A Binocular Vision-Based Method and System for Monitoring the Flatness of Sports Field Plastic Surfaces
By reconstructing the three-dimensional point cloud data of the sports field using binocular vision, extracting local curvature, water accumulation volume, and dynamic acceleration response characteristic parameters, and generating a comprehensive evaluation index, the problem of low evaluation efficiency and poor applicability in traditional methods is solved, and multi-dimensional flatness and safety evaluation of the entire field is realized.
Patent Information
- Application Number
- CN202610654015.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-13
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-05-13
AI Technical Summary
Existing technologies are insufficient for quickly and comprehensively assessing the flatness of sports field plastics, especially in large areas where they cannot reflect the impact response under dynamic scenarios. Furthermore, traditional methods are inefficient and cannot quantify the continuous surface distribution across the entire field.
The system uses binocular vision to reconstruct the three-dimensional point cloud data of the sports field, extracts local curvature, water volume and dynamic acceleration response characteristic parameters, and generates a comprehensive evaluation index through dual reference space anchor points to achieve multi-dimensional adaptive quantitative evaluation of the entire field.
It enables rapid and comprehensive assessment of the flatness of sports fields, overcomes the limitations of traditional methods, provides scientific assessment of athlete safety and drainage performance, and adapts to the needs of different field types and sports.
Smart Images

Figure CN122223012B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image processing technology, and in particular to a method and system for monitoring the flatness of sports field plastic surfaces based on binocular vision. Background Technology
[0002] The flatness of sports field synthetic surfaces (including basketball courts, running tracks, badminton courts, etc.) directly affects athletes' safety and comfort, as well as the field's drainage performance during construction, acceptance, and long-term use. Currently, the flatness of sports field synthetic surfaces mainly relies on traditional measuring tools, such as three-meter rulers, levels, or single-point laser rangefinders. These methods require sampling point by point on the surface, which suffers from problems such as sparse sampling, low testing efficiency, inability to reflect the continuous surface distribution across the entire field, and, in particular, difficulty in quantifying impact response under dynamic conditions.
[0003] With the development of computer vision technology, detection methods based on light stripe projection and monocular 3D reconstruction have been applied to the flatness assessment of small-scale, static scenes. However, this method is only suitable for local detection in narrow slit areas and cannot be extended to full-field monitoring of large sports fields; moreover, it only focuses on static geometric height differences and fails to assess the disturbances caused by surface undulations in dynamic scenes, which are directly related to the athlete's comfort and injury risk. Summary of the Invention
[0004] This application provides a method and system for monitoring the flatness of sports field plastic surfaces based on binocular vision. By reconstructing the three-dimensional surface shape of the entire field through binocular vision, it extracts feature parameters in three dimensions: local curvature, water accumulation volume, and dynamic acceleration response. Based on the ratio of vector space tension coefficients of dual reference space anchor points, it generates a comprehensive evaluation index, realizing a full-field, multi-dimensional, adaptive quantitative assessment of the flatness of sports field plastic surfaces.
[0005] This application provides a method for monitoring the flatness of sports field plastic surfaces based on binocular vision, including: S101, based on the acquisition of texture image pairs of the plastic surface of the sports field by binocular vision, reconstructs the three-dimensional point cloud data of the field through stereo matching and parallax conversion. S102, perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory. S103: Construct dual reference spatial anchors based on historical site datasets, calculate the vector space tension coefficient between the current site feature parameters and the dual reference spatial anchors, and generate a comprehensive evaluation index. S104, compare the comprehensive evaluation index with the preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
[0006] Preferably, the method for generating the local concave volume feature parameter is as follows: Set a preset drainage allowable threshold and identify connected regions in the residual surface that are below the threshold as potential water accumulation areas; Calculate the volume of each potential water accumulation zone and sum them to obtain the total water accumulation volume of the entire field. The drainage performance index is obtained by exponential transformation based on the ratio of the total water volume to the site area, and is used as a characteristic parameter of the local depression volume.
[0007] Preferably, the dynamic response characteristic parameters are generated in the following way: Pre-determine at least one typical motion trajectory based on the type of sports field and set a constant design speed; calculate the second directional derivative of the residual surface in the tangent direction of the trajectory along the trajectory, and obtain the vertical acceleration by combining the square of the velocity; statistically analyze the proportion of trajectory arc lengths where the vertical acceleration exceeds the preset human comfort threshold to obtain the dynamic smoothness index, which is used as a dynamic response characteristic parameter.
[0008] Preferably, the dual reference spatial anchor points include a first reference spatial anchor point and a second reference spatial anchor point; the first reference spatial anchor point is obtained by statistical analysis of the site feature parameters of the sample with the best flatness level in the historical site dataset, and the second reference spatial anchor point is obtained by statistical analysis of the site feature parameters of the sample with the critically qualified flatness level in the historical site dataset.
[0009] Preferably, the vector space tension coefficient is defined as: Calculate the site characteristic parameters of the site to be tested respectively. With the first reference space anchor point and the second reference space anchor point The vector space tension coefficients between the two vector spaces are used to obtain the first vector space tension coefficient and the second vector space tension coefficient, which are defined as the normalized Euclidean distance.
[0010] Preferably, the generation of the comprehensive evaluation index specifically includes: Substituting the two vector space tension coefficients into the following ratio function, the comprehensive evaluation index Q is generated: ; The first vector space tension coefficient, This is the tension coefficient of the second vector space.
[0011] Preferably, the vector space tension coefficient is calculated according to the following formula: ; Where d is the vector space tension coefficient. This is the m-th component of the current site characteristic parameter. For the m-th component of the reference spatial anchor point, It is a preset minimum positive number.
[0012] Preferably, the step of performing trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface specifically includes: The moving least squares surface fitting method is used to separate the macro trend surface from the original 3D point cloud data. The height of the macro trend surface is subtracted from the height of the original point cloud to obtain the local geometric fluctuation residual surface. The discrete residual point cloud is then interpolated using a preset grid to form a continuously gridded residual surface function.
[0013] Preferably, the local curvature feature parameter is defined as: Calculate the second gradient magnitude of the local geometric wave residual surface. ; Calculate its global mean: ,in The total area of the site This is a local curvature characteristic parameter.
[0014] This application also provides a sports field plastic flatness monitoring system based on binocular vision, the system comprising: an acquisition module, an extraction module, and an evaluation module; The acquisition module is used to acquire texture image pairs of the plastic surface of the sports field based on binocular vision, and reconstruct the three-dimensional point cloud data of the field through stereo matching and parallax conversion. The extraction module is used to perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory. The evaluation module is used to construct dual reference spatial anchors based on historical site datasets, calculate the vector spatial tension coefficient between the current site feature parameters and the dual reference spatial anchors, and generate a comprehensive evaluation index. The comprehensive evaluation index is compared with a preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: Employing binocular vision technology, it can reconstruct the three-dimensional point cloud data of a large sports field without the need for active light projection or contact sampling, relying solely on natural textures. It boasts fast detection speed and wide coverage, overcoming the limitations of traditional discrete point sampling methods that cannot reflect the continuous surface distribution of the entire field.
[0016] The flatness assessment is comprehensively evaluated from multiple dimensions. It not only extracts the local curvature feature parameter that characterizes the geometric change of the surface, but also introduces the local depression volume feature parameter that characterizes the drainage performance and the dynamic response feature parameter that characterizes the dynamic user experience. The three feature parameters describe the site quality from the perspectives of geometry, fluid mechanics and human dynamics, respectively, making the flatness assessment more comprehensive and scientific.
[0017] By constructing dual reference space anchor points based on historical best and critical qualified samples, and using the ratio of vector space tension coefficients to generate a comprehensive evaluation index, the subjective problem of manually setting weights in traditional weighted fusion methods is avoided. At the same time, it does not require relying on a large amount of labeled data to train the neural network, and has good portability and site type adaptability.
[0018] The study quantitatively assessed athletic comfort and drainage safety. By simulating vertical acceleration disturbances along typical athletic trajectories, the abstract concept of "smoothness" was transformed into measurable knee joint impact indicators for athletes. The study also quantified the drainage performance of the field by calculating the volume distribution of accumulated water throughout the field. These indicators are directly related to the actual user experience of athletes and the safety of the field, providing a scientific basis for the acceptance and maintenance of sports venues. Attached Figure Description
[0019] Figure 1 This is a flowchart illustrating the method for monitoring the flatness of sports field plastic surfaces based on binocular vision, according to an embodiment of the present invention. Figure 2 This is a structural block diagram of a sports field plastic flatness monitoring system based on binocular vision, according to an embodiment of the present invention. Detailed Implementation
[0020] To facilitate understanding of the present invention, a more complete description of this application will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the present invention.
[0021] It should be noted that the terms "vertical," "horizontal," "up," "down," "left," "right," and similar expressions used in this article are for illustrative purposes only and do not represent the only possible implementation.
[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention; the term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0023] Example 1: During construction, acceptance, and long-term use, the flatness defects of sports field synthetic surfaces not only manifest as static geometric undulations but also directly affect drainage safety, athlete comfort, and injury risk. Traditional detection methods rely on three-point rulers, levels, or single-point laser ranging, which have inherent limitations such as sparse sampling, inability to reflect the overall distribution, and difficulty in quantifying local deformation volume and dynamic impact response. While related technologies can detect local height differences and width uniformity in joint areas, they rely on light stripe projection and monocular 3D reconstruction, making them suitable for small-scale, static joint scenarios. They cannot be extended to full-field monitoring of large sports fields and cannot simulate the vertical acceleration disturbances experienced by athletes during actual movement.
[0024] To this end, this embodiment obtains the overall field shape point cloud through passive 3D reconstruction, and sequentially extracts curvature features that characterize local geometric changes, volume features that characterize water accumulation risk, and acceleration perturbation features that characterize dynamic user experience. It then uses historical reference spatial anchor points for unsupervised adaptive discrimination, ultimately achieving a comprehensive, quantitative, and efficient evaluation of the flatness of the sports field's plastic surface.
[0025] Figure 1 This is a flowchart illustrating the method for monitoring the flatness of sports field plastic surfaces based on binocular vision, according to an embodiment of the present invention.
[0026] like Figure 1 As shown, a method for monitoring the flatness of sports field plastic surfaces based on binocular vision includes the following steps: S101 uses binocular vision to acquire texture image pairs of the plastic surface of the sports field, and reconstructs the three-dimensional point cloud data of the field through stereo matching and parallax transformation.
[0027] Specifically, two industrial-grade cameras, pre-calibrated for both internal and external parameters, are first installed at predetermined locations around the sports field. The angle between their optical axes (i.e., the angle between the projection lines of the two cameras' optical axes on the horizontal plane) is controlled between 15° and 25°. The perpendicular bisector of the baseline direction is used as the axis of symmetry, and the baseline distance is adjusted according to the width of the field (e.g., 0.3 to 0.5 times the width of the field) to ensure full coverage of the field surface, thus forming a binocular stereo vision system. Under uniform lighting conditions, the two cameras simultaneously acquire texture images of the field's plastic surface, obtaining left and right image pairs with parallax. To ensure the accuracy of subsequent stereo matching, random speckle textures (e.g., spraying washable environmentally friendly water-based paint) can be randomly sprayed onto the field surface to increase texture features, but without relying on active light stripe coding or other active light source projection.
[0028] For the acquired left and right image pairs, epipolar correction is first performed using pre-calibrated camera intrinsic parameters, distortion coefficients, and binocular stereo parameters (rotation matrix R, translation vector T) to align the rows of the left and right images. This means reprojecting the image plane onto a common virtual plane so that corresponding points are located on the same horizontal scan line. After epipolar correction, the matching search range is reduced from two dimensions to one dimension, which can significantly improve computational efficiency and matching accuracy.
[0029] A semi-global matching (SGM) algorithm is used for stereo matching to calculate the disparity value d(u,v) for each pixel, where (u,v) are the pixel coordinates. Based on the camera intrinsic parameter matrix, distortion coefficients, and rotation matrix and translation vector between the left and right cameras obtained from camera calibration, the disparity map is converted into a three-dimensional spatial point cloud coordinate set with the site center as the origin. Where N is the total number of point clouds, For horizontal coordinates, This represents the height value relative to a reference plane of the site (e.g., the design elevation plane). For example, based on the camera projection matrix, the parallax value is converted into three-dimensional coordinates (X, Y, Z) with the site center as the origin using the following triangulation formula: , Where B is the baseline distance of the binocular cameras, f is the equivalent focal length of the cameras, and d is the parallax value. Principal point coordinates The focal lengths in the X and Y directions are used to ultimately generate a set of 3D point clouds. .
[0030] This step does not involve any active light projection or phase calculation, relying entirely on passive vision and natural texture to achieve rapid, non-contact 3D reconstruction of large areas. It should be noted that specific details regarding stereo matching and parallax conversion can be found in relevant technical principles; this invention will not elaborate on or limit these aspects.
[0031] S102, perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory.
[0032] In some embodiments, S102 specifically includes: S201, perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface.
[0033] Since sports fields typically have a designed drainage slope (e.g., a 1% cross slope for running tracks, and drainage from the center outwards for basketball courts), it is necessary to separate the macro-topographical trends from local construction defects. Using a pre-defined moving least squares surface fitting method, a macro-trend surface is separated from the original 3D point cloud data. The height of the macro-trend surface is subtracted from the height of the original point cloud to obtain a local geometric fluctuation residual surface. The discrete residual point cloud is then subjected to a pre-defined grid interpolation to form a continuously gridded residual surface function.
[0034] Specifically: Moving least squares surface fitting is used on point cloud P to separate the macro trend surface. With local geometric fluctuation residual surface .
[0035] For example, the moving least squares (MLS) surface fitting method is used to fit each point in the point cloud P. A neighborhood point set with radius R (e.g., 200 mm) is selected, and a weighted least squares fit is performed using quadratic basis functions (polynomial order 2, with 6 basis terms: 1, X, Y, X², XY, Y²). To avoid underdetermining of the point set in the local neighborhood, a regularization term (Tikhonov regularization, with a regularization parameter of 0.001) is introduced to obtain the macro trend surface. (Represents the design slope or natural drainage slope); the fitting weight function is a Gaussian kernel: ,in The horizontal distance from a point in the neighborhood to the point to be fitted. The kernel density is (taken as R / 3); after fitting, the local geometric fluctuation residual surface is: Understandably, for each point to be fitted, a locally weighted least squares normal equation is constructed, and the coefficients of the quadratic polynomial are solved to obtain the value of the macro trend surface at that point.
[0036] It should be noted that after obtaining the discrete point cloud of the local geometric fluctuation residual surface, a bilinear interpolation method is used to interpolate the residual point cloud onto a regular grid corresponding to the horizontal range of the site. The grid spacing is 10 mm to 50 mm (typically 20 mm), forming a continuously gridded residual surface function. This is to facilitate subsequent calculations of partial derivatives, curvature, and integrals. If there are holes (missing data) in the point cloud in a local area, they are filled using neighboring points. If the area of the hole exceeds 0.1 square meters (corresponding to approximately 25 missing grid points), it is marked and excluded from the calculation in subsequent calculations.
[0037] Therefore, the residual surface reflects the unevenness caused by pure construction defects or aging deformation. Positive values indicate areas that are higher than the design slope (bulges), while negative values indicate areas that are lower than the design slope (sinks or puddles).
[0038] S202, based on the local geometric fluctuation residual surface, extracts local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory, and combines them into site feature parameters.
[0039] A1, Local curvature characteristic parameter Extraction.
[0040] To quantify the degree of geometric abrupt changes on the site surface, the residual surface is calculated. The second-order gradient magnitude, i.e., the local curvature characteristic parameter. : Calculate the gradient component of the residual surface in the horizontal direction: Calculate the second-order partial derivative: On a discrete grid, the second-order partial derivatives are calculated using the central difference method: Similarly, calculate and mixed partial derivatives (Using the nearest four-point difference), where The grid spacing is used; forward or backward subtraction can be used instead of central subtraction at the grid boundaries. The local curvature characteristic parameter is defined as the gradient magnitude (i.e., the Frobenius norm of the Hessian matrix): The dimensions of this parameter are It is used to characterize the severity of local bending on the surface; the larger the value, the sharper the protrusion or depression. To obtain a single value for the entire game, calculate its global mean: ,in Total site area, This serves as an input feature for subsequent comprehensive evaluation. The smaller the value, the flatter the surface of the field.
[0041] A2. Local concavity volume characteristic parameter Extraction.
[0042] The local depression volume characteristic parameter is used to evaluate the drainage performance of the site. The local depression volume characteristic parameter is generated as follows: a preset drainage allowable threshold is set, and connected regions in the residual surface below the threshold are identified as potential water accumulation areas; the volume of each potential water accumulation area is calculated, and the total water accumulation volume of the entire site is obtained by summing them; then, based on the ratio of the total water accumulation volume of the entire site to the site area, the drainage performance index is obtained through exponential transformation, which is used as the local depression volume characteristic parameter; the closer the drainage performance index is to 1, the lower the risk of water accumulation.
[0043] Specifically: Preset drainage allowable threshold Based on the regulations regarding water depth limits in sports field construction specifications, this example uses 2mm (meaning water may accumulate if the depression depth exceeds 2mm). Traverse all grid points of the residual surface and identify those that satisfy the condition. Connected regions (connectivity can be determined using 8-neighborhood, including horizontal, vertical, and diagonal adjacencies); for each connected region Calculate its area And filter out isolated noise points with too small an area (preset an area threshold, for example, in this embodiment, take...). (Connected regions below the area threshold are considered noise and ignored), thus obtaining the potential water accumulation area; For each potential flooded area, calculate its water volume: Total water volume of the entire site To normalize the volume dimensions to The interval is then positively correlated with the flatness, and the volumetric characteristic parameter of local depressions is defined as the drainage performance index: ,in This is a preset reference depth constant used to adjust the exponential decay rate. Its value can be preset according to the site type and acceptance requirements. In this embodiment, 5mm is preferred. The closer to 1, the less risk of water accumulation; the closer to 0, the more severe the water accumulation.
[0044] A3. Extract the dynamic response characteristic parameter U along at least one preset typical motion trajectory to evaluate the dynamic use comfort of the site.
[0045] The dynamic response characteristic parameters are generated in the following way: Pre-determine at least one typical motion trajectory based on the type of sports field and set a constant design speed; calculate the second directional derivative of the residual surface in the tangent direction of the trajectory along the trajectory, and obtain the vertical acceleration by combining the square of the velocity; statistically analyze the proportion of trajectory arc lengths where the vertical acceleration exceeds the preset human comfort threshold to obtain the dynamic smoothness index, which is used as a dynamic response characteristic parameter.
[0046] Specifically: Based on the actual use of the sports venue (such as a standard basketball court or a track and field track), several typical sports trajectories are pre-defined. For example, for a standard basketball court, the following can be set: a straight sprint trajectory from the baseline to the opposite baseline, a free throw line shuttle run trajectory, and a three-point line arc cutting trajectory; for a 400-meter track, the following can be set: a 200-meter sprint trajectory on the inner track and a curve cutting trajectory; each trajectory is represented as a series of discrete points (X(s), Y(s)), where s is the arc length parameter accumulated from the starting point, and the step size is 10 millimeters; Assuming the athlete maintains a constant design speed For motion along a trajectory (the velocity value can be determined from a table for the sport, e.g., 6 m / s for basketball sprints and 4 m / s for shuttle runs), the vertical acceleration originates from the surface curvature along the tangent of the trajectory, and the calculation formula is: ,in The second-order directional derivative of the residual surface along the tangent direction of the trajectory is approximated by the central difference method at the discrete trajectory points: ,in Set the trajectory step size (10mm). To obtain from the residual surface mesh via bilinear interpolation; Predefined human comfort vertical acceleration threshold (The value can be determined based on the experience of sports experts; in this embodiment, it is taken as follows) (Exceeding this value will cause significant discomfort or motor imbalance), statistics on each trajectory. arc length segment Calculate the dynamic smoothness index of the trajectory, i.e., the characteristic parameter component of the dynamic response: ,in Let be the total length of the k-th trajectory. If there are multiple trajectories, the average value is taken as the characteristic parameter U of the dynamic response of the entire field. The range of values for U is The closer the value is to 1, the less interference the field has on dynamic movement, and the higher the comfort of the movement.
[0047] Thus, S102 generates three site feature parameters: local curvature feature and global mean. Drainage performance index The three factors—dynamic flatness index U, dynamic smoothness index, and dynamic smoothness index—have different dimensions. After extraction in this step, The larger the values of U and U, the better the site performance. The closer the value is to 1, the lower the risk of flooding; the closer U is to 1, the less dynamic the impact. The smaller the value, the flatter the surface. When constructing reference space anchor points later, the optimal anchor point corresponds to... Small, The combination of large and large U means that the three indicators naturally form a unified direction of advantage or disadvantage in the feature space, without the need for additional conversion.
[0048] S103. Based on the historical site dataset, construct dual reference spatial anchors, including a first reference spatial anchor and a second reference spatial anchor. Calculate the vector space tension coefficients between the current site feature parameters and the first and second reference spatial anchors respectively, and generate a comprehensive evaluation index.
[0049] In some embodiments, the first reference spatial anchor point is obtained by statistically analyzing the corresponding feature parameters of the optimal flatness grade sample, and the second reference spatial anchor point is obtained by statistically analyzing the corresponding feature parameters of the critical acceptable flatness grade sample. The construction of the first and second reference spatial anchor points based on the historical site dataset specifically includes: A large number of accepted sports field samples were collected in advance, each sample containing three pieces of information: measured data... The vector and the flatness rating determined by a professional organization are classified as follows: Optimal level: The maximum gap measured by a 3m ruler (referring to the maximum gap value measured at any detection position) is less than or equal to the preset first gap threshold (e.g., 2mm), and there are no recessed areas with an area greater than or equal to the preset first area threshold (e.g., 0.05m²) and a depth greater than or equal to the preset first depth threshold (e.g., 2mm); Critical acceptance level: Meeting one of the following conditions is sufficient, but all sub-items of that condition must be met simultaneously: Condition A: The maximum gap measured with a 3m ruler is within the preset second gap range (for example, between 3mm and 4mm), and there are no local depressions with a depth exceeding the second depth threshold (e.g., 5mm); Condition B: There is at least one local depression with an area less than or equal to a preset second area threshold (e.g., 0.2m²) and a depth less than or equal to a preset third depth threshold (e.g., 3mm), and the maximum gap measured by a 3m ruler is less than or equal to a preset third gap threshold (e.g., 3mm). Select all samples of the highest level, calculate the arithmetic mean of their three feature parameters, and construct the first reference space anchor point. ;
[0050] For the second reference space anchor point, the following quantitative rules are used to screen critically qualified samples: Samples independently judged as "critically qualified" (i.e., barely meeting the acceptance criteria, and failing if slightly worse) by at least five experts with sports field testing qualifications (at least three of whom hold senior professional titles) are selected; samples unanimously agreed upon by at least four experts are included. If there are no expert judgment criteria, samples with a flatness level of "qualified" and a maximum gap measured with a traditional 3m straightedge between 3mm and 4mm are selected from the historical dataset. The arithmetic mean of the three characteristic parameters of these critically qualified samples is calculated to constitute the second reference space anchor point. .
[0051] The first and second reference space anchor points represent typical positions of the "ideal state" and "barely acceptable state" in the three-dimensional feature space, respectively.
[0052] It should be noted that different sports have different trajectory types, design speeds, and human acceleration thresholds (e.g., basketball courts require rapid changes of direction, while running tracks require sustained long distances). Therefore, independent historical datasets and reference spatial anchors should be constructed for each type of venue (basketball court, standard running track, badminton court, etc.). During detection, the corresponding dataset is automatically selected for matching and judgment based on the current venue type.
[0053] In some embodiments, the vector space tension coefficient is defined as: Calculate the site characteristic parameters of the site to be tested respectively. With the first reference space anchor point and the second reference space anchor point The vector space tension coefficient between them is defined as the normalized Euclidean distance, which aims to eliminate the influence of differences in the dimensions of different feature parameters. The calculation formula is as follows: ; Where d is the vector space tension coefficient. This is the m-th component of the current site characteristic parameter. For the m-th component of the reference spatial anchor point, The minimum positive number is preset; Specifically, for the first reference space anchor point and the second reference space anchor point, the corresponding first vector space tension coefficient and second vector space tension coefficient are as follows: ; in, For a preset minimum positive number (e.g.) This coefficient is used to avoid division by zero errors caused by a zero vector. The physical meaning of this coefficient is: the relative distance of the current point vector from the reference anchor point; the smaller the value, the closer the current site state is to the reference anchor point.
[0054] In some embodiments, generating the comprehensive evaluation index specifically includes: Substituting the two vector space tension coefficients into the following ratio function, the comprehensive evaluation index Q is generated: ; The index ranges from [value range]. ;like If it is very close to the first anchor point (ideal state), then As Q approaches 0, Q approaches 0; if Very close to the second anchor point (critical state). If Q approaches 0, then Q approaches 1; if The value falls between these two, where Q is close to 0.5. Therefore, a smaller Q value indicates better site quality.
[0055] S104, compare the comprehensive evaluation index with the preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
[0056] Specifically, by analyzing the Q-value distributions of "qualified" and "unqualified" sites in a large number of historical samples (e.g., at least 200 historical samples, with no fewer than 80 qualified and 80 unqualified samples), the optimal cutoff point is determined using ROC curves (based on maximizing the Youden index), thus obtaining the preset qualification threshold. (Based on the pre-set acceptance standards for different types of sites, the typical value range is 0.55 to 0.65, depending on the site grade requirements.) when If the flatness of the sports field's plastic surface is deemed acceptable, it is deemed unacceptable; otherwise, it is deemed unacceptable.
[0057] For sites deemed unqualified, the system automatically outputs the following auxiliary information to guide remediation: Map showing the distribution of flooded areas and the total volume of floodwater; Mark the sections where vertical acceleration exceeds the limit and the corresponding peak acceleration on the trajectory diagram; The local curvature characteristic parameter exceeds the global standard deviation of a preset multiplication factor (3 times in this embodiment). The location of the anomaly, among which , The total number of grid points. This represents the average local curvature across the entire field.
[0058] The technical solutions described in the embodiments of this application have at least the following technical effects or advantages: By reconstructing the 3D point cloud of a sports field using binocular vision without active projection, the evaluation of flatness is expanded from "static geometric height difference" to three physically distinct and independent feature parameters: local curvature, water accumulation volume, and dynamic acceleration response. This establishes a multi-dimensional evaluation system in the field of sports field testing, encompassing geometry, fluid dynamics, and human dynamics. By constructing two reference spatial anchor points, "optimal" and "critically qualified," and introducing the ratio of vector space tension coefficients, adaptive judgment without subjective weighting is achieved, solving the technical challenge of varying flatness evaluation standards for different field types and sports. By simulating vertical acceleration disturbances along typical motion trajectories, the abstract concept of "flatness" is transformed into a measurable knee joint impact index for athletes, quantitatively assessing sports comfort.
[0059] Example 2 Figure 2 This is a structural block diagram of a sports field plastic flatness monitoring system based on binocular vision, according to an embodiment of the present invention. like Figure 2As shown, a sports field plastic surface flatness monitoring system based on binocular vision includes: an acquisition module, an extraction module, and an evaluation module. The acquisition module is used to acquire texture image pairs of the plastic surface of the sports field based on binocular vision, and reconstruct the three-dimensional point cloud data of the field through stereo matching and parallax conversion. The extraction module is used to perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory. The evaluation module is used to construct dual reference spatial anchors based on historical site datasets, calculate the vector spatial tension coefficient between the current site feature parameters and the dual reference spatial anchors, and generate a comprehensive evaluation index. The comprehensive evaluation index is compared with a preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
[0060] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for monitoring the flatness of a sports field based on binocular vision, characterized in that, include: S101, based on the acquisition of texture image pairs of the plastic surface of the sports field by binocular vision, reconstructs the three-dimensional point cloud data of the field through stereo matching and parallax conversion. S102, perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concave volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory; the dynamic response feature parameters are generated by: presetting at least one typical motion trajectory according to the type of sports field, and setting a constant design speed. The second directional derivative of the residual surface in the tangent direction of the trajectory is calculated along the trajectory, and the vertical acceleration is obtained by combining it with the square of the velocity. The proportion of trajectory arc lengths where vertical acceleration exceeds a preset human comfort threshold is statistically analyzed to obtain the dynamic smoothness index, which serves as a characteristic parameter of the dynamic response. S103, construct dual reference spatial anchors based on historical site datasets, calculate the vector space tension coefficient between the current site feature parameters and the dual reference spatial anchors respectively, and generate a comprehensive evaluation index; the dual reference spatial anchors include a first reference spatial anchor and a second reference spatial anchor; The first reference spatial anchor point is obtained by statistical analysis of the site feature parameters of the sample with the best flatness level in the historical site dataset, and the second reference spatial anchor point is obtained by statistical analysis of the site feature parameters of the sample with the flatness level of the critical qualified in the historical site dataset. The vector space tension coefficient is defined as follows: calculated respectively the site feature parameters of the current site to be detected. With the first reference space anchor point and the second reference space anchor point The vector space tension coefficients between the two vectors are used to obtain the first and second vector space tension coefficients, which are defined as the normalized Euclidean distance; where... This refers to the local curvature characteristic parameter of the current site. This refers to the local depression volume characteristic parameter of the current site. The dynamic response characteristic parameter of the current site; the vector space tension coefficient is calculated according to the following formula: d is the vector space tension coefficient. This is the m-th component of the current site characteristic parameter. For the m-th component of the reference spatial anchor point, The minimum positive number is preset; S104, compare the comprehensive evaluation index with the preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
2. The method of claim 1, wherein the method further comprises: The method for generating the local concave volume feature parameter is as follows: Set a preset drainage allowable threshold and identify connected regions in the residual surface that are below the threshold as potential water accumulation areas; Calculate the volume of each potential water accumulation zone and sum them to obtain the total water accumulation volume of the entire field. The drainage performance index is obtained by exponential transformation based on the ratio of the total water volume to the site area, and is used as a characteristic parameter of the local depression volume.
3. The method of claim 2, wherein the method further comprises: The generation of the comprehensive evaluation index specifically includes: Substituting the two vector space tension coefficients into the following ratio function, the comprehensive evaluation index Q is generated: ; The first vector space tension coefficient, This is the tension coefficient of the second vector space.
4. The method for monitoring the flatness of sports field plastic surfaces based on binocular vision as described in claim 3, characterized in that, The step of performing trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface specifically includes: The moving least squares surface fitting method is used to separate the macro trend surface from the original 3D point cloud data. The height of the macro trend surface is subtracted from the height of the original point cloud to obtain the local geometric fluctuation residual surface. The discrete residual point cloud is then interpolated using a preset grid to form a continuously gridded residual surface function.
5. The method for monitoring the flatness of sports field plastic surfaces based on binocular vision as described in claim 4, characterized in that, The local curvature feature parameter is defined as follows: Calculate the second gradient magnitude of the local geometric wave residual surface. X and Y are the horizontal coordinates of the local geometric fluctuation residual surface, corresponding to the two horizontal coordinates of the site, respectively. For the residual surface in coordinate The second-order gradient magnitude at that point; Calculate its global mean: ,in The total area of the site This is a local curvature characteristic parameter.
6. A sports field plastic surface flatness monitoring system based on binocular vision, characterized in that, The system for monitoring the flatness of sports field plastic surfaces based on binocular vision, as described in any one of claims 1-5, comprises: an acquisition module, an extraction module, and an evaluation module. The acquisition module is used to acquire texture image pairs of the plastic surface of the sports field based on binocular vision, and reconstruct the three-dimensional point cloud data of the field through stereo matching and parallax conversion. The extraction module is used to perform trend separation on the three-dimensional point cloud data to extract the local geometric fluctuation residual surface, and generate site feature parameters based on the residual surface, including local curvature feature parameters, local concavity volume feature parameters, and dynamic response feature parameters along at least one preset typical motion trajectory. The dynamic response feature parameters are generated as follows: at least one typical motion trajectory is preset according to the type of sports field, and a constant design speed is set; the second directional derivative of the residual surface in the tangent direction of the trajectory is calculated along the trajectory, and the vertical acceleration is obtained by combining the square of the velocity; the proportion of trajectory arc lengths with vertical acceleration exceeding the preset human comfort threshold is statistically analyzed to obtain the dynamic smoothness index, which is used as the dynamic response feature parameter. The evaluation module is used to construct dual reference spatial anchors based on historical site datasets, calculate the vector space tension coefficient between the current site feature parameters and the dual reference spatial anchors, and generate a comprehensive evaluation index. The dual reference spatial anchors include a first reference spatial anchor and a second reference spatial anchor. The first reference spatial anchor is obtained statistically from the site feature parameters of the sample with the best flatness level in the historical site dataset, and the second reference spatial anchor is obtained statistically from the site feature parameters of the sample with a flatness level of near-qualified in the historical site dataset. The vector space tension coefficient is defined as: the calculated site feature parameters of the current site to be tested. With the first reference space anchor point and the second reference space anchor point The vector space tension coefficients between the two vectors are used to obtain the first and second vector space tension coefficients, which are defined as the normalized Euclidean distance; where... This refers to the local curvature characteristic parameter of the current site. This refers to the local depression volume characteristic parameter of the current site. The dynamic response characteristic parameter of the current site; the vector space tension coefficient is calculated according to the following formula: d is the vector space tension coefficient. This is the m-th component of the current site characteristic parameter. For the m-th component of the reference spatial anchor point, The minimum positive number is preset; the comprehensive evaluation index is compared with the preset qualified threshold to determine whether the flatness of the sports field plastic is qualified.
Citation Information
Patent Citations
Double-robot cooperation weakly-rigid component milling rigidity strengthening and flutter suppression method
CN113059571A
Wall flatness detection method and system based on three-dimensional scanning technology
CN119197391A
