A method for detecting the spacing between skeletons based on high-strength alloy template structure

Abstract

The present invention relates to the field of image data processing technology, and in particular to a skeleton spacing detection method based on a high-strength alloy template structure, comprising: obtaining a skeleton grayscale image, constructing three-dimensional point cloud data of the skeleton of the high-strength alloy template structure, obtaining a point cloud skeleton area and a local skeleton area; obtaining the information credibility of the point cloud skeleton area according to the difference between the grayscale fluctuations of each local skeleton area corresponding to the same position in the skeleton grayscale image at different angles, and the visual angle and grayscale information richness; obtaining the shape error coefficient of the point cloud skeleton area according to the regional matching relationship and data difference between each point cloud skeleton area and the standard point cloud data; combining the information credibility and the shape error coefficient, screening the edge feature points according to the discrete feature distribution of each edge feature point, and determining the skeleton spacing detection result according to the screening result. The present invention can obtain a more accurate skeleton spacing detection result.

Description

Technical Field

[0001] The invention relates to the technical field of image data processing, and in particular to a skeleton spacing detection method based on a high-strength alloy template structure. Background Art

[0002] In the field of modern engineering and construction, high-strength alloy formwork structures occupy a pivotal position. With its excellent strength and outstanding corrosion resistance, high-strength alloy formwork structures have been widely used in the construction and maintenance of heavy facilities such as large buildings, bridges, and ships. In recent years, with the continuous development of the economy, the demand for more efficient and durable buildings and infrastructure has also increased. Therefore, the quality control of high-strength alloy formwork is particularly important. The accuracy of the frame spacing detection of the high-strength alloy formwork structure has a great impact on the quality of the high-strength alloy formwork structure.

[0003] At present, the existing technology detects the skeleton spacing through the three-dimensional point cloud data of the high-strength alloy template structure. Among them, the binocular vision method is often used to reconstruct the high-strength alloy template structure in three dimensions. However, this method relies on the grayscale information in the collected image for feature matching and parallax calculation. The existing method does not take into account the areas with sparse texture or redundant data, resulting in matching errors in the three-dimensional construction process, affecting the acquisition accuracy of the three-dimensional point cloud data of the high-strength alloy template structure, and further resulting in inaccurate detection results of the skeleton spacing. Summary of the invention

[0004] In order to solve the technical problem that the accuracy of the three-dimensional point cloud data of the high-strength alloy template structure in the existing method is low, which leads to the inaccurate detection result of the skeleton spacing, the purpose of the present invention is to provide a skeleton spacing detection method based on the high-strength alloy template structure, and the technical scheme adopted is as follows:

[0005] Obtain skeleton grayscale images of the high-strength alloy template structure at different angles, construct three-dimensional point cloud data of the skeleton of the high-strength alloy template structure based on the skeleton grayscale images at different angles, and obtain the point cloud skeleton area in the three-dimensional point cloud data and the local skeleton area corresponding to the same position of the skeleton grayscale images at different angles;

[0006] According to the difference distribution between the grayscale fluctuations of each local skeleton region corresponding to the same position in the skeleton grayscale images at different angles, and the feature distribution between the visual angle and the grayscale information richness, the information credibility of the point cloud skeleton region corresponding to each local skeleton region corresponding to the same position is obtained;

[0007] According to the regional matching relationship and data difference between each point cloud skeleton area and the standard point cloud data of the standard high-strength alloy template structure, the shape error coefficient of each point cloud skeleton area is obtained;

[0008] In combination with the information credibility and the shape error coefficient, the edge feature points are screened according to the discrete feature distribution of each edge feature point in each point cloud skeleton area, and the skeleton spacing detection result is determined according to the screening result.

[0009] Preferably, the step of obtaining the point cloud skeleton region in the three-dimensional point cloud data and the local skeleton region corresponding to the same position of the skeleton grayscale images at different angles specifically includes:

[0010] The three-dimensional point cloud data is segmented to obtain the point cloud skeleton region, and the local skeleton region corresponding to the skeleton grayscale image of the point cloud skeleton region at each angle is recorded as the reference skeleton region of the point cloud skeleton region at each angle.

[0011] Preferably, the information credibility of the point cloud skeleton region corresponding to each local skeleton region corresponding to the same position is obtained based on the difference distribution between the grayscale fluctuations of each local skeleton region corresponding to the same position in the skeleton grayscale images at different angles, and the feature distribution between the visual angle and the grayscale information richness, specifically including:

[0012] For any point cloud skeleton region, the plane disparity factor of the point cloud skeleton region is obtained according to the difference in grayscale fluctuations in the reference skeleton region at each angle of the point cloud skeleton region;

[0013] According to the grayscale information richness of the reference skeleton region at each angle of the point cloud skeleton region and the angle acquisition position information corresponding to the skeleton grayscale image of each reference skeleton region, the feature richness factor of the point cloud skeleton region is obtained;

[0014] The information credibility of the point cloud skeleton region is obtained according to the plane disparity factor and the feature enrichment factor, wherein the plane disparity factor is negatively correlated with the information credibility, and the feature enrichment factor is positively correlated with the information credibility.

[0015] Preferably, obtaining the plane parallax factor of the point cloud skeleton region according to the difference of grayscale fluctuation in the reference skeleton region at each angle of the point cloud skeleton region specifically includes:

[0016] Record any angle as a target angle, and determine the grayscale feature coefficient of the reference skeleton area at the target angle based on the discrete degree of the grayscale values ​​of all pixels in the point cloud skeleton area within the reference skeleton area at the target angle;

[0017] Based on the overall level of difference between the grayscale feature coefficients of the point cloud skeleton region and the reference skeleton region at each two angles, a plane disparity factor of the point cloud skeleton region is determined.

[0018] Preferably, the feature richness factor of the point cloud skeleton region is obtained according to the grayscale information richness of the reference skeleton region at each angle of the point cloud skeleton region and the angle acquisition position information corresponding to the skeleton grayscale image of each reference skeleton region, specifically including:

[0019] Determine the grayscale information coefficient of the reference skeleton region at the target angle based on the grayscale distribution information entropy of all pixels in the reference skeleton region of the point cloud skeleton region at the target angle;

[0020] Determine the spatial information coefficient of the reference skeleton region at the target angle based on the angle cosine value of the skeleton grayscale image at the target angle when the image is acquired, and the negative correlation coefficient of the distance between the device for image acquisition and the plane where the high-strength alloy template is located;

[0021] Based on the overall level of the product between the grayscale information coefficient and the spatial information coefficient of the reference skeleton region at each angle of the point cloud skeleton region, a feature enrichment factor of the point cloud skeleton region is determined.

[0022] Preferably, the shape error coefficient of each point cloud skeleton area is obtained according to the regional matching relationship and data difference between each point cloud skeleton area and the standard point cloud data of the standard high-strength alloy template structure, specifically including:

[0023] Perform point cloud matching on the three-dimensional point cloud data and the standard point cloud data of the standard high-strength alloy template structure to obtain the matching point between each data point in the three-dimensional point cloud data and the standard point cloud data;

[0024] The shape error coefficient of each point cloud skeleton area is obtained according to the data difference between each data point in each point cloud skeleton area and the matching point in the standard point cloud data.

[0025] Preferably, obtaining the shape error coefficient of each point cloud skeleton region according to the data difference between each data point in each point cloud skeleton region and the matching point in the standard point cloud data specifically includes:

[0026] For any point cloud skeleton region, the Euclidean distance between each data point in the point cloud skeleton region and the corresponding matching point is calculated, and the shape error coefficient of the point cloud skeleton region is determined based on the overall level of all Euclidean distances in the point cloud skeleton region.

[0027] Preferably, combining the information credibility and the shape error coefficient, and screening the edge feature points according to the discrete feature distribution of each edge feature point in each point cloud skeleton region, specifically includes:

[0028] Obtaining a feature screening amount for each point cloud skeleton region according to the information credibility and shape error coefficient of each point cloud skeleton region; the information credibility and shape error coefficient are both positively correlated with the feature screening amount;

[0029] According to the data discreteness degree of each edge feature point in each point cloud skeleton area, the feature discrete factor of each edge feature point in each point cloud skeleton area is obtained;

[0030] The edge feature points in each point cloud skeleton region are screened respectively using the feature screening amount and the feature discrete factor.

[0031] Preferably, obtaining the feature discrete factor of each edge feature point in each point cloud skeleton region according to the data discreteness of each edge feature point in each point cloud skeleton region specifically includes:

[0032] For any point cloud skeleton region, the data difference ratio between any edge feature point in the point cloud skeleton region and the mean of all edge feature points in the point cloud skeleton region is used as the feature discrete factor of the any edge feature point.

[0033] Preferably, the step of using the feature screening amount and the feature discrete factor to screen edge feature points in each point cloud skeleton region comprises:

[0034] For any point cloud skeleton region, all edge feature points of the point cloud skeleton region are arranged in ascending order according to the value of the feature discrete factor, and the edge feature points are screened based on the feature screening amount of the point cloud skeleton region in the arrangement order to obtain the screening result.

[0035] The embodiments of the present invention have at least the following beneficial effects:

[0036] The present invention first constructs three-dimensional point cloud data by collecting two-dimensional images at different angles, and obtains the local skeleton area corresponding to the same position in the two-dimensional image when dividing the three-dimensional point cloud data into regions. This step provides a data basis for the subsequent analysis of the information credibility of the point cloud skeleton area by considering the information representation in the region in the skeleton grayscale image at different angles. Then, by comprehensively considering the grayscale fluctuation difference of the local skeleton area of ​​the two-dimensional image at different angles, as well as the visual angle and grayscale information richness of the two-dimensional image, the information credibility of the corresponding point cloud skeleton area is obtained, that is, the grayscale representation difference of the point cloud skeleton area corresponding to different two-dimensional images and the information richness are analyzed, that is, the smaller the difference, the greater the information richness, and the smaller the possibility of error in the point cloud skeleton area when the corresponding three-dimensional point cloud is constructed, that is, the greater the information credibility. Further, consider analyzing the characteristics of the point cloud skeleton area in terms of three-dimensional information, that is, by comparing the feature matching relationship and difference between the point cloud skeleton area and the standard point cloud data, it reflects the difference between the point cloud skeleton area and the standard, and then the shape error coefficient is used to characterize the possibility of error in the three-dimensional size shape of the point cloud skeleton area. Finally, combined with the information credibility of the point cloud skeleton area in the two-dimensional graphics and the shape error coefficient in the three-dimensional data, combined with the characteristic discrete feature distribution of the edge feature points, the edge feature points with smaller discrete features are screened out for skeleton spacing detection, which can effectively avoid the presence of large error data in the three-dimensional point cloud data affecting the accuracy of the detection results. That is, by screening the point cloud data through multiple features, more accurate three-dimensional point cloud data can be obtained for skeleton spacing detection, so that the obtained skeleton spacing detection results are also more accurate, thereby improving the domino assembly accuracy of the high-strength alloy template and ensuring the excellent strength, load-bearing performance and stability of the high-strength alloy template. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0038] Figure 1 It is a flow chart of the steps of a method for detecting the skeleton spacing based on a high-strength alloy template structure provided by the present invention;

[0039] Figure 2 is a grayscale image of the high-strength template structure provided by the present invention;

[0040] Figure 3 This is a top view of the product structure of the high-strength alloy template provided by the present invention;

[0041] Figure 4 is a flowchart of the steps of the method for obtaining the information credibility of each point cloud skeleton area provided by the present invention;

[0042] Figure 5 It is a flowchart of the steps of the method for obtaining the feature enrichment factor of the point cloud skeleton area provided by the present invention;

[0043] Figure 6 It is a flowchart of the steps of the process of screening edge feature points provided by the present invention;

[0044] Figure 7 It is a system block diagram of a skeleton spacing detection system based on a high-strength alloy template structure provided by the present invention;

[0045] Figure 8 It is a structural schematic diagram of a computer device provided by the present invention. DETAILED DESCRIPTION

[0046] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the skeleton spacing detection method based on a high-strength alloy template structure proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination 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.

[0047] 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 invention belongs.

[0048] The following is a detailed description of a specific scheme of a skeleton spacing detection method based on a high-strength alloy template structure provided by the present invention in conjunction with the accompanying drawings.

[0049] See also Figure 1 , which shows a flowchart of a method for detecting the skeleton spacing based on a high-strength alloy template structure provided by an embodiment of the present invention, the method comprising the following steps:

[0050] Step S100, obtain the skeleton grayscale images of the high-strength alloy template structure at different angles, construct three-dimensional point cloud data of the skeleton of the high-strength alloy template structure based on the skeleton grayscale images at different angles, obtain the point cloud skeleton area in the three-dimensional point cloud data and the local skeleton area corresponding to the same position of the skeleton grayscale images at different angles.

[0051] In order to ensure strict control of the quality of production products, a detailed analysis of the skeleton spacing is particularly important. The rationality of the skeleton spacing has a significant impact on the quality strength of the product. Therefore, this embodiment constructs a three-dimensional structure by collecting images of the high-strength alloy template structure at different angles. In order to avoid the problem of low accuracy of the constructed three-dimensional structure affecting the accuracy of the detection results, and considering that the grayscale information contained in the two-dimensional plane image affects the construction accuracy of the three-dimensional structure, the construction quality of the three-dimensional structure can be evaluated by analyzing the grayscale feature distribution in the two-dimensional plane image corresponding to the three-dimensional structure. Based on this, this embodiment first needs to collect image data during feature analysis to perform the three-dimensional structure construction operation, and obtain the corresponding relationship between the two.

[0052] First, it should be noted that the high-strength alloy template structure used in the skeleton detection operation of this embodiment is as follows: Figure 2 As shown in the figure, this high-strength alloy template structure significantly improves the yield strength, tensile strength, elongation and toughness of the steel plate, ensuring the excellent performance of the product. At the same time, its skeleton structure has been carefully designed, using a partial suspension and local compaction treatment method. This innovative design not only effectively enhances the strength of the product, but also significantly reduces the overall weight, thereby greatly improving the portability and transportation efficiency of the product.

[0053] The spacing of the skeleton of the high-strength alloy formwork structure greatly affects the strength and quality of the overall formwork structure. Reasonable domino spacing layout and calibration can ensure that stress is evenly distributed in the high-strength alloy formwork, avoid local stress concentration, and reduce the risk of material fatigue and damage. Excessive or uneven spacing will cause certain areas to be subjected to greater stress, which may cause local deformation or failure. Correct spacing and calibration help to evenly distribute the load, making the formwork more stable and reliable when bearing loads. Prevent cracking, tilting, bending or instability under uneven stress and high loads. Therefore, when the high-strength alloy formwork is assembling dominoes, the spacing of the dominoes needs to be strictly and accurately controlled.

[0054] Based on this, we first need to clarify the components of the skeleton structure, such as Figure 3 This is a top view of the product structure of the high-strength alloy template used in this embodiment. Figure 3 1 represents the longitudinal 42×21×0.8 vertical reinforcement of the frame part in the high-strength alloy formwork structure, 2 and 3 are transverse reinforcements, 2 is an even-numbered transverse reinforcement using 42×21×0.8 transverse reinforcement, 3 is an odd-numbered transverse reinforcement using 61×20×0.8 transverse reinforcement. Among them, the distance distribution between transverse reinforcements can be understood as the transverse frame spacing, and the distance distribution between longitudinal reinforcements can be understood as the longitudinal frame spacing.

[0055] Then, this embodiment needs to collect surface images of the same high-strength alloy template structure to be detected at different angles respectively, so as to provide a data basis for the subsequent construction of the three-dimensional structure of the high-strength alloy template structure. In this embodiment, the binocular vision method is selected to construct the three-dimensional point cloud data. Therefore, at the same height above the high-strength alloy template structure, a CMOS camera is set at different image acquisition angles, and images of the same high-strength alloy template structure placed on the detection platform are collected respectively, and grayscale processing is performed uniformly to obtain the skeleton grayscale images of the high-strength alloy template at two different angles.

[0056] Among them, different acquisition angles can be set by setting the acute angle between the camera and the vertical direction. At the same time, the image acquisition angle corresponding to the image taken at different angles, that is, the acute angle between the camera and the vertical direction, provides a data basis for subsequent data analysis. After the acquired image is grayed, the skeleton grayscale image can also be pre-processed, such as denoising or enhancement, to obtain a skeleton grayscale image with better quality.

[0057] Furthermore, based on the binocular vision method, the skeleton grayscale images collected at two different angles can be used to realize the three-dimensional construction of the high-strength alloy template structure to obtain three-dimensional point cloud data. This method is a well-known technology and will not be introduced in detail here. At the same time, it should be noted that in the process of obtaining the three-dimensional point cloud data, the corresponding relationship between each data point in the three-dimensional point cloud data and the pixel in the skeleton grayscale image at each angle can be obtained.

[0058] In order to detect the skeleton spacing more accurately, it is necessary to perform a single skeleton segmentation operation on the three-dimensional point cloud data. The three-dimensional point cloud data of the high-strength alloy template structure can be segmented using the existing normal vector segmentation method. Each local area can be called a point cloud skeleton area. In the process of constructing the three-dimensional point cloud data, the correspondence between the three-dimensional point cloud data and each pixel in the skeleton grayscale image at each angle can be used to obtain the local range area corresponding to the same position in the skeleton grayscale image of the same point cloud skeleton area at each angle, and recorded as the local skeleton area. Furthermore, the local skeleton area corresponding to the same position in the skeleton grayscale image of the point cloud skeleton area at each angle can be recorded as the reference skeleton area of ​​the point cloud skeleton area at each angle.

[0059] It should be noted that a point cloud skeleton area in the three-dimensional point cloud data can be determined, a local skeleton area corresponding to the point cloud skeleton area in the skeleton grayscale image at one angle, which is also the reference skeleton area, and a local skeleton area corresponding to the point cloud skeleton area in the skeleton grayscale image at another angle, which is also the reference skeleton area.

[0060] It should be further explained that, in the present embodiment, each local range obtained through segmentation in the three-dimensional point cloud data is recorded as a point cloud skeleton area, that is, the effect evaluation of the three-dimensional structure construction needs to be performed indiscriminately to obtain a more accurate data evaluation result.

[0061] Step S200, based on the difference distribution between the grayscale fluctuations of each local skeleton area corresponding to the same position in the skeleton grayscale image at different angles, and the feature distribution between the visual angle and the grayscale information richness, obtain the information credibility of the point cloud skeleton area corresponding to each local skeleton area corresponding to the same position.

[0062] Since the binocular vision system relies on the grayscale information in the image for feature matching and disparity calculation, the system may not be able to accurately find matching points in areas with sparse or repeated textures, resulting in reconstruction errors. At the same time, changes in illumination will also affect the brightness and contrast of the image, thereby affecting the accuracy of stereo matching. Based on this, the credibility of the three-dimensional reconstruction result corresponding to the target part in the two-dimensional image can be quantified by comparing the feature richness and disparity of the local area where the single target part is located under the two acquisition perspectives.

[0063] In this embodiment, if Figure 4 As shown, the method for obtaining the information credibility of each point cloud skeleton region can be implemented by steps S210 to S230.

[0064] Step S210 , for any point cloud skeleton region, a plane parallax factor of the point cloud skeleton region is obtained according to the difference in grayscale fluctuations of the point cloud skeleton region at each angle within the reference skeleton region.

[0065] In this embodiment, the steps of obtaining information credibility are described by taking any point cloud skeleton area as an example. The parallax of the point cloud skeleton area corresponding to a single target part in the three-dimensional point cloud data usually considers analyzing the grayscale feature differences between the reference skeleton areas in the skeleton grayscale images at different angles. However, due to the difference in acquisition viewing angles and the influence of light sources in the acquisition environment, as well as environmental factors such as reflections on the surface of metal structures at different angles, each reference skeleton area corresponding to the point cloud skeleton area in the two-dimensional image will affect the accuracy of the three-dimensional reconstruction result. Based on this, this embodiment quantitatively characterizes the degree of parallax performance of the same point cloud skeleton area by comparing the differences in grayscale feature distributions in different reference skeleton areas corresponding to the same point cloud skeleton area under two different image acquisition viewing angles.

[0066] Specifically, this embodiment is illustrated by taking the local skeleton area in the skeleton grayscale image collected at any angle as an example, that is, any angle is recorded as the target angle, and the grayscale characteristic coefficient of the reference skeleton area at the target angle is determined based on the discrete degree of the grayscale values ​​of all pixels in the reference skeleton area of ​​the point cloud skeleton area at the target angle.

[0067] Among them, in this embodiment, the variance of the grayscale values ​​of all pixels in a reference skeleton area is used as the corresponding grayscale feature coefficient, that is, the variance of the grayscale values ​​of all pixels in the reference skeleton area of ​​the point cloud skeleton area at the target angle represents the discrete degree of grayscale distribution, that is, the grayscale feature coefficient of the reference skeleton area of ​​the point cloud skeleton area at the target angle reflects the fluctuation of grayscale in the corresponding reference skeleton area. The larger its value is, the greater the grayscale fluctuation in the local range is, and the greater the discrete degree of the corresponding grayscale feature is. In other embodiments, the implementer can also use standard deviation, difference and other forms to represent the discrete degree of grayscale values ​​in the local range.

[0068] Further, based on the overall level of the difference between the grayscale feature coefficients of the reference skeleton region at each two angles of the point cloud skeleton region, a plane disparity factor of the point cloud skeleton region is determined.

[0069] In this embodiment, for the point cloud skeleton region, the absolute value of the difference between the grayscale feature coefficients of the corresponding reference skeleton region at two different angles is used as the plane disparity factor of the point cloud skeleton region. The plane disparity factor reflects the representation of the visual grayscale difference of the point cloud skeleton region at different angles in the corresponding two-dimensional image. The smaller its value is, the closer the grayscale feature expression level at the same position between different viewing angles of the current data acquisition is, which further indicates that the possibility of error in the three-dimensional reconstruction of the point cloud skeleton region is smaller.

[0070] It should be noted that in this embodiment, the three-dimensional point cloud data is constructed based on the binocular vision method, so only the skeleton grayscale images at two different angles are obtained, that is, one point cloud skeleton area only corresponds to two different reference skeleton areas, so the difference between the grayscale feature coefficients of the two is used as the plane disparity factor of the corresponding point cloud skeleton area. In other embodiments, if the implementer collects skeleton grayscale images at three or more different angles, then one point cloud skeleton area corresponds to multiple different reference skeleton areas, and then it is necessary to calculate the difference between the grayscale feature coefficients of each two reference skeleton areas respectively, and the mean of all the differences calculated is used as the plane disparity factor of the corresponding point cloud skeleton area. Among them, the mean represents the overall level of the difference between the grayscale feature distributions at all angles.

[0071] Step S220, obtaining a feature richness factor of the point cloud skeleton region according to the grayscale information richness of the reference skeleton region at each angle of the point cloud skeleton region and the angle acquisition position information corresponding to the skeleton grayscale image of each reference skeleton region.

[0072] Considering that there may be certain differences in the angle and distance of the acquisition camera when acquiring skeleton grayscale images under different viewing angles, the grayscale feature richness of different reference skeleton areas in the same point cloud skeleton area may also be different under different viewing angles. In the process of processing three-dimensional reconstruction data, in areas with higher texture richness, the constructed three-dimensional point cloud data can capture more feature points, thereby improving the accuracy of the detection results.

[0073] Based on this, a comprehensive analysis was conducted based on the distance, angle difference, and surface texture between the plane where the reference skeleton area is located and the camera device that collects the image. When the corresponding angle deviation of the reference skeleton area is relatively small, the distance is close, and the surface texture of the area is rich, the feature information representation of the image collected at this angle of the reference skeleton area is relatively rich.

[0074] In this embodiment, any point cloud skeleton region is still used as an example for description. Figure 5 As shown, the method for obtaining the feature enrichment factor of the point cloud skeleton region can be implemented by steps S221 to S223.

[0075] Step S221, based on the grayscale distribution information entropy of all pixels in the reference skeleton area of ​​the point cloud skeleton area at the target angle, determine the grayscale information coefficient of the reference skeleton area at the target angle.

[0076] As a specific example, first, a grayscale co-occurrence matrix is ​​constructed through the grayscale values ​​of all pixels in the reference skeleton area of ​​the point cloud skeleton area at the target angle, and the information entropy of the grayscale co-occurrence matrix is ​​calculated as the grayscale information coefficient of the reference skeleton area of ​​the point cloud skeleton area at the target angle.

[0077] Among them, the information entropy of the grayscale co-occurrence matrix can reflect the degree of confusion of the grayscale distribution information in the corresponding reference skeleton area. The larger its value is, the more chaotic the grayscale information distribution in the reference skeleton area is, and the more grayscale information it contains. This further indicates that the richer the grayscale information in the corresponding reference skeleton area is, the larger the value of the corresponding grayscale information coefficient is.

[0078] In other embodiments, the one-dimensional image entropy of the reference skeleton area of ​​the point cloud skeleton area at the target angle can be directly used as the grayscale information coefficient of the corresponding reference skeleton area, which also characterizes the degree of chaos of the grayscale information distribution in the reference skeleton area and the amount of grayscale information.

[0079] Step S222, based on the angle cosine value of the skeleton grayscale image of the reference skeleton area at the target angle when the image is captured, and the negative correlation coefficient of the distance between the image acquisition device and the plane where the high-strength alloy template is located, determine the spatial information coefficient of the reference skeleton area at the target angle.

[0080] In step S100, the acute angle between the camera device and the vertical direction when the skeleton grayscale image is captured at each different angle is obtained, that is, the angle that can be regarded as the skeleton grayscale image when capturing the image. At the same time, the distance parameter between the camera device and the plane of the detection platform where the high-strength alloy template is located can also be directly obtained.

[0081] As a specific example, taking the i-th angle as the target angle, the spatial information coefficient of the n-th point cloud skeleton area under the i-th angle is It can be expressed as ,in Represents the cosine value of the angle of the skeleton grayscale image where the reference skeleton area is located at the target angle when the image is acquired. Indicates the distance between the camera device for capturing the skeleton grayscale image at the template angle and the plane where the high-strength alloy template is located. That is, the acute angle between the camera device and the vertical direction when the skeleton grayscale image at the target angle is captured.

[0082] The negative correlation coefficient of the distance is obtained in the form of an inverse, and the spatial information coefficient reflects the richness of the information texture when the image is collected at a single angle from the two directions of angle and distance. When the distance is shorter and the angle is smaller, more abundant image information can be collected to the greatest extent possible.

[0083] Step S223, determining a feature enrichment factor of the point cloud skeleton region based on the overall level of the product between the grayscale information coefficient and the spatial information coefficient of the reference skeleton region at each angle of the point cloud skeleton region.

[0084] According to the same method as in step S221 and step S222, the grayscale information coefficient and spatial information coefficient of the reference skeleton area at each angle of each point cloud skeleton area can be obtained, which characterizes the degree of information richness of a point cloud skeleton area in corresponding two-dimensional images from the two directions of the degree of grayscale information confusion and the degree of spatial feature distribution. Therefore, the feature performance of the reference skeleton area in two directions at each different angle corresponding to a point cloud skeleton area can be combined to comprehensively evaluate the information richness in the two-dimensional image.

[0085] Based on this, in this embodiment, taking any point cloud skeleton as an example, for the nth point cloud skeleton area, the product of the grayscale information coefficient and the spatial information coefficient of the reference skeleton area of ​​the nth point cloud skeleton area at the i-th angle is recorded as the information richness of the reference skeleton area of ​​the nth point cloud skeleton area at the i-th angle, and then the average of the information richness of the reference skeleton area of ​​the nth point cloud skeleton area at all angles is the feature richness factor of the nth point cloud skeleton area.

[0086] Among them, the mean of all information richness reflects the overall level of information richness of the reference skeleton area at all angles, and the feature richness factor of the point cloud skeleton area characterizes the information richness contained in the two-dimensional image corresponding to the point cloud skeleton area. The larger the value, the greater the information richness contained in the two-dimensional image corresponding to the point cloud skeleton area, and the smaller the possibility of error in the corresponding three-dimensional feature reconstruction.

[0087] Step S230, obtaining the information credibility of the point cloud skeleton region according to the plane disparity factor and the feature enrichment factor, wherein the plane disparity factor is negatively correlated with the information credibility, and the feature enrichment factor is positively correlated with the information credibility.

[0088] For any point cloud skeleton region, the plane disparity factor characterizes the difference in the grayscale feature distribution of the local area in the two-dimensional image at different angles corresponding to the point cloud skeleton region, and the feature richness factor characterizes the information richness of the local area in the two-dimensional image at different angles corresponding to the point cloud skeleton region.

[0089] The larger the value of the plane parallax factor, the greater the difference in the grayscale feature distribution of the two-dimensional image at different angles. The smaller the value of the feature richness factor, the less grayscale feature information of the two-dimensional image at different angles, and the lower the information richness. The corresponding error phenomenon is more likely to occur when constructing the three-dimensional structure, affecting the accuracy of the subsequent detection results.

[0090] Based on this, the ratio of the feature richness factor and the plane parallax factor of the point cloud skeleton area is calculated, and the ratio is normalized to obtain the information credibility of the point cloud skeleton area. The information credibility reflects the information reference value of the two-dimensional image in the feature construction process of the point cloud skeleton area, and represents the credibility of the corresponding three-dimensional point cloud data in the area. The larger the value, the smaller the possibility of error in the point cloud data construction process, that is, the better the quality and effect of the point cloud skeleton area.

[0091] Step S300, obtaining a shape error coefficient of each point cloud skeleton region according to a region matching relationship and data difference between each point cloud skeleton region and standard point cloud data of a standard high-strength alloy template structure.

[0092] For the spacing and distribution between the skeletons, it is necessary to analyze the product model data of the template structure, such as Figure 3 It can be seen that in order to enhance the overall strength of the template structure and reduce the weight of the template, the corresponding transverse frame, that is, the transverse reinforcement, adopts an alternating structure of hollowing and local compaction. At the same time, the relevant parameters of the transverse reinforcement used in different areas are also different. In order to evaluate the credibility of a single point cloud skeleton area more accurately, the results of the foundation comparison are combined with each point cloud skeleton area and the corresponding template data, reflecting the characterization of the size and shape errors of the local area in the constructed 3D point cloud data.

[0093] Specifically, point cloud matching is performed on the three-dimensional point cloud data and the standard point cloud data of the standard high-strength alloy template structure to obtain the matching points between each data point in the three-dimensional point cloud data and the standard point cloud data. It should be noted that the template matching process between three-dimensional point cloud data is a well-known technology and will not be introduced in detail here. Through the matching results, the point cloud data in the three-dimensional point cloud data of the high-strength alloy template to be detected that has a corresponding matching relationship with the template data can be directly obtained, which is called a matching point.

[0094] It needs to be further explained that Figure 3 That is, an exemplary top view of a standard template is shown. The implementer can also collect standard high-strength alloy template structure to construct standard point cloud data according to the specific implementation scenario. This process is also a well-known technology and will not be introduced in detail here.

[0095] Furthermore, the shape error coefficient of each point cloud skeleton area is obtained according to the data difference between each data point in each point cloud skeleton area and the matching point in the standard point cloud data. That is, by comparing the data difference between each data point in each point cloud skeleton area and the corresponding matching point in the standard point cloud data, the possibility of error in the size and shape of each point cloud skeleton area is quantified.

[0096] Specifically, for any point cloud skeleton region, the Euclidean distance between each data point in the point cloud skeleton region and the corresponding matching point is calculated, and the shape error coefficient of the point cloud skeleton region is determined based on the overall level of all Euclidean distances in the point cloud skeleton region.

[0097] In this embodiment, the mean of the Euclidean distances corresponding to all data points in a point cloud skeleton region is used as the shape error coefficient of the point cloud skeleton region, that is, the mean of all Euclidean distances is used to represent the overall level of all Euclidean distances. In other embodiments, the implementer may also use parameters such as the median and mode to represent the overall level of the data.

[0098] The larger the value of the shape error coefficient, the greater the difference between the data representation of the point cloud skeleton area in the three-dimensional point cloud data and the data corresponding to the standard template, which means that the possibility of errors in the size and shape of the point cloud skeleton area is greater. The smaller the value of the shape error coefficient, the smaller the difference between the data representation of the point cloud skeleton area in the three-dimensional point cloud data and the data corresponding to the standard template, which means that the possibility of errors in the size and shape of the point cloud skeleton area is smaller.

[0099] Step S400, combining the information credibility and the shape error coefficient, screening the edge feature points according to the discrete feature distribution of each edge feature point in each point cloud skeleton area, and determining the skeleton distance detection result according to the screening result.

[0100] The information credibility of the point cloud skeleton area represents the credibility of the feature performance of the corresponding position of the point cloud skeleton area in the two-dimensional image, and the shape error coefficient of the point cloud skeleton area represents the credibility of the shape feature performance of the point cloud skeleton area in the three-dimensional point cloud data. Then, the feature performance of these two aspects can be combined to screen out a certain number of feature points for detection, thereby avoiding the influence of three-dimensional point cloud data with large errors on the detection results and improving the detection accuracy. When screening feature points of three-dimensional point cloud data, on the basis of the credibility of the above two aspects, the representation of edge information in each local feature area in the three-dimensional point cloud data is comprehensively considered to accurately select feature data points.

[0101] In this embodiment, if Figure 6 As shown, the process of screening edge feature points can be implemented by steps S410 to S430.

[0102] Step S410, obtaining a feature screening amount of each point cloud skeleton region according to the information credibility and shape error coefficient of each point cloud skeleton region; the information credibility and shape error coefficient are both positively correlated with the feature screening amount.

[0103] Taking any point cloud skeleton area as an example, the number of feature points in the current point cloud skeleton area during feature analysis is quantified by comprehensively considering the credibility of the feature information of the point cloud skeleton area in the two-dimensional image and the credibility of the feature information in the three-dimensional point cloud data.

[0104] When the credibility of the 3D reconstructed data corresponding to the point cloud skeleton area is smaller, the possibility of errors in the 3D data generated by binocular vision imaging is greater, and the degree of error in the size and shape of the acquired skeleton area will also be affected more. When the size and shape error corresponding to the point cloud skeleton area is higher, more data needs to be considered to comprehensively consider the corresponding edge position information. When the size and shape error corresponding to the point cloud skeleton area is smaller, only the data with a smaller degree of deviation in the edge features of the point cloud skeleton area needs to be considered for feature analysis.

[0105] Specifically, in this embodiment, the normalized value of the product of the information credibility and the shape error coefficient of the point cloud skeleton region is used as the feature screening amount of the point cloud skeleton region. The larger the value of the feature screening amount, the more feature data needs to be obtained for detection, and the smaller the value of the feature screening amount, the less feature data needs to be obtained for detection, that is, the smaller the amount of data for detection.

[0106] Step S420, obtaining a feature discrete factor of each edge feature point in each point cloud skeleton region according to the data discreteness of each edge feature point in each point cloud skeleton region.

[0107] Taking into account that in the skeleton grayscale image, the grayscale values ​​belonging to the edge of the skeleton are relatively uniform and close, and then by analyzing the fluctuation degree and discreteness of the data distribution of each edge feature point in each point cloud skeleton area, the discrete features of each edge feature point are evaluated. If there are large discrete features, it means that the feature performance of the edge feature point is poor, and the edge feature point is not given priority for spacing detection operations.

[0108] It should be noted that, considering that the point cloud skeleton area belongs to a three-dimensional structure, this embodiment uses the data points of each face of a point cloud skeleton area as the edge feature points of the point cloud skeleton area. For example, assuming that the point cloud skeleton area is a rectangular parallelepiped, that is, the data points on each face of the rectangular parallelepiped are the feature edge points of the point cloud skeleton area.

[0109] Furthermore, for any point cloud skeleton region, a data difference ratio between any edge feature point in the point cloud skeleton region and the mean of all edge feature points in the point cloud skeleton region is used as a feature discrete factor of the any edge feature point.

[0110] In this embodiment, taking any point cloud skeleton region as an example, the feature discrete factor of the mth edge feature point in the i-th point cloud skeleton region is It can be expressed as ,in Represents the point cloud data corresponding to the mth edge feature point in the i-th point cloud skeleton area, It represents the mean value of the point cloud data corresponding to all edge feature points in the i-th point cloud skeleton area, which can be obtained by finding the mean value of the data in the same dimension. Norm() represents the linear normalization function.

[0111] The difference between the mth edge feature point in the i-th point cloud skeleton area and the mean of all edge feature points in the i-th point cloud skeleton area can be achieved by calculating the Euclidean distance. At the same time, the difference of the point cloud data in the same dimension can be calculated separately, and then the cumulative sum of the differences in all dimensions can be calculated. The implementer can choose according to the specific implementation scenario.

[0112] Through further normalization processing, the difference ratio of each edge feature point is obtained, which reflects the discreteness of the edge features of the edge feature points in the current point cloud skeleton area. The larger the value, the greater the degree of data dispersion, which means that the feature performance of the edge feature point is poor, and the edge feature point is not given priority for spacing detection operations.

[0113] Step S430: using the feature screening amount and the feature discrete factor, the edge feature points in each point cloud skeleton region are screened respectively.

[0114] The feature screening amount of the point cloud skeleton area represents the quantity size when performing the detection operation, and the feature discrete factor of the edge feature points in the point cloud skeleton area represents the effect of the detection operation. Therefore, the feature data for the detection operation is screened out through quantity characterization and effect evaluation.

[0115] Specifically, for any point cloud skeleton area, all edge feature points of the point cloud skeleton area are arranged in ascending order according to the value of the feature discrete factor, and the edge feature points are screened based on the feature screening amount of the point cloud skeleton area in the arrangement order to obtain the screening result.

[0116] Among them, according to the arrangement order, the detection effect of edge feature points in the point cloud skeleton area is getting better and better, and then the data screening result is obtained through the feature screening amount. Since the feature screening amount of the point cloud skeleton area is a normalized value, the corresponding percentage of edge feature points constituting the feature data points of the screening result can be obtained. For example, if the feature screening amount of the point cloud skeleton area is 0.8, the first 80% of the edge feature points constituting the feature data points of the screening result can be obtained according to the arrangement order.

[0117] In the screening results, each feature data point can avoid the influence of poor representation of two-dimensional image information on three-dimensional point cloud data, which helps to obtain more accurate detection results of skeleton spacing.

[0118] Finally, the skeleton spacing detection result is determined according to the screening result. In step S300, the template structure to be detected is matched with the standard template structure, so the local range area belonging to the horizontal rib part and the vertical rib part in the point cloud data of the current high-strength alloy template structure to be detected can be obtained through the matching result, and then the minimum value of the Euclidean distance between the feature data points of the point cloud skeleton area corresponding to two adjacent horizontal ribs in the three-dimensional point cloud data is calculated to obtain the skeleton spacing between the two adjacent horizontal ribs. Similarly, the minimum value of the Euclidean distance between the feature data points of the point cloud skeleton area corresponding to two adjacent vertical ribs in the three-dimensional point cloud data is calculated to obtain the skeleton distance between the two adjacent values.

[0119] In other embodiments, the point cloud skeleton area of ​​the three-dimensional point cloud data can be re-divided by filtering the characteristic data points obtained in the detection results, and then the skeleton spacing detection is performed based on the division results. This method belongs to the existing technology and will not be introduced in detail here.

[0120] In summary, the spacing and layout of the dominoes will fully consider the global and local stress conditions of the high-strength alloy formwork during the construction process. The present invention accurately extracts the edges of the dominoes during the assembly process through binocular vision, avoiding the problem of inaccurate measurement of the edges and spacing of the dominoes caused by factors such as metal reflection, viewing angle, and blurred edges of plate reinforcements, making the spacing of the assembled dominoes more accurate, reasonable, and more evenly stressed. The production quality of the high-strength alloy formwork is greatly improved, and the excellent performance of the high-strength alloy formwork, such as high yield, high tensile strength, high bearing capacity, and stability, is guaranteed.

[0121] like Figure 7 As shown, the present invention provides a skeleton spacing detection system based on a high-strength alloy template structure, the system comprising:

[0122] A data preprocessing module is used to obtain skeleton grayscale images of the high-strength alloy template structure at different angles, construct three-dimensional point cloud data of the skeleton of the high-strength alloy template structure based on the skeleton grayscale images at different angles, and obtain the point cloud skeleton area in the three-dimensional point cloud data and the local skeleton area corresponding to the same position of the skeleton grayscale images at different angles;

[0123] The credibility analysis module is used to obtain the information credibility of the point cloud skeleton area corresponding to each local skeleton area corresponding to the same position according to the difference distribution between the grayscale fluctuations of each local skeleton area corresponding to the same position in the skeleton grayscale images at different angles, and the feature distribution between the visual angle and the grayscale information richness;

[0124] The error analysis module is used to obtain the shape error coefficient of each point cloud skeleton area according to the area matching relationship and data difference between each point cloud skeleton area and the standard point cloud data of the standard high-strength alloy template structure;

[0125] The feature detection module is used to combine the information credibility and the shape error coefficient, screen the edge feature points according to the discrete feature distribution of each edge feature point in each point cloud skeleton area, and determine the skeleton spacing detection result according to the screening result.

[0126] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the skeleton spacing detection system based on the high-strength alloy template structure and the skeleton spacing detection method embodiment based on the high-strength alloy template structure provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0127] The present application also provides a computer device. Figure 8 , which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, the computer device includes a memory 801, a processor 802, and a computer program 803 stored in the memory 801 and running on the processor 802, wherein when the processor 802 executes the computer program 803, the computer device can execute any of the skeleton spacing detection methods based on the high-strength alloy template structure introduced above.

[0128] The embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute any of the above-mentioned skeleton spacing detection methods based on the high-strength alloy template structure.

[0129] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code is executed on a computer device, the computer device can execute any of the above-mentioned skeleton spacing detection methods based on the high-strength alloy template structure.

[0130] In the embodiments provided in the present application, it should be understood that the provided computer device, computer program product and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0131] The embodiments described above 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, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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.

Claims

1. A method for detecting the skeleton spacing based on a high-strength alloy template structure, characterized in that: The method comprises the following steps: Obtain skeleton grayscale images of the high-strength alloy template structure at different angles, construct three-dimensional point cloud data of the skeleton of the high-strength alloy template structure based on the skeleton grayscale images at different angles, and obtain the point cloud skeleton area in the three-dimensional point cloud data and the local skeleton area corresponding to the same position of the skeleton grayscale images at different angles; According to the difference distribution between the grayscale fluctuations of each local skeleton region corresponding to the same position in the skeleton grayscale images at different angles, and the feature distribution between the visual angle and the grayscale information richness, the information credibility of the point cloud skeleton region corresponding to each local skeleton region corresponding to the same position is obtained; According to the regional matching relationship and data difference between each point cloud skeleton area and the standard point cloud data of the standard high-strength alloy template structure, the shape error coefficient of each point cloud skeleton area is obtained; In combination with the information credibility and the shape error coefficient, the edge feature points are screened according to the discrete feature distribution of each edge feature point in each point cloud skeleton area, and the skeleton spacing detection result is determined according to the screening result.

2. A method for detecting the skeleton spacing based on a high-strength alloy template structure according to claim 1, characterized in that: The step of obtaining the point cloud skeleton region in the three-dimensional point cloud data and the local skeleton region corresponding to the same position of the skeleton grayscale images at different angles specifically includes: The three-dimensional point cloud data is segmented to obtain the point cloud skeleton region, and the local skeleton region corresponding to the skeleton grayscale image of the point cloud skeleton region at each angle is recorded as the reference skeleton region of the point cloud skeleton region at each angle.

3. The method for detecting the skeleton spacing based on the high-strength alloy template structure according to claim 2 is characterized in that: The information credibility of the point cloud skeleton region corresponding to each local skeleton region corresponding to the same position is obtained according to the difference distribution between the grayscale fluctuations of each local skeleton region corresponding to the same position in the skeleton grayscale images at different angles, and the feature distribution between the visual angle and the grayscale information richness, specifically including: For any point cloud skeleton region, the plane disparity factor of the point cloud skeleton region is obtained according to the difference in grayscale fluctuations in the reference skeleton region at each angle of the point cloud skeleton region; According to the grayscale information richness of the reference skeleton region at each angle of the point cloud skeleton region and the angle acquisition position information corresponding to the skeleton grayscale image of each reference skeleton region, the feature richness factor of the point cloud skeleton region is obtained; The information credibility of the point cloud skeleton region is obtained according to the plane disparity factor and the feature enrichment factor, wherein the plane disparity factor is negatively correlated with the information credibility, and the feature enrichment factor is positively correlated with the information credibility.

4. A method for detecting the skeleton spacing based on a high-strength alloy template structure according to claim 3, characterized in that: The plane parallax factor of the point cloud skeleton region is obtained according to the difference of the grayscale fluctuation in the reference skeleton region at each angle of the point cloud skeleton region, specifically including: Record any angle as a target angle, and determine the grayscale feature coefficient of the reference skeleton area at the target angle based on the discrete degree of the grayscale values ​​of all pixels in the point cloud skeleton area within the reference skeleton area at the target angle; Based on the overall level of difference between the grayscale feature coefficients of the point cloud skeleton region and the reference skeleton region at each two angles, a plane disparity factor of the point cloud skeleton region is determined.

5. A method for detecting the skeleton spacing based on a high-strength alloy template structure according to claim 4, characterized in that: The feature richness factor of the point cloud skeleton region is obtained according to the grayscale information richness of the reference skeleton region at each angle of the point cloud skeleton region and the angle acquisition position information corresponding to the skeleton grayscale image of each reference skeleton region, specifically including: Determine the grayscale information coefficient of the reference skeleton region at the target angle based on the grayscale distribution information entropy of all pixels in the reference skeleton region of the point cloud skeleton region at the target angle; Determine the spatial information coefficient of the reference skeleton region at the target angle based on the angle cosine value of the skeleton grayscale image at the target angle when the image is acquired, and the negative correlation coefficient of the distance between the device for image acquisition and the plane where the high-strength alloy template is located; Based on the overall level of the product between the grayscale information coefficient and the spatial information coefficient of the reference skeleton region at each angle of the point cloud skeleton region, a feature enrichment factor of the point cloud skeleton region is determined.

6. The method for detecting the skeleton spacing based on the high-strength alloy template structure according to claim 1 is characterized in that: The shape error coefficient of each point cloud skeleton area is obtained according to the regional matching relationship and data difference between each point cloud skeleton area and the standard point cloud data of the standard high-strength alloy template structure, specifically including: Perform point cloud matching on the three-dimensional point cloud data and the standard point cloud data of the standard high-strength alloy template structure to obtain the matching point between each data point in the three-dimensional point cloud data and the standard point cloud data; The shape error coefficient of each point cloud skeleton area is obtained according to the data difference between each data point in each point cloud skeleton area and the matching point in the standard point cloud data.

7. A method for detecting the skeleton spacing based on a high-strength alloy template structure according to claim 6, characterized in that: The step of obtaining the shape error coefficient of each point cloud skeleton region according to the data difference between each data point in each point cloud skeleton region and the matching point in the standard point cloud data specifically includes: For any point cloud skeleton region, the Euclidean distance between each data point in the point cloud skeleton region and the corresponding matching point is calculated, and the shape error coefficient of the point cloud skeleton region is determined based on the overall level of all Euclidean distances in the point cloud skeleton region.

8. The method for detecting the skeleton spacing based on the high-strength alloy template structure according to claim 1 is characterized in that: The step of combining the information credibility and the shape error coefficient and screening the edge feature points according to the discrete feature distribution of each edge feature point in each point cloud skeleton region specifically includes: Obtaining a feature screening amount for each point cloud skeleton region according to the information credibility and shape error coefficient of each point cloud skeleton region; the information credibility and shape error coefficient are both positively correlated with the feature screening amount; According to the data discreteness degree of each edge feature point in each point cloud skeleton area, the feature discrete factor of each edge feature point in each point cloud skeleton area is obtained; The edge feature points in each point cloud skeleton region are screened respectively using the feature screening amount and the feature discrete factor.

9. A method for detecting the skeleton spacing based on a high-strength alloy template structure according to claim 8, characterized in that: The method of obtaining the feature discrete factor of each edge feature point in each point cloud skeleton region according to the data discreteness of each edge feature point in each point cloud skeleton region specifically includes: For any point cloud skeleton region, the data difference ratio between any edge feature point in the point cloud skeleton region and the mean of all edge feature points in the point cloud skeleton region is used as the feature discrete factor of the any edge feature point.

10. The method for detecting the skeleton spacing based on the high-strength alloy template structure according to claim 8, characterized in that: The step of using the feature screening amount and the feature discrete factor to screen edge feature points in each point cloud skeleton region specifically includes: For any point cloud skeleton region, all edge feature points of the point cloud skeleton region are arranged in ascending order according to the value of the feature discrete factor, and the edge feature points are screened based on the feature screening amount of the point cloud skeleton region in the arrangement order to obtain the screening result.

Images