Waveform guardrail stiffness detection method and system based on microwave radar and spectral clustering

Abstract

The present invention discloses a method and system for detecting the stiffness of a corrugated guardrail based on microwave radar and spectral clustering, belonging to the field of traffic engineering, including: collecting the micro-deformation data of the corrugated guardrail in its natural state by using microwave radar, calculating the similarity matrix of the micro-deformation data of the corrugated guardrail, processing the similarity matrix of the micro-deformation data of the corrugated guardrail through a spectral clustering algorithm to obtain the clustering result of the stiffness level of the corrugated guardrail, and completing the detection of the stiffness of the corrugated guardrail by comparing the characteristics of different clustering groups. The present invention uses microwave radar for non-contact measurement and spectral clustering algorithm for clustering to obtain the stiffness level classification of the corrugated guardrail, complete the stiffness detection of the corrugated guardrail, can detect the stiffness of the corrugated guardrail in real time and accurately, avoid the cumbersome operation of manually contacting the corrugated guardrail, improve the detection efficiency, and avoid the reduction of the accuracy of the detection result caused by human factors.

Description

Technical Field

[0001] The present invention belongs to the field of traffic engineering, and particularly relates to a method and system for detecting the stiffness of corrugated guardrails based on microwave radar and spectral clustering. Background Art

[0002] As an important part of highway traffic safety facilities, the suitability of the stiffness of corrugated guardrails is directly related to the safety performance during vehicle collisions. Traditional methods for detecting the stiffness of corrugated guardrails usually rely on manual measurement or contact sensors, which are not only time-consuming and laborious, have limited accuracy, but also are difficult to achieve continuous monitoring and cannot meet the requirements of real-time monitoring. With the increase in traffic flow and the improvement of vehicle performance, the demand for real-time and accurate detection of the stiffness of corrugated guardrails is becoming increasingly urgent. Summary of the Invention

[0003] Aiming at the above deficiencies in the prior art, the method and system for detecting the stiffness of corrugated guardrails based on microwave radar and spectral clustering provided by the present invention solve the problems of time-consuming and laborious stiffness detection of corrugated guardrails, being easily affected by detection personnel, and being difficult to continuously monitor.

[0004] In order to achieve the above invention purpose, the technical solution adopted by the present invention is: A method for detecting the stiffness of corrugated guardrails based on microwave radar and spectral clustering, comprising the following steps:

[0005] S1: Scanning the corrugated guardrail with a microwave radar to collect the micro-deformation data of the corrugated guardrail in the natural state;

[0006] S2: According to the micro-deformation data, calculating the similarity between each data point in the micro-deformation data, and constructing a similarity matrix of the micro-deformation data of the corrugated guardrail;

[0007] S3: According to the similarity matrix of the micro-deformation data of the corrugated guardrail, using a spectral clustering algorithm for processing to obtain a clustering result of the stiffness levels of the corrugated guardrail;

[0008] S4: According to the clustering result of the stiffness levels of the corrugated guardrail, classifying the stiffness levels of the corrugated guardrail, and determining the stiffness levels corresponding to each clustering group by comparing the characteristics of different clustering groups, so as to realize the detection of the stiffness of the corrugated guardrail.

[0009] The beneficial effects of the present invention are as follows: The present invention uses a microwave radar for non-contact measurement and utilizes the spectral clustering algorithm for clustering to obtain the stiffness level classification of the corrugated guardrail, completing the stiffness detection of the corrugated guardrail. The present invention avoids the cumbersome operations of manually contacting the corrugated guardrail in the traditional method, improves the detection efficiency, and through the high-precision measurement of the microwave radar, can accurately reflect the micro-deformation of the corrugated guardrail and realize the real-time monitoring of the stiffness of the corrugated guardrail, timely discover potential safety hazards, and provide a basis for highway traffic safety management. At the same time, the present invention uses the spectral clustering algorithm for automatic clustering without manual intervention, reduces the influence of human factors on the detection results, and improves the accuracy of the stiffness detection results.

[0010] Further: The similarity between each data point in the micro-deformation data is the Euclidean distance between the data points, and its calculation expression is as follows:

[0011]

[0012] Wherein, and are both data points, is the data point and the data point is the Euclidean distance between them, is the dimension of the data point, is the th data of the data point and is the th data of the data point .

[0013] The beneficial effect of the above further solution is: By calculating the Euclidean distance between each data point through the micro-deformation data, the proximity of each data point in the micro-deformation data can be measured, which is convenient for subsequent cluster classification.

[0014] Further: The specific steps of S3 are as follows:

[0015] S301: According to the similarity matrix of the micro-deformation data of the corrugated guardrail, calculate using the degree matrix to obtain the Laplacian matrix;

[0016] S302: Perform eigenvalue decomposition on the Laplacian matrix to obtain eigenvalues and eigenvectors;

[0017] S303: According to the magnitude of the eigenvalues, sort the eigenvectors corresponding to the eigenvalues in ascending order to obtain the eigenvectors sorted in ascending order;

[0018] S304: According to the preset quantity n, select the first n eigenvectors from the eigenvectors sorted in ascending order as the new data set;

[0019] S305: Perform clustering processing using the K-means algorithm based on the new dataset to obtain the clustering result of the stiffness grades of the corrugated guardrails.

[0020] The beneficial effect of the above further solution is that through the Laplacian matrix and the K-means clustering algorithm, the clustering result of the stiffness grades of the corrugated guardrails can be obtained quickly and efficiently, realizing the real-time monitoring of the corrugated guardrails.

[0021] Further: The calculation expression of the Laplacian matrix in S301 is as follows:

[0022]

[0023] where is the Laplacian matrix, is the degree matrix, is the similarity matrix of the micro-deformation data of the corrugated guardrails.

[0024] The beneficial effect of the above further solution is that converting the similarity matrix of the micro-deformation data into a Laplacian matrix facilitates the subsequent clustering operation of the matrix using the K-means algorithm.

[0025] Further: The specific steps of S305 are as follows:

[0026] S3051: Randomly select K data from the new dataset as the cluster centers;

[0027] S3052: Calculate the Euclidean distance from each data in the dataset to the K cluster centers respectively, and assign it to the initial cluster center with the smallest Euclidean distance to obtain K clusters;

[0028] S3053: Calculate the average value of each of the K clusters respectively, and use the average value as the new cluster center;

[0029] S3054: Determine whether the new cluster centers no longer change, or whether the number of iterations reaches the pre-set number of iterations, or whether the clustering criterion function reaches the minimum value. If so, complete the clustering process of the K-means algorithm, output the K clusters and the center points of each cluster to obtain the clustering result of the stiffness grades of the corrugated guardrails; otherwise, return to S3052.

[0030] The calculation expression of the new cluster center in S3053 is as follows:

[0031]

[0032] where is the cluster center of the th cluster, is the th in the a piece of data, is the number of data in the

[0033] The beneficial effect of the above further solution is that through the K-means clustering algorithm, the Laplacian matrix can be clustered to complete the classification of the stiffness levels of the corrugated guardrail, facilitating the stiffness detection of the corrugated guardrail.

[0034] The present invention also provides a corrugated guardrail stiffness detection system based on microwave radar and spectral clustering, including: a microwave radar measurement module, a data processing module, a clustering module, and a stiffness identification module connected in sequence;

[0035] The microwave radar measurement module is used to scan the corrugated guardrail and collect the micro-deformation data of the corrugated guardrail;

[0036] The data processing module is used to construct a similarity matrix of the micro-deformation data according to the micro-deformation data of the corrugated guardrail;

[0037] The clustering module is used to perform clustering processing on the similarity matrix of the micro-deformation data by using the spectral clustering algorithm;

[0038] The stiffness identification module is used to automatically identify and classify the stiffness levels of the corrugated guardrail according to the clustering results of the clustering module.

[0039] The beneficial effect of the present invention is that the present invention uses a microwave radar measurement module to perform non-contact measurement on the corrugated guardrail, can accurately and real-time detect the micro-deformation data of the corrugated guardrail, and uses the data processing module to convert the micro-deformation data into a similarity matrix of the micro-deformation data, providing a data source for the subsequent clustering module to cluster the stiffness levels of the corrugated guardrail. Finally, the stiffness identification module classifies the stiffness levels of the corrugated guardrail according to the clustering results, completes the monitoring of the stiffness of the corrugated guardrail, can timely discover potential safety hazards, provides a basis for highway traffic safety management. At the same time, the present invention does not require manual intervention, can reduce the influence of human factors on the detection results, and improves the accuracy of the stiffness detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flowchart of a corrugated guardrail stiffness detection method based on microwave radar and spectral clustering;

[0041] Figure 2 is a structural diagram of a corrugated guardrail stiffness detection system based on microwave radar and spectral clustering. DETAILED DESCRIPTION OF THE INVENTION

[0042] The specific embodiments of the present invention will be described below to facilitate those skilled in the art to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0043] Embodiment 1

[0044] As Figure 1 shown, the present invention provides a method for detecting the stiffness of a corrugated guardrail based on microwave radar and spectral clustering, including the following steps:

[0045] S1: Use microwave radar to scan the corrugated guardrail and collect the micro-deformation data of the corrugated guardrail in its natural state;

[0046] S2: According to the micro-deformation data, calculate the similarity between each data point in the micro-deformation data and construct a similarity matrix of the micro-deformation data of the corrugated guardrail;

[0047] S3: According to the similarity matrix of the micro-deformation data of the corrugated guardrail, use the spectral clustering algorithm for processing to obtain the clustering result of the stiffness level of the corrugated guardrail;

[0048] S4: According to the clustering result of the stiffness level of the corrugated guardrail, classify the stiffness level of the corrugated guardrail, and by comparing the characteristics of different clustering groups, determine the stiffness level corresponding to each clustering group to achieve the detection of the stiffness of the corrugated guardrail.

[0049] In an embodiment of the present invention, the specific method for using microwave radar to scan the corrugated guardrail is as follows: The microwave radar emits microwave signals to the corrugated guardrail, and the corrugated guardrail surface reflects the reflected signals to the microwave radar. After the microwave radar receives the reflected signals, the microwave signal propagation time and the phase change of the corrugated guardrail are obtained, and the two are integrated to obtain the micro-deformation data; the transmission frequency and reception sensitivity of the microwave radar are adjusted according to the actual detection needs to ensure the accuracy and reliability of the corrugated guardrail stiffness detection result.

[0050] In S2, the similarity between each data point in the micro-deformation data is the Euclidean distance between the data points, and its calculation expression is as follows:

[0051]

[0052] Among them, and are both data points, is the data point and the data point The Euclidean distance between them, is the dimension of the data point, is the th data of the data point, is the th data of the data point.

[0053] The specific steps of S3 are as follows:

[0054] S301: According to the similarity matrix of the waveform guardrail micro-deformation data, calculate using the degree matrix to obtain the Laplacian matrix. Among them, the calculation expression of the Laplacian matrix is as follows:

[0055]

[0056] Among them, is the Laplacian matrix, is the degree matrix, is the similarity matrix of the waveform guardrail micro-deformation data;

[0057] In an embodiment of the present invention, the degree matrix is a diagonal matrix, and the elements on its diagonal are the degrees of each data point. In an embodiment of the present invention, the expression of the degree matrix is as follows:

[0058]

[0059] The expression of the similarity matrix is as follows:

[0060]

[0061] The expression of the Laplacian matrix is as follows:

[0062]

[0063] Among them, , , and are the elements on the diagonal of the degree matrix , representing the number of all data points similar to the data point, , , , , , , , , , , and are the elements in the similarity matrix , representing the similarity between data points;

[0064] S302: Perform eigenvalue decomposition on the Laplacian matrix to obtain eigenvalues and eigenvectors;

[0065] S303: According to the magnitudes of the eigenvalues, sort the eigenvectors corresponding to the eigenvalues in ascending order to obtain the eigenvectors sorted in ascending order;

[0066] S304: According to the preset quantity n, select the first n eigenvectors from the eigenvectors sorted in ascending order as the new dataset, where n is a positive integer and needs to be adjusted according to the actual stiffness detection requirements;

[0067] S305: According to the new dataset, perform clustering using the K - means algorithm to obtain the clustering result of the stiffness levels of the corrugated guardrails.

[0068] The specific steps of S305 are as follows:

[0069] S3051: Randomly select K data from the new dataset as the cluster centers;

[0070] S3052: Calculate the Euclidean distances from each data in the dataset to the K cluster centers respectively, and assign them to the initial cluster center with the minimum Euclidean distance to obtain K clusters;

[0071] S3053: Calculate the average value of each of the K clusters respectively, and use the average value as the new cluster center. Among them, the calculation expression of the new cluster center is as follows:

[0072]

[0073] Among them, is the cluster center of the th cluster, is the th data in the th cluster, is the th data in the

[0074] S3054: Determine whether the new cluster centers no longer change, or whether the number of iterations reaches the preset number of iterations, or whether the clustering criterion function reaches the minimum value. If so, complete the clustering process of the K - means algorithm, output K clusters and the center points of each cluster to obtain the clustering result of the stiffness levels of the corrugated guardrails; otherwise, return to S3052.

[0075] In an embodiment of the present invention, a microwave radar is used to collect the corrugated guardrails to obtain micro - deformation data. Through a method for detecting the stiffness of corrugated guardrails based on microwave radar and spectral clustering provided by the present invention, after going through S1 to S3, the clustering results of the stiffness levels of K corrugated guardrails are obtained. The following details S4 in combination with this embodiment:

[0076] S4: Classify the stiffness levels of the corrugated guardrails according to the clustering results of the stiffness levels of the corrugated guardrails, and determine the corresponding stiffness levels for each clustering group by comparing the characteristics of different clustering groups, so as to realize the detection of the stiffness of the corrugated guardrails.

[0077] According to historical data and engineering experience, the stiffness levels of the guardrails are predefined in sequence as four levels: A, B, C, and D, where A is the highest stiffness level and D represents the lowest stiffness level.

[0078] Classify the clustering results of the stiffness levels of K corrugated guardrails into the corresponding predefined stiffness levels according to the characteristics of each clustering result, and obtain the classification results of the stiffness levels of the corrugated guardrails, thus completing the detection of the stiffness of the corrugated guardrails. Among them, the characteristics of the clustering results include the average micro-deformation amount, standard deviation, etc.

[0079] In a specific embodiment, by using the method for detecting the stiffness of corrugated guardrails based on microwave radar and spectral clustering provided by the present invention, the stiffness of a 100-meter corrugated guardrail hole is detected, divided into 10 paragraphs each 10 meters long, and the following stiffness level divisions are obtained:

[0080] Paragraphs 1 - 3: The characteristics of the clustering group match the stiffness level A, and the stiffness of the guardrail is excellent.

[0081] Paragraphs 4 - 6: The characteristics of the clustering group match the stiffness level B, and the stiffness of the guardrail is good.

[0082] Paragraphs 7 - 8: The characteristics of the clustering group match the stiffness level C, and the stiffness of the guardrail is average, and monitoring is recommended.

[0083] Paragraphs 9 - 10: The characteristics of the clustering group match the stiffness level D, and the stiffness of the guardrail is poor and needs to be reinforced.

[0084] The beneficial effects of the present invention are as follows: The present invention uses microwave radar for non-contact measurement and uses the spectral clustering algorithm for clustering to obtain the classification of the stiffness levels of the corrugated guardrails and complete the detection of the stiffness of the corrugated guardrails. The present invention avoids the cumbersome operations of manually contacting the corrugated guardrails in the traditional method, improves the detection efficiency, and through the high-precision measurement of the microwave radar, can accurately reflect the micro-deformation of the corrugated guardrails and realize the real-time monitoring of the stiffness of the corrugated guardrails, timely discover potential safety hazards, and provide a basis for highway traffic safety management. At the same time, the present invention uses the spectral clustering algorithm for automatic clustering without manual intervention, reduces the influence of human factors on the detection results, and improves the accuracy of the stiffness detection results.

[0085] Embodiment 2

[0086] As Figure 2As shown in the figure, the present invention provides a waveform guardrail stiffness detection system based on microwave radar and spectral clustering, including: a microwave radar measurement module, a data processing module, a clustering module, and a stiffness identification module connected in sequence;

[0087] The microwave radar measurement module is used to scan the waveform guardrail and collect the micro-deformation data of the waveform guardrail;

[0088] The data processing module is used to construct a similarity matrix of the micro-deformation data according to the micro-deformation data of the waveform guardrail;

[0089] The clustering module is used to perform clustering processing on the similarity matrix of the micro-deformation data by using the spectral clustering algorithm;

[0090] The stiffness identification module is used to automatically identify and classify the stiffness levels of the waveform guardrails according to the clustering results of the clustering module.

[0091] The beneficial effects of the present invention are as follows: The present invention uses a microwave radar measurement module to perform non-contact measurement on the waveform guardrail, can accurately and real-time detect the micro-deformation data of the waveform guardrail, and uses the data processing module to convert the micro-deformation data into a similarity matrix of the micro-deformation data, providing a data source for the subsequent clustering module to cluster the stiffness levels of the waveform guardrail. Finally, the stiffness identification module classifies the stiffness levels of the waveform guardrail according to the clustering results, completing the monitoring of the stiffness of the waveform guardrail, can timely discover potential safety hazards, and provide a basis for highway traffic safety management. At the same time, the present invention does not require manual intervention, can reduce the influence of human factors on the detection results, and improves the accuracy of the stiffness detection results.

Claims

1. A method for detecting stiffness of corrugated guardrail based on microwave radar and spectral clustering, characterized in that: The following steps are involved: S1: Scan the corrugated guardrail using microwave radar to collect micro-deformation data of the corrugated guardrail in the natural state; S2: According to the micro-deformation data, the similarity between each data point in the micro-deformation data is calculated, and a similarity matrix of the micro-deformation data of the corrugated guardrail is constructed; S3: Based on the similarity matrix of the micro-deformation data of the corrugated guardrail, the spectral clustering algorithm is used to process the data and obtain the stiffness level clustering result of the corrugated guardrail; S4: According to the clustering result of the stiffness level of the corrugated guardrail, the stiffness level of the corrugated guardrail is classified, and by comparing the characteristics of different cluster groups, the stiffness level corresponding to each cluster group is determined to realize the detection of the stiffness of the corrugated guardrail; The specific steps of S3 are as follows: S301: performing calculations using a degree matrix according to a similarity matrix of the micro-deformation data of the corrugated guardrail to obtain a Laplace matrix; S302: Perform eigendecomposition on the Laplace matrix to obtain eigenvalues ​​and eigenvectors; S303: Arrange the eigenvectors corresponding to the eigenvalues ​​in ascending order according to the magnitude of the eigenvalues ​​to obtain the eigenvectors arranged in ascending order; S304: According to a preset number n, the first n feature vectors in the feature vectors arranged in ascending order are selected as a new data set; S305: performing clustering processing using a K-means algorithm based on the new data set to obtain a stiffness level clustering result of the corrugated guardrail; The specific steps of S305 are as follows: S3051: Randomly select K data in the new data set as cluster centers; S3052: Calculate the Euclidean distance from each data point in the data set to the centers of the K clusters, and assign the data points to the initial cluster center with the smallest Euclidean distance, thereby obtaining K clusters. S3053: Calculate the average value of each cluster in the K clusters respectively, and use the average value as the new cluster center; S3054: Determine whether the new cluster center no longer changes or whether the number of iterations reaches the preset number of iterations or whether the clustering criterion function reaches the minimum value. If so, complete the clustering process of the K-means algorithm, output K clusters and the center point of each cluster, and obtain the stiffness level clustering result of the corrugated guardrail. Otherwise, return to S3052; The calculation expression of the new cluster center in S3053 is as follows: in, For the The cluster centers of the clusters, For the The first data, For the The number of data in a cluster.

2. The method for detecting stiffness of corrugated guardrail based on microwave radar and spectral clustering according to claim 1 is characterized in that: The similarity between the data points in the micro-deformation data is the Euclidean distance between the data points, and its calculation expression is as follows: in, and are all data points, For data points and data points The Euclidean distance between is the dimension of the data point, For data points No. data, For data points No. data.

3. The method for detecting stiffness of corrugated guardrail based on microwave radar and spectral clustering according to claim 1 is characterized in that: In S301, the calculation expression of the Laplacian matrix is ​​as follows: in, is the Laplace matrix, is the degree matrix, is the similarity matrix of the micro-deformation data of the corrugated guardrail.

4. A corrugated guardrail stiffness detection system based on microwave radar and spectral clustering, characterized in that: The corrugated guardrail stiffness detection system is used to perform the corrugated guardrail stiffness detection method described in any one of claims 1 to 3, comprising: a microwave radar measurement module, a data processing module, a clustering module and a stiffness identification module connected in sequence; The microwave radar measurement module is used to scan the corrugated guardrail and collect micro-deformation data of the corrugated guardrail; The data processing module is used to construct a similarity matrix of micro-deformation data according to the micro-deformation data of the corrugated guardrail; The clustering module is used to perform clustering processing on the similarity matrix of micro-deformation data using a spectral clustering algorithm; The stiffness identification module is used to automatically identify and classify the stiffness level of the corrugated guardrail according to the clustering result of the clustering module.

Images