the noise points can enable a curved surface reconstructed from the surface to be not smooth or to be deformed greatly; and due to the discretization of the point cloud data, holes can be formed in the curved surface reconstructed from the surface by the loopholes. Errors in the point cloud data can cause errors in the point cloud normal vector calculation.
FIG. 2a shows a schematic normal vector structure of a surface of an object with smooth features according to an embodiment of the invention; FIG. 2b shows a schematic of a normal vector structure of the object surface of a sharp feature according to an embodiment of the invention; FIG. 3a shows a schematic of a normal vector structure of an object surface of a convex structure according to an embodiment of the invention; FIG. 3b shows a schematic of a normal vector structure of a surface of an object having a non-convex structure according to an embodiment of the invention.
In one embodiment, the sampling point and the neighboring area curved surface corresponding to the sampling point may be obtained from the acquired point cloud data set by using a curved surface reconstruction technique, and then the normal vector of the neighboring area curved surface of the sampling point may be calculated from the curved model.
As one example, three-dimensional point clouds may be read, surface reconstructed, and normal vectors calculated using three-dimensional model processing software, such as merselab software.
In one embodiment, as shown in fig. 2a, the degree of curvature change of the sampling points can be used to measure the degree of curvature of the curved surface of the surface sheet to be processed. When the curvature value is changed to a smaller extent, the curvature value indicates that the curvature of the curved surface is smaller, that is, the smoothness of the curved surface is higher.
As shown in fig. 2b, in one embodiment, the gradient of the curvature may be used to represent the direction of change of the curvature of the current sampling point, and the gradient value of the curvature may be used to measure the speed of change of the curvature of the current sampling point; when the gradient of the curvature of the sampling point takes an extreme value, the curvature of the object surface at the current sampling point is shown to be suddenly changed. Sharp features generally refer to points in the surface sampling of an object where curvature is abrupt, e.g., sharp features may appear as folds or corners where two or more smooth curved surfaces intersect.
In the embodiment of the invention, the specific orientation of the point cloud normal vector can be determined by a near point verification method. Specifically, since the normal vector of the curved surface of the adjacent area of the sampling point obtained by calculation has direction uniformity, that is, the normal vector of all points in the curved surface of the adjacent area faces the outside of the curved surface. Thus, as shown in FIG. 3a, the higher the smoothness of the object, the higher the accuracy of the normal vector of the sample point. As shown in fig. 3b, if the object surface has a sampling point with a sharp feature, the orientation of the normal vector of other points of the curved surface of the adjacent area where the sampling point is located may be incorrect.
In one embodiment, the specific orientation of the normal vector of the surface on which the sampling point is located may be determined using the point cloud centroid. As an example, for any one sampling point in the point cloud data, the gravity center of the adjacent area curved surface is determined at the adjacent area curved surface corresponding to the sampling point and the adjacent point of the sampling point, and the direction of the gravity center of the adjacent area curved surface of the sampling point towards the sampling point of the object surface is taken as the orientation of the normal vector of the sampling point. The center of gravity of the point cloud may be a point coordinate, for example, an average value of all points in the neighboring area curved surface.
As an example, the method of determining the normal vector of the sampling point of the object surface in the embodiment of the present invention may use the PCL (Point Cloud Library) centroid method.
As shown in fig. 3a, if the curved surface of the adjacent area of the sampling point on the surface of the object is in a convex structure, the steering direction of each point on the curved surface of the adjacent area should be consistent, and the included angle formed by the boundary of the adjacent curved surfaces of each point of the curved surface of the adjacent area is not included, and the direction of the normal vector of the surface where the sampling point is located is consistent with the direction of the actual normal vector of the sampling point by using the center of gravity of the point cloud.
As shown in fig. 3b, the non-convex structure generally refers to a case where the inner angle of a polygonal surface patch corresponding to a plurality of sampling points on the surface of the object is larger than the outer angle, so that the curved surface and the shape including the non-convex portion are complex, and when the normal vector of the surface where the sampling points on the curved surface of the adjacent area with the non-convex structure are located is calculated by using the center of gravity of the point cloud, the direction of the calculated normal vector is easy to deviate from the direction of the actual normal vector of the sampling points.
In order to avoid the problem of inaccurate normal vector calculation of the point cloud in the above embodiment, the embodiment of the present invention provides a processing method of three-dimensional point cloud data, which can still obtain accurate normal vector information when the obtained three-dimensional point cloud data is not nearly perfect, for example, loopholes, noise, and the like occur, and the structural characteristics of the object itself, for example, have a sharp angle structure, a non-convex structure, and the like.
The following describes in detail a method, a device and a system for processing three-dimensional point cloud data according to an embodiment of the present invention with reference to the accompanying drawings.
In the embodiment of the invention, the three-dimensional point cloud data is a three-dimensional point set of the object surface, any point in the three-dimensional point set is used as a first point of the object surface, the surface of a nearby area of the first point can be determined according to the three-dimensional point set nearby the first point, and a normal vector passing through the first point and perpendicular to the surface of the nearby area is used as a normal vector of the first point in the three-dimensional point cloud data.
In one embodiment, the closer two different sampling points in the point cloud data are in space, the higher the similarity between the two sampling points. And determining points most similar to the characteristics of the to-be-calculated algorithm vector points according to the similarity between the point cloud data, and taking the points as a three-dimensional point subset adjacent to the to-be-calculated algorithm vector points.
In one embodiment, matching of the points of the algorithm vector to be calculated and the query of the sample points adjacent to the points of the algorithm vector to be calculated may be performed by a Range query (Range Search) or a K-nearest Neighbor query (K-nearest Search).
As one example, a range query may be understood as: and according to the vector points to be calculated and a threshold value of the preset query distance, finding out all point data with the space distance smaller than the threshold value from the point cloud data.
As one example, a K-nearest neighbor query can be understood as: and according to the vector point to be calculated and the positive integer K, finding out K data closest to the vector point to be calculated from the point cloud data. In this example, distances between the points to be calculated in the point cloud data and other points may be sequentially calculated, and a point with the smallest calculated distance may be selected as a neighboring point of the points to be searched for.
In the embodiment of the invention, if the data volume of the point cloud data is very large, in order to improve the searching efficiency and accuracy, the data index can be constructed by constructing an index tree of all the point cloud data in advance, and the three-dimensional points in the point cloud data. Specifically, points in the point cloud data can be divided into different space regions in advance according to coordinates of points in the point cloud, so that three-dimensional point searching of any point nearby is realized, and a set of nearby points of the point is constructed.
In one embodiment, the space region of the point cloud data is divided, each point in the point cloud data is managed according to the divided space region, and when searching for the adjacent point of the first point on the object surface, rapid matching is performed according to the data index of the point cloud data constructed in advance.
In one embodiment, the data may be spatially partitioned using a tree index structure to achieve efficient data retrieval. As one example, the tree index structure may be divided into two types of cut division clicking and overlap division Overlapping according to whether or not subspaces of the divided three-dimensional space overlap. The former method for cutting and dividing the Clipping divides the subspace obtained by dividing the three-Dimensional space without overlapping, such as KD (K-dimension) tree, which is called as K-D tree for short; the latter method of Overlapping region division divides the subspaces obtained by the three-dimensional space to overlap each other, such as R-tree in the balanced tree.
In the embodiment of the present invention, the K-D tree is a data structure for dividing the K-dimensional data space, and may be used for searching of specified data in the multidimensional space, such as the range search and nearest neighbor search described in the above embodiments. In one embodiment, an N-dimensional data set, such as three-dimensional point cloud data, may be partitioned by using a K-D tree, and when partitioning, the values of the sampling points to be allocated in the point cloud data and the partition nodes on the current dimension are compared at each layer of the tree according to the set partition dimension, and the sampling point data to be allocated is allocated to the left subtree and the right subtree corresponding to the partition nodes according to the comparison result.
For ease of understanding, the construction of Kdtrees is described below in connection with FIGS. 4a and 4b, taking a two-dimensional set of data points as an example. Fig. 4a shows a schematic structural diagram of a Kdtree constructed by a data point set according to an embodiment of the invention, and fig. 4b shows an effect schematic diagram of cutting a plane corresponding to the Kdtree according to an embodiment of the invention.
As shown in fig. 4a, assume a two-dimensional data point data set s= { (2, 3), (5, 4), (9, 6), (4, 7), (8, 1), (7, 2) }. The process of constructing a K-D tree from the two-dimensional data point data set S may include the steps of:
in step S01, a partitioning dimension of the data set is determined.
In one embodiment, the dimensions included in the spatial data set may be determined first, and then the spatial dimensions in which the data is partitioned may be determined according to the degree of dispersion of the data in each dimension.
As an example, since the two-dimensional data point data set S includes two dimensions, the coordinate axes corresponding to the two dimensions are an X coordinate axis and a Y coordinate axis, the coordinate axis number may be denoted as split= {0,1},0 represents the X coordinate axis, and 1 represents the Y coordinate axis, for example.
In one embodiment, the variance of the data represents how close a set of data is to the average of the set of data, with smaller variances indicating that the data is more concentrated to the average and larger variances indicating that the data is more scattered.
Therefore, the variance of the data in the coordinate axis direction corresponding to each dimension can be calculated, and the space dimension corresponding to the coordinate axis with the largest variance is used as the space dimension for dividing the data set. That is, the data set is divided into the spatial dimensions in which the data set is most dispersed.
As an example, for a two-dimensional data point data set S, calculating the variance of the data in the X-direction and the Y-direction, respectively, gives the maximum variance of the data in the X-axis, i.e. a comparison of the data dispersion along the X-axis direction, in which direction the data segmentation will have a better spatial resolution.
Step S01, determining a dividing node and a dividing plane corresponding to the dividing node.
In this step, data points of the data in the data set at intermediate positions of projection positions of the space dimensions where the division is performed may be regarded as division nodes in the data set.
In one embodiment, the data in the data set is ordered in the coordinate direction corresponding to the divided space dimension, and if the number N of the data in the data set is odd, the dividing node may be a data node located in the middle after the ordering; if the number N of data in the data set is even, the dividing node may be the N/2 th largest data node. In this embodiment, the selected data node may be referred to as the median or median in the data set.
In this embodiment, the division plane corresponding to the division node is: the projection positions of the partition nodes on the partition plane in the space dimension are perpendicular to the extending direction corresponding to the partition dimension, and the projection positions of all the data nodes in the data set in the space dimension are intermediate positions of the projection positions of all the data nodes in the data set in the space dimension.
As an example, the data set is sorted according to the value 2,5,9,4,8,7 in the X-axis direction, the number of data in the data set is 6, the median value of the data set is determined to be 7 (the 3 rd largest data node), and therefore the dividing node of the two-dimensional data point data set S can be determined to be (7, 2), and the dividing plane corresponding to the dividing node is: a straight line x=7 passing through the dividing nodes (7, 2) and perpendicular to the X-axis direction.
Step S03, determining a left subspace and a right subspace corresponding to the partition node.
As shown in fig. 4b, the division plane x=7 may divide the entire three-dimensional space into two parts, and the part x < =7 is a left subspace containing 3 nodes { (2, 3), (5, 4), (4, 7) }; the other part is the right subspace, containing 2 nodes { (9, 6), (8, 1) }.
Step S04, determining the partition nodes of the left subspace and the right subspace corresponding to the partition nodes of the two-dimensional data point data set S, and performing iterative processing until the preset space partition limiting condition is met, so as to obtain the two-dimensional data point data set S.
In one embodiment, the preset space division constraint is that each new subspace only includes one three-dimensional point, or that the spatial distance between the three-dimensional point in each new subspace and the new division node is within a preset distance threshold.
As an example, using the partitioning nodes in the two-dimensional data point data set S as root nodes, repeating the processing of the root nodes for the data nodes in the left subspace and the right subspace can obtain the next-level child nodes (5, 4) and (9, 6), and further subdividing the space and the data set by using the next-level child nodes (5, 4) and (9, 6), and finally generating the k-d tree of the two-dimensional data point data set S shown in FIG. 4a and the subspace corresponding to the partitioning nodes in the k-d tree shown in FIG. 4 b.
In the embodiment of the invention, the K points closest to the first point or the points with the distance of the first point in the appointed space dimension within the preset distance threshold can be searched by using the constructed K-D tree. The process of retrieving K points adjacent to the first point in the K-D tree is described below using the nearest neighbor search method of the K-D tree as an example.
For ease of understanding, in the description of the embodiments below, in a given partitioning dimension, the value of a data point in the left subtree branch of a K-D tree node is less than the value of the partitioning node, and the value of a point in the right subtree branch of the partitioning node is greater than the value of the partitioning node.
In one embodiment, the basic principle of the nearest neighbor search method of the K-D tree is: and searching the binary tree of the constructed K-D tree by using the first point.
As an example, a value of the first point in the partition dimension is less than or equal to the value of the partition node, a left sub-tree branch of the partition node is entered, and a value of the first point in the partition dimension is greater than or equal to the value of the partition node, a right sub-tree branch of the partition node is entered, so as to obtain a leaf node of the first point nearest to the first point, that is, in the same subspace as the first point.
In one embodiment, in the process of searching for the nearest point of the first point, the nodes in the K-D tree compared with the first point form a search path according to the comparison sequence. After the nearest point of the first point is obtained, a search path can be traced back, whether data points which are closer to the first point exist in other sub-node spaces of the nodes on the search path or not is judged, if the data points which are closer to the first point exist, the sub-nodes are added into the search path, the nearest point of the first point is searched in the other sub-node spaces of the nodes, and the tracing back process of the search path is repeated until the whole search path is traced back, so that the nearest point of the first point is obtained.
As an example, for a first point of the object surface, the step of determining a subset of three-dimensional points adjacent to the first point may specifically comprise:
step S11, sorting the values of the projection positions of the dividing nodes in the space dimension according to the designated sequence, and obtaining a value set of the projection positions corresponding to the sorted dividing nodes.
Step S12, searching for the specified number of spatial projection values with the space between the first point and the projection position in the spatial dimension meeting the preset space condition in the value set of the projection positions by using a binary search method.
And S13, taking a set of three-dimensional points corresponding to the specified number of spatial projection values meeting the preset spacing condition as a determined three-dimensional point subset adjacent to the first point.
Through the above steps, for any point on the surface of the object, a subset of three-dimensional points adjacent to that point can be determined.
In one embodiment, when three-dimensional point cloud data of an object is acquired using a scanning device, an image of the object needs to be captured using an image acquisition device in a three-dimensional scanner. Fig. 5 shows a schematic view of the viewing direction at any point of the surface of the object.
Taking an image acquisition device as an example of a camera, as shown in fig. 5, for a first point of the object surface, the camera imaging of the first point requires the use of the projection plane and the imaging plane of the camera. The projection plane is characterized in that an O point of the projection plane is an optical center of the image acquisition device, namely a projection center, an Xc axis and a Yc axis are parallel to an X axis and a Y axis of an imaging plane coordinate system, a Zc axis is an optical axis of a camera, and the optical axis is perpendicular to the imaging plane. The rectangular coordinate system formed by the point O, the Xc axis, the Yc axis and the Zc axis can be called a camera coordinate system.
As an example, the spatial position coordinates of the first point P in the point cloud data may be noted as P (x p ，y p ，z p ) The corresponding camera view angle coordinates may be noted as C (x c ，y c ，z c ) The vector of the camera towards the first point P can be expressed as
The direction of the camera towards the first point P is denoted as the viewing angle direction corresponding to the first point P.
Next, a specific procedure of determining a normal vector of the first point using shooting angle of view information of the first point will be described in detail with reference to fig. 6.
Fig. 6 shows a flow diagram of a method of processing point cloud data according to an embodiment of the invention. As shown in fig. 6, in an embodiment, the method 100 for processing point cloud data may specifically include:
step S110, determining a subset of three-dimensional points adjacent to a first point in the three-dimensional point cloud data of the object.
In this step, by the method for finding the three-dimensional points adjacent to the first point described in the above embodiment, a specified number of three-dimensional points adjacent to the first point are acquired, and a subset of the three-dimensional points adjacent to the first point is obtained.
Step S120, determining a subsurface of the object at the first point according to the subset of three-dimensional points adjacent to the first point.
In this step, the sub-surface of the object of the first point may be a curved surface or may be a plane, and the plane may be regarded as one of the curved surfaces. Therefore, the subsurface of the object at the first point can be determined using the surface estimation method in a unified manner.
In one embodiment, the near-area surface may be represented by a spatial feature parameter.
As one example, the near-area surface may be represented as ax+by+cz+d=0, for example. Parameter estimation can be performed on the spatial feature parameters of the subsurface of the object of the first point based on the first point and the subset of three-dimensional points adjacent to the first point.
In one embodiment, the estimated spatial feature parameters satisfy subsurface constraints for the object at the first point, the subsurface constraints being: the offset distance of a point in the three-dimensional point subset of the first point to a corresponding point on the subsurface is minimal.
In this example, each point in the three-dimensional subset of points adjacent to the first point may be used as a sampling point and the corresponding point of the sampling point on the determined subsurface may be used as an estimation point, and the subsurface constraint may be further expressed as: the sum of the squares of the distances from each sampling point to the corresponding point on the subsurface is minimal, i.e. the error of the estimated subsurface is minimal.
In one embodiment, the errors of the subsurface determined by the estimated values of the spatial feature parameters may be represented by a covariance matrix of a subset of three-dimensional points adjacent to the first point.
In particular, the covariance matrix of a subset of three-dimensional points adjacent to a first point may be used to measure how far the subset of three-dimensional points adjacent to the first point deviates from its mean in each dimension. That is, the lower the degree to which the subset of three-dimensional points adjacent to the first point deviates from its mean value in each dimension, the smaller the error of the subsurface determined by the estimated value of the spatial feature parameter, and the smaller the variation of the concavity and convexity of the surface of the adjacent area.
Therefore, in the embodiment of the present invention, the covariance matrix of the three-dimensional point subset adjacent to the first point may reflect the surface roughness change of the sub-surface of the first point, that is, the change of the degree of unevenness. As one example, the degree of unevenness of a sub-surface of a first point may be expressed as the magnitude of curvature of the sub-surface of the first point.
Step S130, determining a first axis passing through the first point and perpendicular to the subsurface.
In embodiments of the present invention, the first axis passing through the first point and perpendicular to the subsurface may be determined in a number of ways.
As one example, a tangent plane at a first point of the subsurface is determined; an axis perpendicular to the tangential plane and passing through the first point is determined as a first axis passing through the first point and perpendicular to the subsurface.
As an example, a straight line direction in which the degree of unevenness of the first point on the sub-surface changes most rapidly, that is, in which the curvature changes most rapidly, may be obtained as the axis of the first point on the sub-surface of the object.
In one embodiment, a first axis passing through the first point and perpendicular to the subsurface may be determined by constructing a covariance matrix of a three-dimensional subset of points of the first point.
Specifically, the covariance matrix may be used to characterize the degree of unevenness of the subsurface of the object at the first point, and the minimum eigenvector of the covariance matrix is calculated, where the direction of the straight line where the minimum eigenvector of the covariance matrix is located is the direction in which the subsurface of the object at the first point changes most rapidly along the curvature of the first point.
In this step, the first point and the subset of three-dimensional points adjacent to the first point may be covariance analyzed by a covariance matrix of the subset of three-dimensional points adjacent to the first point. The feature vector corresponding to the minimum feature value of the covariance matrix, that is, the straight line where the minimum feature vector is located, may be used as a first axis passing through the first point and perpendicular to the sub-surface of the object at the first point. That is, the eigenvector corresponding to the minimum eigenvalue of the covariance matrix may represent the magnitude of the normal vector of the first point, and further determining the specific direction of the normal vector of the first point is required.
In step S140, a vector having an acute angle with respect to the viewing angle direction of the first point on the first axis is determined as a normal vector of the first point.
In one embodiment, with the spatial feature parameters A, B, C of the subsurface of the object of the first point, determining the normal vector of the first point with two possible orientations comprises: first vector
And a second vector
Wherein the first vector->
And a second vector->
On the contrary.
In this embodiment, since the vector (a, B, C) is perpendicular to any straight line passing through the first point in the sub-surface, the straight line along which the vector (a, B, C) is located acts as the first axis, the two directions of which are the two possible orientations of the normal vector to the first point.
In one embodiment, a vector multiplication is performed between any two vectors, if the product is greater than 0, the angle between the two vectors is an acute angle, and if the product is less than 0, the angle between the two vectors is an obtuse angle.
Thus, the first vector
And a second vector->
Vector of camera towards first point P, respectively->
Vector multiplication is performed, and a vector with a vector product greater than 0 is used as a normal vector of the first point. />
In the embodiment of the invention, for the point cloud with any structure, the normal vector of the point can be determined through the view angle direction corresponding to any point in the point cloud, and compared with the traditional method, the method has small calculation complexity and high accuracy, and the calculation efficiency of the normal vector of the point cloud is greatly improved.
In order to better understand the present invention, a method for processing point cloud data according to embodiments of the present invention will be described in detail with reference to the accompanying drawings, and it should be noted that these embodiments are not intended to limit the scope of the disclosure of the present invention.
Fig. 7 shows a flowchart of a method for processing three-dimensional point cloud data according to an embodiment of the present invention. As shown in fig. 7, the method 200 for processing three-dimensional point cloud data in the embodiment of the present invention includes the following steps:
step S210, acquiring a three-dimensional point set of the surface of the object, and determining, for a first point of the surface of the object, a subsurface of the object at the first point.
In one embodiment, the step of determining the sub-surface of the object at the first point in step S210 may specifically include:
in step S211, a subset of three-dimensional points adjacent to the first point is selected.
Step S212, determining a surface including the first point and having the smallest offset distance from the points in the three-dimensional point subset as a sub-surface of the object of the first point.
In this embodiment, the points in the three-dimensional point subset are a specified number of three-dimensional points that are closest to the first point and whose distance from the first point satisfies a preset pitch condition; and the first point and the points in the three-dimensional point subset are in the spatial distribution of the appointed dimension, so that the dispersion degree is highest.
In one embodiment, a sub-surface of the object of the first point is determined using a regression analysis method using the first point and a subset of three-dimensional points adjacent to the first point.
Step S220, determining a first axis passing through the first point and perpendicular to the subsurface.
In one embodiment, step S220 may specifically include:
determining a tangent plane at a first point of the sub-surface; an axis perpendicular to the tangential plane and passing through the first point is determined as a first axis passing through the first point and perpendicular to the subsurface.
In one embodiment, step S220 may specifically include:
in step S221, a subset of three-dimensional points adjacent to the first point is selected.
In step S222, a covariance matrix of the three-dimensional point subset of the first points is constructed, where the covariance matrix is used to characterize the degree of subsurface unevenness of the object of the first points.
In step S223, a minimum eigenvector of the covariance matrix is determined, and a straight line where the minimum eigenvector is located is taken as a first axis, wherein the sub-surface of the object at the first point changes most rapidly along the unevenness of the first point.
In step S230, a viewing angle direction corresponding to the first point is obtained, where the projection position of the first point in the corresponding image capturing device points to the direction of the spatial position of the first point.
In one embodiment, step S230 may specifically include:
in step S231, the projection position of the first point in the corresponding image capturing device, and the line vector pointing to the spatial position of the first point are used as the vector of the image capturing device facing the first point.
In step S232, a vector that forms an acute angle with a vector of the image capturing device toward the first point on the first axis is selected as a normal vector of the first point.
In step S240, a vector that forms an acute angle with the viewing angle direction of the first point on the first axis is determined as a normal vector of the first point.
In one embodiment, the normal vector of the first point is a vector having a vector product of greater than 0 with a vector of the image acquisition device directed toward the first point.
According to the processing method of the three-dimensional point cloud data, the normal vector orientation information can be determined through the camera view angle, the calculation complexity is low, the accuracy is high, and the operation efficiency is greatly improved.
Fig. 8 is a schematic structural diagram of a processing device for point cloud data according to an embodiment of the present invention. As shown in fig. 8, the processing apparatus 800 for point cloud data may include:
a subsurface determining module 810, configured to obtain a three-dimensional point set of a surface of an object, and determine, for a first point of the surface of the object, a subsurface of the object at the first point;
An axis determination module 820 for determining a first axis passing through the first point and perpendicular to the subsurface;
a viewing angle direction obtaining module 830, configured to obtain a viewing angle direction corresponding to the first point, where the viewing angle direction is a direction in which a projection position of the first point in the corresponding image capturing device points to a spatial position of the first point;
the normal vector determining module 840 is configured to determine a vector that forms an acute angle with the viewing angle direction of the first point on the first axis, as a normal vector of the first point.
In one embodiment, the subsurface determination module 810 is specifically configured to:
selecting a subset of three-dimensional points adjacent to the first point; a surface including the first point and having a minimum offset distance from points in the three-dimensional point subset is determined as a subsurface of the object of the first point.
In this embodiment, the points in the three-dimensional point subset are a specified number of three-dimensional points that are closest to the first point and whose distance from the first point satisfies a preset pitch condition; and the first point and the points in the three-dimensional point subset are in the spatial distribution of the appointed dimension, so that the dispersion degree is highest.
In one embodiment, the axis determination module 820 may be specifically configured to:
determining a tangent plane at a first point of the sub-surface; an axis perpendicular to the tangential plane and passing through the first point is determined as a first axis passing through the first point and perpendicular to the subsurface.
In one embodiment, the axis determination module 820 may include:
and the adjacent point selection unit is used for selecting a three-dimensional point subset adjacent to the first point.
The covariance matrix construction unit is used for constructing a covariance matrix of the three-dimensional point subset of the first point, and the covariance matrix is used for representing the subsurface unevenness degree of the object of the first point.
And the minimum feature vector determining unit is used for determining a minimum feature vector of the covariance matrix, and taking a straight line where the minimum feature vector is located as a first axis, wherein the change of the sub-surface of the object at the first point along the unevenness degree of the first point is the fastest.
In one embodiment, normal vector determination module 840 includes:
and the connecting line vector determining unit is used for taking the projection position of the first point in the corresponding image acquisition device and the connecting line vector pointing to the space position of the first point as the vector of the image acquisition equipment facing the first point.
And the normal vector selection unit is used for selecting a vector which is on the first axis and forms an acute angle with a vector of the image acquisition equipment towards the first point, and the vector is a normal vector of the first point.
In one embodiment, the normal vector of the first point is a vector having a vector product of greater than 0 with a vector of the image acquisition device directed toward the first point.
Other details of the device for point cloud data according to the embodiment of the present invention are similar to the method for point cloud data according to the embodiment of the present invention described above in connection with fig. 1 to 7, and are not described here again.
The processing method and apparatus of point cloud data according to the embodiments of the present invention described in connection with fig. 1 to 8 may be implemented by a computing device detachably or fixedly installed on an image processing server.
Fig. 9 is a block diagram illustrating an exemplary hardware architecture of a computing device capable of implementing the method and apparatus for processing point cloud data according to an embodiment of the present invention.
As shown in fig. 9, the computing device 900 includes an input device 901, an input interface 902, a central processor 903, a memory 904, an output interface 905, and an output device 906. The input interface 902, the central processor 903, the memory 904, and the output interface 905 are connected to each other through a bus 910, and the input device 901 and the output device 906 are connected to the bus 910 through the input interface 902 and the output interface 905, respectively, and further connected to other components of the computing device 900. Specifically, the input device 901 receives input information from the outside (e.g., a three-dimensional scanner), and transmits the input information to the central processor 903 through the input interface 902; the central processor 903 processes the input information based on computer-executable instructions stored in the memory 904 to generate output information, temporarily or permanently stores the output information in the memory 904, and then transmits the output information to the output device 909 through the output interface 905; the output device 909 outputs the output information to the outside of the computing device 900 for use by the user.
That is, the computing device shown in fig. 9 may also be implemented to include: a memory storing computer-executable instructions; and a processor that, when executing computer-executable instructions, may implement the method and apparatus for processing point cloud data described in connection with fig. 1-8. Here, the processor may communicate with the image capturing apparatus to execute computer-executable instructions based on related information from point cloud data and the like acquired by the three-dimensional scanner, thereby implementing the processing method and apparatus of point cloud data described in connection with fig. 1 to 8.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be embodied in whole or in part in the form of a computer program product or a computer-readable storage medium. The computer program product or computer-readable storage medium includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.
It should be understood that the invention is not limited to the particular arrangements and instrumentality described above and shown in the drawings. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present invention are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after appreciating the spirit of the present invention.
In the foregoing, only the specific embodiments of the present invention are described, and it will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the systems, modules and units described above may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein. It should be understood that the scope of the present invention is not limited thereto, and any equivalent modifications or substitutions can be easily made by those skilled in the art within the technical scope of the present invention, and they should be included in the scope of the present invention.
Claims (8)
1. A method of processing point cloud data, comprising:
Acquiring a three-dimensional point set of the surface of an object, and determining a subsurface of the object at a first point of the surface of the object;
determining a first axis passing through the first point and perpendicular to the subsurface;
acquiring a viewing angle direction corresponding to the first point, wherein the viewing angle direction is a direction in which a projection position of the first point in corresponding image acquisition equipment points to a space position of the first point;
determining a vector on the first axis at an acute angle to the viewing direction of the first point as a normal vector to the first point;
wherein said determining a sub-surface of the object of the first point comprises:
selecting a subset of three-dimensional points adjacent to the first point;
determining a surface comprising the first point and having the smallest offset distance from the points in the three-dimensional point subset as a sub-surface of the object of the first point;
wherein,,
the points in the three-dimensional point subset are specified number of three-dimensional points which are closest to the first point and have the distance from the first point meeting the preset spacing condition; and is also provided with
The first points and the points in the three-dimensional point subset are in spatial distribution of a designated dimension, and the dispersion degree is highest;
Wherein said determining a vector on said first axis at an acute angle to a viewing direction of said first point as a normal vector to said first point comprises:
the projection position of the first point in the corresponding image acquisition device is directed to a connecting line vector of the spatial position of the first point to be used as a vector of the image acquisition equipment facing the first point;
and selecting a vector which is arranged on the first axis and forms an acute angle with a vector of the image acquisition device towards the first point as a normal vector of the first point.
2. The method of processing point cloud data of claim 1, wherein the determining a first axis passing through the first point and perpendicular to the subsurface comprises:
determining a tangent plane at the first point of the subsurface;
an axis perpendicular to the tangent plane and passing through the first point is determined as a first axis passing through the first point and perpendicular to the subsurface.
3. The method of processing point cloud data of claim 1, wherein the determining a first axis passing through the first point and perpendicular to the subsurface comprises:
selecting a subset of three-dimensional points adjacent to the first point;
Constructing a covariance matrix of the three-dimensional point subset of the first point, wherein the covariance matrix is used for representing the uneven degree of the subsurface of the object of the first point;
and determining a minimum eigenvector of the covariance matrix, and taking a straight line where the minimum eigenvector is located as the first axis, wherein the sub-surface of the object of the first point changes the most rapidly along the unevenness degree of the first point.
4. A processing apparatus for point cloud data, comprising:
the subsurface determining module is used for acquiring a three-dimensional point set of the surface of the object, and determining a subsurface of the object at a first point of the surface of the object;
an axis determination module for determining a first axis passing through the first point and perpendicular to the subsurface;
a viewing angle direction obtaining module, configured to obtain a viewing angle direction corresponding to the first point, where the viewing angle direction is a direction in which a projection position of the first point in a corresponding image capturing device points to a spatial position of the first point;
a normal vector determining module, configured to determine a vector on the first axis, which forms an acute angle with a viewing angle direction of the first point, as a normal vector of the first point;
The sub-surface determining module is specifically configured to:
selecting a subset of three-dimensional points adjacent to the first point;
determining a surface comprising the first point and having the smallest offset distance from the points in the three-dimensional point subset as a sub-surface of the object of the first point;
wherein,,
the points in the three-dimensional point subset are specified number of three-dimensional points which are closest to the first point and have the distance from the first point meeting the preset spacing condition; and is also provided with
The first points and the points in the three-dimensional point subset are in spatial distribution of a designated dimension, and the dispersion degree is highest;
wherein, normal vector determination module includes:
a connection line vector determining unit, configured to use a connection line vector of a projection position of the first point in the corresponding image capturing device, which points to a spatial position of the first point, as a vector of the image capturing device facing the first point;
and the normal vector selection unit is used for selecting a vector which is on the first axis and forms an acute angle with a vector of the image acquisition equipment towards the first point, and the vector is a normal vector of the first point.
5. The processing device of point cloud data according to claim 4, wherein the axis determining module is specifically configured to:
Determining a tangent plane at the first point of the subsurface;
an axis perpendicular to the tangent plane and passing through the first point is determined as a first axis passing through the first point and perpendicular to the subsurface.
6. The processing device of point cloud data according to claim 4, wherein the axis determining module comprises:
a neighboring point selection unit configured to select a three-dimensional point subset neighboring the first point;
a covariance matrix construction unit, configured to construct a covariance matrix of the three-dimensional point subset of the first point, where the covariance matrix is used to characterize a subsurface unevenness degree of the object of the first point;
and a minimum feature vector determining unit, configured to determine a minimum feature vector of the covariance matrix, and take a straight line where the minimum feature vector is located as the first axis, where a sub-surface of the object of the first point changes most rapidly along an unevenness degree of the first point.
7. A processing device for point cloud data, which is characterized by comprising a memory and a processor;
the memory is used for storing executable program codes;
the processor is configured to read executable program code stored in the memory to perform the method of processing point cloud data according to any one of claims 1 to 3.
8. A computer readable storage medium comprising instructions which, when run on a computer, cause the computer to perform the method of processing point cloud data according to any of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810575099.6A CN110570511B (en) | 2018-06-06 | 2018-06-06 | Processing method, device and system of point cloud data and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810575099.6A CN110570511B (en) | 2018-06-06 | 2018-06-06 | Processing method, device and system of point cloud data and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110570511A CN110570511A (en) | 2019-12-13 |
| CN110570511B true CN110570511B (en) | 2023-04-28 |
Family
ID=68772624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810575099.6A Active CN110570511B (en) | 2018-06-06 | 2018-06-06 | Processing method, device and system of point cloud data and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110570511B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021253429A1 (en) * | 2020-06-19 | 2021-12-23 | 深圳市大疆创新科技有限公司 | Data processing method and apparatus, and laser radar and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6732746B2 (en) * | 2014-11-26 | 2020-07-29 | アイロボット・コーポレーション | System for performing simultaneous localization mapping using a machine vision system |
| CN106813568B (en) * | 2015-11-27 | 2019-10-29 | 菜鸟智能物流控股有限公司 | Object measuring method and device |
| WO2017113260A1 (en) * | 2015-12-30 | 2017-07-06 | 中国科学院深圳先进技术研究院 | Three-dimensional point cloud model re-establishment method and apparatus |
| CN106423656B (en) * | 2016-08-11 | 2019-04-02 | 重庆大学 | Automatic spraying system and method based on cloud and images match |
| CN107886561B (en) * | 2017-11-30 | 2021-11-09 | 公安部物证鉴定中心 | Crime scene point cloud data-based storage scheduling and display method |
-
2018
- 2018-06-06 CN CN201810575099.6A patent/CN110570511B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110570511A (en) | 2019-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Özyeşil et al. | A survey of structure from motion*. | |
| Ning et al. | An efficient outlier removal method for scattered point cloud data | |
| Dimitrov et al. | Non‐uniform B‐spline surface fitting from unordered 3D point clouds for as‐built modeling | |
| JP5677798B2 (en) | 3D object recognition and position and orientation determination method in 3D scene | |
| Shi et al. | Adaptive simplification of point cloud using k-means clustering | |
| JP5705147B2 (en) | Representing 3D objects or objects using descriptors | |
| Xu et al. | Pm-pm: Patchmatch with potts model for object segmentation and stereo matching | |
| Li et al. | Improved algorithm for point cloud registration based on fast point feature histograms | |
| WO2015001416A1 (en) | Multi-dimensional data clustering | |
| Ma et al. | A mixture model for robust point matching under multi-layer motion | |
| Lu et al. | Automatic point cloud registration algorithm based on the feature histogram of local surface | |
| CN110415339B (en) | Method and device for calculating matching relation between input three-dimensional shapes | |
| Anderson et al. | Delaunay walk for fast nearest neighbor: accelerating correspondence matching for ICP | |
| CN110570511B (en) | Processing method, device and system of point cloud data and storage medium | |
| Du et al. | MDCS with fully encoding the information of local shape description for 3D Rigid Data matching | |
| Schaefer et al. | A maximum likelihood approach to extract finite planes from 3-D laser scans | |
| Chen et al. | A robust non-rigid point set registration algorithm using both local and global constraints | |
| Markovic et al. | Automatic recognition of cylinders and planes from unstructured point clouds | |
| Ylimäki et al. | Accurate 3-d reconstruction with rgb-d cameras using depth map fusion and pose refinement | |
| Zhao et al. | Accurate extraction of building roofs from airborne light detection and ranging point clouds using a coarse-to-fine approach | |
| Oskarsson | Two-view orthographic epipolar geometry: Minimal and optimal solvers | |
| Li et al. | Contracting medial surfaces isotropically for fast extraction of centred curve skeletons | |
| Vanco | A direct approach for the segmentation of unorganized points and recognition of simple algebraic surfaces | |
| Hauenstein et al. | Curvature determination in range images: new methods and comparison study | |
| CN111915732A (en) | Image generation method and system |
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 |