Method for detecting softening resistance, chlorine salt erosion resistance and frost heaving resistance of aggregate

Abstract

The invention discloses a method for detecting softening resistance, chlorine salt erosion resistance and frost heaving resistance of aggregate, and relates to the field of performance detection of aggregate for pavement engineering, and the method comprises the following steps: obtaining image information of the preprocessed aggregate, carrying out partition processing on the image information by using an image partition processing technology, and obtaining a classification result of the aggregate according to a processing result; carrying out interval sampling on the aggregate based on the classification result, and carrying out identification analysis on the aggregate in combination with a directional tracking algorithm to obtain a grading result of the aggregate; carrying out softening resistance, chlorine salt erosion resistance and frost heaving performance tests on the aggregate, and summarizing the relationship between the aggregate color and the mass loss to obtain the mass loss result of the aggregate; and performing product performance fitting operation on the aggregate to evaluate the aggregate use performance satisfaction degree under the condition of a corresponding quality loss result so as to optimize the aggregate pretreatment process. According to the invention, softening resistance, chlorine salt erosion resistance and frost heaving performance tests are carried out on the aggregate to obtain the mass loss result of the aggregate, and the quality problem can be identified in advance.

Description

Technical Field

[0001] The present invention relates to the field of performance detection of aggregates for pavement engineering. Specifically, it relates to a method for detecting the anti-softening, chloride salt erosion and frost heaving performance of aggregates. Background Technique

[0002] Aggregates are a basic raw material used as a skeleton or filling material in building and engineering materials. They are usually used in construction materials such as concrete, mortar, and asphalt mixtures. Aggregates are mainly divided into two categories: natural aggregates and artificial aggregates, and are further divided into coarse aggregates and fine aggregates according to particle size and shape. Their performance directly affects the strength, durability, workability, etc. of concrete, asphalt and other materials.

[0003] As an important part of building and engineering materials, the performance of aggregates directly determines the final quality and engineering effect of materials such as concrete and asphalt mixtures. By comprehensively understanding the source, classification, performance and application of aggregates, the most suitable aggregates can be selected in design and construction, maximizing the engineering quality and reducing construction risks.

[0004] During the process of engaging in pavement technical services, it is found that there is a phenomenon of "bursting" of pavement aggregates into powder after rainfall or snowfall on existing road sections. Through the relevant test operations in the existing technical specifications, it is impossible to effectively determine what kind of aggregates they are, resulting in slight weathering of such aggregates. After the aggregates are heated at high temperature in the mixing plant and then soaked in water, eroded by chloride salts or frozen, they are prone to expansion and powdering, thereby affecting the pavement quality.

[0005] Regarding the problems in the related technology, no effective solution has been proposed yet. Summary of the Invention

[0006] Regarding the problems in the related technology, the present invention proposes a method for detecting the anti-softening, chloride salt erosion and frost heaving performance of aggregates to overcome the above-mentioned technical problems existing in the existing related technology.

[0007] For this reason, the specific technical solution adopted by the present invention is as follows:

[0008] A method for detecting the anti-softening, chloride salt erosion and frost heaving performance of aggregates, the detection method comprising:

[0009] S1. Obtain the image information of the pretreated aggregates, perform partition processing on the image information by using image partition processing technology, and obtain the classification result of the aggregates according to the processing result;

[0010] S2. Based on the classification result, adopt region extraction technology and morphological reconstruction algorithm to perform interval sampling on the aggregates, and combine the tracking algorithm to perform identification and analysis on the aggregates to obtain the grading result of the aggregates;

[0011] S3. Based on the color classification results and gradation results, conduct tests on the anti-softening, chloride salt erosion, and frost heaving performance of the aggregates, and summarize the relationship between the aggregate color and mass loss according to the test results to obtain the mass loss results of the aggregates;

[0012] S4. Conduct product performance trial mixing operations on the aggregates, evaluate the satisfaction of the use performance of the aggregates under the corresponding mass loss results based on the trial mixing results, and optimize the aggregate pretreatment process according to the satisfaction results.

[0013] Preferably, obtain the image information of the pretreated aggregates, perform partition processing on the image information using image partition processing technology, and obtain the classification results of the aggregates according to the processing results, including:

[0014] S11. After drying and cooling the aggregates, obtain the image information of the aggregates, read the aggregate image using an image reading function, and convert the color space of the aggregate image to a color model based on an image color conversion function;

[0015] S12. Set a color threshold according to the aggregate color. After performing dilation and erosion processing on the aggregate image, use a retrieval function to search for the aggregate particle contours, calculate the bounding box and centroid of each aggregate based on the circumscribed rectangle calculation function and centroid calculation function, and mark each centroid;

[0016] S13. Calculate the average color of each aggregate using an average function according to the marking results, output the color channels of the aggregates based on the calculation results, and summarize and statistically analyze the color channel values of the aggregate colors;

[0017] S14. Calculate the coefficient of variation of the aggregate color channels according to the color channel values and coefficient calculation function, select the color channel with the largest coefficient of variation as the test determination channel value, sort the aggregate particles according to the test determination channel value, and obtain the classification results of the aggregates based on the sorting results.

[0018] Preferably, the calculation formula for the coefficient of variation of the aggregate color channels is:

[0019]

[0020] In the formula, CV R represents the coefficient of variation of the red channel of the aggregate particles, CV G represents the coefficient of variation of the green channel of the aggregate particles, CV B represents the coefficient of variation of the blue channel of the aggregate particles, σ R 、σ G and σ B represent the standard deviations of the red channel, green channel, and blue channel of the aggregate particles respectively, and represent the average values of the red channel, green channel, and blue channel of the aggregate particles respectively.

[0021] Preferably, based on the classification result, the regional extraction technology and the morphological reconstruction algorithm are adopted to perform interval sampling on the aggregates, and the tracking algorithm is combined to identify and analyze the aggregates, and the grading result of the aggregates is obtained, including:

[0022] S21. Based on the color classification result, use the regional extraction technology and the morphological reconstruction algorithm to remove the incomplete aggregate particles at the upper and lower boundaries of the aggregate image;

[0023] S22. Remove the repeatedly appearing aggregate particles in the aggregate image according to the preset rules, and perform interval sampling on the aggregate particles after removing the incomplete and repeated aggregate particles;

[0024] S23. Use the reverse tracking algorithm to track and identify the sampled aggregate particles, and combine the equivalent particle size technology to perform aggregate particle size analysis and processing, and obtain the grading result of the aggregates based on the particle size analysis result.

[0025] Preferably, use the reverse tracking algorithm to track and identify the sampled aggregate particles, and combine the equivalent particle size technology to perform aggregate particle size analysis and processing, and the grading result of the aggregates obtained based on the particle size analysis result includes:

[0026] S231. Extract the feature description points of the aggregate image, use the feature detection technology to perform screening and voting processing on the feature description points, obtain the aggregate feature points based on the processing result, and calculate the homography matrix of the aggregate feature points;

[0027] S232. Substitute the feature description points into the homography matrix, calculate the error sum of each group of homography matrices, and select the homography matrix with the minimum error sum as the optimal solution;

[0028] S233. Use the aggregate feature points corresponding to the optimal solution as the tracking and identifying aggregate points, obtain the particle size information of the tracking and identifying aggregate points, and construct an initial aggregate particle size feature parameter vector in combination with the tracking and identifying aggregate points;

[0029] S234. Use the equivalent particle size technology and the initial aggregate particle size feature parameter vector to construct a particle size feature recognition model, and use the clustering algorithm to analyze the distribution of different particle size particles in the aggregates to obtain the grading result of the aggregates.

[0030] Preferably, use the equivalent particle size technology and the initial aggregate particle size feature parameter vector to construct a particle size feature recognition model, and the grading result of the aggregates obtained by using the clustering algorithm to analyze the distribution of different particle size particles in the aggregates includes:

[0031] S2341. Obtain the statistical ratio of the initial aggregate particle size feature parameter vector, remove the particle size information with the lowest contribution rate based on the statistical ratio result, and obtain the aggregate particle size feature vector set after the removal;

[0032] S2342. Use the aggregate particle size feature vector set as an input parameter to construct a particle size feature recognition model, and randomly select several particle size information from the aggregate particle size feature vector set as the initial clustering centers;

[0033] S2343. Analyze the Euclidean distance from the particle size of all aggregate particles to the initial clustering, and analyze the squared error of the aggregate particle size feature vector set based on the Euclidean distance results;

[0034] S2344. Sort the squared error results in ascending and descending order, analyze the variation law of the aggregate particle size based on the sorting results, and analyze the distribution of particles with different particle sizes in the aggregate according to the law results to obtain the aggregate gradation results.

[0035] Preferably, the calculation formula for the statistical ratio is:

[0036]

[0037] In the formula, T Fisher represents the statistical ratio, c represents the total number of aggregate particle size categories, w t represents the t-th type of aggregate particle size, k represents the dimension of the feature vector, n t represents the number of aggregate particles of the t-th type of aggregate particle size, represents the j-th component of the t-th type of aggregate particle size, represents c j the mean value of the t-th type of aggregate particle size, m j represents c j the mean value of all aggregate particle sizes.

[0038] Preferably, based on the color classification results and gradation results, conduct anti-softening, chloride salt erosion, and frost heaving performance tests on the aggregates, and summarize the relationship between the aggregate color and mass loss according to the test results to obtain the aggregate mass loss results including:

[0039] S31. Weigh the aggregates based on the color classification results and gradation results respectively for drying and cooling. After the cooling is completed, record the color change status of the aggregates, and conduct anti-softening tests on the aggregates using the anti-water detection rules;

[0040] S32. After the anti-softening test is completed, record the mass of the remaining aggregates, and calculate the mass loss of the aggregate particles against water softening based on the remaining mass results;

[0041] S33. Use the anti-chloride salt erosion test rules to conduct chloride salt erosion tests on the remaining aggregates, obtain the mass of the residual aggregates, and calculate the mass loss of the aggregate particles against chloride salt erosion according to the residual mass results and the remaining mass results;

[0042] S34. After soaking the residual aggregates, perform anti-freezing expansion operations, calculate the mass loss of the freezing expansion of the aggregate particles, and count the corresponding relationship between different color classification results and the mass loss. Combine the relationship results with the weighted average principle to obtain the mass loss result of the aggregates.

[0043] Preferably, perform product performance trial mixing operations on the aggregates, evaluate the satisfaction of the use performance of the aggregates under the corresponding mass loss results based on the trial mixing results, and optimize the aggregate pretreatment process according to the satisfaction results, including:

[0044] S41. Select aggregates to prepare asphalt, and perform trial mixing operations on the asphalt for compressive strength, flexural strength, and freeze-thaw resistance performance. Obtain the mechanical properties and durability properties of the asphalt based on the trial mixing results.

[0045] S42. Determine the evaluation criteria and characterization forms of the asphalt. Consider the deviation amount between the adaptation result and the evaluation criteria as the solution margin, and use the satisfaction function to evaluate the solution margin to obtain the qualification rate of the asphalt.

[0046] S43. Use the qualification rate of the asphalt as the judgment criterion for the aggregates. Analyze the satisfaction of the use performance of the aggregates under the corresponding mass loss results based on the judgment criterion, and optimize the aggregate pretreatment process according to the satisfaction results.

[0047] Preferably, determining the evaluation criteria and characterization forms of the asphalt, considering the deviation amount between the adaptation result and the evaluation criteria as the solution margin, and using the satisfaction function to evaluate the solution margin to obtain the qualification rate of the asphalt includes:

[0048] S421. Based on the asphalt use requirements, clarify the evaluation index system, determine the numerical characterization form according to the evaluation index system, and analyze the variation of the trial mixing results of mechanical properties and durability properties with respect to the evaluation index system.

[0049] S422. Set the variation as the solution margin, judge the deviation degree between the solution margin and the evaluation index system, and determine the tolerance deviation amount based on the deviation degree.

[0050] S423. Use the satisfaction function to analyze the correlation coefficient of the tolerance deviation amount, and determine the correlation degree between the trial mixing results of mechanical properties and durability properties and the aggregates according to the correlation coefficient.

[0051] S424. Characterize the correlation coefficient in numerical form, compare it with the evaluation index system, and obtain the qualification rate of the asphalt based on the comparison result.

[0052] The beneficial effects of the present invention are:

[0053] 1. After obtaining the image information of the pretreated aggregate, the present invention uses image partitioning processing technology to partition the image, which can quickly identify and distinguish the surface features of different aggregates, improve the accuracy of aggregate classification, and perform interval sampling on the aggregates based on the classification results to obtain the grading results of the aggregates, enabling the accurate extraction of the particle shape and size distribution of the aggregates. At the same time, based on color classification and grading results, the present invention conducts tests on the anti-softening, chloride salt erosion, and frost heaving performance of the aggregates to obtain the mass loss results of the aggregates, which can identify in advance the quality problems that may be caused by softening, chloride salt erosion, or frost heaving, reduce the risk of engineering failure, and finally directly use the aggregates for actual product trial mixing, which can more intuitively evaluate their performance in the target application and provide comprehensive guarantee for the performance of the aggregates in different environments.

[0054] 2. Based on the color classification results, the present invention uses region extraction technology and morphological reconstruction algorithm to remove the incomplete particles at the upper and lower boundaries of the aggregate image, avoid morphological misjudgment caused by boundary problems, improve the calculation efficiency and accuracy, and remove the particles that repeatedly appear in the aggregate image according to the preset rules. After removing the incomplete and repeated particles, interval sampling is performed to reduce the calculation redundancy caused by over-dense sampling and ensure the overall representativeness of the sampling data. At the same time, the reverse tracking algorithm can quickly and accurately identify the particle position and boundary, which is especially suitable for aggregate particles with complex shapes and ensures the accuracy and credibility of the grading results.

[0055] 3. The aggregate performance detection method proposed by the present invention can quickly and effectively judge whether there are particles in the aggregate that are pulverized when soaked in water, pulverized when eroded by chloride salt, and have poor anti-freezing performance, and avoid using such unqualified aggregates in the pavement engineering entity, further improving the pavement quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0057] Figure 1 is a flowchart of a method for detecting the anti-softening, chloride salt erosion, and frost heaving performance of aggregates according to an embodiment of the present invention;

[0058] Figure 2 is a flowchart of the detection of the anti-softening, chloride salt erosion, and frost heaving performance of aggregates according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.

[0060] According to an embodiment of the present invention, a method for detecting aggregate resistance to softening, chloride salt erosion and frost heave is provided.

[0061] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the method for testing the aggregate's resistance to softening, chloride salt erosion and frost heave performance according to an embodiment of the present invention, the testing method comprises:

[0062] S1. Obtain image information of the pre-processed aggregate, perform partition processing on the image information using image partition processing technology, and obtain classification results of the aggregate according to the processing results.

[0063] In this embodiment, the image information of the pre-processed aggregate is obtained, the image information is partitioned using the image partitioning processing technology, and the classification result of the aggregate is obtained according to the processing result, including:

[0064] S11, after drying and cooling the aggregate, obtaining image information of the aggregate, reading the aggregate image using an image reading function, and converting the aggregate image from a color space to a color model based on an image color conversion function;

[0065] S12, setting a color threshold according to the color of the aggregate, after dilation and corrosion processing of the aggregate image, searching for the outline of the aggregate particles using a retrieval function, calculating the bounding box and the centroid of each aggregate based on the circumscribed rectangle obtaining function and the centroid obtaining function, and marking each centroid;

[0066] S13, calculating the average color of each aggregate using an average function according to the marking result, outputting the color channel of the aggregate based on the calculation result, and summarizing and counting the color channel values ​​of the aggregate color channels;

[0067] S14. Calculate the coefficient of variation of the aggregate color channel according to the color channel value and the coefficient obtaining function, select the color channel with the largest coefficient of variation as the test judgment channel value, sort the aggregate particles according to the test judgment channel value, and obtain the classification result of the aggregate based on the sorting result.

[0068] It should be noted that in specific applications, to ensure the accuracy of sampling, the sampling should be carried out strictly in accordance with the relevant operations in T0301 of the "Test Regulations for Aggregates in Highway Engineering" (JTG 3432-2024). Wash the aggregates clean according to the aggregate specifications, put them into an oven at 105°C ± 5°C and dry for 5h ± 15min, take them out and cool to room temperature for standby.

[0069] Take pictures of the sampled aggregates in batches. When taking pictures, ensure that there is no overlap or occlusion between aggregate particles, and do not take pictures against the light to avoid affecting the accuracy of aggregate color collection. Write a program using the OpenCV library in Python to distinguish the colors of different aggregates and extract the RGB values of each aggregate. The specific steps are as follows:

[0070] 1. Install the openCV library in Python; 2. Use the imread function to read the target photo; 3. Use the vtColor function to convert the color space of the picture to HSV; 4. Set the HSV threshold according to the color of the aggregates; 5. Use the dilate and erode functions to perform dilation and erosion processing on the picture to improve the picture quality; 6. Use the findContours function to find the contours of the aggregate particles; 7. Use the boundingRect and moments functions to calculate the bounding box and centroid of each aggregate respectively, and mark each centroid; 8. Use the average function to calculate the average color of each aggregate; 9. Use the print function to output the RGB values of each aggregate.

[0071] Summarize and analyze the RGB values of the aggregates, and statistically calculate the three color channel values of the aggregates, which are denoted as R i 、G i 、B i , as shown in Table 1, where R i represents the red channel value of the i-th aggregate, G i represents the green channel value of the i-th aggregate, B i represents the blue channel value of the i-th aggregate, n represents the total number of detected aggregate particles, and calculate the coefficient of variation CV R 、CV G 、CV B values of the red channel, green channel, and blue channel of the aggregate particles respectively according to the formulas in Table 1. In the table, σ R 、σ G and σ B represent the standard deviations of the red channel, green channel, and blue channel of the aggregate particles respectively, and represent the average values of the red channel, green channel, and blue channel of the aggregate particles respectively.

[0072] Comparison CV R , CV G , CV B value, select the color channel corresponding to the maximum value among the three as the channel value for test determination. Assume CV R , CV G , CV B value, among CV R is the maximum value, then the red channel is the determination channel. Arrange each aggregate particle in descending order of the R value, find the particles with the maximum and minimum R values, and then divide them into u groups at intervals of 40 - 60 according to the R value, where u ≥ 2.

[0073] Table 1 Channel Value Determination Table

[0074]

[0075] S2. Based on the classification results, use the region extraction technology and morphological reconstruction algorithm to perform interval sampling on the aggregates, and combine the tracking algorithm to identify and analyze the aggregates to obtain the grading results of the aggregates.

[0076] In this embodiment, based on the classification results, use the region extraction technology and morphological reconstruction algorithm to perform interval sampling on the aggregates, and combine the tracking algorithm to identify and analyze the aggregates. The grading results of the aggregates obtained include:

[0077] S21. Based on the color classification results, use the region extraction technology and morphological reconstruction algorithm to remove the incomplete aggregate particles at the upper and lower boundaries of the aggregate image.

[0078] Among them, in the process of removing incomplete aggregate particles at the upper and lower boundaries of the aggregate image by using region extraction technology and morphological reconstruction algorithm based on color classification results, the aggregate particles in the image can be distinguished from the background based on the color characteristics of the aggregate (such as RGB value), and a color segmentation algorithm (such as threshold segmentation or clustering algorithm) is applied to generate a binary image (the particle part is white and the background part is black) to ensure that the particles are effectively separated from the background. The outer contour of the aggregate particles is extracted using an edge detection algorithm (such as Sobel operator or Canny edge detection), and the image-based connected component analysis technology (Connected Component Analysis, CCA) is used to identify the independent region of each particle, mark the boundary of each particle and calculate the minimum bounding box (Bounding Box) of the particle, check whether the bounding box of the particle overlaps with the upper and lower boundaries (top or bottom) of the image, mark the overlapping particles as "boundary particles" and identify them as possibly incomplete particles; dilation (Dilation), erosion (Erosion) and other methods are performed on the boundary particles. Morphological operations reconstruct the complete outline of the particles and analyze whether the boundary particles have complete closed boundaries. If the boundary of the particle is not closed, the particle is confirmed to be an "incomplete particle". The reconstructed particles are size-checked (such as area, length-to-width ratio, etc.) to determine whether the particles meet the integrity standards (for example, particles with an area smaller than a threshold may be incomplete). Based on the results of the outline and morphological reconstruction, particles that are not closed or whose size does not meet the requirements are marked as "incomplete particles". The incomplete particle area is removed from the image and backgrounded (its pixel value is set to the background value) to ensure that only particles that meet the integrity requirements in the image are retained to form the final complete particle data.

[0079] S22, removing repeated aggregate particles in the aggregate image according to preset rules, and performing interval sampling on the aggregate particles after the incomplete and repeated aggregate particles are removed.

[0080] Among them, when removing the aggregate particles that appear repeatedly in the aggregate image according to the preset rules and performing interval sampling on the aggregate particles after the removal of incomplete and repeated aggregate particles, feature information of each particle can be extracted, including the shape, size, position, color, etc. of the particle. Image processing techniques (such as contour extraction, area calculation, centroid position, etc.) are used to obtain the geometric features of the particle, generate the feature vector of each particle, and use it for subsequent comparison and matching. Matching algorithms (such as the nearest neighbor algorithm, feature matching algorithm) are applied to compare the feature vectors of each particle to determine which particles appear repeatedly in the image, that is, the particles whose feature vector similarity exceeds the preset threshold. For the identified repeated particles, one of them is marked as the "reserved particle", and the other repeated particles are marked as "repeated particles". The selection of the reserved particle can be based on certain priority rules, such as selecting the particle with the most forward centroid position or the largest area. The particles marked as "repeated particles" are removed from the image and background processed (setting their pixel values to the background value), and the particles marked as "reserved particles" are retained to form the final image without repeated particles.

[0081] The image is divided into several uniform sub-regions (grids) to ensure that each sub-region contains a certain number of particles. In each sub-region, one or more particles are selected for sampling. The sampling rule can be based on the features of the particles, such as selecting the particle with the largest area in each sub-region, or randomly selecting a particle. The selected particle is marked as the "sampled particle", and its feature information (such as position, size, shape, etc.) is recorded to generate the final list of sampled particles, which is used as the input for subsequent particle size analysis and gradation calculation.

[0082] S23. Use the inverse tracking algorithm to track and identify the aggregate particles obtained by sampling, and combine the equivalent particle size technology to perform aggregate particle size analysis processing, and obtain the gradation result of the aggregate based on the particle size analysis result.

[0083] Among them, using the inverse tracking algorithm to track and identify the aggregate particles obtained by sampling, and combining the equivalent particle size technology to perform aggregate particle size analysis processing, and obtaining the gradation result of the aggregate based on the particle size analysis result includes:

[0084] S231. Extract the feature description points of the aggregate image, use the feature detection technology to perform screening and voting processing on the feature description points, obtain the aggregate feature points based on the processing result, and calculate the homography matrix of the aggregate feature points;

[0085] S232. Substitute the feature description points into the homography matrix, calculate the error sum of each group of homography matrices, and select the homography matrix with the minimum error sum as the optimal solution;

[0086] S233. Take the aggregate feature points corresponding to the optimal solution as the tracking and identifying aggregate points, obtain the particle size information of the tracking and identifying aggregate points, and construct an initial characteristic parameter vector of the aggregate particle size in combination with the tracking and identifying aggregate points;

[0087] S234. Use the equivalent particle size technology and the initial characteristic parameter vector of the aggregate particle size to construct a particle size characteristic recognition model, and adopt a clustering algorithm to analyze the distribution of particles with different particle sizes in the aggregate to obtain the aggregate gradation result.

[0088] Specifically, using the equivalent particle size technology and the initial characteristic parameter vector of the aggregate particle size to construct a particle size characteristic recognition model, and adopting a clustering algorithm to analyze the distribution of particles with different particle sizes in the aggregate to obtain the aggregate gradation result includes:

[0089] S2341. Obtain the statistical ratio of the initial characteristic parameter vector of the aggregate particle size, eliminate the particle size information with the lowest contribution rate based on the statistical ratio result, and obtain the aggregate particle size characteristic vector set after the elimination;

[0090] S2342. Use the aggregate particle size characteristic vector set as the input parameter to construct a particle size characteristic recognition model, and randomly select several particle size information from the aggregate particle size characteristic vector set as the initial clustering centers;

[0091] S2343. Analyze the Euclidean distance from the particle size of all aggregate particles to the initial clustering, and analyze the squared error of the aggregate particle size characteristic vector set based on the Euclidean distance result;

[0092] S2344. Sort the squared error results in ascending and descending order, analyze the change law of the aggregate particle size based on the sorting result, and analyze the distribution of particles with different particle sizes in the aggregate according to the law result to obtain the aggregate gradation result.

[0093] Among them, the calculation formula of the statistical ratio is:

[0094]

[0095] In the formula, T Fisher represents the statistical ratio, c represents the total number of aggregate particle size categories, w t represents the t-th category of aggregate particle size, k represents the dimension of the characteristic vector, n t represents the number of aggregate particles of the t-th category of aggregate particle size, represents the j-th component of the t-th category of aggregate particle size, represents c j at the mean value of the t-th category of aggregate particle size, m j represents c j the mean value of all aggregate particle sizes.

[0096] To facilitate the understanding of the above technical solutions of the present invention, the following will give a detailed description of the operation method for obtaining the gradation result in the actual process of the present invention.

[0097] Step 1: Eliminate incomplete aggregate particles;

[0098] Input the aggregate image data. Assume the image size is 1024×1024 pixels, the resolution is 300 dpi, the background is black with RGB value (0, 0, 0), and the aggregate particles are grayish-white with RGB value range from (120, 120, 120) to (255, 255, 255). Apply the threshold segmentation algorithm to mark the pixels with RGB value greater than (120, 120, 120) as the particle part (white), and the others as the background part (black), generating a binary image where the particle part is white and the background part is black;

[0099] Use the Canny edge detection algorithm to extract the edge contours of the particles. Assume the number of detected particle boundaries is 450. Use the connected component analysis technique (CCA) to label each particle region, obtaining the minimum bounding box (BoundingBox) of the particles. Detect whether the bounding boxes of the particles overlap with the upper and lower boundaries of the image. The result shows that the bounding boxes of 35 particles overlap with the boundaries.

[0100] Apply dilation (structural element size is 3x3) and erosion operations to the particles at the upper and lower boundaries. The detection result shows that the boundaries of 20 particles are not closed, confirming them as incomplete particles. Perform size detection (area threshold is 500 pixels) on the reconstructed particles. Among them, the areas of 15 particles are less than the threshold. Remove the above 20 unclosed particles and 15 particles with too small areas from the image and perform background processing. Finally, the number of remaining particles is 415.

[0101] Step 2: Eliminate duplicate particles and perform interval sampling;

[0102] Extract feature information for each particle, including shape (the contour is approximately circular or elliptical, described as the shape factor), size (calculate the particle area, ranging from 500 to 10000 pixels), centroid position (such as the centroid coordinates of the particle), and color (calculate the average gray value of the particle, ranging from 150 to 255). Use the nearest neighbor algorithm to calculate the similarity of the particle feature vectors. The threshold is set to 0.95. It is found that there are 30 pairs of duplicate particles (feature vector similarity > 0.95). For the duplicate particles, select the particle with the most forward centroid position as the "retained particle", and at the same time remove the other duplicate particles from the image. Finally, the number of remaining particles is 385.

[0103] Divide the image into 10×10 sub-regions (each sub-region size is 102×102 pixels). In each sub-region, select the particle with the largest area as the sampling particle. Finally, 100 particles are sampled.

[0104] Step 3: Conduct gradation analysis based on the reverse tracking algorithm and equivalent particle size technology;

[0105] Use the SIFT (Scale-Invariant Feature Transform) algorithm to extract the feature description points of each sampled particle. Approximately 30 feature points are extracted for each particle. Through screening and voting processing, 15 of the most significant feature points of each particle are retained. Use the RANSAC (Random Sample Consensus) algorithm to calculate the homography matrix. The error sum of each group of homography matrices is between 0.001 and 0.005. Select the homography matrix with the minimum error sum as the optimal solution.

[0106] Take the feature points corresponding to the optimal solution as the tracking and identification points, obtain the particle size information of each particle, and calculate the equivalent particle size of the particle through the equivalent particle size formula (A represents the particle area):

[0107]

[0108] The calculated result shows that the equivalent particle size range is from 2 mm to 20 mm. Use the equivalent particle size and other characteristic parameters (shape factor, gray value, etc.) to construct a particle size feature vector set. Eliminate the particle size information with a contribution rate lower than 5%. Finally, 95 effective particle size characteristic parameters are obtained. Randomly select 5 particle size information from the feature vector set as the initial clustering centers, calculate the Euclidean distance from all particles to the clustering centers, and analyze the squared error of the feature vector set. According to the ascending and descending sorting results of the squared error, the final gradation result is:

[0109] The proportion of particles with a particle size less than 5 mm is 25%; the proportion of particles with a particle size of 5 - 10 mm is 50%; the proportion of particles with a particle size greater than 10 mm is 25%; the total number of aggregate particles is 100; the average particle size is 7.5 mm; the standard deviation is 2.8 mm.

[0110] Furthermore, through the techniques of aggregate image segmentation, morphological reconstruction, and region extraction, the incomplete particles at the upper and lower boundaries and the duplicate particles in the image are eliminated. Subsequently, through the combination of interval sampling and reverse tracking algorithm with equivalent particle size technology, the particle size information of the aggregate particles is accurately obtained, and the gradation result of the aggregate is obtained based on cluster analysis.

[0111] S3. Conduct anti-softening, chloride salt erosion, and frost heaving performance tests on the aggregate based on the color classification result and the gradation result, and summarize the relationship between the aggregate color and mass loss according to the test results to obtain the mass loss result of the aggregate.

[0112] In this embodiment, conducting anti-softening, chloride salt erosion, and frost heaving performance tests on the aggregate based on the color classification result and the gradation result, and summarizing the relationship between the aggregate color and mass loss according to the test results to obtain the mass loss result of the aggregate includes:

[0113] S31. Weigh the aggregates based on the color classification results and gradation results respectively, dry and cool them, record the color change status of the aggregates after cooling is completed, and conduct a softening resistance test on the aggregates using the water resistance detection rules.

[0114] S32. After the softening resistance test is completed, record the mass of the remaining aggregates, and calculate the mass loss of the aggregate particles against water softening based on the remaining mass results.

[0115] S33. Conduct a chloride salt erosion test on the remaining aggregates using the chloride salt erosion test rules, obtain the mass of the residual aggregates, and calculate the mass loss of the aggregate particles against chloride salt erosion based on the residual mass results and the remaining mass results.

[0116] S34. After soaking the residual aggregates, perform anti-freezing expansion operations, calculate the mass loss of the aggregate particles due to freezing expansion, and statistically analyze the corresponding relationship between different color classification results and mass loss. Combine the relationship results with the weighted average principle to obtain the mass loss results of the aggregates.

[0117] It should be noted that in the actual application process, based on the color classification results and gradation results, conduct softening resistance, chloride salt erosion, and freezing expansion performance tests on different grouped color aggregates respectively. As Figure 2 shown, the specific test operations include the following steps:

[0118] Step 1. Preheat the test oven to 200 - 210 °C, then evenly spread the weighed test sample on the tray and place it in the oven, heat for 2 h ± 5 min, take it out and cool to room temperature, and record the color change of the test sample particles.

[0119] Step 2. Inject clean water to ensure that the water surface is 2 cm higher than the test sample, gently stir the test sample by hand until no bubbles escape from the water surface. Immerse it in water at room temperature for 24 h ± 1.5 h, observe the state change of the aggregate particles during the soaking process. If there is particle pulverization or shelling, it is determined that the aggregate is unqualified and the test can be terminated.

[0120] Step 3. If the aggregate particles do not undergo pulverization or shelling during the first soaking process, drain the water in the tray, note that no aggregate particle loss should occur during this operation process, and then place it in an oven at 105 °C ± 5 °C and dry for 5 h ± 15 min.

[0121] Step 4. Repeat the operation in Step 2 to conduct a saturation operation on the test sample and observe the change of the aggregate particles.

[0122] Step 5. Drain the water in the tray, note that no aggregate particle loss should occur during this operation process; then place it in an oven at 105 °C ± 5 °C and dry to a constant weight (when the drying time of the test sample is greater than 5 h, every 1 h later, the difference between the two weighing results before and after is not greater than 0.1%).

[0123] Step 6. After cooling to room temperature, sieve through the lower limit sieve of the sample particle size, and weigh the residue mass and record it as M 1i , and then calculate the mass loss Q of the particle resistance to water softening 1i :

[0124]

[0125] In the formula, Q 1i represents the mass loss percentage (%) of the i-th group of particles against water softening, m 0 represents the dried mass (g) of the sample before the test, and M 1i represents the mass of the residue particles after the water softening test of the i-th group of particles (%).

[0126] Step 7. Put the residue from Step 6 into a tray, pour NaCl solution with a mass percentage concentration of 22% into the tray, ensure that the liquid level is 2 cm higher than the sample, gently stir the sample with a stirrer until no bubbles escape from the water surface. Immerse it in water at room temperature for 24 h ± 1.5 h, observe the state change of the aggregate particles during the soaking process. If there is particle pulverization or shelling, it is determined that the aggregate is unqualified and the test can be terminated.

[0127] Step 8. If no particle pulverization or shelling occurs in the aggregate, perform the drying operation on the aggregate according to Step 3.

[0128] Step 9. Repeat the operation in Step 7 and observe the change of the aggregate particles.

[0129] Step 10. If no particle pulverization or shelling occurs in the aggregate particles, drain the solution in the tray, note that no aggregate particle loss should occur during this operation process; then put it into an oven at 105°C ± 5°C and dry it to a constant weight (when the drying time of the sample is more than 5 h, every 1 h later, the difference between the two weighing results before and after is not more than 0.1%).

[0130] Step 11. After cooling to room temperature, sieve through the lower limit of the sample particle size, and weigh the residue mass and record it as M 2i , and then calculate the mass loss Q of the particle resistance to chloride salt erosion 2i :

[0131]

[0132] In the formula, Q 2i represents the mass loss percentage (%) of the i-th group of particles against chloride salt erosion, M 1i represents the mass of the residue particles after the water softening test of the i-th group of particles (g), and M 2i represents the mass of the residue particles of the i-th group of particle samples after the test (%).

[0133] Step 12: Put the residue from Step 11 into a tray and soak the aggregates in water according to Step 2.

[0134] Step 13: Wipe off the surface water of the soaked specimen with a wet towel, then evenly spread it in the tray and put it into a constant temperature freezer at -18°C ± 2°C for 16 h ± 1 h.

[0135] Step 14: After taking out the specimen, soak it in a constant temperature water bath at 60°C for 12 h ± 45 min, and record the changes of aggregate particles during the soaking process.

[0136] Step 15: Take out the specimen, put it into an oven at 105°C ± 5°C and dry it to a constant weight (when the drying time of the specimen is more than 5 h, every 1 h later, the difference between the weighing results of the previous and the next time is not more than 0.1%). After cooling to room temperature, sieve it with the lower limit sieve of the specimen particle size, and weigh the residue mass and record it as M 3i , and then calculate the mass loss Q of particle frost heaving resistance 3i :

[0137]

[0138] In the formula, Q 3i represents the mass loss percentage (%) of particle frost heaving resistance; M 2i represents the mass (g) of the residue particles after the test of the i-th group of particle specimens; M 3i represents the mass (%) of the residue particles after the test of the specimen.

[0139] After all the grouped tests are completed, count the mass losses of water softening resistance, chloride salt erosion resistance and frost heaving resistance corresponding to different color groups respectively. When large-scale feeding of aggregates from the same quarry and the same parent material, randomly sample according to Step S1, divide the aggregates into u groups according to the color channel, weigh the different groups of aggregates distinguished and record them as m i , and calculate the mass losses of water softening resistance, chloride salt erosion resistance and frost heaving resistance of the incoming materials respectively.

[0140]

[0141] In the formula, Q 1 represents the mass loss of water softening resistance of the incoming materials; Q 2 represents the mass loss of chloride salt erosion resistance of the incoming materials; Q 3 represents the mass loss value of frost heaving resistance of the incoming materials; Q 1u represents the mass loss of water softening resistance of the u-th group of the incoming materials grouped by color; Q 2u represents the mass loss of chloride salt erosion resistance of the u-th group of the incoming materials grouped by color; Q 3uIt represents the mass loss value of the anti-freezing and swelling property of the u-th group of the incoming materials grouped by color; m u It represents the mass (g) of the u-th group of aggregates of the incoming materials grouped by color.

[0142] The mass loss percentage Q of the particles against water softening calculated 1 , the mass loss percentage Q of the particles against chloride salt erosion 2 , the mass loss percentage Q of the particles against freezing and swelling 3 If any one of the results is greater than 3, it means that the corresponding index of the aggregate is unqualified.

[0143] Based on the implementation processes of steps S1 and S3, as shown in Table 2, the 5-10 mm aggregates in a certain project's material yard are detected. According to step S1, the coefficient of variation of the red channel R value is the largest and most sensitive, and the R value is used as the judgment basis. The identification results are sorted according to the R value, and the R value distribution range of the aggregate is 22-94. With an interval of 40, the aggregate is divided into two groups: 22-62 and 63-94, that is, u = 2. For the two groups of aggregates distinguished, the mass losses of the two groups of aggregates against water softening, chloride salt erosion, and freezing and swelling are detected respectively according to step S3.

[0144] For the test results: Q 11 = 0.9%, Q 21 = 1.4%, Q 31 = 1.1%; Q 12 = 1.6%, Q 22 = 3.7%, Q 32 = 2.8%. Samples of the incoming aggregates are taken for testing, and m 1 = 368.3 g, m 2 = 152.6 g. At the same time, the mass losses of the incoming aggregates against water softening are calculated, Q 1 = 1.1%, Q 2 = 2.1%, Q 3 = 1.6%. Then it is judged that the water softening, chloride salt erosion, and freezing and swelling properties of this batch of incoming aggregates meet the requirements.

[0145] Table 2 Aggregate test results

[0146]

[0147]

[0148]

[0149] S4. Conduct product performance trial mixing operations on the aggregates, evaluate the service performance satisfaction of the aggregates under the corresponding mass loss results based on the trial mixing results, and optimize the aggregate pretreatment process according to the satisfaction results.

[0150] In this embodiment, a product performance trial mixing operation is carried out on the aggregate, and based on the trial mixing results, the satisfaction degree of the use performance of the aggregate under the corresponding mass loss result conditions is evaluated. Optimizing the aggregate pretreatment process according to the satisfaction degree results includes:

[0151] S41. Select aggregates to prepare asphalt, and conduct trial mixing operations on the asphalt for compressive strength, flexural strength and freeze-thaw resistance performance. Based on the trial mixing results, obtain the mechanical properties and durability properties of the asphalt.

[0152] Specifically, in the process of selecting aggregates to prepare asphalt, conducting trial mixing operations on the asphalt for compressive strength, flexural strength and freeze-thaw resistance performance, and obtaining the mechanical properties and durability properties of the asphalt based on the trial mixing results, mix coarse aggregates, fine aggregates and mineral powder according to the target gradation ratio, heat the aggregates and asphalt to the specified temperature, mix the aggregates and asphalt evenly according to the design mix ratio, use a Marshall compactor to form specimens, make specimens for compressive strength, flexural strength and freeze-thaw performance tests, conduct compressive strength tests on a pressure testing machine, load until the specimen fails and record the maximum load, test multiple specimens and take the average value as the compressive strength result, conduct flexural strength tests on a three-point bending testing machine, test multiple specimens, record the maximum load at failure, and take the average value as the flexural strength result; conduct multiple freeze-thaw cycle treatments on the specimens, including freezing and soaking steps, test the compressive strength before and after freeze-thaw, calculate the strength loss rate, and evaluate the durability.

[0153] S42. Determine the evaluation criteria and characterization forms of the asphalt, regard the deviation amount between the adaptation result and the evaluation criteria as the solution margin, and use the satisfaction function to evaluate the solution margin to obtain the qualified rate of the asphalt.

[0154] Specifically, determining the evaluation criteria and characterization forms of the asphalt, regarding the deviation amount between the adaptation result and the evaluation criteria as the solution margin, and using the satisfaction function to evaluate the solution margin to obtain the qualified rate of the asphalt includes:

[0155] S421. Based on the asphalt use requirements, clarify the evaluation index system, determine the numerical characterization form according to the evaluation index system, and analyze the variation of the trial mixing results of mechanical properties and durability properties with respect to the evaluation index system;

[0156] S422. Set the variation as the solution margin, judge the deviation degree between the solution margin and the evaluation index system, and determine the tolerance deviation amount based on the deviation degree;

[0157] S423. Use the satisfaction function to analyze the correlation coefficient of the tolerance deviation amount, and determine the correlation degree between the trial mixing results of mechanical properties and durability properties and the aggregates according to the correlation coefficient;

[0158] S424. Characterize the associated common representation in numerical form, compare it with the evaluation index system, and obtain the qualified rate of asphalt based on the comparison result.

[0159] S43. Use the qualified rate of asphalt as the judgment standard for aggregates, analyze the satisfaction degree of the use performance of aggregates under the corresponding quality loss result conditions based on the judgment standard, and optimize the aggregate pretreatment process according to the satisfaction degree result.

[0160] In summary, by means of the above technical solutions of the present invention, after obtaining the image information of the pretreated aggregates, the present invention uses the image partitioning processing technology to partition the image, can quickly identify and distinguish the surface features of different aggregates, improve the accuracy of aggregate classification, and perform interval sampling on the aggregates based on the classification results to obtain the grading results of the aggregates, so that the particle shape and size distribution of the aggregates can be accurately extracted. At the same time, based on the color classification and grading results, the aggregates are tested for anti-softening, chloride salt erosion and frost heaving performance to obtain the quality loss results of the aggregates, which can identify in advance the quality problems that may be caused by softening, chloride salt erosion or frost heaving, reduce the risk of engineering failure, and finally directly use the aggregates for actual product trial mixing, which can more intuitively evaluate their performance in the target application and provide comprehensive guarantee for the performance of aggregates in different environments. Based on the color classification results, the present invention uses the region extraction technology and the morphological reconstruction algorithm to remove the incomplete particles at the upper and lower boundaries of the aggregate image, avoid the morphological misjudgment caused by the boundary problem, improve the calculation efficiency and accuracy, and remove the particles that repeatedly appear in the aggregate image according to the preset rules. After removing the incomplete and repeated particles, interval sampling is performed to reduce the calculation redundancy caused by over-dense sampling, ensure that the sampling data has overall representativeness, and at the same time use the reverse tracking algorithm to quickly and accurately identify the particle positions and boundaries, which is especially suitable for aggregate particles with complex shapes and ensures the accuracy and credibility of the grading results.

[0161] The aggregate performance detection method proposed by the present invention can quickly and effectively judge whether there are particles in the aggregates that are pulverized when soaked in water, pulverized when eroded by chloride salts, and have poor anti-freezing performance, and avoid using such unqualified aggregates in the pavement engineering entity, further improving the pavement quality.

[0162] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for testing the resistance of aggregate to softening, chloride salt erosion and frost heave, characterized in that: The test method includes: S1. Obtain image information of the pre-processed aggregate, perform partition processing on the image information using image partition processing technology, and obtain a classification result of the aggregate according to the processing result; S2. Based on the classification results, the region extraction technology and morphological reconstruction algorithm are used to perform interval sampling on the aggregates, and the direction tracking algorithm is used to identify and analyze the aggregates to obtain the gradation results of the aggregates; S3. Based on the color classification results and grading results, the softening resistance, chloride salt erosion and frost heave performance of the aggregate are tested, and the relationship between the aggregate color and mass loss is summarized according to the test results to obtain the mass loss results of the aggregate; S4. Carry out product performance test mixing of aggregates, evaluate the satisfaction of aggregate performance under the corresponding mass loss conditions based on the test mixing results, and optimize the aggregate pretreatment process according to the satisfaction results.

2. A method for testing the resistance of aggregate to softening, chloride salt corrosion and frost heave according to claim 1, characterized in that: The method of obtaining image information of the preprocessed aggregate, performing partition processing on the image information using image partition processing technology, and obtaining a classification result of the aggregate according to the processing result includes: S11, after drying and cooling the aggregate, obtaining image information of the aggregate, reading the aggregate image using an image reading function, and converting the aggregate image from a color space to a color model based on an image color conversion function; S12, setting a color threshold according to the color of the aggregate, after dilation and corrosion processing of the aggregate image, searching for the outline of the aggregate particles using a retrieval function, calculating the bounding box and the centroid of each aggregate based on the circumscribed rectangle obtaining function and the centroid obtaining function, and marking each centroid; S13, calculating the average color of each aggregate using an average function according to the marking result, outputting the color channel of the aggregate based on the calculation result, and summarizing and counting the color channel values ​​of the aggregate color channels; S14. Calculate the coefficient of variation of the aggregate color channel according to the color channel value and the coefficient obtaining function, select the color channel with the largest coefficient of variation as the test judgment channel value, sort the aggregate particles according to the test judgment channel value, and obtain the classification result of the aggregate based on the sorting result.

3. A method for testing the resistance of aggregate to softening, chloride salt corrosion and frost heave according to claim 2, characterized in that: The calculation formula of the aggregate color channel variation coefficient is: In the formula, CV R Represents the coefficient of variation of the red channel of aggregate particles, CV G Represents the coefficient of variation of the green channel of aggregate particles, CV B represents the coefficient of variation of the blue channel of aggregate particles, σ R , σ G and σ B Respectively represent the standard deviation of the red channel, green channel, and blue channel of aggregate particles. and Represent the average values ​​of the red channel, green channel, and blue channel of aggregate particles, respectively.

4. The method for testing the aggregate's resistance to softening, chloride salt erosion and frost heave according to claim 1, characterized in that: The method adopts the region extraction technology and the morphological reconstruction algorithm based on the classification result to perform interval sampling on the aggregate, and combines the directional tracking algorithm to perform identification and analysis on the aggregate to obtain the aggregate gradation result, including: S21. Based on the color classification results, the region extraction technology and morphological reconstruction algorithm are used to remove incomplete aggregate particles at the upper and lower boundaries of the aggregate image; S22, removing repeated aggregate particles in the aggregate image according to preset rules, and performing interval sampling on the aggregate particles after the incomplete and repeated aggregate particles are removed; S23. Track and identify the aggregate particles obtained by sampling using a reverse tracking algorithm, and perform aggregate particle size analysis in combination with equivalent particle size technology, and obtain aggregate grading results based on the particle size analysis results.

5. A method for testing the resistance of aggregate to softening, chloride salt corrosion and frost heave according to claim 4, characterized in that: The method uses a reverse tracking algorithm to track and identify the aggregate particles obtained by sampling, and combines the equivalent particle size technology to perform aggregate particle size analysis. The gradation results of the aggregate obtained based on the particle size analysis results include: S231, extracting feature description points of the aggregate image, performing screening and voting processing on the feature description points using feature detection technology, obtaining aggregate feature points based on the processing results, and calculating the homography matrix of the aggregate feature points; S232, substituting the feature description points into the homography matrix, calculating the error sum of each group of homography matrices, and selecting the homography matrix with the smallest error sum as the optimal solution; S233, taking the aggregate feature point corresponding to the optimal solution as the tracking and identifying aggregate point, obtaining the particle size information of the tracking and identifying aggregate point, and constructing the initial feature parameter vector of the aggregate particle size in combination with the tracking and identifying aggregate point; S234. Use the equivalent particle size technology and the initial characteristic parameter vector of aggregate particle size to build a particle size feature recognition model, and use the clustering algorithm to analyze the distribution of particles of different particle sizes in the aggregate to obtain the aggregate grading result.

6. A method for testing the resistance of aggregate to softening, chloride salt corrosion and frost heave according to claim 5, characterized in that: The equivalent particle size technology and the initial characteristic parameter vector of the aggregate particle size are used to construct a particle size feature recognition model, and a clustering algorithm is used to analyze the distribution of particles of different particle sizes in the aggregate to obtain the aggregate gradation result, including: S2341, obtaining a statistical ratio of an initial characteristic parameter vector of aggregate particle size, eliminating particle size information with the lowest contribution rate based on the statistical ratio result, and obtaining an aggregate particle size characteristic vector set after the elimination is completed; S2342, using the aggregate particle size feature vector set as an input parameter to construct a particle size feature recognition model, and randomly selecting a number of particle size information from the aggregate particle size feature vector set as initial clustering centers; S2343, analyzing the Euclidean distances of all aggregate particle sizes to the initial cluster, and analyzing the squared error of the aggregate particle size feature vector set based on the Euclidean distance results; S2344. Sort the square error results in ascending and descending order, analyze the variation pattern of aggregate particle size based on the sorting results, and analyze the distribution of particles of different particle sizes in the aggregate according to the regular results to obtain the aggregate grading result.

7. A method for testing the resistance of aggregate to softening, chloride salt corrosion and frost heave according to claim 6, characterized in that: The calculation formula of the statistical ratio is: Where, T Fisher represents the statistical ratio, c represents the total number of aggregate size categories, and w t represents the particle size of the t-th type of aggregate, k represents the dimension of the feature vector, n t represents the number of aggregate particles of the t-th type of aggregate size, represents the jth component of the particle size of the tth type of aggregate, Indicates c j The mean value of the aggregate size in the tth class, m j Indicates c j The mean value for all aggregate sizes.

8. The method for testing the aggregate's resistance to softening, chloride salt erosion and frost heave according to claim 1, characterized in that: The softening resistance, chloride salt erosion and frost heave performance test of the aggregate is performed based on the color classification result and the grading result, and the relationship between the aggregate color and the mass loss is summarized according to the test results, and the mass loss result of the aggregate is obtained, including: S31. Based on the color classification results and the grading results, the aggregates are weighed and dried and cooled respectively. After the cooling is completed, the color change of the aggregates is recorded, and the softening resistance test of the aggregates is performed using the water resistance detection rule; S32. After the anti-softening test is completed, the mass of the remaining aggregate is recorded, and the mass loss of the aggregate particles due to anti-water softening is calculated based on the remaining mass result; S33, performing a chloride salt corrosion test on the remaining aggregate using a chloride salt corrosion resistance test rule, obtaining the mass of the remaining aggregate, and calculating the mass loss of the aggregate particles against chloride salt corrosion based on the residual mass result and the residual mass result; S34. After soaking the residual aggregate in water, follow the anti-frost heave operation to calculate the mass loss of the aggregate particles due to frost heave, and statistically analyze the corresponding relationship between the different color classification results and the mass loss. Combine the relationship result with the weighted average principle to obtain the mass loss result of the aggregate.

9. The method for testing the aggregate's resistance to softening, chloride salt erosion and frost heave according to claim 1, characterized in that: The product performance trial mixing operation is performed on the aggregate, and based on the trial mixing result, the satisfaction of the aggregate performance under the corresponding quality loss result condition is evaluated, and the aggregate pretreatment process is optimized according to the satisfaction result, including: S41. Select aggregates to prepare asphalt, and conduct test mixing of the compressive strength, flexural strength and freeze-thaw resistance of the asphalt, and obtain the mechanical properties and durability of the asphalt based on the test mixing results; S42, determining the evaluation standard and representation form of asphalt, considering the deviation between the adaptation result and the evaluation standard as the solution margin, and using the satisfaction function to evaluate the solution margin to obtain the qualified rate of asphalt; S43. The qualified rate of asphalt is used as the judgment standard for aggregate. Based on the judgment standard, the satisfaction of the aggregate performance under the corresponding quality loss result conditions is analyzed, and the aggregate pretreatment process is optimized according to the satisfaction results.

10. A method for testing the aggregate's resistance to softening, chloride salt erosion and frost heave according to claim 9, characterized in that: The method of determining the evaluation standard and characterization form of asphalt, regarding the deviation between the adaptation result and the evaluation standard as the solution margin, and using the satisfaction function to evaluate the solution margin to obtain the qualified rate of asphalt includes: S421. Clarify the evaluation index system based on the demand for asphalt use, determine the numerical representation form based on the evaluation index system, and analyze the changes in the test results of mechanical properties and durability properties relative to the evaluation index system; S422, setting the change situation as the solution margin, judging the degree of deviation between the solution margin and the evaluation index system, and determining the tolerance deviation amount based on the degree of deviation; S423, using the satisfaction function to analyze the correlation coefficient of the tolerance deviation, and determining the correlation between the mechanical properties and durability performance test results and the aggregate according to the correlation coefficient; S424. Express the associated value in numerical form, compare it with the evaluation index system, and obtain the qualified rate of asphalt based on the comparison result.

Images