Point cloud circle center identification method and device

Abstract

The invention relates to the technical field of point cloud circle center recognition, and provides a point cloud circle center recognition method and device, and the method comprises the steps: collecting point cloud data, and selecting a reference point from the point cloud data; calculating a first distance from each point in the point cloud data to the reference point, and screening out points of which the difference between the first distance and the circumferential radius is smaller than a preset distance threshold to obtain a first point set; screening out points with a first distance between the ring inner diameter and the ring outer diameter in the first point set to obtain a second point set; defining a first error function based on a least square method according to the first point set, obtaining an arc sphere center coordinate estimated value based on the first point set according to the first error function, defining a second error function based on the least square method according to the second point set, and obtaining a ring sphere center coordinate estimated value based on the second point set according to the second error function, and obtaining the center of the point cloud data according to the circular arc center coordinate estimation value and the corresponding error, and the circular ring center coordinate estimation value and the corresponding error.

Description

Technical Field

[0001] The present invention relates to the technical field of point cloud circle center recognition, and in particular to a point cloud circle center recognition method and device. Background Art

[0002] Point cloud data is a collection of a large number of three-dimensional points, usually collected by laser scanners or cameras. In many application fields, such as industrial inspection and object recognition, point cloud data needs to be analyzed and processed. Among them, identifying the center of a circle is an important task. For example, in workpiece inspection, identifying the center of a circle can help locate the position and posture of the workpiece.

[0003] Traditional point cloud circle center recognition methods are usually based on fitting circles or extracting local features. Common techniques include least squares fitting and Hough transform. These methods can identify the center of a circle in a point cloud to a certain extent, but they perform poorly when dealing with a lot of noise or uneven point cloud density. In addition, these methods are often sensitive to initial parameters, require precise parameter settings and adjustments, and are slow when processing large-scale point cloud data.

[0004] Traditional point cloud circle center recognition has the following defects: Sensitivity to noise: Traditional methods are easily disturbed when facing point cloud data with more noise, resulting in inaccurate recognition results; Difficulty in parameter selection: These methods usually need to adjust multiple parameters and are sensitive to parameter selection, which limits their applicability in different scenarios; Slow speed: Traditional methods are slow when processing large-scale point cloud data and cannot meet real-time requirements. Summary of the invention

[0005] In view of the above-mentioned defects, the purpose of the present invention is to propose a point cloud circle center recognition method and device, which aims to accurately identify the center of the circle in the point cloud in various complex environments, improve the accuracy and stability of circle center recognition, and complete the recognition of the center of the circle in the point cloud in a shorter time.

[0006] To achieve this object, the present invention adopts the following technical solutions:

[0007] A point cloud circle center recognition method, the point cloud circle center recognition method comprising:

[0008] Collect point cloud data, select reference points in the point cloud data, and set the circle radius, circle inner diameter and circle outer diameter related to the point cloud data;

[0009] Calculate a first distance from each point in the point cloud data to the reference point, select points whose difference between the first distance and the circle radius is less than a preset distance threshold, and obtain a first point set, wherein the distance threshold is set according to the measurement error and the discreteness of the point cloud;

[0010] Filter out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set;

[0011] A first error function based on the least squares method is defined according to the first point set, and an estimated value of the coordinates of the center of the arc sphere based on the first point set is obtained according to the first error function. A second error function based on the least squares method is defined according to the second point set, and an estimated value of the coordinates of the center of the circle sphere based on the second point set is obtained according to the second error function. The center of the point cloud data is obtained according to the estimated values ​​of the coordinates of the center of the arc sphere and the corresponding errors and the estimated values ​​of the coordinates of the center of the circle sphere and the corresponding errors.

[0012] Preferably, the inner diameter and outer diameter of the ring are set according to the density in the point cloud data.

[0013] Preferably, after collecting the point cloud data, the method further includes removing outliers and downsampling the collected point cloud data.

[0014] Preferably, after obtaining the second point set, the method further includes calculating the normal vector of each point in the first point set and the second point set;

[0015] Let the point in the first point set and the second point set be P, obtain the neighborhood point set of P and set it to {P1,P2,…,P n};

[0016] Fit the plane ax+by+cz+d=0 by the least squares method, and select When the plane normal vector is minimum, as the normal vector of point P.

[0017] Preferably, after obtaining the second point set, the method further includes calculating the curvature of each point in the first point set and the second point set;

[0018] Let the points in the first point set and the second point set be Q, and the coordinate matrix of the domain points of Q be X;

[0019] Calculate the covariance matrix of the domain points to satisfy the relationship: Where n represents the number of domain points, represents the average coordinates of the neighborhood points;

[0020] The curvature of point Q is calculated from the eigenvalues ​​of the covariance matrix.

[0021] Preferably, a first error function based on the least squares method is defined according to the first point set to satisfy the relationship:

[0022]

[0023] Among them, (a, b, c) represents the estimated coordinates of the arc center, r represents the radius of the circle, and A, B, C, D represent the plane equation A of the arc with radius r. x +B y +C z +D=0 equation parameter, d pl,i Indicates the distance from the i-th point in the first point set to the plane where the arc is located, d sp,i Represents the distance from the i-th point in the first point set to the coordinates of the arc center.

[0024] Preferably, a second error function based on the least squares method is defined according to the second point set to satisfy the relationship:

[0025]

[0026] Among them, r represents the radius of the circle, (a, b, c) represents the estimated coordinates of the center of the circle, r1 represents the inner diameter of the ring, r2 represents the outer diameter of the ring, and A, B, C, D represent the plane equation A of the ring with r1 as the inner diameter and r2 as the outer diameter. x +B y +C z +D=0 equation parameter, d pl,i represents the distance from the i-th point in the second point set to the plane where the ring is located, d sp,i Represents the distance from the i-th point in the second point set to the coordinates of the center of the circle.

[0027] Preferably, the center of the point cloud data is obtained based on the estimated value of the arc sphere center coordinates and the corresponding error and the estimated value of the circle sphere center coordinates and the corresponding error to satisfy the relationship:

[0028]

[0029] Among them, (a, b, c) represents the center of the point cloud data, (a1, b1, c1) represents the estimated coordinates of the arc center, (a2, b2, c2) represents the estimated coordinates of the circle center, e1 represents the error of the estimated coordinates of the arc center, and e2 represents the error of the estimated coordinates of the circle center.

[0030] A point cloud circle center recognition device, the point cloud circle center recognition device is applied to the point cloud circle center recognition method as described above, the point cloud circle center recognition device comprises:

[0031] A preprocessing module is used to collect point cloud data, select reference points in the point cloud data, and set the circumference radius, inner diameter of the circle and outer diameter of the circle related to the point cloud data;

[0032] A first screening module is used to calculate a first distance from each point in the point cloud data to the reference point, and screen out points whose difference between the first distance and the circle radius is less than a preset distance threshold to obtain a first point set, wherein the distance threshold is set according to a measurement error and a discreteness of the point cloud;

[0033] A second screening module is used to screen out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set;

[0034] An identification module is used to define a first error function based on the least squares method according to a first point set, obtain an estimated value of the coordinates of the center of an arc sphere based on the first point set according to the first error function, define a second error function based on the least squares method according to a second point set, obtain an estimated value of the coordinates of the center of a circle sphere based on the second point set according to the second error function, and obtain the center of the point cloud data according to the estimated value of the coordinates of the center of the arc sphere and the corresponding error and the estimated value of the coordinates of the center of the circle sphere and the corresponding error.

[0035] One of the above technical solutions has the following advantages or beneficial effects:

[0036] The present invention can focus on the local area related to the center of the circle by collecting point cloud data and setting the circumference radius and the inner and outer diameters of the ring, reducing the amount of calculation and improving efficiency. Then, by extracting the characteristic points of the circumference and the ring and combining the preset distance threshold, the algorithm can effectively remove the interference of noise and irrelevant points, significantly improving the accuracy of circle center recognition; arc and ring fitting based on the least squares method, and the center estimation method combining the two fitting results, further enhance the robustness of the present invention to noise, while meeting the real-time requirements. The present invention is not only suitable for point cloud data of various types and densities, but also has a wide range of versatility and applicability, and can significantly improve the efficiency and reliability of point cloud data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0038] Figure 1 is a flow chart of a point cloud circle center recognition method provided by an embodiment of the present invention;

[0039] Figure 2 It is a structural schematic diagram of a point cloud circle center recognition device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0041] In the present invention, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0042] A point cloud circle center recognition method, such as Figure 1 As shown, in a preferred embodiment of the present invention, the point cloud circle center recognition method comprises the following steps:

[0043] S1: Collect point cloud data, select reference points in the point cloud data, and set the circle radius, inner diameter of the circle, and outer diameter of the circle for the point cloud data;

[0044] In step S1, point cloud data is collected and the scope of local feature extraction is defined to provide a basis for subsequent circle center recognition. Point cloud data consists of points in three-dimensional space and can be collected by laser scanners or cameras. A reference point is selected as the center of local feature extraction, and the circle radius, inner diameter and outer diameter of the ring are set to limit the scope of subsequent processing, reduce the amount of calculation and improve the pertinence of the algorithm. It can effectively focus on the area related to circle center recognition and exclude irrelevant information, laying the foundation for subsequent steps.

[0045] Point cloud data is a collection of points in three-dimensional space. As the input data for circle center recognition, it provides geometric information of the circle center to be recognized. The reference point is the center point of local feature extraction, which is used to define the range of the circumference and the ring. A point with a larger curvature can be selected because it is more likely to be close to the center of the circle. The radius of the circumference is used to define the range of the circumference and to filter out points that may belong to the circumference. The inner and outer diameters of the ring define the range of the ring and are used to filter out points that may belong to the ring. The parameter setting can effectively limit the range of local feature extraction, reduce the interference of noise and irrelevant points, and improve the efficiency and accuracy of the algorithm.

[0046] When implementing this step, point cloud data can be collected in a variety of ways, such as using a high-precision laser scanner to collect point cloud data, which is suitable for scenarios with high precision requirements such as industrial inspection; or using an ordinary camera combined with a depth sensor to collect point cloud data, which is suitable for cost-sensitive application scenarios. When selecting a reference point, you can calculate the curvature of the point cloud and select the point with the largest curvature as the reference point; or use the center of mass of the point cloud as the reference point, which is suitable for situations where the point cloud is relatively evenly distributed. When setting the circle radius and the inner and outer diameters of the ring, you can dynamically adjust according to the density and distribution of the point cloud, or use fixed values ​​as parameters, which is suitable for scenarios with known target sizes.

[0047] By collecting point cloud data and setting relevant parameters, this step provides a basis for subsequent circle center recognition. It can effectively limit the scope of local feature extraction, reduce noise and interference from irrelevant points, and improve the efficiency and accuracy of the algorithm. At the same time, by flexibly selecting reference points and setting parameters, the algorithm can adapt to point cloud data of different densities and types, enhance the versatility and applicability of the algorithm, lay a solid foundation for subsequent circle center recognition, and ensure that the algorithm can run stably in complex environments.

[0048] S2: Calculate a first distance from each point in the point cloud data to the reference point, select points whose difference between the first distance and the circle radius is less than a preset distance threshold, and obtain a first point set, wherein the distance threshold is set according to the measurement error and the discreteness of the point cloud;

[0049] By calculating the distance from each point in the point cloud to the reference point, points close to the circle radius are screened out to form the first point set. The preset distance threshold is used to tolerate measurement errors and the discreteness of the point cloud, ensuring that the screened points are closer to the ideal circular structure. It can effectively remove the interference of noise and irrelevant points, improve the accuracy of subsequent fitting, and provide high-quality input data for subsequent arc fitting.

[0050] The first distance refers to the Euclidean distance from each point in the point cloud to the reference point, which is used to determine whether the point is close to the circumference. The preset distance threshold is a parameter used to tolerate measurement errors and point cloud discreteness, ensuring that the screened points are closer to the ideal circumference. The first point set is the screened point set, which can be regarded as points within the circle, including points that may belong to the circumference, providing input for subsequent arc fitting.

[0051] When implementing this step, the first distance can be calculated in a variety of ways, such as using the Euclidean distance formula to calculate the distance from each point to the reference point, or using a fast nearest neighbor search algorithm (such as KD tree) to accelerate the distance calculation process. When screening the first point set, a fixed distance threshold can be used for screening, which is suitable for evenly distributed point clouds; or the distance threshold can be dynamically adjusted to screen according to the local density of the point cloud, which is suitable for non-uniformly distributed point clouds, which can effectively improve the efficiency and accuracy of screening and adapt to different types of point cloud data.

[0052] S3: Filter out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set;

[0053] Step S3 is to further filter out the points within the ring range to form a second point set. Through double screening, the characteristic points of the circumference and the ring are extracted respectively to provide more accurate input for the subsequent center estimation. This process can further reduce the interference of noise and irrelevant points, improve the stability and accuracy of the algorithm, and provide high-quality input data for the subsequent ring fitting. The second point set can be regarded as a collection of points within the ring.

[0054] In this step, the second point set can be screened in a variety of ways, such as using fixed values ​​as the inner and outer diameters of the ring, which is suitable for scenes with known target sizes; or dynamically adjusting the inner and outer diameters of the ring to screen according to the local density and distribution of the point cloud, which is suitable for complex scenes. In addition, the normal vector and curvature information of the point can be combined to further improve the accuracy of the screening process and reduce the interference of noise points, which can effectively improve the efficiency and accuracy of the screening and adapt to different types of point cloud data.

[0055] S4: Define a first error function based on the least squares method according to the first point set, obtain an estimated value of the coordinates of the center of the arc sphere based on the first point set according to the first error function, define a second error function based on the least squares method according to the second point set, obtain an estimated value of the coordinates of the center of the circle sphere based on the second point set according to the second error function, and obtain the center of the point cloud data according to the estimated value of the coordinates of the center of the arc sphere and the corresponding error and the estimated value of the coordinates of the center of the circle sphere and the corresponding error.

[0056] The arc and the ring are fitted by the least squares method to obtain the estimated coordinates of the center of the arc and the ring. The least squares method optimizes the fitting results by minimizing the error function, improves the accuracy of the center estimation, and finally obtains the final center coordinates by comprehensively considering the two fitting results. This process can effectively make up for the shortcomings of a single fitting method, improve the stability and applicability of the algorithm, and provide the final result for the center identification.

[0057] Among them, the least squares method is a mathematical optimization method. The arc and the ring are fitted by minimizing the error function to obtain the estimated value of the sphere center coordinates. The first error function and the second error function respectively define the error measurement criteria for arc and ring fitting, which are used to optimize the fitting results. The arc sphere center coordinate estimate and the sphere sphere center coordinate estimate represent the circle center coordinate estimate based on arc and ring fitting respectively. Finally, the final circle center coordinates are obtained by weighted averaging or other optimization methods. The weighted average is a method that comprehensively considers the two fitting results. In this step, it plays the role of optimizing fitting and comprehensive estimation to ensure that the final circle center recognition result has high accuracy and high robustness.

[0058] Preferably, the inner diameter and outer diameter of the ring are set according to the density in the point cloud data.

[0059] Specifically, by analyzing the density distribution of point cloud data, the size of the inner and outer diameters of the ring are dynamically adjusted to adapt to point cloud data of different densities. The point cloud density reflects the sparse distribution of points in the point cloud. By adjusting the inner and outer diameters according to the density, the circular feature points can be extracted more accurately, thereby improving the accuracy of circle center recognition. This dynamic adjustment mechanism can reduce the problem of inaccurate feature extraction caused by uneven point cloud density and enhance the robustness of the algorithm.

[0060] Preferably, after collecting the point cloud data, the method further includes removing outliers and downsampling the collected point cloud data.

[0061] Specifically, the collected point cloud data is preprocessed to improve the efficiency and accuracy of subsequent processing. Outlier removal can eliminate abnormal data points caused by measurement errors or noise, and reduce the interference of these points on subsequent center recognition. Downsampling reduces the number of points in the point cloud and reduces the computational complexity while retaining the main geometric features of the point cloud. Therefore, it can significantly improve the efficiency and robustness of the algorithm and provide higher quality input data for subsequent center recognition.

[0062] Preferably, after obtaining the second point set, the method further includes calculating the normal vector of each point in the first point set and the second point set;

[0063] Let the point in the first point set and the second point set be P, obtain the neighborhood point set of P and set it to {P1,P2,…,P n};

[0064] Fit the plane ax+by+cz+d=0 by the least squares method, and select When the plane normal vector is minimum, as the normal vector of point P.

[0065] Specifically, the normal vector is a vector perpendicular to the local surface where the point is located, which can reflect the local geometric structure of the point cloud. By fitting the plane by the least squares method, the normal vector of the point can be obtained, thereby providing more accurate local features for circle center recognition. Specifically, in the process of circle center recognition, points on the circumference usually have similar normal vector directions because they are all located on the same plane. By calculating the normal vector, it is possible to more accurately identify which points belong to the circumference or ring structure, thereby improving the accuracy of circle center recognition. For example: when fitting an arc, the normal vector can help determine the plane where the arc is located, making the fitting process more accurate. When screening feature points, the normal vector can be used as a constraint to exclude points whose normal vector directions are significantly different from the expected ones and reduce noise interference.

[0066] Preferably, after obtaining the second point set, the method further includes calculating the curvature of each point in the first point set and the second point set;

[0067] Let the points in the first point set and the second point set be Q, and the coordinate matrix of the domain points of Q be X;

[0068] Calculate the covariance matrix of the domain points to satisfy the relationship: Where n represents the number of domain points, represents the average coordinates of the neighborhood points;

[0069] The curvature of point Q is calculated from the eigenvalues ​​of the covariance matrix.

[0070] In this step, curvature is a feature that describes the local shape of the point cloud, reflecting the degree of curvature of the point cloud in the local area. The coordinate matrix of the neighborhood points is the coordinate set of the points around point Q, which is used to calculate the curvature of point Q. The covariance matrix is ​​a statistic that describes the distribution of neighborhood points. By calculating the eigenvalues ​​of the covariance matrix, the curvature of point Q can be obtained. The eigenvalues ​​are the eigenvalues ​​of the covariance matrix. Their sizes reflect the distribution of the point cloud in different directions, so they can be used to calculate the curvature. For example, if the eigenvalue of C is calculated as λ1≥λ2≥λ3, then the curvature of point Q is

[0071] In general, the calculation of curvature can provide more geometric information for subsequent circle center recognition, and the calculation of curvature can improve the accuracy of circle center recognition, especially when the point cloud density is uneven or there is noise. By calculating the curvature through the covariance matrix method, more accurate local features can be obtained, thereby improving the stability of circle center recognition.

[0072] Preferably, a first error function based on the least squares method is defined according to the first point set to satisfy the relationship:

[0073]

[0074] Among them, (a, b, c) represents the estimated coordinates of the arc center, r represents the radius of the circle, and A, B, C, D represent the plane equation A of the arc with radius r. x +B y +C z +D=0 equation parameter, d pl,i Indicates the distance from the i-th point in the first point set to the plane where the arc is located, d sp,i Represents the distance from the i-th point in the first point set to the coordinates of the arc center.

[0075] The first error function consists of two parts: the square of the distance from the point to the plane and the square of the difference between the distance from the point to the center of the sphere and the radius. By minimizing this error function, the coordinate estimate of the center of the arc can be obtained, thereby providing more accurate local features for circle center recognition.

[0076] In this embodiment, in the three-dimensional space, the plane equation where the arc (i.e., the circle with the circumference radius as the radius, where the first point set is located) is located can be expressed as A x +B y +C z +D=0, where is the normal vector of the plane, and A 2 +B 2 +C 2 =1, the three-dimensional arc can be regarded as the intersection of the sphere and the plane, and the equation of the sphere is (xa) 2 +(yb) 2 +(zc) 2 =r 2 , where (a, b, c) are the coordinates of the center of the sphere, r is the radius, and for n points on the circumference, the error function E is defined (a,b,c,r,A,B,C,D) , point (x i ,y i ,z i ) to plane A x +B y +C z +D=0 distance d pl,i =|Ax i +By i +Cz i +D|, point (x i ,y i ,z i ) to the center of the sphere (a,b,c) Then the error function is: Then find the partial derivatives and solve them: find the partial derivatives of a, b, c, r, A, B, C, and D respectively and set them to zero, and get a set of nonlinear equations:

[0077]

[0078]

[0079] In actual calculations, the Levenberg-Marquardt algorithm is used to solve this nonlinear equation set and obtain the estimated coordinates of the arc sphere center (a, b, c).

[0080] Preferably, a second error function based on the least squares method is defined according to the second point set to satisfy the relationship:

[0081]

[0082] Among them, r represents the radius of the circle, (a, b, c) represents the estimated coordinates of the center of the circle, r1 represents the inner diameter of the ring, r2 represents the outer diameter of the ring, and A, B, C, D represent the plane equation A of the ring with r1 as the inner diameter and r2 as the outer diameter. x +B y +C z +D=0 equation parameter, d pl,i represents the distance from the i-th point in the second point set to the plane where the ring is located, d sp,i Represents the distance from the i-th point in the second point set to the coordinates of the center of the circle.

[0083] In this embodiment, the three-dimensional ring can be regarded as the intersection of two concentric spheres and a plane. Assume that the radius of the inner sphere is r1, the radius of the outer sphere is r1, the coordinates of the center of the sphere are (a, b, c), and the plane equation is A x +B y +C z +D=0, for m points in the circular point cloud subset, point (x j ,y j ,z j ) to plane A x +B y +C z +D=0 distance d pl,i =|Ax j +By j +Cz j +D|.

[0084] In addition, the point (x j ,y j ,z j ) to the center of the sphere Then the error function is:

[0085] Similarly, partial derivatives of a, b, c, r1, r2, A, B, C, and D are calculated and set to zero to obtain a more complex set of nonlinear equations. In this embodiment, the Levenberg-Marquardt algorithm is also used to solve the set of equations to obtain the coordinates of the center of the three-dimensional ring (a, b, c), the inner and outer radii r1 and r2, and the normal vector of the plane where the ring is located. and constant D.

[0086] Preferably, the center of the point cloud data is obtained based on the estimated value of the arc sphere center coordinates and the corresponding error and the estimated value of the circle sphere center coordinates and the corresponding error to satisfy the relationship:

[0087]

[0088] Among them, (a, b, c) represents the center of the point cloud data, (a1, b1, c1) represents the estimated coordinates of the arc center, (a2, b2, c2) represents the estimated coordinates of the circle center, e1 represents the error of the estimated coordinates of the arc center, and e2 represents the error of the estimated coordinates of the circle center.

[0089] The final coordinates of the center of the point cloud data are obtained by comprehensively considering the estimated coordinates of the arc center and the estimated coordinates of the sphere center, as well as their respective errors. The idea of ​​weighted average is used, where the weight is determined by the error of each estimate. The estimate with a smaller error will have a greater impact on the final result, thereby improving the accuracy and stability of the center recognition. The weighted average method can effectively combine the two fitting results and improve the accuracy and stability of the center recognition. By selecting a suitable optimization algorithm or machine learning method, the accuracy and adaptability of the center recognition can be further improved, providing higher quality input data for the center recognition, significantly improving the efficiency and accuracy of the algorithm, especially when processing large-scale point cloud data.

[0090] A point cloud circle center recognition device, the point cloud circle center recognition device is applied to the point cloud circle center recognition method as described above, such as Figure 2 As shown, the point cloud center recognition device includes:

[0091] Preprocessing module 1, used for collecting point cloud data, selecting reference points in the point cloud data, and setting the circumference radius, inner diameter of the ring and outer diameter of the ring related to the point cloud data;

[0092] A first screening module 2 is used to calculate a first distance from each point in the point cloud data to the reference point, and screen out points whose difference between the first distance and the circumference radius is less than a preset distance threshold to obtain a first point set, wherein the distance threshold is set according to the measurement error and the discreteness of the point cloud;

[0093] A second screening module 3 is used to screen out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set;

[0094] Identification module 4 is used to define a first error function based on the least squares method according to the first point set, obtain an estimated value of the coordinates of the center of the arc sphere based on the first point set according to the first error function, define a second error function based on the least squares method according to the second point set, obtain an estimated value of the coordinates of the center of the circle sphere based on the second point set according to the second error function, and obtain the center of the point cloud data according to the estimated value of the coordinates of the center of the arc sphere and the corresponding error and the estimated value of the coordinates of the center of the circle sphere and the corresponding error.

[0095] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0096] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A point cloud circle center recognition method, characterized in that: The point cloud circle center recognition method comprises: Collect point cloud data, select reference points in the point cloud data, and set the circle radius, circle inner diameter and circle outer diameter related to the point cloud data; Calculate a first distance from each point in the point cloud data to the reference point, select points whose difference between the first distance and the circle radius is less than a preset distance threshold, and obtain a first point set, wherein the distance threshold is set according to the measurement error and the discreteness of the point cloud; Filter out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set; A first error function based on the least squares method is defined according to the first point set, and an estimated value of the coordinates of the center of the arc sphere based on the first point set is obtained according to the first error function. A second error function based on the least squares method is defined according to the second point set, and an estimated value of the coordinates of the center of the circle sphere based on the second point set is obtained according to the second error function. The center of the point cloud data is obtained according to the estimated values ​​of the coordinates of the center of the arc sphere and the corresponding errors and the estimated values ​​of the coordinates of the center of the circle sphere and the corresponding errors.

2. The point cloud circle center recognition method according to claim 1, characterized in that: The inner diameter and outer diameter of the ring are set according to the density in the point cloud data.

3. The point cloud circle center recognition method according to claim 1, characterized in that: After collecting the point cloud data, it also includes removing outliers and downsampling the collected point cloud data.

4. The point cloud circle center recognition method according to claim 1, characterized in that: After obtaining the second point set, the method further includes calculating the normal vector of each point in the first point set and the second point set; Let the point in the first point set and the second point set be P, obtain the neighborhood point set of P and set it to {P1,P2,…,P n }; Fit the plane ax+by+cz+d=0 by the least squares method, and select When the plane normal vector is minimum, as the normal vector of point P.

5. The point cloud circle center recognition method according to claim 1, characterized in that: After obtaining the second point set, the curvature of each point in the first point set and the second point set is calculated; Let the points in the first point set and the second point set be Q, and the coordinate matrix of the domain points of Q be X; Calculate the covariance matrix of the domain points to satisfy the relationship: Where n represents the number of domain points, represents the average coordinates of the neighborhood points; The curvature of point Q is calculated from the eigenvalues ​​of the covariance matrix.

6. The point cloud circle center recognition method according to claim 1, characterized in that: The first error function based on the least squares method is defined according to the first point set to satisfy the relationship: Among them, (a, b, c) represents the estimated coordinates of the arc center, r represents the radius of the circle, and A, B, C, D represent the plane equation A of the arc with radius r. x +B y +C z +D=0 equation parameter, d pl,i Indicates the distance from the i-th point in the first point set to the plane where the arc is located, d sp,i Represents the distance from the i-th point in the first point set to the coordinates of the arc center.

7. The point cloud circle center recognition method according to claim 1, characterized in that: The second error function based on the least squares method is defined according to the second point set to satisfy the relationship: Among them, r represents the radius of the circle, (a, b, c) represents the estimated coordinates of the center of the circle, r1 represents the inner diameter of the ring, r2 represents the outer diameter of the ring, and A, B, C, D represent the plane equation A of the ring with r1 as the inner diameter and r2 as the outer diameter. x +B y +C z +D=0 equation parameter, d pl,i represents the distance from the i-th point in the second point set to the plane where the ring is located, d sp,i Represents the distance from the i-th point in the second point set to the coordinates of the center of the circle.

8. The point cloud circle center recognition method according to claim 1, characterized in that: According to the estimated coordinates of the arc center and the corresponding error and the estimated coordinates of the circle center and the corresponding error, the center of the point cloud data satisfies the relationship: Among them, (a, b, c) represents the center of the point cloud data, (a1, b1, c1) represents the estimated coordinates of the arc center, (a2, b2, c2) represents the estimated coordinates of the circle center, e1 represents the error of the estimated coordinates of the arc center, and e2 represents the error of the estimated coordinates of the circle center.

9. A point cloud circle center recognition device, the point cloud circle center recognition device is applied to the point cloud circle center recognition method according to any one of claims 1 to 9, characterized in that: The point cloud center recognition device comprises: A preprocessing module is used to collect point cloud data, select reference points in the point cloud data, and set the circumference radius, inner diameter of the circle and outer diameter of the circle related to the point cloud data; A first screening module is used to calculate a first distance from each point in the point cloud data to the reference point, and screen out points whose difference between the first distance and the circle radius is less than a preset distance threshold to obtain a first point set, wherein the distance threshold is set according to a measurement error and a discreteness of the point cloud; A second screening module is used to screen out points in the first point set whose first distance is between the inner diameter of the ring and the outer diameter of the ring to obtain a second point set; An identification module is used to define a first error function based on the least squares method according to a first point set, obtain an estimated value of the coordinates of the center of an arc sphere based on the first point set according to the first error function, define a second error function based on the least squares method according to a second point set, obtain an estimated value of the coordinates of the center of a circle sphere based on the second point set according to the second error function, and obtain the center of the point cloud data according to the estimated value of the coordinates of the center of the arc sphere and the corresponding error and the estimated value of the coordinates of the center of the circle sphere and the corresponding error.

Images