Aggregate particle morphology quantification method and system based on Fourier spectrum decoupling

Abstract

The present application relates to the field of aggregate morphology detection and image processing, and discloses a method and system for quantifying aggregate particle morphology based on Fourier spectrum decoupling, including: a three-dimensional point cloud acquisition module for acquiring three-dimensional point cloud data of aggregate particles; a point cloud data processing module for processing the three-dimensional point cloud data to obtain the two-dimensional contour data; a morphological feature analysis module for processing the two-dimensional contour data to obtain aggregate decoupling feature data; and a detection and analysis module for associating various modules. The present invention uses an RGB‑D deep point cloud system to collect multi-view point clouds, and obtains a complete three-dimensional morphological point cloud through noise reduction and registration. Through Monte Carlo simulation sectioning, two-dimensional isoperimetric circle mapping, contour Fourier series fitting and gradient operator definition, a calculation method for decoupling the shape, edges and texture of aggregates using aggregate morphological parameters is established, thereby improving the precision and accuracy of aggregate three-dimensional morphological testing and analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of aggregate morphology detection and image processing, and in particular to a method and system for quantifying aggregate particle morphology based on Fourier spectrum decoupling. Background Art

[0002] Aggregate quality testing and evaluation is a crucial topic in the field of road engineering materials. Aggregate morphology has a crucial influence on the mechanical properties, durability, and service behavior of asphalt mixtures. Aggregates comprise the vast majority of asphalt pavement materials. Their macroscopic shape (e.g., cubic or flaky), microscopic angularity (sharpness), and microscopic texture (surface roughness) collectively form the physical foundation for the skeleton interlocking effect and interfacial adhesion, directly impacting key pavement properties such as high-temperature rutting resistance, low-temperature cracking resistance, and resistance to water damage.

[0003] Traditional inspection techniques, such as the vernier caliper method and the gap ratio method, can only obtain composite indicators of shape and angularity, which are macro-generalized measurements. Two-dimensional visual measurement calculates parameters such as roundness and angularity by projecting contours, but the three-dimensional texture fluctuations are compressed, leading to systematic deviations. Three-dimensional scanning algorithms based on the convex body assumption can cause serious angularity misjudgments in concave areas. CT transmission methods are difficult to promote in engineering due to their high cost, high threshold, and low efficiency. Given the practical application requirements for scientific, accurate, and efficient testing methods, an efficient and scientific analytical testing method is urgently needed. Summary of the Invention

[0004] The purpose of the present invention is to specifically solve the defects existing in the above-mentioned technical status, and to provide a decoupling and quantification method for aggregate particle morphology based on Fourier spectrum descriptors, so as to realize the decoupling and quantitative analysis of the common morphological characteristics of concave and convex aggregates, and provide technical support for the morphological quantification of aggregates and granular materials.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for quantifying aggregate particle morphology based on Fourier spectrum decoupling, comprising the following steps:

[0006] Step 1: Select aggregate particle samples, collect multi-view point clouds of aggregate particles through RGB-D deep point cloud system, perform noise reduction and registration on the multi-view point clouds to obtain the complete 3D point cloud of aggregate particles and obtain the 3D coordinates ;

[0007] Step 2: Reconstruct the point cloud surface, use Monte Carlo simulation to randomly cut the point cloud to obtain a set of contour surfaces, reduce the dimensionality of the points on each cut contour surface, and obtain the coordinates of the two-dimensional contour ;

[0008] Step 3: Perform isoperimetric mapping on the original contour points of the two-dimensional contour coordinates, sample the isoperimetric circles to obtain sampling points, and inversely calculate the sampling mapping coordinates on the original contour based on the sampling mapping;

[0009] Step 4: Construct the aggregate profile descriptor Ω based on Fourier series fitting, and map the coordinates according to the sampling Solve for the aggregate profile descriptor Ω;

[0010] Step 5: Calculate the gradient of the aggregate profile descriptor Ω and gradient direction angle , construct the gradient direction integral, transform it into discrete numerical integral to obtain the two-dimensional aggregate morphological parameters ;

[0011] Step 6: Based on the Monte Carlo sectioning and isoperimetric mapping coordinates and parameters in steps 2, 3, 4, and 5 , construct double integral and perform discretized numerical integration to obtain three-dimensional aggregate morphological parameters ;

[0012] Step 7: Decouple the aggregate morphology according to the Fourier series threshold to obtain the three-dimensional aggregate morphology parameters that characterize the morphological characteristics of aggregate particles at different scales. data.

[0013] Preferably, the nominal maximum particle size range of the aggregate selected for sieving in step 1 is 2.36 mm to 19 mm.

[0014] Preferably, the use of a raster scanning system for depth point cloud acquisition in step 1 has the advantages of high precision and easy operation.

[0015] Preferably, in step 1, based on the point cloud coordinates obtained by the raster scanning system, the point cloud is subjected to noise reduction filtering and smoothing processing using CloudCompare software and then spliced to obtain a complete scanning point cloud of the aggregate outer surface.

[0016] Preferably, in step 2, according to the coordinates of the surface point cloud of the aggregate particles, based on Monte Carlo simulation, first a circumscribed sphere is constructed and Fibonacci grid points are established, and the cross-section of the contour is obtained by randomized sectioning and spatial dimensionality reduction.

[0017] A morphological decoupling and quantification system for aggregate particles based on Fourier spectrum descriptors includes: a three-dimensional point cloud acquisition module for acquiring three-dimensional point cloud data of aggregate particles; a point cloud data processing module for processing the three-dimensional point cloud data to obtain the two-dimensional contour data; a morphological feature analysis module for processing the two-dimensional contour data to obtain aggregate decoupling feature data; a detection and analysis module for associating the modules to realize the functions of aggregate point cloud acquisition, point cloud processing, and morphological analysis, and summarizing and analyzing the data results. The system operates using the above-mentioned detection method.

[0018] The present invention provides a method and system for quantifying aggregate particle morphology based on Fourier spectrum decoupling. This method has the following beneficial effects:

[0019] 1. This invention uses an RGB-D deep point cloud system to collect multi-viewpoint point clouds, which are then registered through noise reduction to produce a complete 3D point cloud. Through Monte Carlo simulation and sectioning, combined with isoperimetric mapping, Fourier series fitting, and gradient calculation, aggregate morphological parameters are accurately constructed from two-dimensional and three-dimensional perspectives, improving test accuracy.

[0020] 2. The method of the present invention can achieve decoupling and quantitative analysis of morphological characteristics through systematic steps, whether it is for concave coarse aggregates with complex shapes or relatively regular convex coarse aggregates, breaking through the limitations of traditional detection and being applicable to aggregates of various shapes.

[0021] 3. The present invention consists of modules such as 3D point cloud acquisition and point cloud data processing. The acquisition module's raster scanning system is highly accurate and easy to operate. Data processing utilizes specialized software, and the modules work collaboratively to facilitate rapid on-site acquisition, processing, and analysis of aggregate data. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Schematic diagram of isoperimetric circle mapping of the contour of the present invention;

[0023] Figure 2 is a flow chart of the method of the present invention;

[0024] Figure 3 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the present specification. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] Please see the attached Figure 1 -Attached Figure 3 The embodiment of the present invention provides a method for quantifying aggregate particle morphology based on Fourier spectrum decoupling, comprising the following steps:

[0027] 1. Aggregate sample selection and scanning:

[0028] The selected aggregate samples should be clean and dry. A binocular structured light depth scanner was used to scan the aggregate particle surface at different viewing angles to generate RGB-D depth point cloud data. The depth point cloud was then processed using CloudCompare software, followed by noise reduction, cropping, splicing, and refined interpolation and smoothing, before being saved as a .pcd file.

[0029] 2.2D Contour Acquisition and Mapping Processing

[0030] According to the surface point cloud obtained by scanning in step 1, the implicit surface of the point cloud is reconstructed using Poisson, and the triangular mesh of the point cloud surface is generated based on the Marching Cubes algorithm. , use triangle weighted average method to calculate the centroid of point cloud ;

[0031] The radius is the distance between the contour point with the longest distance from the surface point cloud to the centroid. , the center of mass is used as the center of the circle to construct a sphere, and the spherical coordinates of the Fibonacci grid points are calculated according to the following formula and and converted to three-dimensional coordinates in the Cartesian system :

[0032]

[0033] Where, is the radius, is the polar angle, is the azimuth, It is the golden ratio;

[0034] Create an index set for Fibonacci grid points Use Fisher-Yates random algorithm to shuffle the index of grid points. Construct the centroid of the grid points As a starting point, The direction vector of the end point ;

[0035] Using the formula Determine the direction vector The plane of the normal vector Passing through the point cloud centroid The cutting plane at the time of traversal of the triangular mesh , for each triangle Calculate the distance from each vertex to the cutting plane according to the following formula :

[0036] Where, is the plane normal vector, is the vertex vector of the triangle, is a point on the cutting plane;

[0037] According to the distance from the vertex of each triangle mesh to the cutting plane , determine any side of the triangle Corresponding vertex-section plane distance If the condition is met, then the edge Intersection point on the cutting plane Satisfy parameters and After the calculation is completed, the points on the contour line after one section can be obtained. ;

[0038] According to the direction vector The plane of the normal vector Passing through the point cloud centroid A set of two-dimensional orthogonal bases are defined on the cutting plane. To construct a local coordinate system, the three-dimensional points on the contour line obtained by cutting The following formula can be used to project onto the cutting plane to achieve dimensionality reduction and obtain the two-dimensional coordinates :

[0039] ①Translate to the origin of the local coordinate system:

[0040] ②Calculation On the base Coordinates on , projected onto the cutting plane:

[0041] Where, are the coordinates of the three-dimensional points on the contour line obtained by sectioning, is the coordinate of any point on the cutting contour line after dimensionality reduction, It is constructed on the cutting surface. is the origin of the local coordinate system of the orthogonal basis.

[0042] Select sampling parameters , usually take , then the number of sampling points ;

[0043] like Figure 1 As shown, the perimeter of the original contour is established Circles of equal circumference have a radius of , in polar angle The starting point on the circle . When sampling, scan the isocircle from the starting point, according to Scan points on the circumference , each sampling point To the starting point Arc length Then the radius of the isoperiphery circle can be Its polar angle calculate;

[0044] Also take the polar angle on the original contour Starting point , calculate each point on the contour With the starting point Distance along the contour .by Distance mark from the starting point Scan each point along the contour in sequence. , then directly record the point Coordinates of , otherwise the record satisfies The adjacent points before and after Linear interpolation coordinate points, denoted as , establish any point on the contour From the sampling points on the isoperimetric circle The unique relationship of the mapping, get the mapping coordinates .

[0045] 3. Decoupling and Quantitative Calculation of Aggregate Particle Morphology

[0046] Calculate the centroid of a 2D contour , with the center of mass as the origin of the Cartesian system , with the horizontal axis The intersection with the contour is the starting point Any point The inclination angle with the X axis is ,point The parameterized representation is ;

[0047] Unique set of sampling maps based on Fourier series and original contour points , construct the aggregate profile descriptor Ω:

[0048] Where, and and and are the coefficients of the Fourier series, is the order of the series ;

[0049] Contour sampling mapping points The undetermined Fourier coefficients, which are much larger than the descriptor Ω, are , substitute the coordinates , solve the Fourier coefficient of the descriptor Ω based on the least squares method and obtain the numerical representation of the descriptor Ω:

[0050]

[0051] Where, and and and are the coefficients of the Fourier series, Points , is the order of the series , the meanings of other symbols are the same as above.

[0052] According to the derivative of the aggregate profile descriptor Ω, the gradient and gradient direction angle It can be calculated by the following formula:

[0053]

[0054]

[0055]

[0056]

[0057] In the formula and is the derivative of the aggregate profile descriptor Ω;

[0058] Gradient direction angle Polar angle only with the isoperimetric circle There is a unique corresponding mapping relationship, that is, there is a unique relationship with the original contour points of the aggregate, which is not constrained by the concavity and convexity of the contour geometry. The degree of contour morphology change can be determined by The integral of the change of is reflected to obtain the two-dimensional aggregate morphology parameters And its discrete numerical integration formula:

[0059]

[0060]

[0061] Where, is the number of points on the contour.

[0062] Three-dimensional aggregate morphological parameters , several section contours of the original particle 3D point cloud can be obtained through the Monte Carlo sectioning process, and each section contour constitutes a point set , and the two-dimensional aggregate morphological parameters of each contour are obtained through steps 3, 4, and 5 Points earned:

[0063]

[0064]

[0065] Where, is the total number of contour surfaces obtained after Monte Carlo sectioning of the 3D point cloud, is the number of points on the contour, is the integration variable.

[0066] The aggregate morphology is divided into three scales: shape, angularity, and texture, which are decoupled by Fourier series thresholding. , Describe shape characteristics, Describe the angular characteristics, Describe texture features;

[0067] The analysis of the three-dimensional shape characteristics of aggregates is based on the three-dimensional aggregate morphological parameters Calculation; Analysis of the three-dimensional angular characteristics of aggregates, based on the three-dimensional aggregate morphological parameters Calculation; Analysis of aggregate three-dimensional shape characteristics, based on three-dimensional aggregate morphological parameters calculate.

[0068] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. Aggregate particle morphology quantification method based on Fourier spectrum decoupling, characterized by: The following steps are involved: S1. Multi-view 3D point cloud acquisition and noise reduction registration: Select aggregate particle samples, collect multi-view point clouds of aggregate particles through RGB-D deep point cloud acquisition system, perform noise reduction and registration on the multi-view point clouds to obtain the complete 3D point cloud of aggregate particles, and obtain the 3D coordinates. ; S2. Point cloud surface modeling and random sectioning dimensionality reduction: Reconstruct the point cloud surface, use Monte Carlo simulation to randomly section the point cloud to obtain a set of contour surfaces, reduce the dimensionality of the points on each section contour surface, and obtain the coordinates of the two-dimensional contour. ; S3, isocircle mapping and reverse sampling coordinate inversion: perform isocircle mapping on the original contour points of the two-dimensional contour coordinates, sample the isocircle to obtain sampling points, and reversely calculate the sampling mapping coordinates on the original contour based on the sampling mapping; S4. Modeling and solving of Fourier series profile descriptor: Constructing aggregate profile descriptor Ω based on Fourier series fitting, and mapping the sampling coordinates according to the Solve for the aggregate profile descriptor Ω; S5. Gradient field integration and two-dimensional morphological parameter extraction: Calculate the gradient of the aggregate contour descriptor Ω and gradient direction angle , construct the gradient direction integral, transform it into discrete numerical integral to obtain the two-dimensional aggregate morphological parameters ; S6. Double integral discretization and 3D morphological parameter generation: Based on the Monte Carlo sectioning, isoperimetric mapping coordinates and parameters in steps 2, 3, 4, and 5 , construct double integral and perform discretized numerical integration to obtain three-dimensional aggregate morphological parameters ; S7. Multi-scale Fourier spectrum decoupling and morphological feature classification: Based on the Fourier series threshold, the aggregate morphology is decoupled and calculated to obtain the three-dimensional aggregate morphological parameters that characterize the morphological characteristics of aggregate particles at different scales. data.

2. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: The surface of the aggregate sample selected in step S1 is cleaned and dried. A binocular structured light depth scanning device is used to scan the aggregate sample. The RGB-D depth point cloud data of the aggregate particle surface is scanned at different viewing angles using the target feature method. The depth point cloud is processed in sequence by denoising, cropping, splicing, and fine-tuning interpolation and smoothing using CloudCompare software to obtain a PCD file.

3. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: In step S2, the collected aggregate surface point cloud is implicitly reconstructed to generate a triangular mesh of the point cloud surface, the centroid data is calculated, the circumscribed sphere is established through the centroid-contour distance, the spherical surface is Fibonacci meshed to establish the Monte Carlo direction vector and the spatial random section surface, and the aggregate surface point cloud is randomly cut and reduced in dimension to obtain a two-dimensional point cloud section contour set.

4. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: In step S3, the original contour length establishes an isoperimetric circle, fixed polar angle sampling is performed on the isoperimetric circle, the polar coordinates of the sampling points are reversely mapped to Cartesian coordinates on the original contour, and the nearest linear interpolation calculation is performed on the points that do not exist after mapping to obtain the sampled mapped contour coordinates of the original contour.

5. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: In step S4, the contour coordinates are parameterized based on the unique sampling mapping relationship to construct the aggregate contour descriptor Ω, and the linear equations are constructed with the help of the point coordinate set to solve the descriptor Ω by the least square method.

6. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: In step S5, the gradient is calculated based on the numerical form of the aggregate profile descriptor Ω. and gradient direction angle , construct the gradient direction angle Integrating differential elements to obtain two-dimensional aggregate morphological parameters , further discrete numerical integration is performed to obtain the easily computable Parameter value.

7. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: Three-dimensional aggregate morphological parameters in step S6 It is based on the Monte Carlo sectioning and isoperimetric mapping coordinates in steps 2, 3, and 4 to calculate the two-dimensional aggregate morphological parameters. Obtained by constructing a double integral and performing discretized numerical integration.

8. The aggregate particle morphology quantification method based on Fourier spectrum decoupling according to claim 1 is characterized in that: In step S7, the aggregate morphology is divided into three scales: shape, angularity, and texture. By limiting the Fourier series threshold, decoupling is performed to further obtain the three-dimensional aggregate morphological parameters calculated for different aggregate morphological characteristics. .

9. A system for quantifying aggregate particle morphology based on Fourier spectrum decoupling, characterized in that: The aggregate particle morphology quantification method based on Fourier spectrum decoupling as claimed in any one of claims 1 to 8 comprises: 3D point cloud acquisition module, used to obtain 3D point cloud data of aggregate particles; A point cloud data processing module, used for processing the three-dimensional point cloud data to obtain two-dimensional contour data; A morphological feature analysis module, configured to process the two-dimensional contour data to obtain aggregate decoupling feature data; The detection and analysis module is used to associate various modules to realize the functions of aggregate point cloud collection, point cloud processing, and morphological analysis, and summarize the analysis data results. The modules include: three-dimensional point cloud collection module, point cloud data processing module, and morphological feature analysis module.

Images