A method for evaluating the degree of abrasion of gravel particles
By constructing cross-sectional images of gravel particles and simulating abrasion, and using machine learning to establish an evaluation model, the problem of inaccurate assessment of gravel particle abrasion degree in traditional methods is solved, achieving abrasion degree assessment with higher accuracy and reliability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- OCEAN UNIV OF CHINA
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-10
AI Technical Summary
Existing technologies, when assessing the degree of gravel abrasion, have limitations. Traditional methods such as percentage roundness and deflattened roundness methods have biases in the calculation results for irregularly shaped particles, and there is a non-linear mapping relationship between the degree of abrasion and roundness, leading to inaccurate assessments.
By constructing cross-sectional images of gravel particles of various shapes and sizes, abrasion simulation is performed. Machine learning methods are used to establish the correspondence between gravel particle profile parameters and abrasion counts, and an evaluation model is trained to assess the degree of abrasion.
It improves the accuracy and reliability of gravel particle abrasion assessment, overcomes the influence of morphology and size on the assessment, and the number of abrasions is closer to the degree of abrasion, thus solving the nonlinear bias of traditional methods.
Smart Images
Figure CN122367907A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of testing and analysis technology, for example to a method for assessing the degree of gravel abrasion. Background Technology
[0002] In geology, gravel refers to rock fragments with a diameter greater than 2 mm. After detaching from the parent rock mass, gravel particles initially often have sharp edges. Subsequently, under external dynamic forces, such as hydrodynamics, they undergo migration. During this migration, the gravel particles are abraded, and their sharp edges are blunted. Under sufficient abrasion, the shape of the gravel particles often becomes round or elliptical. Larger, more abraded gravel particles are commonly known as pebbles. This process is called gravel abrasion.
[0003] The degree of abrasion of gravel particles reflects the migration process they undergo after detaching from the parent rock. Understanding this process is crucial for understanding the source of the gravel, its transport process, and the transport distance, which in turn helps in understanding the source, formation history, and formation process of the strata to which it belongs. Furthermore, understanding the source, formation history, and formation process of strata is of great value for mineral exploration.
[0004] The conventional process for analyzing the degree of gravel abrasion is as follows: (1) Cut and grind the rock or core to obtain a smooth surface; (2) Take a picture of the smooth surface using a camera or microscope to obtain a two-dimensional image; (3) Obtain the outer contour of the gravel particles using an automatic analysis program or a manual drawing method; (4) Calculate the roundness based on the outer contour of the gravel particles, and then evaluate the degree of gravel abrasion. In step (4), the roundness is usually calculated using two methods: percentage roundness or deflattened roundness. The percentage roundness method mainly uses the perimeter and area of the particle contour to obtain the result.
[0005] The percentage roundness formula is relatively simple and is mainly applicable to particles with a near-circular shape. For particles of other shapes, the calculation results have a large deviation.
[0006] The deflattening roundness method takes into account the influence of the flattening degree of the particles on the roundness calculation. Its calculation steps are as follows: (1) Calculate the minimum bounding rectangle of the gravel particle profile; (2) Transform the minimum bounding rectangle into a square, simultaneously transforming the particle profile; (3) Calculate the percentage roundness of the deformed particle profile, using this value to indicate the roundness of the particles. Although the above methods have initially achieved the assessment of the degree of gravel particle abrasion, these methods still have the following shortcomings: (1) The percentage roundness method is mainly applicable to particles with a near-circular shape, while the deflattening roundness method is mainly applicable to elongated particles. However, for particles with other shapes (such as triangles, trapezoids, and polygons), the calculation results show obvious deviations. (2) Generally speaking, for gravel particles with the same shape but different sizes, when achieving the same roundness, larger particles undergo a longer number of abrasions or a longer abrasion distance, or in other words, larger particles are more difficult to round. Therefore, assessing the roundness of gravel particles cannot accurately determine their degree of abrasion. (3) There is a non-linear mapping relationship between roundness and wear degree. Directly using roundness to refer to wear degree has a non-linear deviation.
[0007] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention
[0008] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments, but rather as a prelude to the detailed description that follows.
[0009] This disclosure provides a method for evaluating the degree of gravel abrasion, so as to more accurately reflect the degree of gravel abrasion.
[0010] In some embodiments, the method for assessing the degree of gravel abrasion includes: constructing cross-sectional images of gravel particles of various shapes and sizes; performing abrasion simulation on the cross-sectional images to obtain abrasion cross-sectional images at different abrasion stages; analyzing the contours of the gravel particles in the abrasion cross-sectional images and calculating the contour parameters of the gravel particles; after the abrasion simulation is completed, using machine learning methods to establish the correspondence between the contour parameters of the gravel particles and the number of abrasions to obtain a trained assessment model; and using the assessment model to assess the degree of abrasion of the gravel particles to be analyzed.
[0011] In some embodiments, the gravel abrasion assessment device includes: a cross-sectional image construction module configured to construct cross-sectional images of gravel particles of various shapes and sizes; a gravel abrasion simulation module configured to perform abrasion simulation on the cross-sectional images to obtain abrasion cross-sectional images at different abrasion stages; a gravel parameter calculation module configured to analyze the contours of gravel particles in the abrasion cross-sectional images and calculate the contour parameters of the gravel particles; an assessment model training module configured to establish a correspondence between the contour parameters of the gravel particles and the number of abrasions using machine learning methods after the abrasion simulation is completed, thereby obtaining a trained assessment model; and a gravel abrasion assessment module configured to assess the abrasion degree of the gravel particles to be analyzed using the assessment model.
[0012] The gravel particle abrasion assessment method provided in this disclosure can achieve the following technical effects: First, cross-sectional images of gravel particles of various shapes and sizes were designed. Abrasion simulation was then performed on these images to model the abrasion process at the edges of the gravel particles, resulting in abrasion cross-sectional images of the gravel particles at different abrasion stages. Next, the contours of the gravel particles in the abrasion cross-sectional images were analyzed, and their contour parameters were calculated. After the abrasion simulation, a machine learning model was trained based on the contour parameters and the number of abrasion cycles to obtain an evaluation model. This evaluation model characterizes the correspondence between the contour parameters of the gravel particles and the number of abrasion cycles. Using this evaluation model to analyze and evaluate the gravel particles effectively overcomes the influence of gravel particle shape and size on the degree of abrasion, improving the accuracy and reliability of the analysis. Furthermore, compared to the traditional method of using roundness parameters to evaluate the degree of abrasion, the number of abrasion cycles calculated by this method more closely reflects the degree of abrasion. Using the number of abrasion cycles to evaluate the degree of abrasion also overcomes the deficiency of roundness parameters, which change significantly in the early stages of abrasion and less significantly in the later stages, indicating a non-linear relationship between roundness and the degree of abrasion.
[0013] The above general description and the description below are exemplary and illustrative only and are not intended to limit this application. Attached Figure Description
[0014] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations and drawings do not constitute a limitation on the embodiments. Elements having the same reference numerals in the drawings are shown as similar elements. The drawings are not to be scaled. And wherein: Figure 1 This is a schematic diagram of a method for evaluating the degree of gravel particle abrasion provided in an embodiment of this disclosure; Figure 2 These are cross-sectional images of gravel particles of various shapes and sizes provided in the embodiments of this disclosure; Figure 3 This is a schematic diagram of a method for simulating abrasion on a cross-sectional image provided in an embodiment of this disclosure; Figure 4 This is a schematic diagram of calculating the prominence of edge pixels according to an embodiment of the present disclosure. Green indicates the cross-sectional pixels of gravel particles, red indicates edge pixel samples A and B to be calculated, and the circular dashed lines around pixels A and B indicate the calculation range. Figure 5 This is a schematic diagram illustrating the process by which square particles gradually become rounded after repeated abrasion, as provided in an embodiment of this disclosure. Figure 6 This is a schematic diagram of the abrasion process of triangular, arbitrary quadrilateral, trapezoidal and pentagonal gravel particles provided in the embodiments of this disclosure; Figure 7This is a schematic diagram of the calculation results of the method for calculating the distance from the centroid to the edge of the particle every two degrees provided in the embodiments of this disclosure for the cross section of a pentagonal particle at the beginning of abrasion. Figure 8 This is a schematic diagram showing the change in the percentage roundness of gravel particles of different shapes with the number of abrasion cycles, as provided in the embodiments of this disclosure. Figure 9 This is a schematic diagram showing the change in the percentage roundness of gravel particles of different sizes with the number of abrasion cycles provided in the embodiments of this disclosure. The top value in the figure is the side length of the square. Figure 10 This is a schematic diagram showing the fitting relationship between the number of abrasions calculated by the method provided in this embodiment and the actual number of abrasions. Detailed Implementation
[0015] To provide a more detailed understanding of the features and technical content of the embodiments of this disclosure, the implementation of the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. The accompanying drawings are for illustrative purposes only and are not intended to limit the embodiments of this disclosure. In the following technical description, for ease of explanation, several details are used to provide a full understanding of the disclosed embodiments. However, one or more embodiments may still be implemented without these details. In other cases, well-known structures and devices may be simplified in their depiction to simplify the drawings.
[0016] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.
[0017] Unless otherwise stated, the term "multiple" means two or more.
[0018] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.
[0019] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.
[0020] The term "correspondence" can refer to an association or binding relationship. The correspondence between A and B means that there is an association or binding relationship between A and B.
[0021] Combination Figure 1 As shown, this disclosure provides a method for evaluating the degree of gravel particle abrasion, including: S10, construct cross-sectional images of gravel particles of various shapes and sizes.
[0022] S20, perform abrasion simulation on the cross-sectional image to obtain abrasion cross-sectional images at different abrasion stages.
[0023] S30, analyze the contours of gravel particles in the abrasion cross-section image and calculate the contour parameters of the gravel particles.
[0024] S40. After the abrasion simulation is completed, machine learning methods are used to establish the correspondence between the contour parameters of gravel particles and the number of abrasions, and a trained evaluation model is obtained.
[0025] S50 uses an evaluation model to assess the degree of abrasion of the gravel particles to be analyzed.
[0026] First, cross-sectional images of gravel particles of various shapes and sizes are designed and constructed. The area of the gravel particles can be 0.01 to 1 times the maximum area of the cross-sectional image, and the maximum diameter of the gravel particles can be 0.01 to 1 times the length of the cross-sectional image.
[0027] Abrasion simulations are performed on various cross-sectional images. A complete abrasion simulation consists of multiple abrasion processes, each of which yields an abrasion cross-sectional image of the gravel particles, thus providing abrasion cross-sectional images at different abrasion stages.
[0028] Subsequently, the contours of gravel particles in the abrasion cross-sectional images were analyzed, and a series of parameters characterizing the shape of the gravel particles were calculated. After the abrasion simulation of all cross-sectional images was completed, the abrasion count for each cross-sectional image could be obtained. Using machine learning methods, such as random forest, the correspondence between the contour parameters of the gravel particles and the abrasion count was established, thereby obtaining a trained evaluation model. The evaluation model was then used to assess the degree of abrasion of the gravel particles under analysis.
[0029] In summary, the evaluation method provided in this embodiment designs cross-sectional images of gravel particles of various shapes and sizes. By simulating the abrasion process of the gravel particle edges using these cross-sectional images, abrasion cross-sectional images of the gravel particles at different abrasion stages are obtained. The contours of the gravel particles in the abrasion cross-sectional images are then analyzed, and their contour parameters are calculated. After the abrasion simulation, a machine learning model is trained based on the contour parameters and the number of abrasion cycles to obtain the evaluation model. The evaluation model can characterize the correspondence between the contour parameters of the gravel particles and the number of abrasion cycles. Using the evaluation model to analyze and evaluate the gravel particles effectively overcomes the influence of gravel particle shape and size on the degree of abrasion, improving the accuracy and reliability of the analysis. Furthermore, compared with the traditional roundness parameter used to evaluate the degree of abrasion, the number of abrasion cycles calculated by this method is closer to the degree of abrasion. Using the number of abrasion cycles to evaluate the degree of abrasion also overcomes the deficiency that the roundness parameter changes significantly in the early stages of abrasion and changes less in the later stages, i.e., there is a non-linear correspondence between roundness and the degree of abrasion.
[0030] Optionally, such as Figure 2 As shown, the cross-sectional shapes of the gravel particles include regular polygons, non-regular polygons, convex polygons, and concave polygons. These include, but are not limited to, triangles, rectangles, squares, trapezoids, arbitrary quadrilaterals, and pentagons. The side lengths of the gravel particles include, but are not limited to, 250, 300, 350, 400, 450, and 500 pixels. The dimensions of the cross-sectional image, i.e., length and width, are calculated in pixels; this embodiment uses 500 pixels, but other sizes are acceptable. This ensures the diversity of the abrasion samples, better reflects the actual shape of the gravel particles, and helps improve the accuracy of the evaluation method.
[0031] Optionally, combined Figure 3 As shown, S20, the abrasion simulation of the cross-sectional image includes: S21, extract the edge pixels of the cross-sectional image and calculate the prominence of the edge pixels.
[0032] S22, based on the degree of prominence of the edge pixels, simulates abrasion on the cross-sectional image.
[0033] S23, this process continues until the gravel particles are completely eroded or the number of erosions reaches the threshold.
[0034] After designing the cross-sectional images of gravel particles, a program is developed to simulate the abrasion of various gravel cross-sectional image samples. First, the cross-sectional images of the gravel particles are read, and edge pixels are extracted. The prominence of the edge pixels is calculated. First, the number of cross-sectional pixels and blank pixels within a preset range around a contour pixel is calculated. Then, based on the ratio of blank pixels to cross-sectional pixels, the prominence of that edge pixel is determined. For example, in... Figure 4 The edge pixel A, located at the top left corner of the square, is calculated within a dashed circle surrounding A; within this circle, approximately 3 / 4 of the pixels are located in the blank area (white), and approximately 1 / 4 of the pixels are located on the gravel section (green); its prominence is approximately 3. Figure 3 The edge pixel B, located near the center of the top edge of the square, is calculated within a dashed circle surrounding B. Within this circle, approximately half of the pixels are located in the blank area (white), and approximately half are located on the gravel cross-section (green); its prominence is approximately 1. Thus, the prominence of a pixel's location can be reflected by the ratio of blank pixels to cross-section pixels.
[0035] When abrading a cross-sectional image, the first step is to determine whether abrasion has occurred based on the degree of prominence of the edge pixel's location. If the degree of prominence of the edge pixel's location is greater than or equal to the abrasion threshold, the edge pixel is abraded and is removed during this abrasion process.
[0036] If the prominence of the edge pixel is insufficient, i.e., less than the abrasion threshold, the cross-sectional image is further abraded based on the prominence and robustness of the edge pixel. Specifically, this includes determining the reduction in pixel robustness based on the prominence of the edge pixel. The greater the prominence, the greater the reduction in pixel robustness. In each abrasion process, the reduction in pixel robustness is subtracted from the current robustness of the edge pixel in the cross-sectional image. When the real-time robustness of the edge pixel is less than the robustness threshold, the edge pixel is removed from the cross-sectional image, thus forming a new edge. The robustness of the edge pixels is uniformly assigned a fixed initial value at the start of the simulation. Optionally, at the beginning of the abrasion simulation, the initial robustness of the edge pixels is set to 100, and the robustness threshold when the edge pixel undergoes abrasion is set to less than or equal to 0. This completes one abrasion cycle.
[0037] After one round of abrasion, the degree of pixel prominence of the newly formed edge after abrasion is used to determine whether the gravel particles will be worn away in the next abrasion process. If the determination result is yes, meaning the cross-sectional image can still be abraded, a new round of abrasion is performed according to S20. Abrasion of the current cross-sectional image ends when the gravel particles are completely abraded, or when the number of abrasions reaches a threshold. Optionally, the threshold value is 5000 times. For an image with a length and width of 500 pixels set in this embodiment, the simulated abrasion using this program resulted in a very smooth surface after 5000 abrasions.
[0038] Optionally, the erosion of the cross-sectional image is performed by a first main program, a spur subroutine, and an erosion subroutine. The first main program is used to call the spur subroutine and the erosion subroutine. The spur subroutine is used to calculate the spur degree of each edge pixel. The erosion subroutine is used to determine whether each edge pixel will be worn away in the next erosion process based on its spur degree.
[0039] The abrasion strategy primarily sets an abrasion threshold based on the prominence of edge pixels. Abrasion occurs when the prominence of an edge pixel is greater than or equal to the abrasion threshold. When the prominence of an edge pixel is less than the abrasion threshold, its robustness is reduced, with the reduction in robustness being positively correlated with its prominence. Optionally, this abrasion threshold is set to 2. To better reflect actual abrasion conditions, the calculation of the reduction in robustness does not entirely depend on the prominence but incorporates a certain degree of randomness. Abrasion occurs when the robustness drops to 0.
[0040] for Figure 4 If pixel A, located at the top left corner of the square, has a protrusion level of 3, then it will definitely experience wear, meaning that this pixel will change from a cross-sectional pixel to a blank pixel in the next round of wear. Figure 5 If pixel B, located at the top left corner of the square, has a protrusion level of 1, then the amount of reduction in its robustness is determined based on its protrusion level, and the current robustness level is used to determine whether wear occurs.
[0041] See Figure 5 This image illustrates the process of a square particle gradually becoming rounder after multiple abrasions. Pixels protruding outwards are more prone to abrasion, such as those near the four corners of the square, while pixels along the four straight sides are less susceptible to wear. The closer to the corners, the more easily abrasion occurs. A is the original image of the square particle, with green indicating cross-sectional pixels and gray indicating blank pixels; B is the image of the square particle after initial abrasion, primarily occurring at the four corners; C is the image of the square particle after a certain number of abrasions, with greater abrasion at the corners and less abrasion on the sides; D is the image of the square particle becoming rounder after sufficient abrasion.
[0042] See Figure 6 It demonstrates the abrasion process of triangular, arbitrary quadrilateral, trapezoidal, and pentagonal gravel particles. Taking triangular particles as an example (…). Figure 6 The following is an explanation of section A): Initially, it is a complete triangular shape (the edge shape is shown by light blue particles). During the etching process, the outer edge pixels are gradually etched (the degree of blueness of the pixels corresponds to the number of etching times, with lighter blue indicating fewer etching times and darker blue indicating more etching times). The etching stops after 5000 times, and the final outer contour is shown in green.
[0043] Optionally, S30, analyze the profile of the gravel particles and calculate the profile parameters of the gravel particles, including: Find the centroid of the cross-sectional image.
[0044] Then, starting from the centroid, the distance from the centroid to the edge of the particle outline is calculated at preset angles to obtain the radius of the gravel particle in that direction.
[0045] Calculate the minimum radius, maximum radius, particle area, particle flatness, particle perimeter, radius ratio, and percentage roundness of the gravel particles based on their radius in that direction.
[0046] Analyze the images of gravel particles obtained in S20 after different abrasion cycles. First, locate the centroid of the cross-sectional image. Then, starting from the centroid, calculate the distance from the centroid to the edge of the particle outline (the radius of the particle in that direction) at preset angles. Optionally, the preset angle is two degrees. Then, based on the radius of the gravel particles in that direction, calculate the minimum radius (MinR), maximum radius (MaxR), particle area (A), particle flatness (Ar, the ratio of particle area to the area corresponding to the maximum radius), particle perimeter (P), radius ratio (RRatio, the ratio of the maximum radius to the minimum radius), and percentage roundness (Yb), etc.
[0047] Optionally, the calculation of gravel particle parameters is performed by subroutines such as the second main program, distance and angle subroutines, parameter subroutines, edge point subroutines, and centroid subroutines. The second main program is responsible for calling these subroutines. The distance and angle subroutines are used to calculate the distance from the centroid to the edge of the particle outline; the parameter subroutines are used to calculate parameters such as minimum radius, maximum radius, particle area, particle flatness, particle perimeter, radius ratio, and percentage roundness; the other subroutines serve the above subroutines.
[0048] Figure 7 The distance and angle subroutine provided in this embodiment calculates the distance from the centroid to the edge of the cross-section obtained after the pentagonal particle has been worn 1 to 100 times every two degrees. The minimum value among these distances is the minimum radius, and the maximum value is the maximum radius. The particle area is obtained by accumulating the sector areas corresponding to these distances. The ratio of the particle area to the area corresponding to the maximum radius is the particle flatness. Other parameters are also calculated based on these data, and will not be described in detail here.
[0049] See Figure 8This study demonstrates the variation of the percentage roundness of gravel particles with the number of abrasion cycles after processing the abrasion results of different morphologies using the aforementioned procedure. The results show that in the most primitive state (when the particles have just detached from the parent material and have sharp edges), pentagonal particles have a higher percentage roundness (0.87). Triangular particles have a lower percentage roundness (0.64), while rectangular and square particles have intermediate percentage roundness. These results also indicate that even in the most primitive state, before abrasion occurs, the percentage roundness of the particles is relatively high, with pentagonal particles reaching 0.87, close to 1. Therefore, using percentage roundness to assess the degree of gravel particle abrasion has significant uncertainty. As the number of abrasion cycles increases, the percentage roundness of the gravel particles gradually increases, with triangular particles showing the largest increase rate, while other particles show relatively smaller increases. This result not only demonstrates the different variation processes between the percentage roundness of gravel particles of different morphologies and the number of abrasion cycles, but also specifically illustrates that this process has non-linear characteristics.
[0050] See Figure 9 This study demonstrates the change in the percentage roundness of gravel particles of different sizes with the number of abrasion cycles. The results show that in the most primitive state (when the particles have just detached from the parent material and have sharp edges), the percentage roundness of the six different square sizes is essentially the same, at 0.82. With increasing abrasion cycles, the percentage roundness of the gravel particles gradually increases, with the smallest particles showing the largest increase rate, while the increase rate for larger particles gradually decreases. This result demonstrates a different nonlinear relationship between the percentage roundness of gravel particles of different sizes and the number of abrasion cycles; smaller particles abrade faster, while larger particles abrade slower.
[0051] In summary, the results indicate that the change in the percentage roundness of gravel particles is not only related to their morphology and size, but also has a non-linear relationship with the number of abrasion cycles. At the beginning of abrasion, the percentage roundness increases rapidly, but the rate of increase slows down with each abrasion cycle.
[0052] In addition to the comparative analysis described above, this method also compared the minimum radius (MinR), maximum radius (MaxR), particle area (A), particle flatness (Ar), particle perimeter (P), and radius ratio (RRatio, the ratio of the maximum radius to the minimum radius) with the actual number of abrasions. The results show that these parameters all exhibit complex nonlinear relationships with the actual number of abrasions. Further, a multiple regression analysis was employed to establish a fitting formula between all the parameters obtained above and the actual number of abrasions. Although the number of abrasions calculated using this fitting formula showed good agreement with the actual number of abrasions, it was still inferior to the results calculated by the model established using machine learning methods, as described below.
[0053] Based on the above analysis and comparison, this embodiment uses machine learning methods to establish the relationship between the aforementioned parameters and the actual number of abrasion cycles. These parameters include minimum radius (MinR), maximum radius (MaxR), particle area (A), particle flatness (Ar, the ratio of particle area to the area corresponding to the maximum radius), particle perimeter (P), radius ratio (RRatio, the ratio of the maximum radius to the minimum radius), and percentage roundness (Yb). In S40, using these parameters, a machine learning model is trained using the random forest method based on the Python language machine learning library (sklearn). Figure 10 The correlation between the number of erosions and the actual number of erosions is calculated using a trained machine learning model based on the aforementioned parameters. The two show a good linear fit, with a coefficient of determination (R²) of 1 / 2. 2 The value is as high as 0.99.
[0054] Then, in step S50, the degree of abrasion of the gravel particles being analyzed is assessed using an evaluation model, including: The outer contour parameters of the gravel particles to be analyzed are calculated according to S30. The calculated parameters are the same as those mentioned above, including minimum radius (MinR), maximum radius (MaxR), particle area (A), particle flatness (Ar, the ratio of particle area to the area corresponding to the maximum radius), particle perimeter (P), radius ratio (RRatio, the ratio of the maximum radius to the minimum radius), and percentage roundness (Yb). Substituting these parameters into the evaluation model trained by S40, the number of abrasion cycles is calculated.
[0055] Further standardization of the abrasion degree is performed by dividing the calculated number of abrasions by the aforementioned abrasion number threshold (e.g., 5000 times). The final abrasion degree is a ratio between 0 and 1, where 0 represents the worst abrasion degree and 1 represents the best abrasion degree.
[0056] This disclosure also provides a gravel particle abrasion assessment device, comprising: a cross-sectional image construction module, a gravel abrasion simulation module, a gravel parameter calculation module, an assessment model training module, and a gravel abrasion assessment module. The cross-sectional image construction module is configured to construct cross-sectional images of gravel particles of various shapes and sizes. The gravel abrasion simulation module is configured to simulate abrasion on the cross-sectional images to obtain abraded cross-sectional images at different abrasion stages. The gravel parameter calculation module is configured to analyze the contours of gravel particles in the abraded cross-sectional images and calculate the contour parameters of the gravel particles. The assessment model training module is configured to, after the abrasion simulation is completed, use machine learning methods to establish the correspondence between the contour parameters of the gravel particles and the number of abrasion cycles, obtaining a trained assessment model. The gravel abrasion assessment module is configured to use the assessment model to assess the abrasion degree of the gravel particles to be analyzed.
[0057] The specific implementation process of the device can be found in the description of the above method embodiments, and will not be repeated here.
[0058] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0059] While the specific embodiments of the present invention have been described above, they are not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.
Claims
1. A method for assessing the degree of gravel particle abrasion, characterized in that, include: Construct cross-sectional images of gravel particles of various shapes and sizes; Abrasion simulation was performed on the cross-sectional image to obtain abrasion cross-sectional images at different abrasion stages; Analyze the contours of the gravel particles in the abrasion cross-section image and calculate the contour parameters of the gravel particles; After the abrasion simulation is completed, machine learning methods are used to establish the correspondence between the contour parameters of the gravel particles and the number of abrasions, and a trained evaluation model is obtained. The degree of abrasion of the gravel particles to be analyzed is evaluated using the aforementioned evaluation model.
2. The method for evaluating the degree of gravel particle abrasion according to claim 1, characterized in that, The cross-sectional shape of the gravel particles includes: regular polygons, non-regular polygons, convex polygons, and concave polygons; The dimensions of the cross-sectional image are measured in pixels.
3. The method for assessing the degree of gravel particle abrasion according to claim 1, characterized in that, The process of simulating the abrasion of the cross-sectional image includes: Extract the edge pixels of the cross-sectional image and calculate the prominence of the edge pixels at their locations; Based on the degree of protrusion of the edge pixels, the cross-sectional image is subjected to abrasion simulation; This process continues until the gravel particles are completely worn away or the number of wears reaches a threshold.
4. The method for evaluating the degree of gravel particle abrasion according to claim 3, characterized in that, The calculation of the prominence of the edge pixel's location includes: Calculate the number of cross-sectional pixels and blank pixels within a preset range around a contour pixel; The prominence of the edge pixel's location is determined based on the ratio of the number of blank pixels to the number of cross-sectional pixels.
5. The method for evaluating the degree of gravel particle abrasion according to claim 3, characterized in that, The step of simulating abrasion on the cross-sectional image based on the protrusion of the edge pixels includes: When the protrusion of the edge pixel is greater than or equal to the abrasion threshold, the edge pixel is removed during this abrasion process, i.e., abrasion is performed. When the protrusion of the edge pixel is less than the abrasion threshold, the cross-sectional image is subjected to abrasion simulation based on the protrusion of the edge pixel's location and the edge pixel's robustness; the edge pixel's robustness is uniformly assigned a fixed initial value at the beginning of the simulation.
6. The method for evaluating the degree of gravel particle abrasion according to claim 5, characterized in that, The step of simulating abrasion on the cross-sectional image based on the protrusion and rigidity of the edge pixels includes: The reduction in the rigidity of the edge pixel is determined based on the degree of protrusion at the location of the edge pixel; During each abrasion process, the decrease in strength is subtracted from the current strength of the edge pixel; When the real-time robustness of the edge pixel is less than the robustness threshold, the edge pixel is removed during this abrasion process.
7. The method for evaluating the degree of gravel particle abrasion according to claim 6, characterized in that, The step of determining the reduction in the rigidity of the edge pixel based on the degree of protrusion at the location of the edge pixel includes: The greater the protrusion of the edge pixel, the greater the reduction in rigidity.
8. The method for evaluating the degree of gravel particle abrasion according to claim 1, characterized in that, The analysis of the gravel particle profile and the calculation of the gravel particle profile parameters include: Locate the centroid of the cross-sectional image; Then, starting from the centroid, the distance from the centroid to the edge of the particle outline is calculated at preset angles to obtain the radius of the gravel particle in that direction; Based on the radius of the gravel particles in this direction, calculate the minimum radius, maximum radius, particle area, particle flatness, particle perimeter, radius ratio, and percentage roundness of the gravel particles.
9. A device for assessing the degree of gravel particle abrasion, characterized in that, include: The cross-sectional image construction module is configured to construct cross-sectional images of gravel particles of various shapes and sizes; The gravel abrasion simulation module is configured to perform abrasion simulation on the cross-sectional image to obtain abrasion cross-sectional images at different abrasion stages. The gravel parameter calculation module is configured to analyze the contours of gravel particles in the abrasion cross-section image and calculate the contour parameters of the gravel particles. The evaluation model training module is configured to establish the correspondence between the contour parameters of the gravel particles and the number of abrasions after the abrasion simulation is completed, and obtain the trained evaluation model by using machine learning methods. The gravel abrasion assessment module is configured to assess the degree of abrasion of the gravel particles to be analyzed using the assessment model.