A method and system for constructing a three-dimensional model of a concrete precast component

Abstract

The present application relates to the technical field of three-dimensional construction of prefabricated components, and particularly relates to a method and system for constructing a three-dimensional model of a concrete prefabricated component. The method comprises: obtaining point cloud data corresponding to the concrete prefabricated component; obtaining a sorting sequence corresponding to each point in the point cloud data; obtaining the local variation degree of a point according to the sorting sequence corresponding to the point, and then calculating the required detail degree of the point; determining the K value and the adjacent point set of each point; calculating the curvature value of each point according to the adjacent point set of each point, and obtaining suspected noise points; dividing adjacent suspected noise points into the same connected domain, obtaining a first sequence, and then calculating the structural characteristic value of each suspected noise point; screening the suspected noise points to obtain real noise points; denoising the real noise points in the point cloud data to obtain denoised point cloud data; and constructing a three-dimensional model by using the denoised point cloud data. The present application can improve the accuracy of constructing a three-dimensional model of a concrete prefabricated component.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of three-dimensional construction of prefabricated components, and in particular to a three-dimensional model construction method and system for concrete prefabricated components. BACKGROUND

[0002] The three-dimensional model of a concrete prefabricated component refers to the conversion of information such as the shape and structure of a concrete prefabricated component in reality into a three-dimensional coordinate model through three-dimensional digital technology. Using the three-dimensional model of a concrete prefabricated component can realize digital quality detection, virtual assembly simulation, collision checking, etc. of the prefabricated component, and can significantly improve the precision and efficiency of the component in the entire process of design, manufacturing, construction and operation. The three-dimensional point cloud technology is a commonly used method in the three-dimensional model of a concrete prefabricated component. However, when scanning a concrete prefabricated component, the reflected signal is easily disturbed by other factors, resulting in the point cloud data obtained by scanning often containing noise points. Therefore, denoising the obtained point cloud data is a key step to ensure the accuracy of the constructed three-dimensional model.

[0003] The existing denoising method for point cloud data often cannot effectively distinguish noise points from holes, corners and other structures on the surface of the prefabricated component, resulting in the loss of part of the real structure information on the surface of the prefabricated component after denoising, which affects the accuracy of the results of further quality detection, collision simulation, etc. of the prefabricated component. SUMMARY

[0004] To solve the above technical problems, the purpose of the present application is to provide a three-dimensional model construction method and system for concrete prefabricated components, and the technical solution adopted is as follows:

[0005] In the first aspect, an embodiment of the present application provides a three-dimensional model construction method for a concrete prefabricated component, which comprises:

[0006] scanning the concrete prefabricated component to obtain point cloud data; constructing a window with a point in the point cloud data as the center, and sorting the points in the window to obtain a sorting sequence;

[0007] obtaining the local variation intensity of a point according to the coordinate difference between every two adjacent points in the sorting sequence corresponding to the point; and calculating the required detail degree of the point according to the local variation intensity of the point and the adjacent points in the sorting sequence corresponding to the point;

[0008] determining the K value of each point according to the required detail degree of each point, and respectively obtaining the neighboring point set of each point; calculating the curvature value of each point according to the neighboring point set of each point, and obtaining suspected noise points;

[0009] The adjacent suspected noise points are divided into the same connected domain, and a point on the edge of the connected domain is taken as a starting point to sort the suspected noise points in the connected domain by using a curvature minimum path to obtain a first sequence; and structural characteristic values of the suspected noise points in the connected domain are obtained based on the first sequence;

[0010] The suspected noise points are screened based on the structural characteristic values and the curvature values of the suspected noise points to obtain real noise points; the real noise points in the point cloud data are denoised to obtain denoised point cloud data; and a three-dimensional model is constructed by using the denoised point cloud data.

[0011] Preferably, the points in the window are sorted to obtain a sorting sequence, including:

[0012] A point closest to a top position of the window in a window corresponding to a point is taken as a starting point, distances between each point in the window and the starting point are calculated, and a point with the minimum distance is taken as a second point; distances between other points in the window except the starting point and the second point are calculated, and a point with the minimum distance is taken as a third point, and the sorting of the points in the window is completed by taking the above steps until all the points in the window corresponding to the point are traversed, and a sorting sequence corresponding to the point is obtained.

[0013] Preferably, the local variation degree of a point is obtained according to a coordinate difference between each two adjacent points in a sorting sequence corresponding to the point, including:

[0014] A sum of absolute values of coordinate differences between two adjacent points in a sorting sequence corresponding to a point is obtained, and the sum is denoted as a coordinate difference between the two adjacent points; and a mean value of the coordinate differences between each two adjacent points in the sorting sequence corresponding to the point is obtained to obtain the local variation degree of the point.

[0015] Preferably, a required detail degree of a point is calculated according to the local variation degree of the point and adjacent points in a sorting sequence corresponding to the point, including:

[0016] A sum of distances between the point and adjacent points in a sorting sequence corresponding to the point is obtained, and the sum is multiplied by the local variation degree of the point to obtain the required detail degree of the point.

[0017] Preferably, a K value of each point is determined according to the required detail degree of the point, and a neighbor point set of each point is obtained, including:

[0018] A K value of a point is obtained by multiplying a set initial K value by a first preset value and a difference between the required detail degree of the point, and then taking an integer greater than or equal to the product; and then a neighbor point set of each point is obtained by using the K value of the point and a KNN algorithm.

[0019] Preferably, the curvature value of each point is calculated according to the set of neighboring points of the point, and the suspected noise points are obtained, comprising:

[0020] The covariance matrix is calculated according to the set of neighboring points of a point, and the eigenvalue decomposition is performed on the covariance matrix to obtain a normal eigenvalue, a first eigenvalue and a second eigenvalue; the curvature value of the point is obtained by dividing the normal eigenvalue by the sum of the normal eigenvalue, the first eigenvalue and the second eigenvalue; the curvature values of each point are clustered to obtain different clusters, and the maximum curvature value in the largest cluster is selected as a judgment threshold, and the curvature value of a point is greater than the judgment threshold, the point is a suspected noise point.

[0021] Preferably, the structural eigenvalue of each suspected noise point in the connected domain is obtained based on the first sequence, comprising:

[0022] The difference between the curvature value of the latter suspected noise point and the curvature value of the former suspected noise point in two adjacent suspected noise points in the first sequence corresponding to a connected domain is calculated, and is recorded as the curvature difference value of the latter point in the two adjacent suspected noise points; the curvature difference values of each two adjacent suspected noise points are sorted according to the order of the first sequence to obtain a curvature difference value sequence; the least square method is used to linearly fit the curvature difference value sequence to obtain a fitting curve; the root mean square error of the fitting curve is calculated according to the curvature difference value in the curvature difference value sequence and the corresponding curvature difference value on the fitting curve.

[0023] The reciprocal of the sum of the root mean square error and a first preset value is calculated, and is recorded as a first structural eigenvalue; the absolute value of the difference between the curvature difference value of a suspected noise point in the connected domain and the corresponding curvature difference value of the point on the fitting curve is calculated, and is added to the first preset value and then the reciprocal is calculated to obtain a second structural eigenvalue; the first structural eigenvalue and the second structural eigenvalue are multiplied and normalized to obtain the structural eigenvalue of the suspected noise point in the connected domain.

[0024] Preferably, the suspected noise points are screened based on the structural eigenvalue and the curvature value of each suspected noise point to obtain the real noise points, comprising:

[0025] The reciprocal of the curvature value and the structural eigenvalue of a suspected noise point are multiplied and normalized to obtain the real noise possibility of the suspected noise point; if the real noise possibility of the suspected noise point is greater than a first preset threshold, the suspected noise point is a real noise point.

[0026] In a second aspect, the present application also provides a concrete prefabricated component three-dimensional model construction system, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the computer program implements the steps of a concrete prefabricated component three-dimensional model construction method when executed by the processor.

[0027] The embodiment of the present application has at least the following beneficial effects: the point cloud data of the prefabricated concrete component is obtained, a window is constructed with a point in the point cloud data as the center, the points in the window are sorted to obtain a sorting sequence, and then the required detail degree of each point is calculated according to the sorting sequence corresponding to each point, and the K value of each point in the curvature calculation is adaptively adjusted (the neighboring point set is obtained by using the K nearest neighbor method), so that the interference of the noise points on the non-noise points is minimized, the characteristics of the detail positions such as the hole edges and the component corners are also preserved as much as possible, and the noise points are prevented from being smoothed out, which improves the accuracy of noise point identification and provides more accurate data for subsequent prefabricated component structure analysis.

[0028] After the suspected noise points are obtained through the initial screening, the adjacent suspected noise points are divided into the same connected domain, a point on the edge of the connected domain is taken as the starting point, the suspected noise points in the connected domain are sorted by using the minimum curvature path to obtain a first sequence, the structural characteristic values of the suspected noise points in the connected domain are obtained, the suspected noise points are screened based on the structural characteristic values and the curvature values of the suspected noise points to obtain the real noise points, the structural characteristics of the positions of the suspected noise points are combined to further distinguish the real noise points and the structural points on the component, the real structure on the prefabricated component is prevented from being smoothed out, more component characteristic information can be preserved, and the accuracy of the denoising processing is improved; finally, the real noise points in the point cloud data are denoised to obtain denoised point cloud data, and the accuracy of subsequent three-dimensional modeling is improved. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0030] Figure 1 A method flowchart of a prefabricated concrete component three-dimensional model construction method provided by the embodiment of the present application;

[0031] Figure 2 A window schematic diagram of a prefabricated concrete component three-dimensional model construction method provided by the embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined object of the application, the following describes in detail the specific implementation, structure, features and effects of the three-dimensional model construction method and system of a concrete prefabricated component according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0034] The specific scheme of the three-dimensional model construction method and system of a concrete prefabricated component provided by the embodiments of the present application is specifically described below in combination with the accompanying drawings.

[0035] Embodiment 1:

[0036] The main application scenarios of the present application are:

[0037] In the scanning process of the concrete prefabricated component, the laser scanning method is often used. When the surface of the concrete prefabricated component is rough, and the scanning process is shaken, etc., noise points may be contained in the obtained point cloud data. Because the noise points and the holes and corners of the prefabricated component surface have similar characteristics, the traditional denoising method will blur the corner and hole information of the prefabricated component, so that the accuracy of the constructed three-dimensional model is insufficient. The present application distinguishes the noise points from the surface hole and corner points according to the structural feature difference between the noise points and the real prefabricated component surface points, filters out the real noise points, and performs noise reduction processing, so as to improve the accuracy of the three-dimensional model construction.

[0038] Please refer to Figure 1 which shows a method flowchart of the three-dimensional model construction method of a concrete prefabricated component provided by the embodiments of the present application. The method comprises the following steps:

[0039] Step S1, scanning the concrete prefabricated component to obtain point cloud data; constructing a window with a point in the point cloud data as the center, and sorting the points in the window to obtain a sorting sequence.

[0040] The concrete prefabricated component is scanned from multiple angles by using a laser scanner, the laser reflected by the surface of the prefabricated component is received, the distance between each collection point and the scanner is calculated, the spatial coordinates of the collection points on the prefabricated component are generated, and the original point cloud data of the prefabricated component is obtained. Before scanning, the scanning precision related parameters are set according to the size of the prefabricated component and the detection requirements, and here the angular resolution is set to 0.1° and the point spacing is set to 0.5mm according to experience. During the scanning process, it is necessary to ensure that there is no other object to block, and the scanner is placed on the ground as flat as possible to reduce the shaking during the scanning process. Thus, the point cloud data of the concrete prefabricated component is obtained.

[0041] Because the noise points are caused by unstable changes of reflected laser due to interference factors such as reflection and shaking, the noise point coordinates change greatly compared with other normal point coordinates, therefore, the curvature detection is used to obtain points with large curvature first, i.e. points with large change degree relative to the neighborhood surface, as suspected noise points. Because the surface of the concrete prefabricated component usually has holes, corners and other structures, the points at these positions also show the characteristics of large curvature, so the suspected noise points obtained may also include points at the hole and corner positions, and direct denoising processing may blur the structure information of these positions, so it is necessary to screen out the real noise points from the suspected noise points. Because the holes and corners are the real structures of the prefabricated component, i.e. the points at the corresponding positions have certain structural characteristics, and the noise points have great randomness in causes and manifestations, i.e. the noise points do not have structural characteristics, so the real noise points are screened out according to the structural characteristic values of the positions of the suspected noise points.

[0042] The curvature of the points in the collected point cloud data reflects the geometric change degree of the points relative to the surrounding points, the normal surface of the concrete prefabricated component is relatively flat, so the curvature of the corresponding points is relatively low, and the noise points usually appear floating in the air or drilling into the object, obviously deviating from the real surface of the prefabricated component, so the curvature of the noise points is relatively large, therefore, the noise manifestation degree of each point is calculated according to the size of the curvature.

[0043] Firstly, for each point, the corresponding neighboring point set is obtained by using the KNN algorithm. Because the selection of K value directly affects the curvature value obtained by calculation, if the K value is too large, the curvature will be smoothed, and the edges, corners, detailed structures and noise points of the component cannot be accurately identified, and if the K value is too small, the curvature will be seriously affected by the noise, and the normal points near the noise points may be misidentified as noise, therefore, the K value needs to be adaptively adjusted according to the distribution characteristics of each point and the surrounding points.

[0044] Thus, the initial K value is set to 30 according to the current scanning accuracy. Further, the required detail degree of each point needs to be calculated according to the variation degree of the coordinates of the position where the point is located. Specifically, a window is constructed with a point as the center, the window is a three-dimensional window, and the size of the window is preferably 9x9x9. If a point is located at the position of the edge, the detail structure of the component, the coordinate variation of the points in the corresponding window is relatively large, while the coordinate variation of the points at the flat position is usually continuous variation with small amplitude or continuous variation of the coordinate value in only one direction. The construction of the window needs to be constructed in a unified manner, for example, the upper and lower surfaces and the side surfaces of all windows are respectively parallel to the XOY plane, the XOZ plane and the YOZ plane of the three-dimensional rectangular coordinate system, as shown in Figure 2 the figure, X, Y and Z respectively represent the three coordinate axes, O represents the coordinate origin, and the cube in the figure is the constructed window, and the upper and lower surfaces and the side surfaces thereof are respectively parallel to the XOY plane, the XOZ plane and the YOZ plane of the three-dimensional rectangular coordinate system.

[0045] Thus, the points in the window need to be sorted. Specifically, one of the points closest to the upper vertex position of the window in the window is taken as the starting point, the distances between each point in the window and the starting point are calculated, and the point with the smallest distance is taken as the second point. The distances between the other points in the window and the second point are calculated, and the point with the smallest distance is taken as the third point. In this way, the sorting of the points in the window is completed, and the sorting sequence corresponding to the window is obtained, that is, the sorting sequence corresponding to the point.

[0046] It should be noted that the previously sorted points do not participate in subsequent sorting, that is, the sorting is not put back. In addition, if a point corresponds to multiple points with the smallest distance, one of the points with the smallest distance can be taken. After obtaining the sorting sequence corresponding to each point, the variation degree of the points around each point is analyzed according to the sorting sequence.

[0047] In step S2, the local variation degree of a point is obtained according to the coordinate difference between each two adjacent points in the sorting sequence corresponding to the point, and the required detail degree of the point is calculated according to the local variation degree of the point and the adjacent points in the sorting sequence corresponding to the point.

[0048] In step S1, the sorting sequence corresponding to each point in the point cloud data is obtained, and then the local variation degree of each point is analyzed in the sequence. The local variation degree of a point is obtained according to the coordinate difference between each two adjacent points in the sorting sequence corresponding to the point.

[0049] Specifically, the sum of the absolute values of the coordinate differences of two adjacent points in the permutation sequence corresponding to the point is obtained, denoted as the coordinate difference of the two adjacent points, and the average of the coordinate differences of every two adjacent points in the permutation sequence corresponding to the point is calculated to obtain the local change intensity of the point.

[0050] The calculation model of the local change intensity is specifically:

[0051] ,

[0052] represents the local change intensity corresponding to the i-th point, that is, the coordinate change intensity of the points in the window corresponding to the i-th point; represents the number of points contained in the window corresponding to the i-th point, that is, the number of points in the permutation sequence; represents the coordinate of the j-th point in the permutation sequence corresponding to the i-th point, represents the coordinate of the j+1-th point in the permutation sequence corresponding to the i-th point, represents the absolute value of the difference between the coordinates of the two adjacent points, that is, the X-axis coordinates, and the greater the value, the more intense the change of the coordinates of the points in the window, that is, the stronger the local change, represents the coordinate difference between two adjacent points in the permutation sequence, that is, the coordinate difference between the j-th point and the j+1-th point in the permutation sequence corresponding to the i-th point; because the number of points contained in the window corresponding to each point may be different, the average value of the coordinate change degree of the adjacent points in the window is used to represent the overall coordinate change intensity of the window. Thus, the local change intensity of each point in the point cloud data can be obtained.

[0053] Because the single noise point appearing on the relatively flat plane of the component may have a low coordinate change intensity of other points in the corresponding window, in order to avoid smoothing the curvature of the noise point when calculating the curvature, the corresponding K value should also be reduced, and therefore, the required detail degree of the point is calculated in combination with the distance of the current analysis point relative to the adjacent points before and after it.

[0054] Specifically, the sum of the distances of the adjacent points before and after the point in the permutation sequence corresponding to the point from the point is obtained, multiplied by the local change intensity of the point, and then normalized to obtain the required detail degree of the point.

[0055] The calculation model of the required detail degree is specifically:

[0056] ,

[0057] represents the required detail degree of the i-th point, represents the local change intensity corresponding to the i-th point, and​​ respectively, the greater the sum of the two distances, the greater the mutation degree of the ith point relative to its adjacent points, i.e. it is likely to be noise data, in order to avoid being smoothed out, the point needs to obtain a more accurate curvature value; norm represents a normalization function.

[0058] Thus, the required detail degree of each point can be obtained, when obtaining the corresponding neighbor point set of each point by using the KNN algorithm, the greater the required detail degree of the point, the smaller the corresponding K value should be.

[0059] Step S3, determine the K value of each point according to the required detail degree of each point, and obtain the neighbor point set of each point respectively; calculate the curvature value of each point according to the neighbor point set of each point, and obtain the suspected noise point.

[0060] The greater the local variation of a point, the greater the distance between the adjacent points in its corresponding sorting sequence, which indicates that it needs more attention and more attention to its details, so when obtaining the neighbor point set of the point by using the KNN algorithm, the corresponding K value should be smaller.

[0061] Further, the K value of each point is determined according to the required detail degree of each point. Specifically, the initial K value set is multiplied by the difference between the first preset value and the required detail degree of a point and rounded up to obtain the K value of the point.

[0062] The calculation model of the K value of each point is:

[0063] ,

[0064] wherein, K(i) represents the K value of the ith point, K0 represents the initial K value set, D(i) represents the required detail degree of the ith point, the first preset value is 1, D(i) represents the required detail degree of the ith point, the first preset value is 1, .

[0065] Then, the neighbor point set of each point is obtained by using the K value of each point combined with the KNN algorithm.

[0066] ​After obtaining the set of neighboring points of each point, a covariance matrix is calculated according to the set of neighboring points of a point, and eigenvalue decomposition is performed on the covariance matrix to obtain a normal eigenvalue, a first eigenvalue and a second eigenvalue; a curvature value of the point is obtained by dividing the normal eigenvalue by a sum of the normal eigenvalue, the first eigenvalue and the second eigenvalue.

[0067] The calculation model of the curvature value is:

[0068]

[0069] wherein, denotes the curvature value of the i th point, and respectively denote the normal eigenvalue, the first eigenvalue and the second eigenvalue corresponding to the i th point, and the three eigenvalues represent the spatial distribution of the point in different principal directions, wherein corresponds to the principal normal direction, i.e. the minimum direction.

[0070] The curvature value of a noise point relative to a collection point is higher, and since most positions on the surface of the concrete prefabricated component have higher flatness, i.e. the corresponding curvature value is smaller and closer, the curvature values of each point are clustered by using the DBSCAN clustering algorithm to obtain different clusters, and the maximum curvature value in the largest cluster is selected as a judgment threshold, i.e. when the curvature value of a point is greater than the judgment threshold, the point is considered as a suspected noise point. Thus, the suspected noise points in the point cloud data are obtained.

[0071] In step S4, adjacent suspected noise points are divided into the same connected domain, and a point on the edge of the connected domain is taken as a starting point to sort the suspected noise points in the connected domain by using the minimum curvature path to obtain a first sequence; and structural eigenvalues of the suspected noise points in the connected domain are obtained based on the first sequence.

[0072] Since the curvature values of the corners, hole edges and other positions of the prefabricated component are also higher than those of other flat positions, the suspected noise points obtained above may also include points at the corners, hole edges and other positions, and these points reflect the size, shape and quality characteristics of the prefabricated component, so the characteristics of these points should be preserved as much as possible during denoising, i.e. it is necessary to distinguish noise points from points of the real structure on the prefabricated component. Although the curvature of the corners, hole edges and other positions is higher than that of other flat positions, since these points are real structures on the prefabricated component, these points have strong structural characteristics with their surrounding points, while noise points have strong randomness and do not have structural characteristics, so according to the structural eigenvalues of the suspected noise points, noise points and non-noise points are distinguished.

[0073] ​​Firstly, adjacent suspected noise points are divided into the same connected domain, because the curvature of the preform edge point is continuously changed, while the curvature value of the noise point is random, so the structural characteristics of each connected domain are calculated according to the difference degree of the curvature change between adjacent points in the connected domain.

[0074] Further, the curvature gradient minimum path is used to generate the edge trend curve starting from one point on the edge of the connected domain, and the points on the connected domain are sorted according to the curvature gradient minimum path, that is, the path connecting the points with continuously changed curvature on the surface of the component, so as to obtain the sorted sequence of the suspected noise points in a connected domain, that is, the first sequence. It should be noted that this part is prior art and will not be described in detail here.

[0075] Thus, the first sequence corresponding to each connected domain composed of adjacent suspected noise points can be obtained. Further, the difference between the curvature value of the latter suspected noise point and the curvature value of the former suspected noise point in the first sequence corresponding to a connected domain is calculated, denoted as the curvature difference value of the latter point between two adjacent suspected noise points; the curvature difference values of each two adjacent suspected noise points are sorted according to the order of the first sequence to obtain the curvature difference value sequence; the least square method is used to linearly fit the curvature difference value sequence to obtain the fitting curve; and the root mean square error of the fitting curve is calculated according to the curvature difference value in the curvature difference value sequence and the corresponding curvature difference value on the fitting curve.

[0076] If a suspected noise point is adjacent to an edge point, it will be divided into the same connected domain, because the structural characteristics of other points are strong, which may weaken the randomness of the suspected noise point, so the structural characteristic value of each suspected noise point is calculated according to the difference degree of the curvature difference value corresponding to each suspected noise point relative to the corresponding point on the fitting curve.

[0077] Specifically, the reciprocal of the sum of the root mean square error and the first preset value is calculated, denoted as the first structural characteristic, the absolute value of the difference between the curvature difference value of a suspected noise point in the connected domain and the corresponding curvature difference value on the fitting curve is calculated, and the first preset value is added and then the reciprocal is obtained to obtain the second structural characteristic, the first structural characteristic and the second structural characteristic are multiplied and normalized to obtain the structural characteristic value of the suspected noise point in the connected domain. Wherein, the curvature difference value corresponding to the first point in the first sequence is the curvature difference value of the second point in the first sequence.

[0078] The calculation model of the structural characteristic value is specifically:

[0079] ,

[0080] Wherein, The structural characteristic value of the ith suspected noise point is denoted as represents the i-th suspected noise point, represents the root mean square error value of the fitting curve after fitting the curvature difference value sequence corresponding to the connected domain where the i-th suspected noise point is located, is a first structural feature, indicating that the smaller the root mean square error value corresponding to the connected domain where the i-th suspected noise point is located, the stronger the structure of the connected domain, i.e. the stronger the structure of the suspected noise point, and the first preset value 1 is added to ensure that the fraction is meaningful; represents the curvature difference value of the i-th suspected noise point, represents the curvature difference value of the i-th suspected noise point on the corresponding fitting curve, is a second structural feature, indicating that the closer the actual curvature difference value of the point to the fitting value, the stronger the structure of the point; norm represents a normalization operation. Thus, the structural feature value of each point can be obtained.

[0081] In step S5, the suspected noise points are screened based on the structural feature values and curvature values of the suspected noise points to obtain real noise points; the real noise points in the point cloud data are denoised to obtain denoised point cloud data; and a three-dimensional model is constructed using the denoised point cloud data.

[0082] In step S4, the structural feature of the suspected noise point is analyzed based on the structure of the preform, and the structural feature value of each suspected noise point is obtained. Because the curvature value of the noise point is often higher than that of the edge point, and the structure is poor, the curvature value and the structural feature value of the suspected noise point are combined to calculate the possibility of the suspected noise point being a real noise point.

[0083] The suspected noise points are screened based on the structural feature values and curvature values of the suspected noise points to obtain real noise points. Specifically, the reciprocal of the curvature value and the structural feature value of a suspected noise point are multiplied and normalized to obtain the real noise possibility of the suspected noise point.

[0084] The greater the curvature value and the smaller the structural feature value indicate that the suspected noise point is more likely to be a real noise point. Because the structural difference between the noise point and the edge point is large, i.e. the real noise possibility of the noise point and the edge point is large. Therefore, the median 0.5 of the value range of the real noise possibility can be set as a first preset threshold value, and if the real noise possibility of the suspected noise point is greater than the first preset threshold value, the suspected noise point is a real noise point.

[0085] After the real noise points in the point cloud data are obtained based on the above steps, in order to avoid the problem that direct deletion of the noise points may cause incomplete structure, the noise points are smoothed by using a local surface projection (MLS) method. The search radius is set to 3 mm, and the fitting surface order is set to 2. The three-dimensional model of the concrete prefabricated component is constructed according to the denoised point cloud data.

[0086] To sum up, the curvature detection method is used to identify the suspected noise points in the data, and then the structural characteristics of the suspected noise points are analyzed according to the structural characteristics of the surface holes and corners of the prefabricated component and the random and irregular characteristics of the noise points. Finally, the real noise points are identified and denoised by comprehensively considering the noise performance degree and the structural characteristic value of the suspected noise points. The noise data in the point cloud data of the concrete prefabricated component can be effectively processed, and the accuracy of the three-dimensional model construction of the concrete prefabricated component is improved.

[0087] Embodiment 2

[0088] The embodiment provides a three-dimensional model construction system of a concrete prefabricated component, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program implements the steps of a three-dimensional model construction method of a concrete prefabricated component when executed by the processor. Since the three-dimensional model construction method of a concrete prefabricated component has been described in detail in Embodiment 1, it will not be described in detail here.

[0089] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or may be advantageous.

[0090] Each embodiment in the present specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the differences from other embodiments.

[0091] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for constructing a three-dimensional model of a precast concrete component, characterized in that, The method includes: Scan the precast concrete components to obtain point cloud data; construct a window centered on a point in the point cloud data, and sort the points within the window to obtain a sorted sequence; The degree of local change of a point is obtained by the difference in coordinates between every two adjacent points in the sorted sequence corresponding to that point; the required level of detail for that point is calculated based on the degree of local change of that point and the adjacent points in its sorted sequence. Determine the K value for each point based on the required level of detail, and obtain the set of neighboring points for each point; calculate the curvature value for each point based on the set of neighboring points, and obtain suspected noise points; Adjacent suspected noise points are divided into the same connected domain. Starting from a point on the edge of the connected domain, the suspected noise points in the connected domain are sorted using the path of minimum curvature to obtain the first sequence. Based on the first sequence, the structural feature values ​​of each suspected noise point in the connected domain are obtained. Based on the structural feature values ​​and curvature values ​​of each suspected noise point, the suspected noise points are screened to obtain the real noise points; the real noise points in the point cloud data are denoised to obtain denoised point cloud data; and a 3D model is constructed using the denoised point cloud data. The step of obtaining the structural feature values ​​of each suspected noise point in the connected component based on the first sequence includes: Calculate the difference between the curvature value of the latter suspected noise point and the curvature value of the former suspected noise point in the first sequence corresponding to a connected component. This difference is denoted as the curvature difference of the latter point in the first sequence. Sort the curvature difference values ​​of every two adjacent suspected noise points according to the order of the first sequence to obtain a curvature difference value sequence. Use the least squares method to perform linear fitting on the curvature difference value sequence to obtain a fitting curve. Calculate the root mean square error of the fitting curve based on the curvature difference values ​​in the curvature difference value sequence and the corresponding curvature difference values ​​on the fitting curve. The reciprocal of the sum of the root mean square error and the first preset value is taken and denoted as the first structural feature. The absolute value of the difference between the curvature difference of a suspected noise point in the connected domain and the curvature difference of the suspected noise point on the fitted curve is taken and added to the first preset value and then the reciprocal is taken to obtain the second structural feature. The first structural feature and the second structural feature are multiplied and normalized to obtain the structural feature value of the suspected noise point in the connected domain.

2. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The step of sorting the points within the window to obtain a sorted sequence includes: Starting from the point closest to the top vertex of the window corresponding to a given point, calculate the distance between the starting point and all other points within the window. Select the point with the smallest distance as the second point. Calculate the distance between the second point and all other points within the window, except for the starting point and the second point. Select the point with the smallest distance as the third point, and so on, until all points within the window corresponding to that point have been traversed. This process sorts the points within the window and yields the sorted sequence for that point.

3. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The step of obtaining the degree of local change of a point based on the coordinate difference between every two adjacent points in the sorted sequence corresponding to that point includes: To obtain the sum of the absolute values ​​of the differences between the coordinates of two adjacent points in the permutation sequence corresponding to a given point, we record the coordinate difference between the two adjacent points. We then calculate the mean of the coordinate differences between any two adjacent points in the permutation sequence corresponding to that point to obtain the degree of drastic local change at that point.

4. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The calculation of the required level of detail for a point based on the degree of local change and its adjacent points in the sorted sequence includes: The sum of the distances between a point and its immediate and next adjacent points in its corresponding sorted sequence is obtained, multiplied by the degree of local change of the point, and then normalized to obtain the required level of detail for the point.

5. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The process of determining the K value for each point based on the required level of detail, and obtaining the neighbor set for each point, includes: The initial K value is multiplied by the difference between the first preset value and the required level of detail for a point, and then rounded up to obtain the K value for that point. Then, the K value of each point is combined with the KNN algorithm to obtain the set of neighboring points for each point.

6. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The step of calculating the curvature value of each point based on the set of neighboring points and obtaining suspected noise points includes: The covariance matrix is ​​calculated based on the set of neighboring points of a point, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain the normal eigenvalue, the first eigenvalue, and the second eigenvalue. The curvature value of the point is obtained by dividing the normal eigenvalue by the sum of the normal eigenvalue, the first eigenvalue, and the second eigenvalue. The curvature values ​​of each point are clustered to obtain different clusters. The maximum curvature value in the largest cluster is selected as the judgment threshold. When the curvature value of a point is greater than the judgment threshold, the point is a suspected noise point.

7. The method for constructing a three-dimensional model of a precast concrete component according to claim 1, characterized in that, The process of filtering suspected noise points based on their structural feature values ​​and curvature values ​​to obtain actual noise points includes: The true noise probability of a suspected noise point is obtained by multiplying the curvature value and the reciprocal of its structural feature value and normalizing the result. If the true noise probability of a suspected noise point is greater than a first preset threshold, then the suspected noise point is a true noise point.

8. A three-dimensional model building system for precast concrete components, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for constructing a three-dimensional model of a precast concrete component as described in any one of claims 1-7.

Images