Asphalt temperature segregation detection method, system and device

Abstract

The invention belongs to the technical field of asphalt paving segregation detection, and particularly relates to an asphalt temperature segregation detection method, system and device. The method comprises the following steps: S1, acquiring an asphalt infrared image; s2, performing super-pixel processing on the asphalt infrared image by adopting a multi-scale morphological watershed algorithm to obtain a super-pixel image; s3, performing clustering processing on the super-pixel image by adopting a K-Means clustering algorithm to obtain a plurality of segmented regions; s4, processing each segmented region by adopting an FCM clustering algorithm to generate a fuzzy label, and merging the fuzzy label into the superpixel image to serve as a segmentation result; and S5, extracting temperatures of different points in the subareas, and judging whether segregation exists in the corresponding subareas or not according to the temperatures. According to the invention, the positioning precision can be improved, so that the asphalt temperature detection requirement in the paving process is met.

Description

Technical Field

[0001] The present invention belongs to the technical field of asphalt paving segregation detection, and in particular relates to an asphalt temperature segregation detection method, system and device. Background Art

[0002] Temperature segregation is one of the important reasons for the decline in pavement construction quality. Due to the complexity of its distribution and the inability to be identified by the naked eye, the detection and positioning of temperature segregation becomes increasingly important.

[0003] Traditional methods for detecting asphalt temperature segregation mainly include the following two methods: One is the mercury thermometer method, which has the advantages of simple structure and low cost. However, its measurement time is as long as tens of seconds, the response speed is slow, and the temperature reading is subject to human error. It is also dangerous and can easily cause harm to the human body. Therefore, it is usually used as an instrument for on-site monitoring. The other is the thermocouple digital thermometer method. The thermocouple digital thermometer mainly consists of two parts: a measuring terminal and a reference terminal. When there is a temperature difference between the two ends, a current is generated in the circuit. Since the magnitude of the current is related to the temperature difference between the two ends, the temperature of the measuring terminal can be determined based on the temperature difference. However, this method suffers from rapid material loss and the reference terminal temperature cannot be kept constant, which affects measurement accuracy.

[0004] With the development of infrared imaging technology, temperature measurement methods based on infrared thermal imagers have gradually been used in various non-destructive testing fields. Infrared thermal imaging technology, with its many advantages, such as non-contact temperature measurement, small instrument size, weak environmental impact, and fast response speed, has gradually been used in highway construction, mainly for temperature detection of hot-mix asphalt to detect whether there is temperature segregation during the asphalt paving process. However, in actual engineering use, because asphalt infrared images are affected by the complex texture of the asphalt pavement, their temperature distribution is very dispersed. Therefore, the commonly used asphalt temperature segregation detection method can only detect temperature segregation in small areas. In addition, because existing detection methods are usually based on data fitting, they require a large amount of data post-processing and cannot achieve real-time detection during the asphalt paving process, and thus cannot suppress the occurrence of asphalt temperature segregation at the source.

[0005] In summary, while infrared thermal imaging-based asphalt temperature segregation detection methods can address the "spot measurement" and "high risk" issues inherent in traditional asphalt temperature segregation detection methods, the results are typically localized and discontinuous. This prevents real-time measurement of pavement temperature during asphalt paving, nor can it accurately locate areas experiencing temperature segregation. Therefore, research on improving the refinement and online nature of asphalt temperature segregation detection methods is crucial for extending road life and ensuring driving safety. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method, system and device for detecting asphalt temperature separation, which can improve the positioning accuracy and thus meet the asphalt temperature detection requirements during the paving process.

[0007] The present invention is implemented as follows: The present invention provides an asphalt temperature segregation detection method, comprising the following steps: S1. Obtain infrared image of asphalt; S2. Using a multi-scale morphological watershed algorithm to perform super-pixel processing on the asphalt infrared image to obtain a super-pixel image; S3, clustering the superpixel image using a K-Means clustering algorithm to obtain multiple segmented regions; S4. Processing each segmented region using an FCM clustering algorithm to generate a fuzzy label, and merging the fuzzy label into the superpixel image as a segmentation result; S5. Extract the temperatures of different points in the partition, and determine whether segregation exists in the corresponding partition based on the temperatures.

[0008] Furthermore, the step S2 includes the following sub-steps: S201, constructing a multi-scale morphological gradient image based on the asphalt infrared image; S202, dividing the multi-scale morphological gradient image into different regions, and selecting a water injection point in each region; S203, assigning a unique label to each water injection point in the area and performing a simulated water injection operation to obtain an expanded area; S204: When there is a boundary between the extended areas corresponding to different areas, a super-pixel image is obtained.

[0009] Furthermore, the calculation formula for constructing the multi-scale morphological gradient image is: ; In the above formula, Represents the result of multi-scale morphological gradient operation, R f represents morphological operation, g represents asphalt infrared image, r represents morphological convolution kernel size, r∈(r1,r2); The calculation formula of the water injection point is: ; In the above formula, M represents the water injection point, I represents the morphological gradient image, p represents the current pixel point, and q represents the point within the set neighborhood N(p); The calculation formula of the extended area is: ; In the above formula, R k (h) represents the extended area, pmk Indicates p to m k The path, m k represents the minimum value of each area, indicating the water level; The boundary is calculated as follows: ; In the above formula, B represents the boundary, L(p) represents the pixel p belongs to the local minimum m k Starting area.

[0010] Furthermore, the S3 includes the following sub-steps: S301, calculating the average grayscale value of each superpixel in the superpixel image; S302: For each superpixel, calculate the position feature of its set center; S303, randomly selecting several initial centers from the features of superpixels as initial clustering centers; S304, for each superpixel in the superpixel image, calculating the distance between its feature and each cluster center, and assigning the superpixel to the cluster center closest to it based on the calculation result; S305, recalculate each cluster center by extracting the feature average of all superpixels belonging to the cluster to update. If the center reaches a preset range or reaches a preset number of iterations, go to S306, otherwise go to S304; S306 , aggregating the superpixel image of each cluster center into a region to obtain the segmented region.

[0011] Furthermore, the grayscale average value is calculated as follows: ; In the above formula, C avg represents the average value, c i represents the grayscale value of the i-th pixel, and N represents the number of superpixels in the entire image; The calculation formula of the position feature is: ; In the above formula, (x centroid ,y centroid ) represents the position feature, (x i ,y i ) represents the position of the i-th pixel in the superpixel; The distance is calculated as follows: ; In the above formula, d(j,k) represents the distance between the jth superpixel and the kth center, f j represents the feature of the jth superpixel, f k represents the characteristics of the k-th center, (xcj ,y cj ) represents the center position feature of the jth superpixel, (x ck ,y ck ) represents the central position feature of the k-th center point; The calculation formula of the cluster center is: ; In the above formula, μ k represents the cluster center, S k represents the set of all superpixels assigned to k cluster centers.

[0012] Furthermore, the S4 includes the following sub-steps: S401, setting clustering parameters, wherein the clustering parameters include the number of clusters and the fuzzy coefficient; S402: For each segmented region obtained by K-Means, several points are selected from the region as initial cluster centers for FCM clustering. S403, calculating the membership degree of each point in each segmented area to each cluster center; S404, update the centroid by maximizing the membership so that the centroid position has the best representation for the data points in the current cluster. If the membership center change is less than a preset threshold after iteration, go to S405, otherwise repeat this step; S405: The cluster labels with the best membership are combined and displayed in the target image as the segmentation result.

[0013] Furthermore, the calculation formula of the membership degree is: ; In the above formula, x i Represents the superpixel feature vector including color and space in each area, that is, the horizontal coordinate of the i-th pixel, v j represents the cluster center of the jth class, v k Indicates the k The cluster centers of each class, c represents the number of clusters, and m represents the fuzzy coefficient; The calculation formula of the cluster center is: ; In the above formula, N represents the number of super pixels in this category.

[0014] Furthermore, the step S5 includes the following sub-steps: S501, calculating the average temperature value of each superpixel center point in the segmented area, and taking it as the temperature representative value of the segmented area; S502, calculating the temperature difference between different segmented areas according to the temperature representative value; S503: Determine whether the temperature interpolation exceeds a temperature difference threshold for temperature segregation; if so, determine that temperature segregation occurs; otherwise, determine that temperature segregation does not occur.

[0015] Based on the same inventive concept, the present invention also provides an asphalt temperature segregation detection system, comprising: An acquisition module, used for acquiring infrared images of asphalt; A super-pixel processing module, configured to perform super-pixel processing on the asphalt infrared image using a multi-scale morphological watershed algorithm to obtain a super-pixel image; A segmentation module, configured to perform clustering processing on the superpixel image using a K-Means clustering algorithm to obtain a plurality of segmented regions; A merging module, configured to process each segmented region using an FCM clustering algorithm, generate fuzzy labels, and merge the fuzzy labels into the superpixel image as a segmentation result; The judgment module is used to extract the temperature of different points in the partition and judge whether the corresponding partition has segregation according to the temperature.

[0016] Based on the same inventive concept, the present invention also provides an asphalt temperature segregation detection device, comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the above method.

[0017] The beneficial effects brought by the present invention are: 1) This paper improves the watershed superpixel algorithm with the help of multi-scale morphology, which reduces the number of temperature regions in asphalt infrared images while ensuring the correctness of boundaries and improves the usability of images.

[0018] 2) This paper uses K-Means and FCM clustering methods to achieve superpixel segmentation. Starting from the pixel, while optimizing segmentation accuracy and efficiency, it abandons the reliance of existing image segmentation methods on large data sets, thereby improving the practical engineering application value of infrared image processing technology in temperature segregation detection.

[0019] 3) The present invention helps to realize online detection of asphalt temperature segregation, thereby helping to fundamentally reduce the probability of asphalt segregation events. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a demonstration image of infrared images of asphalt with temperature segregation detected using the method of the present invention; Figure 3 This is a demonstration diagram of infrared images of asphalt without temperature segregation detected by the method of the present invention; Figure 4 Schematic diagram of the detection results in the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0022] Example 1: Figure 1 As shown, this embodiment discloses a method for detecting asphalt temperature segregation, comprising the following steps:

[0023] S1. Obtain infrared images of asphalt.

[0024] S2. Use a multi-scale morphological watershed algorithm to perform super-pixel processing on the asphalt infrared image to obtain a super-pixel image. Specifically, this step includes the following sub-steps: S201. Construct a multi-scale morphological gradient image based on the asphalt infrared image. This step includes the following sub-steps: First, set the morphological convolution kernel to a circular shape of size r, where r∈(r1,r2). Based on cross-validation, r1=2 and r2=50. Then, calculate the morphological gradient image based on morphological operations and obtain the union of the morphological gradient images. The specific calculation formula is: ; In the above formula, Represents the result of multi-scale morphological gradient operation, R f represents morphological operation, and g represents asphalt infrared image.

[0025] S202: Divide the multi-scale morphological gradient image into different regions according to the watershed algorithm, and select the minimum value of each region as the water injection point M. The regional minimum value can be defined as the set of pixels in the morphological gradient image I that meet the following conditions: ; In the above formula, p represents the current pixel point, and q represents the point within the set neighborhood N(p).

[0026] S203, assign a unique label to each water filling point in each area and define a label image L. Then simulate water filling from the minimum value of the label, first set the maximum water level h max = 256, starting from the lowest grayscale value and gradually increasing. For each water level h, the expansion area R k The calculation formula for (h) is: ; In the above formula, pm k Indicates p to m k The path, m k Indicates the minimum value for each region.

[0027] S204: When there is a shared pixel p in the extended areas corresponding to different areas, it is recorded as a boundary B, and a super-pixel image is obtained. The calculation formula of the boundary B is: ; In the above formula, L(p) indicates that pixel p belongs to the local minimum m k Starting area.

[0028] S3. Clustering the superpixel image using the K-Means clustering algorithm to obtain multiple segmented regions. This step includes the following sub-steps: S301. Calculate the average grayscale value of each superpixel in the superpixel image. Since the image after superpixel is a grayscale image, it has only one average grayscale value, so the average value C avg The calculation formula is: ; In the above formula, c i represents the grayscale value of the i-th pixel, and N represents the number of superpixels in the entire image.

[0029] S302: For each superpixel, calculate the position feature of its set center. The calculation formula of the position feature is: ; In the above formula, (x centroid ,y centroid ) represents the position feature, (x i ,y i ) represents the position of the i-th pixel in the superpixel.

[0030] S303. Randomly select several initial centers from the superpixel features as initial cluster centers. In this example, since the main purpose is to determine whether the asphalt infrared image has temperature segregation, the final segmentation area is 2. Therefore, k = 2 initial centers are randomly selected. These centers are the superpixel features.

[0031] S304: For each superpixel in the superpixel image, calculate the distance between its feature and each cluster center, and assign the superpixel to the cluster center closest to it based on the calculation result. The distance calculation formula is: ; In the above formula, d(j,k) represents the distance between the jth superpixel and the kth center, f j represents the feature of the jth superpixel, f k represents the characteristics of the k-th center, (x cj ,y cj ) represents the center position feature of the jth superpixel, (xck ,y ck ) represents the central position feature of the kth center point.

[0032] S305, recalculate each cluster center by extracting the feature average of all superpixels belonging to the cluster to update. If the center reaches the preset range or reaches the preset number of iterations, go to S306, otherwise go to S304. The cluster center calculation formula is: ; In the above formula, μ k represents the cluster center, S k represents the set of all superpixels assigned to k cluster centers. In this example, the number of iterations is 200.

[0033] S306 , aggregating the superpixel image of each cluster center into one region to obtain two segmented regions.

[0034] S4. Process each segmented region using the FCM clustering algorithm to generate fuzzy labels, and merge the fuzzy labels into the superpixel image as the segmentation result. This step includes the following sub-steps: S401. Set clustering parameters, set the number of clustering layers c=2, and the fuzzy coefficient m=2.

[0035] S402: Select two points in each segmented area as cluster centers.

[0036] S403, calculating the membership degree of each point in each segmented area to each cluster center. The calculation formula of the membership degree is: ; In the above formula, x i Represents the superpixel feature vector including color and space in each area, that is, the horizontal coordinate of the i-th pixel, v j represents the cluster center of the jth class, v k represents the k-th cluster center, c represents the number of clusters, and m represents the fuzzy coefficient.

[0037] S404, by maximizing the membership to update the centroid, so that the centroid position has the best representation of the data points in the current cluster, if the membership center change is less than the preset threshold, then go to S405, otherwise repeat this step. The cluster center is calculated as follows: In the above formula, N represents the number of superpixels in this category. In this example, the preset threshold is 0.005.

[0038] S405: The cluster labels with the best membership are combined and displayed in the target image as the segmentation result.

[0039] S5. Extract the temperature of different points in the partition and determine whether there is segregation in the corresponding partition based on the temperature. This step includes the following sub-steps: S501 , calculating the average temperature value of each superpixel center point in the segmented area, and using it as the temperature representative value of the segmented area.

[0040] S502: Calculate the temperature difference between different segmented areas according to the temperature representative value.

[0041] S503: Determine whether the interpolated temperature exceeds the temperature difference threshold for temperature segregation. If so, determine that temperature segregation has occurred; otherwise, determine that temperature segregation has not occurred. Based on the physical parameters of the asphalt mixture, the temperature difference threshold for determining that temperature segregation has occurred is determined to be 10°C. The asphalt mixture parameters are shown in Table 1: Table 1: Temperature requirements for asphalt and aggregate project Temperature requirement (℃) Asphalt heating temperature 170 Aggregate heating temperature 180 Discharge temperature range 170-180 Temperature range of the mixture when transported to the site 165-175 Temperature requirements during paving >60 Figure 2 and Figure 3 Demonstration images of infrared images with temperature segregation and infrared images without temperature segregation are shown respectively. In order to verify the effectiveness of the method of the present invention, 200 infrared images of asphalt are selected for temperature segregation detection, including 100 infrared images with temperature segregation and 100 infrared images without temperature segregation. Figure 4 As shown in the figure, the labels of temperature segregation and non-temperature segregation are represented by the numbers -1 and 1 respectively. Among them, 195 samples out of 200 test images were correctly classified, and the accuracy of this experiment was 97.5%.

[0042] Example 2: Based on the same inventive concept, the present invention also provides an asphalt temperature segregation detection system, comprising: An acquisition module, used for acquiring infrared images of asphalt; A super-pixel processing module, configured to perform super-pixel processing on the asphalt infrared image using a multi-scale morphological watershed algorithm to obtain a super-pixel image; A segmentation module, configured to perform clustering processing on the superpixel image using a K-Means clustering algorithm to obtain a plurality of segmented regions; A merging module, configured to process each segmented region using an FCM clustering algorithm, generate fuzzy labels, and merge the fuzzy labels into the superpixel image as a segmentation result; The judgment module is used to extract the temperature of different points in the partition and judge whether the corresponding partition has segregation according to the temperature.

[0043] Example 3: Based on the same inventive concept, the present invention also provides an asphalt temperature segregation detection device, including a processor and a memory, the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the above method.

[0044] The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting asphalt temperature segregation, characterized in that: The steps include: S1. Obtain infrared image of asphalt; S2. Using a multi-scale morphological watershed algorithm to perform super-pixel processing on the asphalt infrared image to obtain a super-pixel image; S3, clustering the superpixel image using a K-Means clustering algorithm to obtain multiple segmented regions; S4. Processing each segmented region using an FCM clustering algorithm to generate a fuzzy label, and merging the fuzzy label into the superpixel image as a segmentation result; S5. Extract the temperatures of different points in the partition, and determine whether segregation exists in the corresponding partition based on the temperatures.

2. The asphalt temperature segregation detection method according to claim 1, characterized in that: The S2 includes the following sub-steps: S201, constructing a multi-scale morphological gradient image based on the asphalt infrared image; S202, dividing the multi-scale morphological gradient image into different regions, and selecting a water injection point in each region; S203, assigning a unique label to each water injection point in the area and performing a simulated water injection operation to obtain an expanded area; S204: When there is a boundary between the extended areas corresponding to different areas, a super-pixel image is obtained.

3. The asphalt temperature segregation detection method according to claim 2, characterized in that: The calculation formula for constructing the multi-scale morphological gradient image is: ； In the above formula, Represents the result of multi-scale morphological gradient operation, R f represents morphological operation, g represents asphalt infrared image, r represents morphological convolution kernel size, r∈(r1,r2); The calculation formula of the water injection point is: ； In the above formula, M represents the water injection point, I represents the morphological gradient image, p represents the current pixel point, and q represents the point within the set neighborhood N(p); The calculation formula of the extended area is: ； In the above formula, R k (h) represents the extended area, pm k Indicates p to m k The path, m k represents the minimum value of each area, indicating the water level; The boundary is calculated as follows: ； In the above formula, B represents the boundary, L(p) represents the pixel p belongs to the local minimum m k Starting area.

4. The asphalt temperature segregation detection method according to claim 1, characterized in that: The S3 includes the following sub-steps: S301, calculating the average grayscale value of each superpixel in the superpixel image; S302: For each superpixel, calculate the position feature of its set center; S303, randomly selecting several initial centers from the features of superpixels as initial clustering centers; S304, for each superpixel in the superpixel image, calculating the distance between its feature and each cluster center, and assigning the superpixel to the cluster center closest to it based on the calculation result; S305, recalculate each cluster center by extracting the feature average of all superpixels belonging to the cluster to update. If the center reaches a preset range or reaches a preset number of iterations, go to S306, otherwise go to S304; S306 , aggregating the superpixel image of each cluster center into a region to obtain the segmented region.

5. The asphalt temperature segregation detection method according to claim 4, characterized in that: The calculation formula of the grayscale average value is: ； In the above formula, C avg represents the average value, c i represents the grayscale value of the i-th pixel, and N represents the number of superpixels in the entire image; The calculation formula of the position feature is: ； In the above formula, (x centroid ,y centroid ) represents the position feature, (x i ,y i ) represents the position of the i-th pixel in the superpixel; The distance is calculated as follows: ； In the above formula, d(j,k) represents the distance between the jth superpixel and the kth center, f j represents the feature of the jth superpixel, f k represents the characteristics of the k-th center, (x cj ,y cj ) represents the center position feature of the jth superpixel, (x ck ,y ck ) represents the central position feature of the k-th center point; The calculation formula of the cluster center is: ； In the above formula, μ k represents the cluster center, S k represents the set of all superpixels assigned to k cluster centers.

6. The asphalt temperature segregation detection method according to claim 1, characterized in that: The S4 includes the following sub-steps: S401, setting clustering parameters, wherein the clustering parameters include the number of clusters and the fuzzy coefficient; S402: For each segmented region obtained by K-Means, several points are selected from the region as initial cluster centers for FCM clustering. S403, calculating the membership degree of each point in each segmented area to each cluster center; S404, update the centroid by maximizing the membership so that the centroid position has the best representation for the data points in the current cluster. If the membership center change is less than a preset threshold after iteration, go to S405, otherwise repeat this step; S405: The cluster labels with the best membership are combined and displayed in the target image as the segmentation result.

7. The asphalt temperature segregation detection method according to claim 6, characterized in that: The calculation formula of the membership degree is: ； In the above formula, x i Represents the superpixel feature vector including color and space in each area, that is, the horizontal coordinate of the i-th pixel, v j represents the cluster center of the jth class, v k Indicates the k The cluster centers of each class, c represents the number of clusters, and m represents the fuzzy coefficient; The calculation formula of the cluster center is: ； In the above formula, N represents the number of super pixels in this category.

8. The asphalt temperature segregation detection method according to claim 1, characterized in that: The S5 comprises the following sub-steps: S501, calculating the average temperature value of each superpixel center point in the segmented area, and taking it as the temperature representative value of the segmented area; S502, calculating the temperature difference between different segmented areas according to the temperature representative value; S503: Determine whether the temperature interpolation exceeds a temperature difference threshold of temperature segregation; if so, determine that temperature segregation occurs; otherwise, determine that temperature segregation does not occur.

9. An asphalt temperature segregation detection system, characterized in that: include: An acquisition module, used for acquiring infrared images of asphalt; A super-pixel processing module, configured to perform super-pixel processing on the asphalt infrared image using a multi-scale morphological watershed algorithm to obtain a super-pixel image; A segmentation module, configured to perform clustering processing on the superpixel image using a K-Means clustering algorithm to obtain a plurality of segmented regions; A merging module, configured to process each segmented region using an FCM clustering algorithm, generate fuzzy labels, and merge the fuzzy labels into the superpixel image as a segmentation result; The judgment module is used to extract the temperature of different points in the partition and judge whether the corresponding partition has segregation according to the temperature.

10. An asphalt temperature separation detection device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the method according to claim 1.

Images