A method for detecting the flatness of concrete components used in highway construction

By calculating the possibility of flatness anomalies and normal roughness in the point cloud data of concrete components, the feature vector is constructed, and the problem of interference from the detection results in the prior art is solved, and the accuracy of flatness detection of concrete components is improved.

CN118730031BActive Publication Date: 2024-11-12GUOJI CONSTR ENG CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411230456.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-11-12
Estimated Expiration
2044-09-04

AI Technical Summary

Technical Problem

The prior art is susceptible to interference from abnormal data points in the flatness detection of concrete components, resulting in inaccurate fitting planes and inaccurately reflecting the plane conditions of concrete components, affecting the accuracy of the detection results.

Method used

By obtaining point cloud data of concrete components, calculate the possibility of flat anomalies and normal roughness of each data point, build feature vectors, determine in-site and external points, improve the representativeness of the fitting plane, and reduce the defects of unevenness. Effect on flatness detection.

Benefits of technology

The accuracy of flatness detection of concrete components is improved, the impact of uneven defects on the overall flatness is reduced, and the reliability of the detection results is enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118730031B_ABST
    Figure CN118730031B_ABST
Patent Text Reader

Abstract

The present application relates to the field of flatness detection technology, and specifically to a method for detecting the flatness of concrete components for highway construction, including: obtaining point cloud data of concrete components for highway construction; obtaining the flatness anomaly probability of each data point in the point cloud data; obtaining the feature vector of each data point according to the normal roughness probability and the corrected flatness anomaly probability of each data point; obtaining the preferred degree of each data point according to the feature vector; and performing flatness detection on the concrete component according to the preferred degree. The present application constructs a feature vector by analyzing the features of point cloud data to obtain the flatness detection result of the concrete structure, thereby improving the accuracy of the flatness detection result of the concrete component during highway construction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of flatness detection, and in particular to a method for detecting the flatness of concrete components used in highway construction. Background Art

[0002] Concrete components are an indispensable and important part of highway construction, such as highway bridge construction. Therefore, concrete components need to be safety tested, and flatness is an important indicator for evaluating the quality of highway bridge construction, reflecting the flatness of the road surface or bridge wall. Concrete components that do not meet the flatness standards will cause highway buildings or bridges to deform or tilt, affecting the safety and stability of the overall construction. Therefore, it is important and necessary to check the flatness of concrete components.

[0003] The prior art usually uses the least square fitting method to obtain the fitting plane for the flatness detection of concrete components during highway construction, and determines the flatness of the concrete components according to the distance relationship of the fitting plane. However, this method is a global fitting based on all data points of the concrete component, which is easily interfered by abnormal data points, resulting in inaccurate fitting plane. Therefore, it is necessary to use the RANSAC (Random Sample Consensus) algorithm to select some data points for obtaining the fitting plane. If the point cloud data of the concrete component contains a large number of data points that appear uneven, the data points that appear as defects will be erroneously added to the local points, resulting in the plane fitted according to the local points being obtained by fitting a large number of abnormal data, and thus failing to accurately reflect the plane condition of the concrete component, resulting in inaccurate flatness detection results. Summary of the invention

[0004] In order to solve the above problems, the present application provides a method for detecting the flatness of concrete components for highway construction, the method comprising:

[0005] Acquire point cloud data of a concrete component, wherein the point cloud data includes a plurality of data points;

[0006] Obtain a neighborhood data point set for each data point; obtain the leveling anomaly possibility of each data point based on the distribution of data points in the neighborhood data point set of each data point;

[0007] According to the flattening anomaly probability and the position distribution difference between the data points in the neighborhood data point set of each data point, the normal roughness probability of each data point is obtained; according to the flattening anomaly probability and the distribution similarity between the data points in the neighborhood data point set of each data point, the corrected flattening anomaly probability of each data point is obtained; according to the corrected flattening anomaly probability and the normal roughness probability, the characteristic vector of each data point is obtained;

[0008] The flatness of concrete components is detected based on the feature vector.

[0009] Preferably, the specific method of obtaining the neighborhood data point set of each data point includes:

[0010] Preset a neighborhood parameter For any data point in the point cloud data of the concrete component, a neighborhood parameter is preset with the data point as the window center point. For any data point in the point cloud data of the concrete component, take the data point as the center point and construct a point cloud data with a radius of The spherical region is taken as the neighborhood data point set of each data point.

[0011] Preferably, the method of obtaining the possibility of leveling anomaly of each data point according to the distribution of data points in the neighborhood data point set of each data point includes the following specific methods:

[0012] For any data point in the point cloud data of the concrete component, a neighborhood data point set of the data point is fitted using a minimum quadratic surface fitting method to obtain a neighborhood fitting surface of the data point, and the Gaussian curvature of all data points in the neighborhood data point set of the data point is obtained through the neighborhood fitting surface according to Weyl's curvature theorem;

[0013] The variance of the Gaussian curvature of all data points in the neighborhood data point set of the data point is recorded as the first variance; the mean of the Gaussian curvature of all data points in the neighborhood data point set of the data point is recorded as the first mean; and the ratio of the first mean to the first variance is used as the possibility of smoothing anomaly of the data point.

[0014] Preferably, the normal roughness possibility of each data point is obtained according to the flattening anomaly possibility and the position distribution difference between data points in the neighborhood data point set of each data point, and the specific method includes:

[0015] For any data point in the point cloud data of the concrete component, any data point in the neighborhood data point set of the data point is recorded as the target data point; the vertical distance from the target data point to the neighborhood fitting surface of the data point is taken as the neighborhood data distance of the target data point;

[0016] The ratio of the variance of the neighborhood data distances of all data points in the neighborhood data point set of the data point to the total number of all data points in the neighborhood data point set of the data point is recorded as the first ratio; the variance of the neighborhood data distances of all data points in the neighborhood data point set of the data point is recorded as the first ratio. The probability of a flattened anomaly of a data point is related to the probability of the first The product of the distances of the neighborhood data of the data point is recorded as the first product of the data points; the cumulative sum of the first products of all data points in the neighborhood data point set of the data point is recorded as the first cumulative sum; the sum of the first cumulative sum and the first ratio is recorded as the first sum value; the inverse proportional value of the first sum value is used as the normal rough possibility of the data point.

[0017] Preferably, the method of obtaining the corrected leveling anomaly possibility of each data point according to the leveling anomaly possibility and the distribution similarity between data points in the neighborhood data point set of each data point includes the following specific methods:

[0018] For any data point in the point cloud data of the concrete component, record any data point in the neighborhood data point set of the data point as the target data point; obtain the plane similarity of the target data point;

[0019] According to the plane similarity and the possibility of flattening anomalies, the corrected anomaly possibility factor of each data point is obtained;

[0020] The corrected anomaly possibility factors of all data points in the point cloud data are obtained, and each corrected anomaly possibility factor after linear normalization of all corrected anomaly possibility factors is recorded as the corrected flattening anomaly possibility.

[0021] Preferably, the obtaining of the plane similarity of the target data point includes the following specific methods:

[0022] Obtain the positive normal vector of the neighborhood fitting surface corresponding to the data point and the positive normal vector of the neighborhood fitting surface corresponding to the target data point; and use the cosine similarity between the positive normal vector of the neighborhood fitting surface corresponding to the target data point and the positive normal vector of the neighborhood fitting surface corresponding to the data point as the plane similarity of the target data point.

[0023] Preferably, the method of obtaining the corrected abnormal possibility factor of each data point according to the plane similarity and the possibility of flattening abnormality includes the following specific methods:

[0024] Obtaining a position constant of a neighborhood fitting surface corresponding to the data point and a position constant of a neighborhood fitting surface corresponding to a target data point;

[0025] The reciprocal of the sum of the minimum absolute value of the plane similarity of all data points in the neighborhood data point set of the data point and the constant parameter is recorded as the first reciprocal; the reciprocal of the sum of the minimum absolute value of the plane similarity of all data points in the neighborhood data point set of the data point is recorded as the first reciprocal; The absolute value of the difference between the position constant of the neighborhood fitting surface corresponding to the data point and the position constant of the neighborhood fitting surface corresponding to the data point is recorded as The first absolute value of the difference of the data points; the first absolute value of the difference of the neighborhood data points of the data point The sum of the plane similarity of data points and the constant parameter is used as the denominator; The ratio of the absolute value of the first difference of the data point to the denominator is recorded as The second ratio of the data points; The second ratio of the data point to the first The product of the smoothing anomaly probability of data points is recorded as the second product of the data point; the cumulative sum of the second products of all data points in the neighborhood data point set of the data point is recorded as the second cumulative sum; the product of the second cumulative sum and the first reciprocal is used as the corrected abnormal possibility factor of the data point.

[0026] Preferably, the method of obtaining the feature vector of each data point according to the possibility of correcting the abnormal smoothing and the possibility of normal roughness includes the following specific methods:

[0027] For any data point in the point cloud data of the concrete component, the difference between 1 and the corrected flatness anomaly possibility of the data point is recorded as the flatness anomaly feature of the data point; the vector formed by the flatness anomaly feature and the normal roughness possibility of the data point is used as the feature vector of the data point.

[0028] Preferably, the flatness detection of the concrete component according to the characteristic vector includes the following specific methods:

[0029] According to the feature vector, obtain the preference degree of each data point;

[0030] Preset a threshold parameter and a quantity parameter For any data point in the point cloud data of the concrete component, if the preference degree of the data point is less than the threshold parameter , the data point is recorded as an outlier point; if the ratio of the total number of all outliers in the point cloud data of the concrete component to the total number of all data points in the point cloud data of the concrete component is greater than or equal to the quantity parameter , the flatness test results of building concrete components are marked as unqualified.

[0031] Preferably, the specific method of obtaining the preference degree of each data point according to the feature vector is as follows:

[0032] The reciprocal of the vertical distance from the data point to the neighborhood fitting surface is recorded as the second reciprocal; the product of the modulus of the eigenvector of the data point and the second reciprocal is used as the optimization factor of the data point;

[0033] The preferred factors of all data points in the point cloud data are obtained, and each preferred factor after linear normalization of all preferred factors is recorded as the preferred degree.

[0034] The beneficial effects of the technical solution of this application are:

[0035] First, the present application obtains the corrected flatness anomaly possibility of each data point according to the flatness anomaly possibility and the distribution similarity between data points in the neighborhood data point set of each data point; utilizes the corrected flatness anomaly possibility; then, obtains the characteristic vector of each data point according to the corrected flatness anomaly possibility and the normal roughness possibility, determines the in-liquid point and the conditions for adding the data point to the in-liquid point through the characteristic vector of the data point, and improves the representativeness of the in-liquid point to the plane of the concrete component; then, determines the preference degree of each data point based on the characteristic vector of each data point and the vertical distance of each data point to the neighborhood fitting surface; so as to facilitate the subsequent calculation of the flatness detection result of the concrete component; finally, calculates the flatness detection result of the concrete component during the highway construction process based on the preference degree of all data points; Flatness anomaly possibility Flatness anomaly possibility Flatness anomaly possibility Normal roughness possibility The present application constructs a characteristic vector by processing point cloud data, so that the plane obtained by fitting the in-liquid point can better reflect the actual situation of the plane of the concrete component, reduces the influence of uneven defects on the overall flatness of the concrete component, and thus improves the accuracy of the flatness detection result of the concrete component. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flowchart of the steps of a method for detecting the flatness of concrete components used in highway construction is provided for this application;

[0038] Figure 2 This application discloses a specific implementation flow chart of a method for detecting the flatness of concrete components used in highway construction. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the method for detecting the flatness of concrete components for highway construction proposed in the present application, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0041] The specific scheme of the method for detecting the flatness of concrete components for highway construction provided by the present application is described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1 , which shows a flowchart of a method for detecting the flatness of a concrete component for highway construction provided by an embodiment of the present application, the method comprising the following steps:

[0043] Step S001: Acquire point cloud data of concrete components during highway construction.

[0044] During the highway construction process, concrete components are one of the construction materials that can effectively speed up the progress of highway construction. Considering that the flatness detection accuracy and detection efficiency of concrete components will affect the progress of subsequent highway construction. Therefore, in this application, the three-dimensional point cloud data of the concrete component is first obtained, and the point cloud data includes a number of data points. The use of point cloud data for detection improves the detection efficiency, and the flatness data of large-area components can be quickly obtained, saving time and labor costs, and there is no need to contact the surface of the component, which can avoid possible damage or interference to the surface of the component, and also reduces the influence of human factors and subjective factors on the flatness judgment results. Among them, the acquisition method of the point cloud data includes but is not limited to time-of-flight sensor acquisition, structured light projection acquisition, and laser scanner acquisition. The acquisition of point cloud data is a well-known technology, and the specific process will not be repeated.

[0045] At this point, the point cloud data of the concrete component is obtained through the above method, which is used for the subsequent calculation of the flatness detection result.

[0046] Step S002: Obtain the leveling anomaly probability of each data point in the point cloud data.

[0047] In this application, a preliminary judgment of flatness is made based on Gaussian curvature information between point cloud data of concrete components.

[0048] It should be noted that if there are obvious uneven defects in the concrete component, the point cloud data will have more obvious Gaussian curvature information. If the concrete component appears to be a flat plane, all data points with the same single dimension in the point cloud data should be on the same plane. Multiple planes constitute regular point cloud data, and the point cloud data points will hardly have obvious Gaussian curvature. Therefore, the Gaussian curvature information of the data point is used to judge the possibility of flatness abnormality of the data point.

[0049] It should be further explained that if there are uneven defects in concrete components, there will be a certain degree of concave and convex areas, and the Gaussian curvature information of the data points can be used to describe this feature; however, the surface of the concrete component is not absolutely smooth, but has a certain degree of roughness, which is manifested as the presence of normal concave and convex areas, resulting in some data points that normally show roughness to obtain a larger Gaussian curvature during the calculation process, making the Gaussian curvature information of the data points inaccurate; because the area of ​​the concave and convex defects is larger than that of the normal rough areas, all data points in the neighborhood data point set of the data points that appear in the concave and convex defect area are also data points that show the characteristics of concave and convex defects. Therefore, the difference in Gaussian curvature between the neighborhood data point sets of the data points is used to determine the possibility of flatness anomalies of the data points.

[0050] Preset a neighborhood parameter , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It depends on the specific implementation situation.

[0051] Specifically, for any data point in the point cloud data of the concrete component, take each data point as the center point and construct a point cloud data with a radius of A spherical region is defined, and a data point set formed by all data points in the spherical region is used as a neighborhood data point set of each data point; the neighborhood data point set of the data point is fitted by using a minimum quadratic surface fitting method to obtain a neighborhood fitting surface of the data point, and the Gaussian curvature of all data points in the neighborhood data point set of the data point is obtained by using the neighborhood fitting surface and Weyl's curvature theorem.

[0052] Among them, the minimum quadratic surface fitting method and Weyl's curvature theorem are existing technologies, and are not described in detail in this embodiment.

[0053] Furthermore, the possibility of a flattening anomaly for each data point is obtained according to the distribution of data points in a neighborhood data point set of each data point.

[0054] As an example, the calculation method for obtaining the possibility of flatness anomaly of each data point in the point cloud data of a concrete component is as follows:

[0055] For any data point in the point cloud data of the concrete component, the variance of the Gaussian curvature of all data points in the neighborhood data point set of each data point is recorded as the first variance; the mean of the Gaussian curvature of all data points in the neighborhood data point set of each data point is recorded as the first mean; the ratio of the first mean to the first variance is taken as the possibility of smoothing anomaly for each data point.

[0056] It should be noted that the smaller the variance is, the less obvious fluctuation there is in the Gaussian curvature of all data points in the neighborhood data point set of the data point, and the greater the possibility that the data point is a defective data point in the concrete component; however, the Gaussian curvature corresponding to the data point that appears as a flat area may also remain stable. By utilizing the difference in curvature size corresponding to the flat area and the area with obvious concave-convex defects, a larger Gaussian curvature is required to represent the flat area, and the Gaussian curvature of most data points in the corresponding neighborhood data point set is smaller, and the first mean will also be smaller; that is, the larger the first mean, the greater the possibility that the data point is a defective data point in the concrete component, and the higher the probability of showing flatness anomalies at the data point.

[0057] So far, the possibility of flatness anomaly of each data point in the point cloud data of the concrete component is obtained through the above method.

[0058] Step S003: Obtain the feature vector of each data point according to the normal roughness possibility and the modified smoothing abnormality possibility of each data point.

[0059] It should be noted that the possibility of flatness anomaly of data points can be obtained through Gaussian curvature. It is not accurate to judge whether a data point should be added to the inliers of the RANSAC algorithm only by the possibility of flatness anomaly. This is because during the highway construction process, the surface of the concrete component will have a certain degree of roughness due to the composition and production process. Therefore, it is necessary to determine the inliers in combination with the characteristics of the concrete component when the flatness is normal; when the concrete component is flat, the data points on a single surface of the component should be on the same plane. If all data points are fitted on a plane, it is easy to be affected by defective data points, resulting in the fitting plane not accurately reflecting the plane information of the real concrete component; therefore, it is necessary to use the plane information of the local manifestation of the data point to determine the normal characteristics of the data point, thereby determining the inliers of the RANSAC algorithm. Specifically including steps 301 and 302:

[0060] 301. Get the normal-coarse likelihood of each data point in point cloud data.

[0061] It should be noted that the Gaussian curvature of data points is easily affected by noise and curvature parameters, resulting in distortion of the calculated Gaussian curvature. On the other hand, the defect features of depressions or protrusions are not ideal spheres, resulting in large differences between Gaussian curvatures that also appear as defect features. As a result, it is inaccurate to use only Gaussian curvature features that represent abnormal data points. Therefore, it is necessary to combine the plane features that represent normal data points to determine the inliers of the RANSAC algorithm.

[0062] It should be further explained that the plane information shown in the neighborhood data point set of the data point is used to determine the normal roughness possibility of the data point; on the one hand, the reason is that the smaller the distance between the data point and the plane where some data points in the neighborhood data point set are located, the greater the possibility that the data point and some data points in the neighborhood data point set are in the same plane, and the distance difference between the data points and the same plane is consistent, indicating that the fluctuation amplitude of the data points is basically consistent, and the greater the possibility that they belong to the normal rough surface of the concrete component, the greater the possibility that the data point is in the normal rough area of ​​the concrete component; on the contrary, if there is an obvious fluctuation in the distance difference, it means that some data points in the neighborhood data point set of the data point have a poor fitting effect on the corresponding plane of the neighborhood data point set, and there are uneven features, and the greater the possibility of being defective data points, the smaller the possibility that some data points in the neighborhood data point set are in the same plane; on the other hand, the reason is that the smaller the distance between the data point and the plane, the greater the possibility of being a normal and flat data point, but the fitting plane is easily disturbed by abnormal data points, resulting in the plane not reflecting the real situation, so that the distance is inaccurate, so the flat abnormal possibility of the data point is used as the weight of the distance for adjustment.

[0063] Specifically, for any data point in the point cloud data of the concrete component, any data point in the neighborhood data point set of each data point is recorded as the target data point; the vertical distance from the target data point to the neighborhood fitting surface of each data point is taken as the neighborhood data distance of the target data point.

[0064] The normal roughness likelihood of each data point is obtained based on the smoothed anomaly likelihood and the location distribution differences between data points in the neighborhood data point set of each data point.

[0065] As an example, the normal rough likelihood of getting each data point is calculated as:

[0066]

[0067] In the formula, Represents the normal-roughness probability of any data point in the point cloud data; Represents the variance of the neighborhood data distances of all data points in the neighborhood data point set of the data point; Represents the total number of all data points in the neighborhood data point set of the data point; The neighborhood data point set represents the The probability of a flattened anomaly for each data point; The neighborhood data point set represents the The neighborhood data distance of each data point; represents an exponential function with a natural constant as the base, and the embodiment adopts Model to present inverse proportional relationship and normalization, As the input of the model, the implementer can choose the inverse proportional function and the normalization function according to the actual situation.

[0068] Among them, the higher the consistency of the vertical distances from all data points in the neighborhood data point set to the neighborhood fitting surface, the closer the values ​​of the neighborhood data distances of all data points in the neighborhood data point set of the data point are, and the smaller the value of Vard is; It represents the overall impact trend of the flatness anomaly of data points at different positions in the neighborhood data point set on the data points. The flatness anomaly possibility is used as the weight for adjustment, so that the impact of the data points in the neighborhood data point set is different; that is, the larger the value of Z, the greater the possibility that the data point is in the normal rough area of ​​the concrete component surface.

[0069] At this point, the normal-rough likelihood of each data point in the point cloud data is obtained.

[0070] 302. Obtain the corrected leveling anomaly probability of each data point in the point cloud data.

[0071] It should be noted that if the concave-convex defect is too large or the number of data points in the neighborhood data point set of the data point is too small, the local data points with defects can still construct a better plane, which means that these data points should be abnormal and also have a greater possibility of normal roughness. Therefore, the above-mentioned normal roughness possibility alone cannot accurately reflect the defects or normal characteristics of the data points, resulting in the defective data points being judged as normal data points and added to the local points during the iteration process of the RANSAC algorithm. The fitting plane obtained cannot be used as a reference standard for judging the flatness of concrete components. Therefore, the flatness of concrete components needs to be adopted as a global feature. If the planes obtained by the local features have directional similarity in the case of flatness, the constructed planes are close in position. On the contrary, if there is a large difference in the direction of the plane when a concave-convex defect occurs, although the tip of the defect may be similar to the overall plane direction, there is a certain position difference, thereby correcting the possibility of abnormal flatness of the data point.

[0072] Specifically, for any data point in the point cloud data of the concrete component, obtain the positive normal vector of the neighborhood fitting surface corresponding to each data point, the positive normal vector of the neighborhood fitting surface corresponding to the target data point, the position constant of the neighborhood fitting surface corresponding to each data point, and the position constant of the neighborhood fitting surface corresponding to the target data point; and take the cosine similarity between the positive normal vector of the neighborhood fitting surface corresponding to the target data point and the positive normal vector of the neighborhood fitting surface corresponding to the data point as the plane similarity of the target data point.

[0073] Among them, obtaining the normal vector and the position constant, and the cosine similarity are well-known contents, and will not be described in detail in this embodiment.

[0074] According to the possibility of the smoothing anomaly and the distribution similarity between the data points in the neighborhood data point set of each data point, the corrected possibility of the smoothing anomaly of each data point is obtained.

[0075] As an example, the calculation method for obtaining the corrected abnormal possibility factor for each data point is:

[0076]

[0077] In the formula, Indicates the possible factor of correction anomaly of any data point in the point cloud data; Represents the minimum value of the plane similarity of all data points in the neighborhood data point set of the data point; Represents the total number of all data points in the neighborhood data point set of the data point; The neighborhood data point set represents the The position constant of the neighborhood fitting surface corresponding to each data point; Represents the position constant of the neighborhood fitting surface corresponding to the data point; The neighborhood data point set represents the The plane similarity of data points; The neighborhood data point set represents the The probability of a flattened anomaly for each data point; Indicates taking the absolute value, is a constant parameter used to prevent the denominator from being 0. The size is set to 0.01.

[0078] The smaller the similarity between the normal vectors of the neighborhood fitting surface corresponding to the data point in its neighborhood data point set, the greater the flatness change in the local area where the data point is located. The smaller the value of The greater the possibility of flattening anomaly of the data point, the smaller the plane similarity, which means that the jth data point has a higher degree of influence on the data point; and the greater the difference in position constant between the data point and the jth data point in its neighborhood data point set, the greater the influence of the jth data point on the flattening possibility of the data point. represents the combined effect of all data points in the neighborhood data point set on the possibility of smoothing the data point; that is, The larger the value of , the greater the correction anomaly probability factor of the data point, and the higher the degree to which the smoothing probability of the data point is corrected.

[0079] The corrected anomaly possibility factors of all data points in the point cloud data are obtained, and each corrected anomaly possibility factor after linear normalization of all corrected anomaly possibility factors is recorded as the corrected flattening anomaly possibility.

[0080] At this point, the possibility of corrected flattening anomaly for each data point in the point cloud data is obtained.

[0081] Specifically, for any data point in the point cloud data of the concrete component, the difference between 1 and the corrected flatness anomaly possibility of the data point is recorded as the flatness anomaly feature of the data point; and the vector formed by the flatness anomaly feature and the normal roughness possibility of the data point is used as the feature vector of the data point.

[0082] At this point, the feature vector of each data point in the point cloud data is obtained through the above method.

[0083] Step S004: Obtain the degree of preference of each data point according to the characteristic vector; and perform flatness detection on the concrete component according to the degree of preference.

[0084] It should be noted that in the RANSAC algorithm, the initial inlier point determines the subsequent addition of inlier points and the detection effect of the flatness of the concrete component. Therefore, it is necessary to ensure that the initial inlier point and the subsequently added data points are as normal data points as possible; the larger the modulus of the eigenvector of the data point, the greater the possibility that the data point belongs to a normal data point; the smaller the distance from the data point to the neighborhood fitting surface, the greater the possibility that the data point belongs to a normal data point; and then use the RANSAC algorithm to construct the fitting plane. Since the more normal data points there are inlier points, the more accurate the fitting effect, it is necessary to add normal data points to the inlier points as much as possible during the iteration process of the algorithm.

[0085] Specifically, the preference degree of each data point is obtained according to the feature vector.

[0086] As an example, the calculation method for obtaining the preference of each data point is:

[0087] For any data point in the point cloud data of the concrete component, the reciprocal of the vertical distance from the data point to the neighborhood fitting surface is recorded as the second reciprocal; the product of the modulus of the eigenvector of the data point and the second reciprocal is used as the optimization factor of the data point.

[0088] The preferred factors of all data points in the point cloud data are obtained, and each preferred factor after linear normalization of all preferred factors is recorded as the preferred degree.

[0089] Preset a threshold parameter and a quantity parameter , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It depends on the specific implementation situation.

[0090] Furthermore, for any data point in the point cloud data of the concrete component, if the preference degree of the data point is less than the threshold parameter , the data point is recorded as an outlier point; if the ratio of the total number of all outliers in the point cloud data of the concrete component to the total number of all data points in the point cloud data of the concrete component is greater than or equal to the quantity parameter , the flatness test results of concrete components during highway construction were marked as unqualified.

[0091] At this point, this embodiment is complete; please refer to Figure 2 , which shows a specific implementation flow chart of a method for detecting the flatness of concrete components used in highway construction.

[0092] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0093] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting the flatness of concrete components for highway construction, characterized in that: The method comprises the following steps: Acquire point cloud data of a concrete component, wherein the point cloud data includes a plurality of data points; Obtain a neighborhood data point set for each data point; obtain the leveling anomaly possibility of each data point based on the distribution of data points in the neighborhood data point set of each data point; According to the flattening anomaly probability and the position distribution difference between the data points in the neighborhood data point set of each data point, the normal roughness probability of each data point is obtained; according to the flattening anomaly probability and the distribution similarity between the data points in the neighborhood data point set of each data point, the corrected flattening anomaly probability of each data point is obtained; according to the corrected flattening anomaly probability and the normal roughness probability, the characteristic vector of each data point is obtained; Performing flatness detection on concrete components based on feature vectors; The method of obtaining the possibility of leveling anomaly of each data point according to the distribution of data points in the neighborhood data point set of each data point includes the following specific methods: For any data point in the point cloud data of the concrete component, a neighborhood data point set of the data point is fitted using a minimum quadratic surface fitting method to obtain a neighborhood fitting surface of the data point, and the Gaussian curvature of all data points in the neighborhood data point set of the data point is obtained through the neighborhood fitting surface according to Weyl's curvature theorem; The variance of the Gaussian curvature of all data points in the neighborhood data point set of the data point is recorded as the first variance; the mean of the Gaussian curvature of all data points in the neighborhood data point set of the data point is recorded as the first mean; the ratio of the first mean to the first variance is used as the possibility of smoothing anomaly of the data point; The method of obtaining the normal roughness possibility of each data point according to the flattening abnormality possibility and the position distribution difference between the data points in the neighborhood data point set of each data point includes the following specific methods: For any data point in the point cloud data of the concrete component, any data point in the neighborhood data point set of the data point is recorded as the target data point; the vertical distance from the target data point to the neighborhood fitting surface of the data point is taken as the neighborhood data distance of the target data point; The ratio of the variance of the neighborhood data distances of all data points in the neighborhood data point set of the data point to the total number of all data points in the neighborhood data point set of the data point is recorded as the first ratio; the variance of the neighborhood data distances of all data points in the neighborhood data point set of the data point is recorded as the first ratio. The probability of a flattened anomaly of a data point is related to the probability of the first The product of the distances of the neighborhood data of the data point is recorded as The first product of the data point; the cumulative sum of the first products of all data points in the neighborhood data point set of the data point is recorded as the first cumulative sum; the sum of the first cumulative sum and the first ratio is recorded as the first sum value; the inverse proportional value of the first sum value is used as the normal roughness possibility of the data point; The method of obtaining the corrected leveling anomaly possibility of each data point according to the leveling anomaly possibility and the distribution similarity between data points in the neighborhood data point set of each data point includes the following specific methods: For any data point in the point cloud data of the concrete component, record any data point in the neighborhood data point set of the data point as the target data point; obtain the plane similarity of the target data point; According to the plane similarity and the possibility of flattening anomalies, the corrected anomaly possibility factor of each data point is obtained; Obtain the corrected anomaly possibility factors of all data points in the point cloud data, and record each corrected anomaly possibility factor after linearly normalizing all corrected anomaly possibility factors as the corrected flattening anomaly possibility; The specific method of obtaining the plane similarity of the target data point includes: Obtaining the positive normal vector of the neighborhood fitting surface corresponding to the data point and the positive normal vector of the neighborhood fitting surface corresponding to the target data point; taking the cosine similarity between the positive normal vector of the neighborhood fitting surface corresponding to the target data point and the positive normal vector of the neighborhood fitting surface corresponding to the data point as the plane similarity of the target data point; The specific method of obtaining the corrected abnormal possibility factor of each data point according to the plane similarity and the possibility of flattening abnormality includes: Obtaining a position constant of a neighborhood fitting surface corresponding to the data point and a position constant of a neighborhood fitting surface corresponding to a target data point; The reciprocal of the sum of the minimum absolute value of the plane similarity of all data points in the neighborhood data point set of the data point and the constant parameter is recorded as the first reciprocal; the reciprocal of the sum of the minimum absolute value of the plane similarity of all data points in the neighborhood data point set of the data point is recorded as the first reciprocal; The absolute value of the difference between the position constant of the neighborhood fitting surface corresponding to the data point and the position constant of the neighborhood fitting surface corresponding to the data point is recorded as The first absolute value of the difference of the data points; the first absolute value of the difference of the neighborhood data points of the data point The sum of the plane similarity of data points and the constant parameter is used as the denominator; The ratio of the absolute value of the first difference of the data point to the denominator is recorded as The second ratio of the data points; The second ratio of the data point to the first The product of the smoothing anomaly probability of data points is recorded as The second product of the data point; the cumulative sum of the second products of all data points in the neighborhood data point set of the data point is recorded as the second cumulative sum; the product of the second cumulative sum and the first reciprocal is used as the corrected abnormal possibility factor of the data point; The specific method of obtaining the feature vector of each data point according to the possibility of correcting the abnormal flattening and the possibility of normal roughness is as follows: For any data point in the point cloud data of the concrete component, the difference between 1 and the possibility of the corrected flatness anomaly of the data point is recorded as the flatness anomaly feature of the data point; the vector formed by the flatness anomaly feature and the normal roughness possibility of the data point is used as the feature vector of the data point; The specific method of detecting the flatness of the concrete component according to the characteristic vector is as follows: According to the feature vector, obtain the preference degree of each data point; Preset a threshold parameter and a quantity parameter For any data point in the point cloud data of the concrete component, if the preference degree of the data point is less than the threshold parameter , the data point is recorded as an outlier point; if the ratio of the total number of all outliers in the point cloud data of the concrete component to the total number of all data points in the point cloud data of the concrete component is greater than or equal to the quantity parameter , the flatness test results of building concrete components are marked as unqualified.

2. A method for detecting the flatness of concrete components for highway construction according to claim 1, characterized in that: The specific method of obtaining the neighborhood data point set of each data point includes: Preset a neighborhood parameter For any data point in the point cloud data of the concrete component, take the data point as the center point and construct a point cloud data with a radius of A spherical region is formed, and a data point set formed by all data points in the spherical region is used as a neighborhood data point set of the data point.

3. A method for detecting the flatness of concrete components for highway construction according to claim 1, characterized in that: The specific method of obtaining the preference degree of each data point according to the feature vector is as follows: The reciprocal of the vertical distance from the data point to the neighborhood fitting surface is recorded as the second reciprocal; the product of the modulus of the eigenvector of the data point and the second reciprocal is used as the optimization factor of the data point; The preferred factors of all data points in the point cloud data are obtained, and each preferred factor after linear normalization of all preferred factors is recorded as a preferred degree.