Quantitative evaluation method for adhesion between aggregate and high viscosity pitch based on rotating bottle method

Abstract

The application discloses a kind of aggregate and high viscosity pitch adhesion quantitative evaluation method based on rotary bottle method, it is related to pitch-aggregate adhesion evaluation technical field.The interface contact condition of water-pitch-aggregate under actual driving condition is simulated by using rotary bottle, the physical peeling of high viscosity pitch from aggregate surface can be effectively accelerated, the resolution ability of adhesion performance difference of high viscosity system is significantly improved;By using non-negative matrix factorization algorithm for high light repair, the highlight artifact of asphalt coating area on aggregate surface caused by mirror reflection can be accurately separated and inhibited, to avoid the high light area in image binarization process misidentification as pitch peeling area, so as to guarantee the accuracy of pitch peeling area calculation.The method has high repeatability, can be quantitatively evaluated, in high-grade highway, bridge pavement and heavy traffic section, accurately select high viscosity pitch suitable high-quality aggregate, and support asphalt mixture water stability design and life prediction such as engineering practice has important application value.

Description

Technical Field

[0001] This invention relates to the field of asphalt-aggregate adhesion evaluation technology, and more specifically, to a quantitative evaluation method for the adhesion of aggregates to high-viscosity asphalt based on the rotating bottle method. Background Technology

[0002] The adhesion between asphalt and aggregate is one of the key factors in ensuring the water stability of asphalt mixtures. The current standard, "Test Procedures for Asphalt Mixtures in Highway Engineering" (JTG 3410—2025), proposes using the boiling water method to qualitatively evaluate the water stripping resistance of coarse aggregates. The boiling water method involves placing aggregates coated with asphalt into gently boiling water for 3 minutes, and then having two or more experienced testers assess the coating rate of the same aggregate (as shown in Table 1) to determine the adhesion grade of the aggregate.

[0003] Table 1 Adhesion grades between asphalt and aggregate

[0004] However, in practical applications, the boiling water method test and evaluation have the following problems: First, relying on experienced personnel to judge the results of the boiling water method test is highly subjective, and the test results can only be evaluated qualitatively; in addition, since the boiling water method test evaluates the adhesion of aggregates by boiling water, boiling water cannot accelerate the peeling of asphalt from the aggregate surface. Aggregates with good adhesion or the adhesion between aggregates and high-viscosity asphalt are all at level 5, but the adhesion between aggregates with poor adhesion and high-viscosity asphalt is also at level 5, which makes it impossible to accurately evaluate the adhesion of aggregates, thus affecting the effect of asphalt pavement.

[0005] In view of this, the present invention is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide a quantitative evaluation method for the adhesion of aggregates to high-viscosity asphalt based on the rotating bottle method.

[0007] This invention is implemented as follows: In a first aspect, the present invention provides a method for quantitative evaluation of the adhesion between aggregates and high-viscosity asphalt based on the rotating bottle method, comprising: taking images of the initial material at a fixed height, and then obtaining an image of the initial material after uniform background processing; the initial material includes aggregates with asphalt coated on the surface.

[0008] The initial material is loaded into a rotating flask and rotated on a rotating flask tester for more than 24 hours. The material inside the rotating flask is then removed to obtain the experimental material.

[0009] Images of the experimental material are captured at a fixed height to obtain the initial experimental image. After the initial experimental image is processed to unify the background, a non-negative matrix factorization algorithm is used to repair the highlights, separate and remove the specular reflection component in the initial experimental image, and retain the diffuse reflection component in the initial experimental image to obtain the de-glare experimental image.

[0010] The initial material image and the de-lighting experimental image were binarized, and the number of black pixels in the two images was counted to calculate the asphalt stripping rate of the experimental material relative to the initial material.

[0011] Secondly, the present invention provides an application of the method as described in any of the foregoing embodiments in the field of roads.

[0012] The present invention has the following beneficial effects: This invention provides a quantitative evaluation method for the adhesion between aggregates and high-viscosity asphalt based on the rotating bottle method. By using a rotating bottle to simulate the long-term peeling process of the water-asphalt-aggregate interface under dynamic shear and repeated wetting coupling under actual traffic loads, it can effectively accelerate the physical peeling of high-viscosity asphalt from the aggregate surface and significantly improve the ability to distinguish differences in the adhesion performance of high-viscosity systems. By using a non-negative matrix factorization algorithm for highlight restoration, it can accurately separate and suppress the high-brightness artifacts caused by specular reflection in the asphalt-coated area on the aggregate surface, avoiding the misidentification of highlight areas as asphalt peeling areas during image binarization, thus ensuring the accuracy of asphalt peeling area calculation. By combining fixed-height shooting with uniform background processing, it achieves an objective, repeatable, and quantifiable evaluation of the adhesion between aggregates and high-viscosity asphalt. It has important application value in engineering practices such as accurately selecting high-quality aggregates suitable for high-viscosity asphalt in high-grade highways, bridge pavement, and heavy traffic sections, and supporting the water stability design and life prediction of asphalt mixtures. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a physical image of the initial material provided in Embodiment 1 of the present invention; Figure 2 This is the initial material image provided in Embodiment 1 of the present invention; Figure 3 This is a diagram of the rotating bottle experiment provided in Embodiment 1 of the present invention; Figure 4 The initial experimental image provided in Embodiment 1 of the present invention; Figure 5 These are experimental images of light removal provided in Embodiment 1 of the present invention; Figure 6 This is the pixel histogram provided in Embodiment 1 of the present invention; Figure 7 This is a comparison image of the initial material image and the light-removing experimental image after binarization processing by the method provided in Embodiment 1 of the present invention; Figure 8 The initial experimental image provided in Embodiment 2 of the present invention; Figure 9 This is the pixel histogram provided in Embodiment 2 of the present invention; Figure 10 The image provided in Embodiment 2 of the present invention is a photo of a light removal experiment. Figure 11 This is a comparison image of the initial material image and the light-removing experimental image after binarization processing by the method provided in Embodiment 2 of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Where specific conditions are not specified in the embodiments, conventional conditions or conditions recommended by the manufacturer shall apply. Reagents or instruments whose manufacturers are not specified are all conventional products that can be purchased commercially.

[0016] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0017] In a first aspect, the present invention provides a method for quantitative evaluation of the adhesion between aggregates and high-viscosity asphalt based on the rotating bottle method, comprising: taking images of the initial material at a fixed height, and then obtaining an image of the initial material after uniform background processing; the initial material includes aggregates with asphalt coated on the surface.

[0018] The initial material is loaded into a rotating flask and rotated on a rotating flask tester for more than 24 hours. The material inside the rotating flask is then removed to obtain the experimental material.

[0019] Images of the experimental material are captured at a fixed height to obtain the initial experimental image. After the initial experimental image is processed to unify the background, a non-negative matrix factorization algorithm is used to repair the highlights, separate and remove the specular reflection component in the initial experimental image, and retain the diffuse reflection component in the initial experimental image to obtain the de-glare experimental image.

[0020] The initial material image and the de-lighting experimental image were binarized, and the number of black pixels in the two images was counted to calculate the asphalt stripping rate of the experimental material relative to the initial material.

[0021] This invention employs a rotating bottle to simulate the long-term peeling process of the water-asphalt-aggregate interface under dynamic shearing and repeated wetting coupling under actual traffic loads. This effectively accelerates the physical peeling of high-viscosity asphalt from the aggregate surface and significantly improves the ability to distinguish differences in the adhesion performance of high-viscosity systems. By using a non-negative matrix factorization algorithm for highlight restoration, it can accurately separate and suppress the high-brightness artifacts caused by specular reflection in the asphalt-coated area on the aggregate surface, avoiding misidentification of highlight areas as asphalt peeling areas during image binarization, thus ensuring the accuracy of asphalt peeling area calculation. Through fixed-height shooting and unified background processing, it achieves an objective, repeatable, and quantifiable evaluation of the adhesion between aggregates and high-viscosity asphalt. This invention has significant application value in engineering practices such as accurately selecting high-quality aggregates suitable for high-viscosity asphalt in high-grade highways, bridge pavements, and heavy-load traffic sections, and supporting the water stability design and life prediction of asphalt mixtures.

[0022] For ease of understanding, the present invention provides a quantitative evaluation method for the adhesion of aggregates to high-viscosity asphalt based on the rotating bottle method, comprising the following steps: S01, Initial Material Pretreatment Images of the initial material are taken at a fixed height and then processed with a uniform background to obtain an image of the initial material; the initial material includes aggregates coated with asphalt.

[0023] In an optional embodiment, the method for preparing the initial material includes mixing aggregate and asphalt for 100-150 seconds.

[0024] Preferably, the mixing can be carried out in an asphalt mixture mixing tank.

[0025] In an optional embodiment, the mass ratio of aggregate to asphalt in the initial material is 510:(16~18).

[0026] Preferably, the aggregate has a particle size of 8~11.2mm.

[0027] In an optional implementation, the captured image has a resolution of ≥10 million pixels to ensure the effectiveness of the asphalt adhesion analysis.

[0028] Preferably, the fixed height is 10~30cm. As long as the initial material and the experimental material are photographed at the same height, the specific height can be selected according to the actual situation.

[0029] S02, Rotary Bottle Experiment The initial material is loaded into a rotating flask and rotated on a rotating flask tester for more than 24 hours. The material inside the rotating flask is then removed to obtain the experimental material.

[0030] In an optional embodiment, the rotation time of the rotating bottle on the rotating bottle tester is 24~48h, preferably 24h, to ensure the accuracy of the experimental results and save experimental time.

[0031] S03, Highlight Removal Processing Images of the experimental material are captured at a fixed height to obtain the initial experimental image. After the initial experimental image is processed to unify the background, a non-negative matrix factorization algorithm is used to repair the highlights, separate and remove the specular reflection component in the initial experimental image, and retain the diffuse reflection component in the initial experimental image to obtain the de-glare experimental image.

[0032] Preferably, the uniform background processing includes adjusting the image background color to a single grayscale value to make the pixel histogram set uniform.

[0033] Because some surface asphalt detached and thinned after the rotating bottle experiment, the asphalt loss in some areas could cause specular reflection during subsequent imaging, affecting the asphalt area recognition. Therefore, the initial experimental images needed to be de-highlighted. However, the initial material images did not require de-highlighting because the surface asphalt layer was thick, the asphalt layer itself was non-reflective, and the initial image information needed to be preserved for subsequent comparison.

[0034] Specifically, the color of an image is composed of two parts: diffuse reflection and specular reflection, as shown in Equation (I).

[0035] (I) In equation (Ⅰ), I is the color intensity matrix of the original image, I diffuse Let I be the diffuse reflection component matrix. specular This is the specular reflection component matrix.

[0036] Diffuse reflection refers to the scattering of light in all directions when it hits a rough surface. Diffuse reflection does not focus light at a single point, so it does not produce noticeable bright spots or areas of high brightness. Specular reflection, on the other hand, refers to the reflection of light rays at the same angle when they hit a smooth surface. This creates bright areas, i.e., highlights. This step is to remove the highlights from the initial experimental image, i.e., I. specular .

[0037] Furthermore, in equation (Ⅰ), I represents the color intensity value at a certain point in the image, which is composed of R, G, and B. Among them, R is the intensity value of the red channel, G is the intensity value of the green channel, and B is the intensity value of the blue channel.

[0038] I diffuse The color intensity of diffuse reflection is given by R. d G d and Bd Composition; I specular The color intensity of specular reflection is given by R. s G s and B s composition.

[0039] Equation (I) can be decomposed into Equation (II) using a nonnegative matrix factorization algorithm: (II) In equation (II), W is a non-negative basic pattern matrix, K is a non-negative coefficient matrix, and the submatrix ks(x) in K corresponding to the specular reflection component is a sparse matrix.

[0040] Specifically, based on the color intensity value, equation (II) can be expressed as equation (III): (III).

[0041] In equation (Ⅲ), x and y are the x and y positions of a pixel in the image. The diffuse reflectance coefficient is... The specular reflection coefficient, , , .

[0042] Since each initial experimental image contains only a portion of the highlight region, ks(x) is a sparse matrix. Furthermore, the matrix parameters in equation (Ⅲ) are all non-negative, so a non-negative matrix factorization algorithm with sparsity constraints can be used to decompose and calculate the diffuse and specular reflection intensities.

[0043] Alternatively, the nonnegative matrix factorization algorithm calculates W through error iteration, and the calculation steps are as follows: S1: Input image color intensity matrix I; set the sparsity target of the sparse matrix: L1 norm sparsity (ksh1) is 0.6~0.8, L2 norm sparsity (ksh2) is 1.0, and the convergence error err is set to 0.001. Use the software to automatically generate random non-negative matrices W and K.

[0044] S2: For matrix W, use equation (Ⅳ) to iteratively update the matrix coefficients; (IV) In formula (Ⅳ), The Kronecker product has the following operation rules: if A is an m×n matrix and B is a p×q matrix, then A B is an mp×nq block matrix, where each element a ij Replaced with block a ij B.

[0045] Dividing matrices element by element follows these rules: .

[0046] V represents the input image matrix; For the first i Transpose of the coefficient matrix of the basic pattern in the next iteration; For the first i The basic pattern matrix of the next iteration For the first i Transpose of the coefficient matrix of the basic pattern in the next iteration; For the first i The basic pattern matrix for +1 iterations.

[0047] S3: For matrix K, respectively for... and The matrix coefficients are iteratively updated; S31: The matrix coefficients are updated using equation (V); (V) S32: For matrix coefficients, first set all negative elements to 0, then update the calculation using L1 norm sparsity and L2 norm sparsity constraints, and finally update the coefficients using formula (V). Furthermore, the coefficient update is based on the L1 norm sparsity constraint index: First, the L1 norm sparsity (ksh1) constraint is calculated on the matrix, the current L1 norm (current1) of the vector is calculated, current1 = sum(abs(x)), and the scaling factor (scale) is calculated, scale = ksh1 / current1. Each element of the original vector is multiplied by this scaling factor, i.e. = scale× Complete the coefficient update of L1 norm sparsity.

[0048] Coefficient updates based on L2 norm sparsity constraints: First, calculate a reference point m that satisfies the L1 norm sparsity constraint and has the most uniform distribution (currently, the zero component is fixed at 0, and the remaining components equally divide the L1 norm), i.e., m i = ksh1 / (n-|Z|) (n is a vector k) s The total number of elements, where |Z| is the number of zero components). Scaling along the direction vector: Starting from the reference point m, scale along the direction vector. d = - m By solving the quadratic equation (Saving the L2 norm sparsity constraint requirements), solve for α, and calculate the vector. Update using error control ,comprehensive and Update matrix K i+1 .

[0049] S4: Calculate the error (err) using formula (VI).

[0050] (VI) If err≤0.001, the calculation can be considered complete; otherwise, repeat steps S2~S4 until the error requirement is met.

[0051] In an optional implementation, after eliminating the specular reflection color intensity using a non-negative matrix factorization algorithm, the aggregate image is obtained by smoothing using neighboring pixels and removing highlights.

[0052] S04, Calculate the asphalt stripping rate The initial material image and the de-lighting experimental image were binarized, and the number of black pixels in the two images was counted to calculate the asphalt stripping rate of the experimental material relative to the initial material.

[0053] In an optional implementation, the method for calculating the asphalt stripping rate includes: calculating the first asphalt coating area (S) in the initial material image based on the image physical size and the total number of pixels. c ) and the second bitumen coating area (S) in the light removal experiment image x The asphalt spalling area (S) is obtained from the difference between the two. b The ratio of the asphalt stripping area to the first asphalt coating area is used as the asphalt stripping rate (r). b (This is used to quantitatively evaluate the adhesion between aggregates and asphalt.)

[0054] Secondly, the present invention provides an application of the method as described in any of the foregoing embodiments in the field of roads.

[0055] The asphalt used in the following examples and comparative examples includes ordinary viscosity asphalt (No. 70 asphalt) and two types of high viscosity asphalt (SBS asphalt and rubber asphalt). The properties of the three types of asphalt are shown in Table 2.

[0056] Table 2 Properties of Asphalt

[0057] The following examples and comparative examples use aggregates from 10 different sources or lithologies, including basalt from 3 sources, tuff from 2 sources, diabase from 2 sources, limestone from 2 sources, and granite from 1 source. The aggregate properties are shown in Table 3.

[0058] Table 3. Performance of Aggregates

[0059] Example 1 This embodiment provides a quantitative evaluation method for the adhesion between aggregates and high-viscosity asphalt based on the rotating bottle method. This embodiment uses basalt No. 1 aggregate and SBS asphalt as an example for illustration. Other combinations of asphalt in Table 2 and aggregates in Table 3 can also be evaluated using this method. The specific steps are as follows: S01, Initial Material Pretreatment 510g of aggregate (particle size 8~11.2mm) and 17g of asphalt were used as raw materials. These were added to an asphalt mixing tank and mixed for 120 seconds to obtain the initial material, such as... Figure 1 As shown.

[0060] Images of the initial material are taken from a fixed height of 20cm, with each image having at least 10 megapixels. After background processing, an image of the initial material is obtained, such as... Figure 2 As shown.

[0061] S02, Rotary Bottle Experiment like Figure 3 As shown, the initial material was loaded into a rotating flask, and distilled water was poured into the flask to the mark. The flask was rotated on a rotating flask tester for 24 hours. The material inside the rotating flask was then removed to obtain the experimental material.

[0062] S03, Highlight Removal Processing Images of the experimental material were taken at a fixed height of 20cm. After background processing, the initial experimental images were obtained, such as... Figure 4 As shown, the initial experimental image was subjected to specular retouching using a non-negative matrix factorization algorithm. The specular reflection component in the initial experimental image was separated and removed, while the diffuse reflection component was retained, resulting in a de-glazed experimental image, as shown. Figure 5 As shown.

[0063] The background unification process uses image processing software to unify the background color of both the initial experimental image and the initial material image. After unifying the background color, the pixel histogram of the image (e.g., ...) is then analyzed. Figure 6 , Figure 6 'a' represents the unified background. Figure 6 (b) with a unified background is more focused, making it easier to calculate the asphalt stripping rate on the aggregate surface through image processing.

[0064] S04, Calculate the asphalt stripping rate The initial material image and the delighting experiment image were binarized separately to obtain, as shown below. Figure 7 ( Figure 7 a is the initial material image. Figure 7Image b shows the results of the light removal experiment. The number of black pixels in both images was counted, and the asphalt stripping rate of the experimental material relative to the initial material was calculated.

[0065] The method for calculating the asphalt stripping rate includes: calculating the first asphalt coating area (S) in the initial material image based on the image's physical size and total number of pixels. c ) and the second bitumen coating area (S) in the light removal experiment image x The asphalt spalling area (S) is obtained from the difference between the two. b The ratio of the asphalt stripping area to the first asphalt coating area is used as the asphalt stripping rate (r). b (This is used to quantitatively evaluate the adhesion between aggregates and asphalt.)

[0066] Example 2 This embodiment provides a quantitative evaluation method for the adhesion between aggregates and high-viscosity asphalt based on the rotating bottle method. The specific steps are the same as in Embodiment 1, except that the asphalt used as the initial material in this embodiment is rubber asphalt and the aggregate is basalt No. 1.

[0067] The image results obtained during the evaluation process in this embodiment are as follows: Figures 8-11 As shown.

[0068] Figure 8 This is the initial experimental image obtained in step S03 of this embodiment, which contains some highlight points to be processed.

[0069] Figure 9 This is a comparison image of the initial experimental image and the initial material image before and after background unification, in which... Figure 9 'a' represents the unified background. Figure 9 b represents a unified background. By unifying the background color, the image pixel histogram becomes more concentrated, facilitating image processing for calculating the asphalt stripping rate on the aggregate surface.

[0070] Figure 10 The light removal experimental image obtained in step S03 of this embodiment is compared to... Figure 8 The highlights have been eliminated, which is beneficial for accurately calculating the asphalt stripping rate.

[0071] Figure 11 The results of binarizing the initial material image and the delighting experiment image provided in this embodiment are as follows: Figure 11 a is the initial material image. Figure 11 b is the image from the light removal experiment. This image allows for a direct comparison of the rotating bottle method provided in this embodiment, which provides a more accurate evaluation of the adhesion effect between asphalt and aggregate.

[0072] Comparative Example 1 This comparative example provides a quantitative evaluation method for the adhesion between aggregates and high-viscosity asphalt, specifically the boiling water method, which involves testing the initial materials, such as those in Example 1, according to the "Test Procedure for Asphalt Mixtures in Highway Engineering" (JTG 3410-2025).

[0073] Experimental Example 1 The initial materials obtained from different asphalt and different aggregate combinations in Tables 2 and 3 were evaluated for viscosity according to the methods provided in Example 1 and Comparative Example 1, respectively, and the results are shown in Table 4.

[0074] In Table 4, the parameter corresponding to Comparative Example 1 is the level number of the adhesion evaluation index shown in Table 1, in units of levels; the parameter corresponding to Example 1 is the asphalt stripping rate, in units of %.

[0075] Table 4. Viscosity evaluation test results for different initial materials

[0076] As shown in Table 4, aggregates of the same adhesion grade obtained by the boiling water method can be more precisely distinguished from asphalt adhesion after the rotating bottle test provided in this embodiment. In particular, for high-viscosity asphalt, after being compounded with basalt, tuff, diabase and limestone, the test results of the boiling water method are all grade 5. However, after evaluation by the method provided in Embodiment 1 of this invention, it was found that the adhesion of basalt and limestone is much higher than that of tuff.

[0077] It should be noted that granite with very poor adhesion is not allowed to be used in asphalt pavement. However, under the test conditions of high viscosity asphalt + boiling water method, a test result of level 4 can be obtained (the best is level 5). The granite obtained by the method provided in this embodiment of the invention has a high asphalt stripping rate under high viscosity asphalt, which can prove that it is not allowed to be used in asphalt pavement. Therefore, the method provided by this invention can quantitatively evaluate the adhesion performance between asphalt and aggregate, and the evaluation results are accurate and have excellent application value.

[0078] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. 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 quantitative evaluation method for the adhesion of aggregates to high-viscosity asphalt based on the rotating bottle method, characterized in that, include: The initial material is photographed at a fixed height, and then a uniform background is processed to obtain an image of the initial material; the initial material includes aggregates with asphalt coated on the surface. The initial material is loaded into a rotating flask and rotated on a rotating flask tester for more than 24 hours. The material inside the rotating flask is then removed to obtain the experimental material. An initial experimental image is obtained by taking an image of the experimental material at the fixed height. After the initial experimental image is processed to unify the background, a non-negative matrix factorization algorithm is used to repair the highlights, separate and remove the specular reflection component in the initial experimental image, and retain the diffuse reflection component in the initial experimental image to obtain a de-glazed experimental image. The initial material image and the light-removing experimental image are binarized respectively. The number of black pixels in the two images is counted, and the asphalt stripping rate of the experimental material relative to the initial material is calculated.

2. The method according to claim 1, characterized in that, The mass ratio of aggregate to asphalt in the initial material is 510:(16~18); Preferably, the particle size of the aggregate is 8~11.2mm.

3. The method according to claim 1, characterized in that, The method for preparing the initial material includes mixing the aggregate and the asphalt for 100-150 seconds.

4. The method according to claim 1, characterized in that, The rotation time of the rotating bottle on the rotating bottle tester is 24~48h.

5. The method according to claim 1, characterized in that, The nonnegative matrix factorization algorithm is based on the image color intensity decomposition model of equation (I): （Ⅰ） In equation (Ⅰ), I is the color intensity matrix of the original image, I diffuse Let I be the diffuse reflection component matrix. specular This is the matrix of mirror reflection components; The nonnegative matrix factorization algorithm decomposes I in equation (Ⅰ) into equation (Ⅱ): （Ⅱ） In equation (II), W is a non-negative basic pattern matrix and K is a non-negative coefficient matrix; the submatrix ks(x) in K corresponding to the specular reflection component is a sparse matrix.

6. The method according to claim 5, characterized in that, The nonnegative matrix factorization algorithm applies L1 norm sparsity constraints and L2 norm sparsity constraints to K; the L1 norm sparsity value is 0.6~0.8, and the L2 norm sparsity value is 1.0~1.

1.

7. The method according to claim 1, characterized in that, The captured image has a resolution of ≥10 million pixels; Preferably, the fixed height is 10~30cm.

8. The method according to claim 1, characterized in that, The unified background processing includes adjusting the image background color to a single grayscale value to make the pixel histogram set.

9. The method according to claim 1, characterized in that, The method for calculating the asphalt stripping rate includes: calculating the first asphalt coating area in the initial material image and the second asphalt coating area in the light-removing test image based on the physical size of the image and the total number of pixels; obtaining the asphalt stripping area based on the difference between the two; and using the ratio of the asphalt stripping area to the first asphalt coating area as the asphalt stripping rate to quantitatively evaluate the adhesion between aggregate and asphalt.

10. The application of the method as described in any one of claims 1 to 9 in the field of roads.

Images