An asphalt pavement construction gradation alarm system and method

Abstract

The application provides an asphalt pavement construction grading alarm system and method. In the method, first, the real-time obtained asphalt paving layer image is subjected to grading feature analysis, the aggregate grading feature data of the target region is extracted, and then the grading risk situation analysis of the current paving layer is carried out based on a preset construction grading prediction model, the sieve hole passing rate of each particle size interval and the local space proportion are output in real time, so as to realize the rapid quantitative evaluation of the grading state, so that the grading deviation can be captured in time without waiting for the post-detection result. On this basis, further deviation feature analysis is carried out according to the passing rate deviation feature and the spatial distribution dispersion, and the particle size deviation score is calculated, and finally the grading abnormality alarm can be triggered according to the particle size deviation score or the local space proportion of each particle size interval, realizing the real-time active perception and immediate feedback of the abnormal state. Thus, the timeliness and accuracy of the asphalt pavement construction quality control are effectively improved.

Description

Technical Field

[0001] This application relates to the field of gradation monitoring technology, and in particular to an alarm system and method for gradation monitoring in asphalt pavement construction. Background Technology

[0002] Current methods for alarming the gradation of asphalt pavement during construction typically rely on periodic sampling and testing. However, the cumbersome and time-consuming process of sampling aggregates and testing materials means that only a low sampling frequency can be maintained during construction, making it difficult to achieve real-time alarms covering the entire process. More importantly, current methods only provide post-construction data recording or analysis, lacking a mechanism for immediate alarms when gradation anomalies occur. Construction workers cannot promptly detect gradation deviations and take corrective measures during paving. This inability to provide timely alarms means that gradation anomalies are often only discovered in subsequent tests, severely impacting quality control during pavement construction. Summary of the Invention

[0003] In view of the above problems, and in order to address the high lag in the current construction gradation alarm methods, this application provides an asphalt pavement construction gradation alarm system and method.

[0004] The embodiments of this application disclose the following technical solutions:

[0005] In a first aspect, embodiments of this application provide an alarm method for asphalt pavement construction gradation, including:

[0006] Gradation feature analysis is performed on the target area corresponding to the asphalt paving layer image to obtain the aggregate gradation feature data of the target area;

[0007] Based on the preset construction gradation prediction model, the gradation risk situation analysis of the current paving layer is performed according to the aggregate gradation characteristic data to obtain gradation risk characteristic data; the gradation risk characteristic data includes: the sieve passing rate of each particle size range and the local space ratio of each particle size range.

[0008] Based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval, deviation characteristic analysis is performed to obtain the gradation deviation characteristic vector corresponding to each particle size interval.

[0009] The particle size deviation is calculated based on the gradation deviation feature vector of each particle size interval to obtain the particle size deviation score of each particle size interval.

[0010] An alarm is triggered for gradation abnormalities during the construction of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval.

[0011] In one possible implementation, the step of alarming for gradation anomalies during the construction of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each of the particle size intervals includes:

[0012] If the particle size deviation score is less than the score threshold, the gradation status of the current paving layer is determined to be abnormal, and a gradation abnormality alarm signal is generated for alarm purposes.

[0013] If the particle size deviation score is not less than the score threshold, an alarm is triggered for gradation abnormalities during the construction of the target asphalt pavement based on the local spatial proportion of each particle size interval.

[0014] In one possible implementation, the step of alarming for gradation anomalies during the construction of the target asphalt pavement based on the local spatial proportion of each of the particle size intervals includes:

[0015] Based on the local spatial proportion of each particle size range, calculate the coefficient of variation of the proportion of each particle size range, and the Pearson correlation coefficient between any two particle size ranges; the coefficient of variation of the proportion is used to characterize the degree of dispersion of aggregate particles in the target area in the corresponding space of the particle size range.

[0016] If the coefficient of variation of the proportion of any of the particle size ranges is greater than the first coefficient threshold, or if any of the Pearson correlation coefficients is less than the second threshold, the gradation state of the current paving layer is determined to be abnormal, and the gradation abnormality alarm signal is generated for alarm purposes.

[0017] In one possible implementation, after determining that the gradation state of the current paving layer is abnormal, the method further includes:

[0018] If the current gradation state of the paved layer is determined to be abnormal, historical sieve aperture passing rate data and historical spatial distribution data of the abnormal particle size range in the current aggregate paving process are obtained.

[0019] Based on the historical sieve aperture passing rate data and the historical spatial distribution data, a passing rate trend curve and a set of spatial distribution curves for the abnormal particle size range are determined; the passing rate trend curve is used to characterize the changing trend of the sieve aperture passing rate with the construction mileage, and the set of spatial distribution curves includes multiple spatial distribution curves divided according to a preset construction mileage segment; the spatial distribution curves are used to characterize the changing trend of the local spatial proportion with the physical coordinates of the road surface.

[0020] The gradation anomaly region is determined based on the pass rate trend curve and the set of spatial distribution curves for the abnormal particle size range.

[0021] In one possible implementation, determining the gradation anomaly region based on the pass rate trend curve and the set of spatial distribution curves for the abnormal particle size range includes:

[0022] Identify the abnormal data segment within the pass rate trend curve; the pass rate change rate of the abnormal data segment is greater than the warning threshold.

[0023] Based on the set of spatial distribution curves, determine the abnormal spatial distribution curves that are in the same construction mileage segment as the abnormal data segment;

[0024] Based on the road surface physical coordinates covered by the abnormal spatial distribution curve, the initial gradation abnormal area is determined, and the drift trend analysis is performed on the abnormal data segment.

[0025] If it is determined that the abnormal data segment has a continuous drift trend, the diffusion trend of the abnormal area is determined according to the distribution direction of the grid in the initial gradation abnormal area.

[0026] Based on the working condition data of the paving equipment and the diffusion trend of the abnormal area, the initial gradation abnormal area is adjusted for regional diffusion to obtain the gradation abnormal area.

[0027] If it is determined that the abnormal data segment does not exhibit the continuous drift trend, the initial gradation abnormal region is identified as the gradation abnormal region.

[0028] In one possible implementation, the step of performing gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area includes:

[0029] The asphalt paving layer image is segmented according to the aggregate type to determine the morphological parameters of each aggregate particle in the target area, as well as the aggregate distribution image of the target area; the aggregate distribution image is used to characterize the spatial distribution of aggregate particles of different sizes in the target area.

[0030] Based on the morphological parameters of each aggregate particle, the aggregate particles are screened and statistically analyzed according to a preset particle size range screening rule to obtain the particle size range category covered by the target area.

[0031] The particle size range categories covered by the target area, the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area are determined as the aggregate gradation characteristic data.

[0032] In one possible implementation, the aggregate type includes AC type aggregate;

[0033] The step of segmenting the asphalt paving layer image according to aggregate type to determine the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area, includes:

[0034] When the aggregate type is AC type aggregate, the asphalt paving layer image is white-balance corrected to obtain a first pre-processed asphalt paving layer image; the first pre-processed asphalt paving layer image is used to enhance the contrast between aggregate particles and asphalt binder.

[0035] Based on the bimodal grayscale features of the first preprocessed asphalt paving layer image, segmentation parameters are analyzed to determine the double-layer segmentation threshold; the double-layer segmentation threshold is used to distinguish the aggregate particles from the asphalt binder and to classify aggregate particles of different particle size ranges.

[0036] The first preprocessed asphalt paving layer image is segmented into aggregate particles using the double-layer segmentation threshold to obtain an aggregate particle outline map.

[0037] Based on the aggregate particle outline map, the morphological parameters of each aggregate particle in the target area are extracted, and the aggregate particles belonging to different particle size ranges are labeled to obtain the aggregate distribution image.

[0038] In one possible implementation, the dual-layer segmentation threshold includes a first grayscale threshold and a second grayscale threshold, wherein the first grayscale threshold is greater than the second grayscale threshold.

[0039] The step of performing aggregate particle segmentation processing on the first preprocessed asphalt paving layer image using the dual-layer segmentation threshold to obtain an aggregate particle outline map includes:

[0040] Using the first grayscale threshold, aggregate particles with grayscale values ​​greater than the first grayscale threshold in the first preprocessed asphalt paving layer image are extracted to obtain the first coarse aggregate region.

[0041] The first coarse aggregate area is removed from the first pre-processed asphalt paving layer image to obtain the second pre-processed asphalt paving layer image.

[0042] By using the second grayscale threshold, aggregate particles with grayscale values ​​less than the first grayscale threshold and greater than the second grayscale threshold in the second preprocessed asphalt paving layer image are extracted to obtain the first fine aggregate region.

[0043] The first coarse aggregate region and the first fine aggregate region are merged and optimized to obtain the aggregate particle outline diagram.

[0044] In one possible implementation, the aggregate type includes SMA-type aggregate; the step of performing aggregate particle segmentation processing on the asphalt paving layer image according to the aggregate type to determine the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area, includes:

[0045] When the aggregate type is SMA type aggregate, the asphalt paving layer image is subjected to histogram equalization processing to obtain a third pre-processed asphalt paving layer image; the third pre-processed asphalt paving layer image is used to remove non-aggregate areas in the asphalt paving layer image.

[0046] The third preprocessed asphalt paving layer image is divided into regions using a marker-controlled watershed algorithm to obtain a fourth preprocessed asphalt paving layer image. The fourth preprocessed asphalt paving layer image includes a second coarse aggregate region, a second fine aggregate region, and a mastic slurry region. The grayscale value of the second coarse aggregate region is greater than a third grayscale threshold, the grayscale value of the second fine aggregate region is between a fourth grayscale threshold and the third grayscale threshold, and the grayscale value of the mastic slurry region is less than the fourth grayscale threshold. The third grayscale threshold is greater than the fourth grayscale threshold.

[0047] Extract aggregate particles whose particle edge spacing is not greater than the preset aggregate adhesion judgment threshold in the second coarse aggregate region, and remove aggregate particles whose particle edge spacing is greater than the preset aggregate adhesion judgment threshold in the second coarse aggregate region to obtain the skeleton aggregate region.

[0048] Based on the skeleton aggregate region and the second fine aggregate region, the morphological parameters of each aggregate particle in the target region are extracted, and the skeleton aggregate region, the second fine aggregate region and the mastic paste region are labeled respectively to obtain the aggregate distribution image.

[0049] In a second aspect, embodiments of this application provide an asphalt pavement construction gradation alarm system, the system being used to implement any possible asphalt pavement construction gradation alarm method in the first aspect, the system comprising:

[0050] The feature analysis module is used to perform gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area;

[0051] The risk analysis module is used to perform a gradation risk situation analysis on the current paving layer based on a preset construction gradation prediction model and the aggregate gradation characteristic data, and obtain gradation risk characteristic data; the gradation risk characteristic data includes: the sieve passing rate of each particle size range and the local space ratio of each particle size range.

[0052] The feature processing module is used to perform deviation feature analysis based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval to obtain the gradation deviation feature vector corresponding to each particle size interval.

[0053] The particle size analysis module is used to calculate the particle size deviation based on the gradation deviation feature vector of each particle size interval, and obtain the particle size deviation score of each particle size interval.

[0054] The alarm processing module is used to issue an alarm for gradation abnormalities during the construction process of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval.

[0055] Compared with existing technologies, this application has the following advantages: The asphalt pavement construction gradation alarm system and method provided in this application first analyzes the gradation features of real-time acquired asphalt paving layer images, extracting aggregate gradation feature data of the target area. Then, based on a preset construction gradation prediction model, it analyzes the gradation risk situation of the current paving layer, instantly outputting the sieve passing rate and local spatial proportion of each particle size interval. This achieves rapid quantitative assessment of the gradation state, enabling gradation deviations to be captured promptly without waiting for post-construction testing results. Furthermore, based on the passing rate deviation characteristics and spatial distribution dispersion, deviation feature analysis is performed, and particle size deviation scores are calculated. Finally, based on the particle size deviation score or the local spatial proportion of each particle size interval, a gradation anomaly alarm can be triggered, achieving real-time proactive perception and immediate feedback of abnormal states. Therefore, gradation deviation information can be obtained immediately during the paving process, and corresponding adjustment measures can be taken in a timely manner, effectively improving the timeliness and accuracy of asphalt pavement construction quality control. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating an alarm method for asphalt pavement construction gradation provided in this application embodiment;

[0058] Figure 2 A flowchart illustrating another method for alarming the gradation of asphalt pavement construction provided in this application embodiment;

[0059] Figure 3 A flowchart illustrating a method for determining aggregate gradation characteristic data provided in an embodiment of this application;

[0060] Figure 4 A schematic diagram of aggregate distribution for AC aggregates provided in this application embodiment;

[0061] Figure 5 A flowchart illustrating a method for determining gradation anomaly regions provided in an embodiment of this application;

[0062] Figure 6 A schematic diagram of a pass rate trend curve provided for an embodiment of this application;

[0063] Figure 7 A schematic diagram of a spatial distribution curve provided in an embodiment of this application;

[0064] Figure 8 A flowchart illustrating another method for determining gradation anomaly regions provided in this application embodiment;

[0065] Figure 9 This application provides a schematic diagram of a process for particle segmentation of AC type aggregates.

[0066] Figure 10 This application provides a schematic diagram of a process for particle segmentation of SMA type aggregates.

[0067] Figure 11 This is a schematic diagram of the structure of an asphalt pavement construction gradation alarm system provided in an embodiment of this application. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and accompanying drawings. It should be particularly noted that the embodiments described in this application are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0069] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0070] As mentioned earlier, current methods for alarming the construction gradation of asphalt pavements typically rely on periodic sampling and testing. However, the process of sampling aggregates and testing materials is cumbersome and time-consuming, resulting in a low sampling and testing frequency during construction. This makes it difficult to achieve high-frequency, full-process coverage of real-time alarms for construction gradation. Current construction gradation alarms suffer from high latency, which seriously affects the timeliness of quality control during pavement construction.

[0071] To address this issue, this application provides an asphalt pavement construction gradation alarm system and method. This method first analyzes the gradation features of real-time acquired asphalt paving layer images, extracting aggregate gradation feature data for the target area. Then, based on a preset construction gradation prediction model, it analyzes the gradation risk situation of the current paving layer, instantly outputting the sieve passing rate and local spatial proportion of each particle size interval. This achieves rapid quantitative assessment of the gradation status, enabling timely detection of gradation deviations without waiting for post-construction testing results. Furthermore, it analyzes deviation features based on passing rate deviation characteristics and spatial distribution dispersion, calculating particle size deviation scores. Finally, based on the particle size deviation score or the local spatial proportion of each particle size interval, a gradation anomaly alarm is triggered, achieving real-time proactive perception and immediate feedback of abnormal states. Therefore, gradation deviation information can be obtained immediately during the paving process, allowing for timely adjustments and effectively improving the timeliness and accuracy of asphalt pavement construction quality control.

[0072] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0073] See Figure 1 The figure is a flowchart illustrating an asphalt pavement construction gradation alarm method provided in an embodiment of this application, which specifically includes the following steps:

[0074] S101: Perform gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area.

[0075] In the asphalt paving layer image acquisition stage, industrial cameras or vision sensors can be symmetrically mounted on the front, sides, and rear of the paving equipment to acquire asphalt paving layer images in real time. During the image acquisition process, the lens is pointed towards the paving layer surface and adjusted to a suitable focal length to ensure coverage of the entire paving width. Simultaneously, anti-vibration and dustproof / waterproof protective devices are used to avoid the impact of dust or equipment vibration in the construction environment on image quality, achieving high-definition real-time image capture during the paving process.

[0076] The target area within the asphalt paving layer image is determined using the paving equipment's operating data and a pre-defined aggregate type. Specifically, the paving equipment's operating data can be acquired through the paving equipment's built-in monitoring system or external sensors. This data includes parameters such as paving speed, paving thickness, vibration frequency, hopper discharge, equipment trajectory, and machine levelness. In determining the target area requiring focused gradation monitoring analysis based on the operating data, non-paved areas at the image edges can be cropped according to the equipment's trajectory and paving width. Abnormal areas that are too thin or too thick can be eliminated based on a pre-defined paving thickness threshold. Simultaneously, non-aggregate interference areas such as road markings and impurities can be excluded by combining the grayscale feature range corresponding to the aggregate type. Finally, the core monitoring area, covering the effective paving range and free from interference, is identified and designated as the target area. Through the above method, step S101 achieves rapid conversion from raw image data to structured gradation feature data, providing accurate data support for subsequent gradation risk situation analysis based on the preset construction gradation prediction model, fundamentally ensuring that the entire gradation alarm method can run in real time during the paving process, rather than relying on data with lag properties.

[0077] S102: Based on the preset construction gradation prediction model, the gradation risk situation of the current paving layer is analyzed according to the aggregate gradation characteristic data to obtain gradation risk characteristic data; the gradation risk characteristic data includes: the sieve passing rate of each particle size range and the local space ratio of each particle size range.

[0078] In this embodiment, the preset construction gradation prediction model is an intelligent model trained on a large number of asphalt pavement construction samples. The input data covers the particle size range categories of the target area, the morphological parameters of each aggregate, and the aggregate distribution image (i.e., the data included in the aggregate gradation feature data). The final output includes gradation risk feature data containing the sieve passing rate and local spatial proportion of each particle size range. Taking AC type aggregate (Dense-graded Asphalt Concrete mixture) as an example, the target area is set to cover the fine, medium, and coarse particle size ranges in the example above. The model first extracts the number and morphological characteristics of particles in each range. The average equivalent particle size of the 230 particles in the fine aggregate range is 3.5 mm, and the roundness is 0.72. The average equivalent particle size of the 186 particles in the medium aggregate range is 7.8 mm, and the aspect ratio is 2.4. The average equivalent particle size of the 124 particles in the coarse aggregate range is 12.3 mm, and the outline integrity is 0.81. At the same time, the spatial coordinates and aggregation state of particles of different sizes in the aggregate distribution image are read. Based on this, the model uses a built-in algorithm to correlate morphological parameters with pre-set sieve aperture passing rate standards, predicting the sieve aperture passing rate of particles in each interval. Specifically, the passing rate of fine aggregates is 38% for 4.75mm sieve apertures, 29% for medium aggregates for 9.5mm sieve apertures, and 19% for coarse aggregates for 13.2mm sieve apertures, thus completing the prediction of sieve aperture passing rate. At the same time, combined with the spatial grid division of the aggregate distribution image, the area proportion of particles in different sub-regions of each interval is statistically analyzed, resulting in the local spatial proportion of fine aggregates fluctuating between 32% and 40%, medium aggregates between 27% and 31%, and coarse aggregates between 17% and 21%, forming local spatial proportion data and outputting complete gradation risk characteristic data.

[0079] In one possible implementation, the pre-defined construction gradation prediction model adopts an RF-MLP ensemble model. The random forest module in the model is responsible for mining the nonlinear correlation between morphological parameters and particle size range categories, while the multilayer perceptron module focuses on capturing the spatial features of the aggregate distribution image.

[0080] S103: Based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval, deviation characteristic analysis is performed to obtain the gradation deviation characteristic vector corresponding to each particle size interval.

[0081] In the previous step, the sieve passing rate and local spatial proportion of each particle size interval were obtained through the preset construction gradation prediction model. This characterized the current gradation state of the paving layer from two dimensions: quantity proportion and spatial distribution. However, these two indicators only represent absolute measurement values ​​and cannot directly determine whether the gradation deviates from the design requirements. Therefore, this step performs a passing rate deviation characteristic analysis on the sieve passing rate of each particle size interval. The actual measured sieve passing rate is compared with the preset sieve passing rate interval to determine the direction and magnitude of deviation of each interval relative to the standard range. Simultaneously, a spatial distribution dispersion analysis is performed on the local spatial proportion of each particle size interval. By calculating the degree of variation in the proportion of each particle size interval at different spatial locations within the target area, the uniformity level of aggregate spatial distribution is quantitatively assessed. Based on this, the above analysis results are further integrated to generate a gradation deviation feature vector corresponding to each particle size interval. This gradation deviation feature vector is a multi-dimensional feature set that can comprehensively characterize the abnormal state of the particle size interval in both the passing rate deviation and spatial dispersion dimensions.

[0082] S104: Calculate the particle size deviation based on the gradation deviation feature vector of each particle size interval to obtain the particle size deviation score for each particle size interval.

[0083] Specifically, the particle size deviation score is calculated through the following two steps:

[0084] Step 1: Based on the particle size range categories covered by the target area, determine the preset sieve aperture passing rate range associated with each particle size range, and the gradation influence weight associated with each particle size range.

[0085] In the gradation monitoring process of this embodiment, the pre-set sieve passing rate range is the standard for judging whether the aggregates in each particle size range meet the design requirements. Different types of particle size ranges correspond to different sieve passing rate benchmarks, thereby improving the accuracy of gradation monitoring. Taking common AC type aggregate as an example, its particle size range is set as three categories: fine aggregate, medium aggregate, and coarse aggregate as in the example above. The sieve sizes corresponding to the three particle size ranges are 4.75mm, 9.5mm, and 13.2mm, respectively. The preset passing rate range is set based on design requirements. For example, if the passing rate of fine aggregate is too high, it will cause the road surface to bleed and the anti-skid performance to decrease; if it is too low, it will affect the adhesion between asphalt and aggregate. Therefore, the corresponding preset sieve passing rate range needs to be set within the range of 25%-35%. If the passing rate of coarse aggregate is too high, it will destroy the stability of the skeleton structure; if it is too low, it will increase the porosity and cause rainwater infiltration problems. The corresponding preset sieve passing rate range is set to 15%-25%. Therefore, the essence of pre-setting the pass rate range is to define a safe range for the pass rate of aggregate gradation, ensuring that its performance meets the basic requirements of pavement design.

[0086] The gradation influence weight is a key parameter for measuring the contribution of each particle size range to the overall pavement performance. The weight must be determined based on construction experience, statistical data, or machine learning models. The logic behind setting the gradation influence weight is that the particle size range with the greater impact on pavement performance receives a higher weight. For example, in AC type aggregates, coarse aggregate is the core of the pavement skeleton, directly determining strength and stability, thus receiving the highest weight (approximately 0.4); medium aggregate connects coarse and fine aggregates, affecting gradation continuity, and has the next highest weight (approximately 0.35); fine aggregate fills voids, affecting density and skid resistance, and has the lowest weight (approximately 0.25). Based on this, when calculating the particle size deviation score according to the gradation influence weight, the deviation value of each range is multiplied by its corresponding weight and then summed. For example, if the coarse aggregate pass rate deviation is 10% (weight 0.4) and the fine aggregate deviation is 5% (weight 0.25), the final deviation score will highlight the anomaly of the coarse aggregate, thereby ensuring the targeted nature of gradation monitoring. Deviations in sections that significantly impact pavement performance (such as coarse aggregate) are more clearly reflected in the scoring, preventing minor deviations in secondary sections from masking major problems in primary sections, thus accurately identifying the severity of gradation anomalies.

[0087] Step 2: Based on the gradation deviation feature vector of each particle size interval, the preset sieve aperture passing rate interval, and the gradation influence weight, the deviation is calculated to obtain the particle size deviation score of each particle size interval.

[0088] This step builds upon the previously generated gradation deviation feature vector, transforming it into a particle size deviation score that comprehensively reflects the severity of gradation anomalies through weighted deviation calculation. Specifically, the gradation deviation feature vectors for each particle size interval have been generated. This vector integrates deviation information in two dimensions: sieve aperture passing rate deviation and spatial distribution dispersion. However, before proceeding to the anomaly assessment stage, this multi-dimensional deviation information needs to be further compressed into a straightforward score value for comparison with a preset scoring threshold. Therefore, this step extracts the deviation component related to sieve aperture passing rate from the gradation deviation feature vectors of each particle size interval and performs weighted deviation calculation based on the preset sieve aperture passing rate interval and gradation influence weights. Since different particle size intervals have varying degrees of influence on the pavement performance of asphalt mixtures—for example, coarse aggregate forms the core of the pavement skeleton, directly determining high-temperature stability and load-bearing capacity, while fine aggregate mainly fills voids and improves density, contributing relatively little to overall performance—it is necessary to use gradation influence weights to highlight the deviation contribution of key intervals, making the scoring results more accurately reflect the true degree of harm caused by gradation anomalies to pavement quality. Taking the fine, medium, and coarse particle size ranges in AC type aggregate as an example: the preset passing rate ranges for the three are 35%-45%, 25%-35%, and 15%-25%, respectively, with corresponding weights of 0.25, 0.35, and 0.4. If the passing rate deviation components extracted from the gradation deviation feature vectors of the three ranges are 2% (actual 38% from the midpoint and 40% from the midpoint), 1% (actual 29% from the midpoint and 30% from the midpoint), and 2% (actual 18% from the midpoint and 20% from the midpoint), then the weighted sum is 2%×0.25+1%×0.35+2%×0.4=1.65, and the particle size deviation score is 100-1.65=98.35 points.

[0089] S105: Based on the particle size deviation score or the local spatial proportion of each particle size interval, an alarm is triggered for gradation abnormalities during the construction process of the target asphalt pavement.

[0090] In this step, the process of issuing an alarm for gradation abnormalities during the construction of the target asphalt pavement based on particle size deviation scores or the proportion of various local spatial factors is achieved through the following two steps:

[0091] Step 1: If the particle size deviation score is less than the score threshold, determine that the gradation status of the current paving layer is abnormal, and generate a gradation abnormality alarm signal to trigger an alarm.

[0092] The particle size deviation score is a comprehensive score obtained by weighting the gradation deviation feature vectors of each particle size range. Its advantage lies in its ability to intuitively reflect the overall degree of deviation of the current paving layer's gradation from the design standard in terms of overall quantity proportions, facilitating rapid initial screening and determination of gradation status. Based on this, this step first compares the calculated particle size deviation score with a preset scoring threshold: when the particle size deviation score is less than the threshold, it indicates that the current paving layer has shown a significant overall deviation in the sieve passing rate of the key particle size range, and this deviation, after weighting, has reached or exceeded the preset warning line, posing a clear threat to pavement quality. At this point, no further spatial analysis is needed; the abnormal gradation status of the current paving layer is directly determined, and a gradation anomaly alarm signal is generated. This allows for the fastest possible transmission of the abnormal information to on-site operators for timely process adjustment measures.

[0093] In one possible implementation, the alarm signal can be presented as a pop-up window on the display terminal in the paver's cab, simultaneously triggering an audible and visual alarm to emit a buzzer and flashing warning, drawing the operator's immediate attention. The alarm information can also be simultaneously pushed to the mobile terminal of on-site quality management personnel or a remote monitoring platform, including the specific time, location, abnormal particle size range, and deviation details of the anomaly, facilitating timely understanding of the abnormal situation and targeted responses by management personnel. Based on this, a tiered alarm mechanism can be set according to the severity of the anomaly. For minor deviations, only a warning is issued, reminding on-site personnel to pay attention to the gradation change trend and prepare for adjustments; for severe deviations, a mandatory alarm is triggered, requiring immediate suspension of paving operations until the deviation is effectively corrected. Through the generation and transmission of the above alarm signals, operators can obtain gradation anomaly information immediately during the paving process without waiting for the next day's sampling and sieving results, thus allowing for timely adjustments to the mixing plant's mix proportions or the paver's operating parameters.

[0094] Step 2: If the particle size deviation score is not less than the score threshold, an alarm is triggered for gradation abnormalities during the construction of the target asphalt pavement based on the local spatial proportion of each particle size interval.

[0095] However, meeting the particle size deviation score only indicates that the sieve passing rate of each particle size range meets the numerical requirements, but it cannot reflect the spatial uniformity of the aggregate. This is because the essence of gradation is not only about meeting the proportion of each particle size, but also about the aggregate forming a stable and uniform spatial structure within the paving layer. For example, coarse aggregate needs to be evenly distributed to form a skeleton support, while fine aggregate needs to fill the voids to ensure density. Therefore, even if the aggregate particle size meets the numerical requirements, the presence of coarse aggregate clusters or fine aggregate concentrations in local areas (e.g., coarse aggregate accounting for more than 40% in the left half and less than 15% in the right half) will lead to significant performance differences between different areas of the pavement, resulting in structural hazards. Such structural hazards cannot be identified through a weighted scoring mechanism, but may cause localized cracking of the pavement under long-term vehicle loads. Therefore, after ensuring that the particle size deviation score is not less than the scoring threshold, it is necessary to further monitor the gradation based on the local spatial proportion distribution of aggregate particles within each particle size range in the target area, thereby ensuring that the paving layer also meets the requirements in local dimensions.

[0096] Next, based on the accompanying drawings of specific embodiments, the process of triggering an alarm for gradation anomalies during the construction of the target asphalt pavement according to the local spatial proportion of each particle size range will be described. See also... Figure 2 The figure is a flowchart illustrating another asphalt pavement construction gradation alarm method provided in this application embodiment, which specifically includes the following steps:

[0097] S1046: Calculate the coefficient of variation of the proportion of each particle size interval and the Pearson correlation coefficient between any two particle size intervals based on the local spatial proportion of each particle size interval; the coefficient of variation of the proportion is used to characterize the degree of dispersion of aggregate particles in the target area in the corresponding space of the particle size interval.

[0098] In the local spatial analysis stage of gradation monitoring, the coefficient of variation (COP) is a key indicator for quantifying the dispersion of aggregates within a single particle size range. Before calculation, the asphalt pavement layer needs to be divided into several small spatial units (such as a 10cm × 10cm grid). Image recognition technology is used to obtain the mass percentage of aggregates within a certain particle size range in each unit (i.e., the local spatial percentage). For example, for coarse aggregates of 10-15mm, each grid has corresponding percentage data, such as 18%, 22%, 19%, 25%, 17%, 23%, etc. Subsequently, statistical methods are used to calculate the overall average percentage and standard deviation of this particle size range in the target area, and the ratio of the standard deviation to the mean is used as the COP. This effectively avoids the misleading effect of absolute standard deviation due to differences in particle size mean. For example, when the mean of coarse aggregate is 20% and the standard deviation is 4%, the COP is 0.2. However, when the mean of fine aggregate is 30% and the standard deviation is also 4%, the COP is only 0.13, indicating that the spatial distribution of fine aggregate is more uniform.

[0099] The coefficient of variation (CVA) is used to identify hidden problems where the overall proportion is acceptable but localized aggregate aggregation is present. If the CVA of coarse aggregate exceeds a threshold, it means that the proportion of coarse aggregate in some meshes may far exceed the average. Excessive aggregation of coarse aggregate in these areas can lead to abnormalities such as an overly dense skeleton structure and a sharp decrease in porosity. This can prevent asphalt from effectively coating the aggregate surface, easily causing problems such as asphalt stripping and bleeding. A single sieve pass rate test can only verify whether the overall proportion of coarse aggregate meets the preset standard, but it cannot identify such localized imbalances. The CVA fills this gap, preventing potential quality problems caused by uneven local distribution.

[0100] Unlike the coefficient of variation, which focuses on the local spatial proportion of a single particle size range, the Pearson correlation coefficient focuses on the synergy in the spatial distribution of two particle size ranges. In calculating the Pearson correlation coefficient, the local spatial proportion data of any two particle size ranges in each grid cell needs to be extracted. Then, the linear correlation between the two is calculated using statistical methods, with the result expressed as a value between -1 and 1. A positive value indicates that both increase or decrease together, such as an area with more coarse aggregate also having more fine aggregate. A negative value indicates an opposite trend, such as an area with more coarse aggregate and less fine aggregate. The larger the absolute value, the stronger the correlation.

[0101] The Pearson correlation coefficient is used to identify hidden segregation problems where the proportions are acceptable but the structure fails. For example, if the correlation coefficient between coarse and fine aggregates is -0.7 (strong negative correlation), it indicates that areas with high coarse aggregate concentrations are severely deficient in fine aggregates, and the voids in the aggregate skeleton are not filled, leading to excessive porosity in the mixture and easy water accumulation and water damage. Conversely, areas with high fine aggregate concentrations lack the coarse aggregate skeleton, resulting in reduced load-bearing capacity and rutting. Even if the overall proportions of both meet design requirements, this spatial mismatch will still undermine the structural rationality of the gradation. Therefore, the significance of the Pearson correlation coefficient lies in transforming the synergistic logic of aggregate distribution into a quantifiable digital signal. Even if the proportions of all particle sizes meet the standards, if the spatial distribution does not meet the set synergistic requirements, the Pearson correlation coefficient can be used to identify and ensure that the aggregate distribution truly conforms to the pre-defined structural intent.

[0102] S1047: Determine whether the coefficient of variation of the proportion of any particle size range is less than the first coefficient threshold, and whether any Pearson correlation coefficient is greater than the second coefficient threshold.

[0103] S1048: If the coefficient of variation of the proportion of any of the particle size ranges is less than the first coefficient threshold and any of the Pearson correlation coefficients is greater than the second coefficient threshold, the gradation status of the current paving layer is determined to be normal.

[0104] S1049: If the coefficient of variation of the proportion of any of the particle size ranges is greater than the first coefficient threshold, or if any of the Pearson correlation coefficients is less than the second coefficient threshold, the gradation state of the current paving layer is determined to be abnormal, and the gradation abnormality alarm signal is generated for alarm purposes.

[0105] After calculating the coefficient of variation for the proportion of each particle size interval and the Pearson correlation coefficient between any two particle size intervals, the system determines whether each particle size interval meets the requirements based on a pre-set first and second coefficient thresholds. In the judgment logic of S1047, it is explicitly stated that the coefficient of variation for the proportion of each particle size interval must be less than the pre-set first coefficient threshold to ensure that no single particle size is excessively aggregated or missing in space. For example, the proportion of coarse aggregate cannot suddenly increase in a certain area, leading to an overly dense skeleton, nor can the proportion of fine aggregate suddenly decrease in a certain area, resulting in unfilled voids. Simultaneously, the Pearson correlation coefficient between any two particle size intervals must be greater than the pre-set second coefficient threshold to ensure that the distribution of different particle sizes follows the set logic. For example, in areas with abundant coarse aggregate, fine aggregate should be appropriately filled, rather than completely absent, and medium aggregate should connect the two to form a continuous gradation structure.

[0106] If any condition is not met, the gradation state of the current paving layer is determined to be abnormal, and a gradation abnormality alarm signal is generated. This abnormality originates from two aspects. First, the coefficient of variation of the proportion of a certain particle size range exceeds the standard. For example, if the coefficient of variation of the proportion of coarse aggregate exceeds the first coefficient threshold, it indicates that it is excessively aggregated in a local area, resulting in an overly dense skeleton structure, reduced porosity, and insufficient asphalt coating. Second, the Pearson correlation coefficient between two particle size ranges is lower than the second coefficient threshold. For example, if the correlation coefficient between the coarse aggregate range and the fine aggregate range is too low, it indicates that the two are negatively correlated in spatial distribution. The coarse aggregate aggregation area lacks fine aggregate filling, leading to excessive porosity and susceptibility to water damage, while the fine aggregate aggregation area lacks a coarse aggregate skeleton, resulting in reduced load-bearing capacity and rutting. Even if the overall proportion meets the design, this spatial mismatch can still cause implicit segregation, leading to a decline in pavement performance.

[0107] When both conditions are met simultaneously, it indicates that the spatial distribution of aggregates has reached an ideal state: uniform dispersion of a single particle size, without local aggregation or gaps, ensures consistent composition of the mixture in each area; different particle sizes complement each other, with coarse aggregates forming a stable skeleton, fine aggregates fully filling voids, and medium aggregates connecting the skeleton and the filler layer, resulting in continuous overall gradation that conforms to the design structural intent. At this point, the mechanical properties of the mixture can be fully utilized; for example, its resistance to deformation, water damage, and fatigue can all meet the specifications, and the pavement quality is under control.

[0108] The above is an introduction to the construction gradation monitoring method in this application embodiment. Next, the process of generating aggregate gradation characteristic data in step S101 will be described in conjunction with the accompanying drawings of a specific process embodiment. See also... Figure 3 The figure is a flowchart illustrating a method for determining aggregate gradation characteristic data according to an embodiment of this application, specifically including the following steps:

[0109] S1011: Perform aggregate particle segmentation processing on the asphalt paving layer image according to the aggregate type to determine the morphological parameters of each aggregate particle in the target area, as well as the aggregate distribution image of the target area; the aggregate distribution image is used to characterize the spatial distribution of aggregate particles of different sizes in the target area.

[0110] In actual asphalt pavement construction scenarios, due to significant differences in structural characteristics, particle morphology, and surrounding media among different aggregates, general aggregate particle segmentation methods are insufficient to guarantee accuracy. Taking AC-type aggregate and SMA-type aggregate (SMA-type aggregate, used in discontinuously graded asphalt mastic macadam mixtures) as examples, AC-type aggregate is a continuously graded aggregate with a uniform particle size distribution, interspersed coarse and fine aggregates, and moderate grayscale contrast with asphalt mastic, exhibiting a continuous and gradual grayscale distribution in the image. Therefore, for AC-type aggregate, using segmentation algorithms suitable for cohesive particles can easily lead to over-segmentation, misclassifying a single aggregate particle as multiple particles. Thus, a targeted double-layer grayscale threshold segmentation method is needed. This method utilizes the bimodal grayscale characteristics of the image to determine two levels of thresholds, thereby separating the coarse and fine aggregate regions within the target area and avoiding missegmentation. SMA-type aggregates are discontinuously graded aggregates, characterized by a high content of coarse aggregates, large particle size, and easy interlocking and adhesion, easily forming continuous high-grayscale areas. Furthermore, their grayscale difference from mastic mortar is significant. For SMA-type aggregates, using the same grayscale threshold segmentation method as for AC-type aggregates will not only fail to solve the coarse aggregate adhesion problem but will also confuse the boundaries between fine aggregates and mastic mortar. Therefore, for SMA-type aggregates, it is necessary to eliminate interference in the graph through histogram equalization, extract seed points of coarse aggregates and segment adhered particles, and then screen out the coarse aggregates that truly constitute the skeleton. This ensures that the segmentation results accurately reflect the structural characteristics of SMA-type aggregates, providing the true morphological contours of the particles for subsequent analysis.

[0111] In this step, morphological parameters are a quantitative description of the core characteristics of aggregate particles. Different aggregate types exhibit certain differences in morphological parameters. For AC type aggregates, morphological parameters mainly include equivalent particle size, particle roundness, and particle area. These parameters can intuitively show the number and proportion of particles belonging to different particle size ranges within the aggregate, thus replacing the cumbersome process of traditional physical screening. For SMA type aggregates, in addition to equivalent particle size, key morphological parameters include the number of contact points (counting the number of contacts between a single coarse aggregate and other coarse aggregates; if the contact exceeds a certain value, it is judged as a skeleton particle), angularity, and skeleton gap ratio. These parameters can characterize the skeleton structure characteristics of SMA aggregates, providing a quantitative standard for judging whether the gradation meets load-bearing requirements. The aggregate distribution image is a visual representation of the morphological parameters, aiming to intuitively present the spatial distribution of particles through differentiated annotations. For details, please refer to... Figure 4 The disclosed diagram illustrates an aggregate distribution image for AC-type aggregates. As shown in the image, in this example of AC-type aggregate aggregate distribution, blue indicates coarse aggregate areas, green indicates fine aggregate areas, and gray indicates asphalt mastic. This transforms abstract particle size distribution data into an intuitive image, providing a visual reference for judging gradation uniformity.

[0112] S1012: Based on the morphological parameters of each aggregate particle, each aggregate particle is screened and statistically analyzed using a preset particle size range screening rule to obtain the particle size range category covered by the target area;

[0113] S1013: The particle size range category covered by the target area, the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area are determined as the aggregate gradation characteristic data.

[0114] In this step, the purpose of the preset particle size range screening rules is to classify aggregate particles with different morphological parameters and different particle sizes into different particle size ranges. This facilitates subsequent statistical analysis of the proportion of different types of aggregate particles within the aggregate, ensuring the accuracy of gradation monitoring. Specifically, taking AC type aggregate as an example, the corresponding particle size range categories (i.e., the preset particle size range screening rules) for this type of aggregate include three ranges: fine aggregate (2-5mm), medium aggregate (5-10mm), and coarse aggregate (10-15mm). After completing the aggregate particle segmentation process, the morphological parameters of all aggregates within the target area are extracted and matched against the rules one by one. For example, if a particle has an equivalent particle size of 3.8mm, a projected area of ​​9.2mm², an aspect ratio of 2.5, and a roundness of 0.71, all parameters meet the requirements of the fine aggregate range, and therefore this aggregate particle can be classified into the fine aggregate category. Another particle, with an equivalent particle size of 9.4 mm, an aspect ratio of 2.7, and a roundness of 0.66, meets the standards for medium aggregates and is therefore classified into the medium aggregate particle size range. Finally, after screening and statistically analyzing all aggregate particles within the target area, it can be determined that the area covers three particle size ranges: fine, medium, and coarse. The number of aggregate particles within each particle size range can also be determined, facilitating the subsequent determination of the gradation ratio. It should be noted that the preset particle size range screening rules in this step are strongly correlated with the type of aggregate in the actual application scenario. There is no unified naming convention for different particle size ranges as fine, medium, and coarse aggregates; the specific classification depends on the actual type of aggregate. This embodiment does not impose such a limitation.

[0115] In practical gradation monitoring scenarios, after determining that the gradation state within the current paving layer is abnormal, quickly identifying the abnormal gradation areas within the paving layer is equally important for maintaining pavement construction quality. Based on this, this application embodiment also provides a method for determining abnormal gradation areas, facilitating rapid location of abnormal gradation areas when gradation anomalies are detected. Next, the method for determining abnormal gradation areas provided by this application embodiment will be described in detail with reference to the accompanying drawings of specific process embodiments.

[0116] See Figure 5 The figure is a flowchart illustrating a method for determining gradation anomaly regions provided in an embodiment of this application, specifically including the following steps:

[0117] S201: If the gradation state of the current paving layer is determined to be abnormal, obtain the historical sieve aperture passing rate data and historical spatial distribution data of the abnormal particle size range during the aggregate paving process.

[0118] When determining that the gradation status of the current paving layer is abnormal, it is necessary to analyze the historical data of the current construction process to identify the abnormal gradation area. Among them, the abnormal particle size range is the particle size range that is determined to be abnormal in the above judgment rules. The above includes three judgment rules: (1) whether the sieve passing rate of the aggregate particles in the particle size range is within the corresponding preset sieve passing rate range; (2) whether the coefficient of variation of the proportion corresponding to the particle size range is less than the first coefficient threshold; (3) whether the Pearson correlation coefficient between the particle size range and other ranges is greater than the second coefficient threshold. For these three judgment rules, if the particle size range does not meet any of the rules, the historical data of the particle size range is included in the process of determining the gradation abnormal area.

[0119] Specifically, the historical data corresponding to the abnormal particle size range includes historical sieve passing rate and historical spatial distribution data. The former is the aggregate screening results collected in segments by mileage (e.g., a fixed 50-meter interval) during construction, recording the overall proportion changes of each particle size range at different construction stages. For example, the passing rate of 10-15mm coarse aggregate is 20% in the first 0-50 meters, and becomes 22% in the next 50-100 meters. The trend of the passing rate reflects whether the amount of this particle size gradually deviates from the set passing rate range as construction progresses. The latter is the spatial distribution information of aggregate in previous mileage segments, recording the local proportion of each particle size in each small spatial unit. For example, in the 300-350 meter mileage segment, the proportion of 10-15mm coarse aggregate in each grid is 18%, and other changes will occur in the next mileage segment. This data reflects whether the distribution of this particle size in space is uniform. In this embodiment, the reason for analyzing the gradation abnormality area based on historical sieve passing rate data and historical spatial distribution data is that gradation abnormalities often do not occur in isolation. For example, if the coefficient of variation of a certain particle size range exceeds the standard, it may be because its throughput gradually increases with mileage, resulting in excessive local usage and thus accumulating too many aggregate particles in that particle size range; while if the correlation coefficient of two types of particle sizes is too low, it may be because their spatial distribution shows a negative correlation with mileage, such as when the throughput of coarse aggregate increases while the throughput of fine aggregate decreases, causing the two to be unable to coordinate in space.

[0120] S202: Based on the historical sieve aperture passing rate data and the historical spatial distribution data, determine the passing rate trend curve and spatial distribution curve set for the abnormal particle size range; the passing rate trend curve is used to characterize the changing trend of the sieve aperture passing rate with the construction mileage, and the spatial distribution curve set includes multiple spatial distribution curves divided according to the preset construction mileage segment, and the spatial distribution curve is used to characterize the changing trend of the local spatial proportion with the physical coordinates of the road surface.

[0121] After obtaining historical data on abnormal particle size ranges, the historical sieve aperture passing rate data and historical spatial distribution data are transformed into visualized curves to more intuitively analyze the development process of the anomalies. See also... Figure 6 and Figure 7 , Figure 6 This is a schematic diagram of a pass rate trend curve provided for an embodiment of this application. Figure 7 This is a schematic diagram of the spatial distribution curve. As shown in the figure, the throughput trend curve is formed by connecting historical sieve throughput data in sequence according to construction mileage. The horizontal axis represents the construction mileage, and the vertical axis represents the throughput of the abnormal particle size range. The GIA curve can show the trend of usage changes for this particle size. For example, if the curve rises continuously from mileage 300 meters, and the throughput increases from 20% to 24%, it indicates that the usage of aggregate of this particle size gradually increases in the later stage of construction, which may be due to uneven material feeding from the silo or insufficient mixing time. If the curve fluctuates sharply, it indicates that the usage of aggregate of this particle size is unstable, which may be due to large fluctuations in gradation when the aggregate arrives at the site. The set of spatial distribution curves consists of multiple curves divided according to preset construction mileage segments. Each curve corresponds to a mileage segment. The horizontal axis is the physical coordinate of the road surface (e.g., horizontal from left to right, vertical from the start to the end of the segment), and the vertical axis is the local spatial proportion of this abnormal particle size. For example, in the 300-350 meter gradation range, the proportion of certain grids is much higher than the average, while adjacent grids are much lower than the average, indicating that the distribution of this particle size is extremely uneven within this range, with localized aggregation anomalies. In the 200-250 meter gradation range, the curve is flat, indicating a more uniform distribution. Thus, these curve sets can intuitively reflect the spatial distribution differences of this particle size in different gradation ranges, thereby helping to quickly identify areas of gradation anomalies.

[0122] S203: Determine the gradation anomaly region based on the passing rate trend curve and the set of spatial distribution curves for the abnormal particle size range.

[0123] Finally, based on the obtained pass rate trend curve and spatial distribution curve set, the pass rate trend curve is used to pinpoint the mileage range where the anomaly occurred, and the spatial distribution curve is used to locate the specific position of the anomaly, thereby quickly determining the gradation anomaly area. Next, the process of determining the gradation anomaly area based on the pass rate trend curve and spatial distribution curve set in step S203 will be described in conjunction with the accompanying drawings of a specific process embodiment.

[0124] See Figure 8 The figure is a flowchart illustrating another method for determining gradation anomaly regions provided in an embodiment of this application, specifically including the following steps:

[0125] S2031: Identify the abnormal data segment within the pass rate trend curve; the pass rate change rate of the abnormal data segment is greater than the warning threshold.

[0126] First, identifying abnormal data segments is the starting point for location analysis. The throughput trend curve records the dynamic changes in the sieve throughput of each particle size range with the construction mileage. By calculating the throughput change rate of adjacent mileage segments (e.g., if the throughput rate of a certain mileage segment increases from 20% to 25%, with a mileage difference of 50 meters, the change rate is 0.1% / meter), mileage segments with change rates exceeding the warning threshold can be screened out. For example, in the throughput trend curve of 10-15mm coarse aggregate for a certain AC type aggregate, the change rate of the 300-350 meter mileage segment reaches 0.25% / meter. With a warning threshold set at 0.15% / meter, this indicates a rapid increase in the amount of coarse aggregate used in this segment, possibly due to blockage at the hopper discharge port or uneven mixing. This mileage segment is then marked as an abnormal data segment, thus providing a mileage range for subsequent spatial location analysis.

[0127] S2032: Based on the set of spatial distribution curves, determine the abnormal spatial distribution curves that are in the same construction mileage segment as the abnormal data segment.

[0128] Next, it is necessary to find the abnormal spatial distribution curves corresponding to the abnormal data segments. The set of spatial distribution curves is divided according to preset mileage segments. Each curve reflects the local spatial proportion of abnormal particle size within that mileage segment as a function of the road surface physical coordinates (lateral / longitudinal). For example, the spatial distribution curve corresponding to the abnormal data segment of 300-350 meters above shows that in the region 1-3 meters to the left laterally, the local proportion of coarse aggregate reaches 30% (the average for this segment is set at 22%), while the adjacent region to the right is only 15%, with fluctuations far exceeding the spatial distribution uniformity threshold. This curve characteristic of local aggregation indicates that the spatial distribution of coarse aggregate within this segment is extremely uneven, therefore this curve is identified as an abnormal spatial distribution curve. In this step, the anomaly is narrowed down from the mileage range to a specific spatial location, for example, the abnormal spatial distribution curve is determined to be: 300-350 meter mileage segment, 1-3 meters to the left laterally, forming the initial spatial coordinate range.

[0129] S2033: Based on the physical coordinates of the road surface covered by the abnormal spatial distribution curve, determine the initial gradation abnormal area and perform drift trend analysis on the abnormal data segment.

[0130] After obtaining the abnormal data segments and the abnormal spatial distribution curves, the initial gradation anomaly area can be determined by the range of road surface physical coordinates covered by the abnormal spatial distribution curves. However, the initial gradation anomaly area often fails to accurately cover the anomaly areas in the actual scenario. Therefore, it is necessary to perform drift trend analysis on the abnormal data segments to determine whether the anomaly area is spreading. In the specific drift trend judgment process, it is necessary to determine whether the anomaly occurs in a continuous mileage segment. For example, if the rate of change in the throughput of the 350-400 meter segment is still 0.2% / meter (exceeding the warning threshold), it indicates that the anomaly is continuously drifting along the construction direction, that is, the abnormal coarse aggregate usage is extending to subsequent mileage segments; if the rate of change in the 350-400 meter segment recovers to 0.1% / meter (below the warning threshold), it indicates that the anomaly has not drifted and is only concentrated in the 300-350 meter segment. The purpose of drift trend analysis is to determine whether the anomaly is spreading, providing a basis for subsequent area adjustments.

[0131] S2034: If it is determined that the abnormal data segment has a continuous drift trend, the diffusion trend of the abnormal area is determined according to the distribution direction of the grid in the initial gradation abnormal area;

[0132] S2035: Based on the working condition data of the paving equipment and the diffusion trend of the abnormal area, the initial gradation abnormal area is adjusted to obtain the gradation abnormal area;

[0133] S2036: If it is determined that the abnormal data segment does not exhibit the continuous drift trend, the initial gradation abnormal region is determined as the gradation abnormal region.

[0134] For anomalous data segments exhibiting a continuous drift trend, it is necessary to further determine the diffusion trend of the anomalous area. The diffusion trend is determined based on the distribution direction of the grid within the initial anomalous area. For example, if the initial area is on the left side laterally and the drift direction is longitudinal, then the diffusion trend is longitudinal extension superimposed on the left side laterally. Subsequently, by combining the working data of the paving equipment, such as the paver's travel speed and the auger distributor's rotation speed, the initial area is adjusted for diffusion. For example, if the paver's travel speed is 4 meters per minute, and the anomalous data segment drifts from 300 meters to 400 meters, then the diffusion process of the initial gradation anomaly area is: "300-350 meters + 1-3 meters on the left" will diffuse 50 meters into the subsequent mileage, forming a gradation anomaly area of ​​"300-450 meters + 1-3 meters on the left". For anomalous data segments that do not exhibit a continuous drift trend, it indicates that the anomaly is only concentrated in the initial area and has not diffused; therefore, the initial area is directly determined as the final gradation anomaly area.

[0135] The above is a flowchart of the asphalt pavement construction gradation monitoring method in the embodiments of this application. Next, the process of dividing aggregate particles according to aggregate type in step S102 above will be described in conjunction with the accompanying drawings of the specific flowchart embodiments.

[0136] In this embodiment, the aggregate types specifically include AC type aggregate and SMA type aggregate. The particle segmentation processing method for AC type aggregate will be described first. (See...) Figure 9 The figure is a schematic diagram of a process for particle splitting of AC type aggregates according to an embodiment of this application, specifically including the following steps:

[0137] S1021: When the aggregate type is AC type aggregate, perform white balance correction on the asphalt paving layer image to obtain a first pre-processed asphalt paving layer image; the first pre-processed asphalt paving layer image is used to enhance the contrast between aggregate particles and asphalt mastic.

[0138] In asphalt pavement construction, AC-type aggregate is a common type of aggregate due to its uniform particle size distribution and the interweaving of coarse and fine aggregates. However, although the grayscale contrast between this type of aggregate and asphalt mortar is moderate, uneven lighting and dust adhesion in the construction environment can easily cause color temperature deviations in the asphalt paving layer image, resulting in blurred boundaries between the aggregate and the mortar. To address this issue, before performing aggregate particle segmentation processing on the asphalt paving layer image of AC-type aggregate, white balance correction is necessary. This correction compensates for color deviations caused by ambient light, restoring the highlight areas of the aggregate particles in the image to true white, thereby enhancing the grayscale contrast between the aggregate particles and the asphalt mortar.

[0139] S1022: Based on the grayscale bimodal features of the first preprocessed asphalt paving layer image, segmentation parameters are analyzed to determine the double-layer segmentation threshold; the double-layer segmentation threshold is used to distinguish the aggregate particles from the asphalt mortar and to classify aggregate particles of different particle size ranges.

[0140] Another characteristic of AC-type aggregates is their bimodal grayscale distribution. The coarse aggregate, due to its larger particle size and rougher surface, reflects more light and has the highest grayscale value. The fine aggregate, with its smaller particle size and tighter bond with the asphalt mortar, has the next highest grayscale value. The asphalt mortar, due to its greater light absorption, has the lowest grayscale value, forming a trough between the two peaks. Therefore, this step utilizes this characteristic to determine the two-layer segmentation thresholds for aggregate segmentation processing: a first grayscale threshold and a second grayscale threshold. The first grayscale threshold, defined as the grayscale range above the first peak, distinguishes coarse aggregate from other components. The second grayscale threshold, located between the second peak and the trough, distinguishes fine aggregate from asphalt mortar. For example, the grayscale histogram of the AC-type aggregate asphalt paving layer image shows that the grayscale peak of coarse aggregate is at 190, that of fine aggregate is at 130, and that of asphalt mortar is at 80. At this point, the first threshold can be set to 170, which is higher than the peak value of fine aggregates and lower than the lower limit of the peak value of coarse aggregates, ensuring that only coarse aggregates with an ash value > 170 are extracted; the second threshold is set to 100, which is higher than the trough of the binder and lower than the lower limit of the peak value of fine aggregates, ensuring that fine aggregates with an ash value between 100 and 170 are extracted. This effectively matches the continuous gradation and interwoven particle structure characteristics of AC-type aggregates, avoiding the problem of misclassifying fine aggregates and binder as the same type in traditional single-layer threshold segmentation, thus ensuring the accuracy of aggregate particle segmentation.

[0141] S1023: The first preprocessed asphalt paving layer image is processed by aggregate particle segmentation using the double-layer segmentation threshold to obtain an aggregate particle outline map.

[0142] After determining the two-layer segmentation thresholds used for aggregate particle segmentation, the aggregate particles can be segmented from the preprocessed image. This process of segmenting aggregate particles based on the first and second grayscale thresholds is implemented through the following four steps:

[0143] Step 1: Using the first grayscale threshold, extract aggregate particles with grayscale values ​​greater than the first grayscale threshold from the first preprocessed asphalt paving layer image to obtain the first coarse aggregate region.

[0144] Step 2: Remove the first coarse aggregate area from the first pre-processed asphalt paving layer image to obtain the second pre-processed asphalt paving layer image.

[0145] In the aggregate segmentation process based on a first grayscale threshold and a second grayscale threshold, coarse aggregate is first extracted from the white-balance corrected image using the first grayscale threshold. This type of coarse aggregate has large particle size, rough surface, and reflects a lot of light, resulting in a significantly higher grayscale value than fine aggregate and binder. Therefore, areas with grayscale values ​​greater than the first grayscale threshold can be identified as the first coarse aggregate region. Subsequently, this region is removed from the image to obtain the second pre-processed asphalt paving layer image.

[0146] Step 3: Using the second grayscale threshold, extract aggregate particles in the second preprocessed asphalt paving layer image whose grayscale value is less than the first grayscale threshold and greater than the second grayscale threshold to obtain the first fine aggregate region;

[0147] Step 4: Merge and optimize the first coarse aggregate region and the first fine aggregate region to obtain the aggregate particle outline diagram.

[0148] Subsequently, the second pre-processed asphalt paving layer image is processed using a second grayscale threshold. At this point, the grayscale value of the fine aggregate lies between that of the binder and the coarse aggregate, forming a transition layer connecting the coarse aggregate and the binder. Therefore, the area with grayscale values ​​between the first and second grayscale thresholds can be marked as the first fine aggregate area. These particles have small particle sizes and are tightly bound to the binder, but the boundary between the fine aggregate and the binder is distinguished by the second threshold. After extraction, the first coarse aggregate area and the first fine aggregate area are merged. In one possible implementation, a series of noises present in the image can be optimized. For example, noise points with an area less than 0.5 mm² can be removed, or broken particle edges can be connected through morphological operations to make the contours more complete. The final aggregate particle contour map clearly shows the boundary of each aggregate particle. The coarse aggregate is distributed in blocks, the fine aggregate fills the spaces between the coarse aggregates, and the binder serves as the background, effectively restoring the continuous gradation and interwoven spatial structure of the AC type aggregate. The aggregate particle outline map is a key foundation for subsequent extraction of morphological parameters and generation of aggregate distribution images, providing a visual and quantitative basis for gradation monitoring.

[0149] S1024: Based on the aggregate particle outline map, extract the morphological parameters of each aggregate particle in the target area, and label the aggregate particles belonging to different particle size range categories to obtain the aggregate distribution image.

[0150] Finally, based on the aggregate particle profile map, the morphological parameters of each aggregate particle within the area covered by the profile map are extracted. These morphological parameters include equivalent particle size, projected area, aspect ratio, and roundness. Subsequently, according to the morphological parameters and a preset particle size range screening rule, each aggregate particle is labeled with its particle size range. For example, coarse aggregate is labeled in red, fine aggregate in blue, and medium aggregate in green. The labeling results are then superimposed on the original image to obtain an aggregate distribution image. This aggregate distribution image can intuitively display the spatial distribution of aggregates of different particle sizes within the target area. For example, red coarse aggregate is evenly distributed in the image, blue fine aggregate fills the spaces between coarse aggregate, and green medium aggregate acts as a transition layer connecting the two, thus reflecting the spatial distribution of aggregate particles within different particle size ranges.

[0151] The following section will introduce the particle size distribution method for SMA type aggregates. (See also...) Figure 10The figure is a schematic diagram of a process for particle splitting SMA type aggregates according to an embodiment of this application, specifically including the following steps:

[0152] S1025: When the aggregate type is the SMA type aggregate, the asphalt paving layer image is subjected to histogram equalization processing to obtain a third pre-processed asphalt paving layer image; the third pre-processed asphalt paving layer image is used to remove non-aggregate areas in the asphalt paving layer image.

[0153] SMA-type aggregates, as a typical example of discontinuously graded asphalt mixtures, have a high content of coarse aggregates in their asphalt paving layer images. These coarse aggregates are easily interlocked, forming continuous high-grayscale areas, while mastic slurry fills the gaps between the coarse aggregates with a dark gray color. However, non-aggregate areas in the construction scene (such as the shadows left by the paver body and residual markings) can severely interfere with aggregate segmentation. For example, the dark gray of the shadows can be easily confused with mastic slurry, and the highly reflective high-grayscale areas may be misjudged as coarse aggregates. These interferences can cause subsequent segmentation results to deviate from the true aggregate structure. Therefore, for SMA-type aggregates, histogram equalization processing of the asphalt paving layer image is necessary. By stretching the grayscale distribution range of the image, the grayscale values ​​originally concentrated in the middle range are dispersed to the full range of 0-255, enhancing the contrast between different areas. For example, in the original image, the grayscale value of mastic slurry is 80-100, while the grayscale value of the shadow is 70-90, with significant overlap between the two. After histogram equalization, the grayscale value of mastic slurry remains at 80-100, while the grayscale value of the shadow is stretched to 50-70, forming a clear boundary. At this point, by setting a grayscale threshold (e.g., <70), non-aggregate areas such as shadows and markings can be removed, resulting in the third pre-processed asphalt paving layer image. This image retains only the areas of coarse aggregate, fine aggregate, and mastic slurry, completely eliminating interfering factors and laying the foundation for accurate segmentation later.

[0154] S1026: The third preprocessed asphalt paving layer image is divided into regions using a marker-controlled watershed algorithm to obtain a fourth preprocessed asphalt paving layer image. The fourth preprocessed asphalt paving layer image includes a second coarse aggregate region, a second fine aggregate region, and a mastic slurry region. The gray value of the second coarse aggregate region is greater than a third gray value threshold, the gray value of the second fine aggregate region is between the fourth gray value threshold and the third gray value threshold, and the gray value of the mastic slurry region is less than the fourth gray value threshold. The third gray value threshold is greater than the fourth gray value threshold.

[0155] To address the issue of SMA-type aggregates easily agglomerating, this embodiment employs a marker-controlled watershed algorithm for region segmentation. The basic principle of the watershed algorithm is to treat the image as a terrain surface, with areas of low grayscale values ​​representing valleys (corresponding to mastic mortar) and areas of high grayscale values ​​representing peaks (corresponding to coarse aggregates), segmenting the regions through the water injection process. However, the ordinary watershed algorithm is prone to oversegmentation due to grayscale changes in agglomerated particles, failing to accurately reflect the skeletal structure of SMA-type aggregates. Therefore, a marker-controlled optimization algorithm is needed for region segmentation.

[0156] First, based on the grayscale characteristics of SMA-type aggregates, a third and fourth grayscale threshold are set, with the third threshold being greater than the fourth. Regions in the third preprocessed image with grayscale values ​​greater than the third threshold are marked as foreground seed points (corresponding to the high grayscale regions of coarse aggregates, forming the core skeleton of SMA-type aggregates), while regions with grayscale values ​​less than the fourth threshold are marked as background seed points (corresponding to the low grayscale regions of mastic mortar). The intermediate grayscale range is the transition region (corresponding to fine aggregates, connecting the skeleton and the filler). Then, watershed segmentation is performed starting from these seed points. The foreground seed points gradually expand, dividing the high grayscale coarse aggregate group into independent peaks (corresponding to the second coarse aggregate region). The background seed points gradually expand, dividing the low grayscale mastic mortar into valleys (corresponding to the mastic mortar region). The transition region naturally forms a slope (corresponding to the second fine aggregate region). Finally, the fourth preprocessed asphalt paving layer image is obtained, clearly defining the three regions. This division method effectively matches the structural characteristics of SMA-type aggregates, which consist of coarse aggregates as the skeleton and fine aggregates and mastic as fillers. It provides an accurate regional basis for the subsequent extraction of the skeleton aggregate area and ensures the authenticity of the morphological parameter extraction.

[0157] Specifically, the purpose of setting a fourth grayscale threshold is to accurately separate the mastic slurry area from the fine aggregate area. The fourth grayscale threshold is determined based on the statistical characteristics of the low grayscale segment of the grayscale histogram of the SMA-type aggregate paving layer image. Due to the strong light absorption of the asphalt binder, the mastic slurry forms an independent low grayscale secondary peak in the 40-80 grayscale range of the histogram, while the fine aggregate, being partially exposed on the surface, has a grayscale response distributed in the 80-130 range. There is a clear valley boundary between the two between the 70-90 grayscale values. Based on this, the grayscale value corresponding to this valley is used as the base value of the fourth grayscale threshold, ensuring that more than 90% of the area of ​​the low grayscale secondary peak is classified into the mastic slurry area, while avoiding the accidental removal of fine aggregate particles. Further feedback verification is performed using the designed asphalt-aggregate ratio of the SMA mixture (typically 5.5%-6.5%): the percentage of pixel area below the fourth threshold is statistically analyzed and converted into the mortar area coverage. If this value deviates from the theoretical mortar area calculated from the designed asphalt-aggregate ratio by more than a set value, the threshold is finely adjusted in a gradient until the deviation converges to within the allowable range. Furthermore, considering the differences in near-infrared light reflectivity of different lithological aggregates, the fourth grayscale threshold can also employ a normalized dynamic bias strategy based on the histogram peak position, which will not be elaborated upon here.

[0158] On the other hand, the purpose of setting the third grayscale threshold is to completely extract the coarse aggregate skeleton particles and suppress excessive segmentation of the weakly reflective edge areas. SMA-type aggregates have a high coarse aggregate content, and their surface grayscale response forms a distinct high grayscale peak in the 140-200 range of the histogram. Fine aggregates, due to their smaller particle size and partial coating by the binder, have a grayscale response concentrated in the 100-140 range. A transition valley is formed between the two peaks at a grayscale value of 130-150. The grayscale value corresponding to the bottom of this valley is the theoretical basis for the third threshold, ensuring that more than 85% of the area of ​​the coarse aggregate main peak is completely segmented. After determining the initial threshold, further constraints are imposed based on the design requirements for the aggregate gap ratio of the SMA mixture. Areas exceeding the third threshold are extracted, and their skeleton gap ratio is calculated. This value is compared with the target value (usually ≥17%). If the deviation exceeds the set range, a threshold sliding search is performed along both sides of the valley bottom until the skeleton gap ratio meets the design requirements.

[0159] S1027: Extract aggregate particles in the second coarse aggregate region whose particle edge spacing is not greater than the preset aggregate adhesion judgment threshold, and remove aggregate particles in the second coarse aggregate region whose particle edge spacing is greater than the preset aggregate adhesion judgment threshold to obtain the skeleton aggregate region.

[0160] As discussed earlier regarding SMA-type aggregates, their advantage lies in the skeleton structure formed by the interlocking of coarse aggregates. However, in asphalt paving layer images, coarse aggregates often adhere due to construction vibration or the shape of the aggregates themselves. Some coarse aggregates, although located in the second coarse aggregate zone, do not form effective interlocking with other coarse aggregates. These particles cannot participate in skeleton formation, and if not removed, it will lead to an artificially high skeleton ratio, affecting the accuracy of gradation monitoring.

[0161] Therefore, the purpose of this step is to screen out the coarse aggregates that truly constitute the skeleton from the second coarse aggregate region. The contour edge of each coarse aggregate is extracted using an edge detection algorithm, the minimum distance between the edges of adjacent particles is calculated, and compared with a preset aggregate adhesion judgment threshold. If the edge distance is not greater than the threshold, it indicates that two coarse aggregates are in contact or interlocked, and are part of the skeleton, so they are retained. If the edge distance is greater than the threshold, it indicates that the particles are isolated and do not participate in skeleton formation, so they are discarded. For example, in a certain second coarse aggregate region, there is an adhesion group consisting of 5 coarse aggregates. The edge distance of 4 of these particles is less than 2mm, forming a chain of interlocking particles, while the edge distance of the 5th particle is 2.5mm from the other particles, not forming effective contact. Therefore, the 5th particle is discarded, and the remaining 4 particles are retained as the skeleton aggregate region.

[0162] S1028: Based on the skeleton aggregate region and the second fine aggregate region, extract the morphological parameters of each aggregate particle in the target region, and label the skeleton aggregate region, the second fine aggregate region and the mastic mortar region respectively to obtain the aggregate distribution image.

[0163] Finally, the visual contours of the skeleton aggregate region are transformed into quantitative parameters and a visual image. First, based on the skeleton aggregate region and the second fine aggregate region, the morphological parameters of each aggregate particle are extracted. For skeleton aggregate, the focus is on the number of contact points (i.e., the number of times each skeleton particle contacts other skeleton particles), the degree of sharpness, and the skeleton gap ratio. For the second fine aggregate, the equivalent particle size and particle roundness are mainly extracted. For example, if a skeleton particle has 3 contact points (meeting the requirement of ≥2), a sharpness of 0.85 (obvious sharpness), and a skeleton gap ratio of 18% (meeting the design requirement of 15%-20%), it indicates good interlocking effect. Subsequently, differential annotation is applied to the skeleton aggregate region, the second fine aggregate region, and the mastic mortar region to generate an aggregate distribution image.

[0164] This application provides an asphalt pavement construction gradation alarm system and method. The method first analyzes the gradation features of real-time acquired asphalt paving layer images, extracting aggregate gradation feature data for the target area. Then, based on a preset construction gradation prediction model, it analyzes the gradation risk situation of the current paving layer, instantly outputting the sieve passing rate and local spatial proportion of each particle size interval. This achieves rapid quantitative assessment of the gradation status, allowing gradation deviations to be captured promptly without waiting for post-construction testing results. Furthermore, it analyzes deviation features based on passing rate deviation characteristics and spatial distribution dispersion, calculating particle size deviation scores. Finally, based on the particle size deviation score or the local spatial proportion of each particle size interval, a gradation anomaly alarm is triggered, achieving real-time proactive perception and immediate feedback of abnormal states. Therefore, gradation deviation information can be obtained immediately during the paving process, allowing for timely adjustments and effectively improving the timeliness and accuracy of asphalt pavement construction quality control.

[0165] The following describes an asphalt pavement construction gradation alarm system provided in the embodiments of this application. The asphalt pavement construction gradation alarm system described below can be referred to in correspondence with the asphalt pavement construction gradation alarm method described above.

[0166] See Figure 11 The figure is a structural schematic diagram of an asphalt pavement construction gradation alarm system provided in an embodiment of this application, which specifically includes the following modules:

[0167] The feature analysis module 100 is used to perform gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area;

[0168] The risk analysis module 200 is used to perform a gradation risk situation analysis on the current paving layer based on a preset construction gradation prediction model and the aggregate gradation characteristic data, and obtain gradation risk characteristic data; the gradation risk characteristic data includes: the sieve passing rate of each particle size range and the local space ratio of each particle size range.

[0169] The feature processing module 300 is used to perform deviation feature analysis based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval to obtain the gradation deviation feature vector corresponding to each particle size interval.

[0170] The particle size analysis module 400 is used to calculate the particle size deviation based on the gradation deviation feature vector of each particle size interval, and obtain the particle size deviation score of each particle size interval.

[0171] The alarm processing module 500 is used to issue an alarm for gradation abnormalities during the construction process of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval.

[0172] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for methods and systems, since they are basically similar to the method embodiments, the description is relatively simple, and relevant parts can be referred to the description of the method embodiments. The methods and systems described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components indicated as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the solution in this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0173] The above description is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for alarming the gradation of asphalt pavement construction, characterized in that, include: Gradation feature analysis is performed on the target area corresponding to the asphalt paving layer image to obtain the aggregate gradation feature data of the target area; Based on the preset construction gradation prediction model, the gradation risk situation of the current paving layer is analyzed according to the aggregate gradation characteristic data to obtain gradation risk characteristic data; The gradation risk characteristic data includes: the sieve aperture passing rate of each particle size range and the local space ratio of each particle size range; Based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval, deviation characteristic analysis is performed to obtain the gradation deviation characteristic vector corresponding to each particle size interval. The particle size deviation is calculated based on the gradation deviation feature vector of each particle size interval to obtain the particle size deviation score of each particle size interval. An alarm is triggered for gradation abnormalities during the construction of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval.

2. The method according to claim 1, characterized in that, The method of alarming for gradation abnormalities during the construction of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval includes: If the particle size deviation score is less than the score threshold, the gradation status of the current paving layer is determined to be abnormal, and a gradation abnormality alarm signal is generated for alarm purposes. If the particle size deviation score is not less than the score threshold, an alarm is triggered for gradation abnormalities during the construction of the target asphalt pavement based on the local spatial proportion of each particle size interval.

3. The method according to claim 2, characterized in that, The method of alarming for gradation abnormalities during the construction of the target asphalt pavement based on the local spatial proportion of each of the aforementioned particle size ranges includes: Based on the local spatial proportion of each particle size range, calculate the coefficient of variation of the proportion of each particle size range, and the Pearson correlation coefficient between any two particle size ranges; the coefficient of variation of the proportion is used to characterize the degree of dispersion of aggregate particles in the target area in the corresponding space of the particle size range. If the coefficient of variation of the proportion of any of the particle size ranges is greater than the first coefficient threshold, or if any of the Pearson correlation coefficients is less than the second threshold, the gradation state of the current paving layer is determined to be abnormal, and the gradation abnormality alarm signal is generated for alarm purposes.

4. The method according to claim 2 or 3, characterized in that, After determining that the gradation state of the current paving layer is abnormal, the method further includes: If the current gradation state of the paved layer is determined to be abnormal, historical sieve aperture passing rate data and historical spatial distribution data of the abnormal particle size range in the current aggregate paving process are obtained. Based on the historical sieve aperture passing rate data and the historical spatial distribution data, a passing rate trend curve and a set of spatial distribution curves for the abnormal particle size range are determined; the passing rate trend curve is used to characterize the changing trend of the sieve aperture passing rate with the construction mileage, and the set of spatial distribution curves includes multiple spatial distribution curves divided according to a preset construction mileage segment; the spatial distribution curves are used to characterize the changing trend of the local spatial proportion with the physical coordinates of the road surface. The gradation anomaly region is determined based on the pass rate trend curve and the set of spatial distribution curves for the abnormal particle size range.

5. The method according to claim 4, characterized in that, The step of determining the gradation anomaly region based on the passing rate trend curve and the set of spatial distribution curves for the abnormal particle size range includes: Identify the abnormal data segment within the pass rate trend curve; the pass rate change rate of the abnormal data segment is greater than the warning threshold. Based on the set of spatial distribution curves, determine the abnormal spatial distribution curves that are in the same construction mileage segment as the abnormal data segment; Based on the road surface physical coordinates covered by the abnormal spatial distribution curve, the initial gradation abnormal area is determined, and the drift trend analysis is performed on the abnormal data segment. If it is determined that the abnormal data segment has a continuous drift trend, the diffusion trend of the abnormal area is determined according to the distribution direction of the grid in the initial gradation abnormal area. Based on the working condition data of the paving equipment and the diffusion trend of the abnormal area, the initial gradation abnormal area is adjusted for regional diffusion to obtain the gradation abnormal area. If it is determined that the abnormal data segment does not exhibit the continuous drift trend, the initial gradation abnormal region is identified as the gradation abnormal region.

6. The method according to claim 1, characterized in that, The step of performing gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area includes: The asphalt paving layer image is segmented according to the aggregate type to determine the morphological parameters of each aggregate particle in the target area, as well as the aggregate distribution image of the target area; the aggregate distribution image is used to characterize the spatial distribution of aggregate particles of different sizes in the target area. Based on the morphological parameters of each aggregate particle, the aggregate particles are screened and statistically analyzed according to a preset particle size range screening rule to obtain the particle size range category covered by the target area. The particle size range categories covered by the target area, the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area are determined as the aggregate gradation characteristic data.

7. The method according to claim 6, characterized in that, The aggregate type includes AC type aggregate; The step of segmenting the asphalt paving layer image according to aggregate type to determine the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area, includes: When the aggregate type is AC type aggregate, the asphalt paving layer image is white-balance corrected to obtain a first pre-processed asphalt paving layer image; the first pre-processed asphalt paving layer image is used to enhance the contrast between aggregate particles and asphalt mastic. Based on the bimodal grayscale features of the first preprocessed asphalt paving layer image, segmentation parameters are analyzed to determine the double-layer segmentation threshold; the double-layer segmentation threshold is used to distinguish the aggregate particles from the asphalt binder and to classify aggregate particles of different particle size ranges. The first preprocessed asphalt paving layer image is segmented into aggregate particles using the double-layer segmentation threshold to obtain an aggregate particle outline map. Based on the aggregate particle outline map, the morphological parameters of each aggregate particle in the target area are extracted, and aggregate particles belonging to different particle size ranges are labeled to obtain the aggregate distribution image.

8. The method according to claim 7, characterized in that, The dual-layer segmentation threshold includes a first grayscale threshold and a second grayscale threshold, wherein the first grayscale threshold is greater than the second grayscale threshold. The step of performing aggregate particle segmentation processing on the first preprocessed asphalt paving layer image using the dual-layer segmentation threshold to obtain an aggregate particle outline map includes: Using the first grayscale threshold, aggregate particles with grayscale values ​​greater than the first grayscale threshold in the first preprocessed asphalt paving layer image are extracted to obtain the first coarse aggregate region. The first coarse aggregate area is removed from the first pre-processed asphalt paving layer image to obtain the second pre-processed asphalt paving layer image. By using the second grayscale threshold, aggregate particles with grayscale values ​​less than the first grayscale threshold and greater than the second grayscale threshold in the second preprocessed asphalt paving layer image are extracted to obtain the first fine aggregate region. The first coarse aggregate region and the first fine aggregate region are merged and optimized to obtain the aggregate particle outline diagram.

9. The method according to claim 6, characterized in that, The aggregate type includes SMA type aggregate; the step of performing aggregate particle segmentation processing on the asphalt paving layer image according to the aggregate type, determining the morphological parameters of each aggregate particle in the target area, and the aggregate distribution image of the target area, includes: When the aggregate type is SMA type aggregate, the asphalt paving layer image is subjected to histogram equalization processing to obtain a third pre-processed asphalt paving layer image; the third pre-processed asphalt paving layer image is used to remove non-aggregate areas in the asphalt paving layer image. The third preprocessed asphalt paving layer image is divided into regions using a marker-controlled watershed algorithm to obtain a fourth preprocessed asphalt paving layer image. The fourth preprocessed asphalt paving layer image includes a second coarse aggregate region, a second fine aggregate region, and a mastic slurry region. The grayscale value of the second coarse aggregate region is greater than a third grayscale threshold, the grayscale value of the second fine aggregate region is between a fourth grayscale threshold and the third grayscale threshold, and the grayscale value of the mastic slurry region is less than the fourth grayscale threshold. The third grayscale threshold is greater than the fourth grayscale threshold. Extract aggregate particles whose particle edge spacing is not greater than the preset aggregate adhesion judgment threshold in the second coarse aggregate region, and remove aggregate particles whose particle edge spacing is greater than the preset aggregate adhesion judgment threshold in the second coarse aggregate region to obtain the skeleton aggregate region. Based on the skeleton aggregate region and the second fine aggregate region, the morphological parameters of each aggregate particle in the target region are extracted, and the skeleton aggregate region, the second fine aggregate region and the mastic paste region are labeled respectively to obtain the aggregate distribution image.

10. An alarm system for asphalt pavement construction gradation, characterized in that, The system is used to implement the asphalt pavement construction gradation alarm method as described in any one of claims 1-9, the system comprising: The feature analysis module is used to perform gradation feature analysis on the target area corresponding to the asphalt paving layer image to obtain aggregate gradation feature data of the target area; The risk analysis module is used to perform a gradation risk situation analysis on the current paving layer based on a preset construction gradation prediction model and the aggregate gradation characteristic data, and obtain gradation risk characteristic data; the gradation risk characteristic data includes: the sieve passing rate of each particle size range and the local space ratio of each particle size range. The feature processing module is used to perform deviation feature analysis based on the throughput deviation characteristics and spatial distribution dispersion of each particle size interval to obtain the gradation deviation feature vector corresponding to each particle size interval. The particle size analysis module is used to calculate the particle size deviation based on the gradation deviation feature vector of each particle size interval, and obtain the particle size deviation score of each particle size interval. The alarm processing module is used to issue an alarm for gradation abnormalities during the construction process of the target asphalt pavement based on the particle size deviation score or the local spatial proportion of each particle size interval.

Images