Road detection method and system based on ground penetrating radar and electronic equipment

Abstract

The invention provides a road detection method and system based on a ground penetrating radar and electronic equipment, and relates to the technical field of radar detection, and the method comprises the steps: scanning a to-be-detected road through the ground penetrating radar, and obtaining radar image data; preprocessing the radar image data to obtain target image data; performing feature extraction on the target image data to obtain a group of features, the group of features including amplitude, frequency, phase, event continuity and multiple reflection features; identifying the group of features to obtain an identification result, the identification result being used for representing whether an underground disease body exists in the to-be-detected road, and determining a disease body type of the underground disease body under the condition of determining that the underground disease body exists in the to-be-detected road; and determining a target risk level of the to-be-detected road by using a road collapse risk assessment model in combination with the identification result and the surrounding environment factors. By implementing the technical scheme provided by the invention, the effect of improving the detection efficiency and accuracy is achieved.

Description

Technical Field

[0001] The present application relates to the field of radar detection technology, and in particular to a road detection method, system and electronic equipment based on ground-penetrating radar. Background Art

[0002] Road projects, when exposed to natural environmental factors such as rainfall and temperature fluctuations, may face risks such as landslides and collapses. Furthermore, underground urban areas often harbor geological structures (or defects) such as voids, holes, and loose bodies that threaten urban safety. To ensure the normal operation of highway projects and urban safety, rapid and convenient detection of roadbed and subsurface defects is essential. In recent years, ground-penetrating radar (GPR) technology, due to its high efficiency and non-destructive nature, has gained widespread application in road inspection, providing a crucial tool for road health monitoring. Related technologies using GPR for road inspection primarily rely on manual interpretation of radar image data to identify defect types and locations. Specifically, technicians typically rely on observing features in radar images and combining experience to determine the presence and type of defects. This method suffers from low detection efficiency and is prone to inaccuracies due to human factors, making it difficult to meet the growing demand for road maintenance. Summary of the Invention

[0003] In order to solve the above technical problems, the present application provides a road detection method, system and electronic equipment based on ground penetrating radar.

[0004] In the first aspect, the present application provides a road detection method based on ground penetrating radar, comprising: scanning the road to be detected by ground penetrating radar to obtain radar image data; preprocessing the radar image data to obtain target image data; extracting features from the target image data to obtain a set of features, wherein a set of features includes amplitude, frequency, phase, event axis continuity and multiple reflection features; identifying a set of features to obtain an identification result, wherein the identification result is used to indicate whether there is an underground disease body on the road to be detected, and to determine the type of the underground disease body when it is determined that there is an underground disease body on the road to be detected; combining the identification results and surrounding environmental factors, using a road collapse risk assessment model to determine the target risk level of the road to be detected.

[0005] By adopting the above technical solution, the purpose of automatic detection and risk assessment of underground road damage bodies can be achieved. Specifically, by obtaining radar image data through ground-penetrating radar scanning, the geological structure information under the road can be fully captured, thereby improving the detection range and accuracy; the radar image data is preprocessed to effectively remove noise and normalize the data to ensure the accuracy of subsequent feature extraction; the amplitude, frequency, phase, event axis continuity and multiple reflection features in the target image data are extracted to provide multi-dimensional data support for the identification of underground damage bodies and improve recognition reliability; the feature recognition method is used to determine the presence and type of underground damage bodies, which greatly reduces the workload of manual interpretation and improves detection efficiency and accuracy; the risk assessment is carried out by combining the identification results with surrounding environmental factors, comprehensively considering the characteristics of the damage body and external influences, making the risk assessment more scientific and reasonable, and providing an important basis for road safety maintenance.

[0006] Optionally, the target risk level of the road to be inspected is determined using a road collapse risk assessment model in combination with the identification results and surrounding environmental factors, including: determining a weight vector corresponding to an evaluation factor set, the evaluation factor set including the type of diseased body, the scale of the diseased body, the depth of the diseased body, the distance to the building, the distribution of underground pipelines, the soil type and the groundwater level, the weight vector including the weight values ​​corresponding to each evaluation factor in the evaluation factor set, wherein the identification results include the type of diseased body, the scale of the diseased body and the depth of the diseased body, and the surrounding environmental factors include the distance to the building, the distribution of underground pipelines, the soil type and the groundwater level; constructing a fuzzy relationship matrix, and obtaining an evaluation result vector based on the weight vector and the fuzzy relationship matrix, wherein the fuzzy relationship matrix is ​​used to represent the fuzzy relationship between each evaluation factor in the evaluation factor set and each evaluation level in the evaluation level set; and determining the target risk level of the road to be inspected based on the evaluation result vector.

[0007] By adopting the above technical solution, it is possible to combine the identification results with the surrounding environmental factors, and use the weight vector and fuzzy relationship matrix to evaluate the road collapse risk, thereby determining the target risk level of the road to be inspected. Specifically, it includes: First, by combining the identification results of the type, scale, depth, etc. of the diseased body with the surrounding environmental factors such as the distance to the building and the distribution of underground pipelines, the various factors affecting the road collapse are fully considered, thereby improving the accuracy of the risk assessment; second, the fuzzy relationship matrix is ​​used to describe the fuzzy relationship between the evaluation factors and the evaluation level, which effectively solves the problem of inaccurate evaluation caused by the complex relationship between factors in the traditional method; third, the evaluation result vector is calculated based on the weight vector and the fuzzy relationship matrix, which realizes the quantitative analysis of the road collapse risk and provides a scientific basis for the subsequent targeted maintenance measures.

[0008] Optionally, a fuzzy relationship matrix is ​​constructed, and an evaluation result vector is obtained based on the weight vector and the fuzzy relationship matrix, including: determining the membership of each evaluation factor in the evaluation factor set to each evaluation level in the evaluation level set to obtain the fuzzy relationship matrix; multiplying the weight vector and the fuzzy relationship matrix to obtain an evaluation result vector; determining the target risk level of the road to be inspected based on the evaluation result vector, including: determining the evaluation level corresponding to the maximum value of the element in the evaluation result vector as the target risk level.

[0009] By adopting the above technical solution, a fuzzy relationship matrix is ​​constructed by determining the membership of each evaluation factor in the evaluation factor set to the evaluation level set, making the evaluation process more scientific and reasonable, and being able to fully consider the impact of multiple factors on the risk level; the evaluation result vector is obtained by multiplying the weight vector with the fuzzy relationship matrix, which realizes quantitative analysis and improves the accuracy and reliability of the evaluation results; the target risk level is determined according to the evaluation level corresponding to the maximum value of the element in the evaluation result vector, which simplifies the decision-making process, ensures that the evaluation results are intuitive and clear, and facilitates the subsequent implementation of targeted measures. Describing the relationship between evaluation factors and evaluation levels based on membership can more accurately reflect the impact of each factor on the risk level, avoid the errors caused by simple classification or weighted average in traditional methods, and thus improve the accuracy of road risk level assessment. This technical solution can achieve a more accurate assessment of road collapse risk level.

[0010] Optionally, the road to be inspected is scanned by a ground-penetrating radar to obtain radar image data, including: transmitting a detection signal to the road to be inspected by the ground-penetrating radar; receiving an echo signal, wherein the echo signal is a signal generated by the detection signal being transmitted through the underground medium interface of the road to be inspected; and obtaining radar image data based on the echo signal.

[0011] By adopting the above technical solution, the ground-penetrating radar transmits detection signals to the road to be inspected, covering the geological structures of different depths and ranges below the road, ensuring the comprehensiveness of the detection; receiving the echo signal and generating radar image data based on it can accurately reflect the reflection characteristics of the underground medium interface, thereby providing a reliable data basis for subsequent feature extraction and identification.

[0012] Optionally, preprocessing the radar image data to obtain target image data includes: performing denoising and normalization processing on the radar image data to obtain target image data.

[0013] By adopting the above technical solution, radar image data can be denoised and normalized, which can effectively reduce noise interference and improve data quality. At the same time, normalization processing helps to unify the data range and provide more reliable basic data for subsequent feature extraction and identification, thereby improving the accuracy and efficiency of road damage detection.

[0014] Optionally, identifying a set of features to obtain an identification result includes: using a target recognition model to identify a set of features to obtain an identification result; wherein the target recognition model is trained in the following manner: obtaining a training sample set, wherein each training sample in the training sample set contains a set of sample features and corresponding road disease results; using the training sample set to train a network model to be trained until an end condition is met, and determining the network model obtained after the training as the target recognition model, wherein the end condition includes that the loss value of the network model to be trained meets a preset convergence condition or the number of iterations reaches a preset threshold.

[0015] By employing this technical solution and utilizing a target recognition model to identify features, the accuracy and efficiency of road damage identification can be significantly improved. By training on a training set containing sample features and road damage results, the network model learns the mapping relationship between different features and damage bodies, enabling more accurate determination of the presence and type of underground damage bodies on the road under inspection. Furthermore, by setting appropriate termination conditions to ensure sufficient model training and avoid overfitting, the reliability of the recognition results is further enhanced.

[0016] Optionally, before scanning the road to be inspected by the ground penetrating radar, the above method also includes: using a group of sensors to collect the working environment parameters of the ground penetrating radar in real time, wherein the working environment parameters include soil moisture and electromagnetic interference intensity; according to the working environment parameters, automatically adjusting the transmission frequency, power and sampling rate of the ground penetrating radar through a preset algorithm.

[0017] By adopting this technical solution, sensors can collect real-time environmental parameters, including soil moisture and electromagnetic interference intensity, before the ground-penetrating radar (GPR) begins scanning roads. Based on these parameters, a pre-set algorithm automatically adjusts the GPR's transmission frequency, power, and sampling rate. Dynamically adjusting GPR parameters reduces environmental interference with detection signals, improves the radar's adaptability in diverse environments, and mitigates the impact of external factors on detection results, thereby enhancing the accuracy and reliability of detection data.

[0018] Optionally, after determining the target risk level of the road to be inspected using the road collapse risk assessment model, the above method also includes: when the target risk level is determined to be a preset risk level, issuing a warning message, the warning message including target position information, identification results and target risk level, wherein the target position is used to indicate the current location of the ground penetrating radar.

[0019] By adopting the above technical solution, when it is determined that the target risk level of the road to be inspected reaches the preset risk level, early warning information containing target location information, identification results and target risk level can be issued in a timely manner; this technical solution improves the initiative of road detection and can quickly notify relevant personnel when the potential risk is high, so that effective measures can be taken to prevent accidents; because the early warning information contains target location information and identification results, it helps to quickly locate the problem area and understand the specific situation of the diseased body, thereby improving the efficiency and accuracy of emergency response.

[0020] In the second aspect of the present application, a road detection system based on ground penetrating radar is also provided, including: a scanning module, used to scan the road to be detected by using a ground penetrating radar to obtain radar image data; a processing module, used to preprocess the radar image data to obtain target image data; an extraction module, used to extract features from the target image data to obtain a set of features, wherein a set of features includes amplitude, frequency, phase, phase axis continuity and multiple reflection features; an identification module, used to identify a set of features to obtain an identification result, wherein the identification result is used to indicate whether there is an underground disease body on the road to be detected, and to determine the type of the underground disease body when it is determined that there is an underground disease body on the road to be detected; a determination module, used to combine the identification result and surrounding environmental factors, and use the road collapse risk assessment model to determine the target risk level of the road to be detected.

[0021] In a third aspect of the present application, an electronic device is provided, comprising a memory and a processor, wherein a computer program is stored in the memory, and the processor implements any one of the above method steps when executing the program.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is further provided. The computer-readable storage medium stores instructions. When the instructions are executed, any one of the above method steps is performed.

[0023] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: 1. It can achieve the purpose of automatic detection and risk assessment of underground road damage bodies, and use feature recognition methods to determine the presence and type of underground damage bodies, greatly reducing the workload of manual interpretation and improving detection efficiency and accuracy; 2. Risk assessment is conducted by combining identification results with surrounding environmental factors, taking into account the characteristics of the disease body and external influences, making the risk assessment more scientific and reasonable, and providing an important basis for road safety maintenance; 3. Describing the relationship between evaluation factors and evaluation levels based on membership can more accurately reflect the impact of each factor on the risk level, avoiding the errors caused by simple classification or weighted average in traditional methods, thereby improving the accuracy of road risk level assessment; 4. By dynamically adjusting the parameters of the ground penetrating radar, the interference of environmental factors on the detection signal can be reduced, the adaptability of the ground penetrating radar in different environments can be improved, and the impact of external factors on the detection results can be reduced, thereby improving the accuracy and reliability of the detection data. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of a road detection method based on ground penetrating radar provided in an embodiment of the present application; Figure 2 This is a structural block diagram of a road detection system based on ground penetrating radar provided in an embodiment of the present application; Figure 3 This is a schematic structural diagram of an electronic device disclosed in an embodiment of the present application.

[0025] Description of reference numerals: 300 - electronic device; 301 - processor; 302 - communication bus; 303 - user interface; 304 - network interface; 305 - memory. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0027] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0028] In the description of the embodiments of the present application, the term "plurality" means two or more. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of such features. The terms "include," "comprise," "have" and their variations all mean "including but not limited to," unless otherwise specifically emphasized.

[0029] The following is combined with Figure 1-Figure 3 The embodiments of the present application are described.

[0030] This application provides a road detection method based on ground penetrating radar. Figure 1 , Figure 1 : is a flow chart of a road detection method based on ground penetrating radar provided in an embodiment of the present application, the method comprising: Step S101, scanning the road to be inspected by ground penetrating radar to obtain radar image data; Step S102, preprocessing the radar image data to obtain target image data; Step S103, extracting features from the target image data to obtain a set of features, wherein the set of features includes amplitude, frequency, phase, event continuity, and multiple reflection features; Step S104: Identify the set of features to obtain an identification result, wherein the identification result is used to indicate whether an underground diseased body exists on the road to be inspected, and to determine the type of the underground diseased body if an underground diseased body exists on the road to be inspected; Step S105 , combining the recognition result and surrounding environmental factors, and using the road collapse risk assessment model to determine the target risk level of the road to be detected.

[0031] Through the above steps, the purpose of automatic detection and risk assessment of underground road damage bodies can be achieved. Specifically, by obtaining radar image data through ground-penetrating radar scanning, the geological structure information under the road can be fully captured, thereby improving the detection range and accuracy; the radar image data is preprocessed to effectively remove noise and normalize the data to ensure the accuracy of subsequent feature extraction; the amplitude, frequency, phase, event axis continuity and multiple reflection features in the target image data are extracted to provide multi-dimensional data support for the identification of underground damage bodies and improve identification reliability; the feature recognition method is used to determine the presence and type of underground damage bodies, which greatly reduces the workload of manual interpretation and improves detection efficiency and accuracy; the risk assessment is carried out by combining the identification results with surrounding environmental factors, comprehensively considering the characteristics of the damage body and external influences, making the risk assessment more scientific and reasonable, and providing an important basis for road safety maintenance.

[0032] Use ground-penetrating radar to scan the road to be inspected and obtain radar image data. Ground-penetrating radar can penetrate the ground and obtain electromagnetic wave reflection signals from underground structures. These signals can reflect the physical properties of the underground medium; pre-process the acquired radar image data, such as removing noise, enhancing signals, etc., to obtain clearer and easier-to-analyze target image data; extract a set of features from the target image data, including amplitude, frequency, phase, phase axis continuity and multiple reflection features. These features can reflect the physical characteristics of underground diseased bodies. For example, amplitude reflects the intensity of electromagnetic wave reflection, frequency indicates target depth, and also reflects the attenuation of different frequency bands, and phase characterizes interface properties and reflects the dielectric properties between the two media. The system uses the difference in properties, event continuity to determine structural integrity, and the distribution of anomalies. Multiple reflections reflect the reflective capacity of the underlying medium. The extracted features are then identified to determine the presence of underground damage bodies on the road under inspection and, if present, to determine their type. For example, machine learning or deep learning algorithms can be used to map features to damage body types by learning from a large amount of known damage body data. A road collapse risk assessment model is then used to determine the target risk level for the road under inspection, combining the identification results with surrounding environmental factors (such as building distance, underground pipeline distribution, soil type, and groundwater level). The risk assessment model comprehensively assesses the impact of damage body type, size, depth, and environmental factors on road collapse risk. This avoids the low detection efficiency and accuracy inherent in related technologies, which primarily rely on manual interpretation of geological radar images. This embodiment reduces manual intervention through automated image processing and feature recognition, can quickly process large amounts of radar image data, and improve detection efficiency. By using machine learning or deep learning algorithms to identify features, it can more accurately determine the presence and type of diseased objects, thereby improving detection accuracy. Combining the recognition results with surrounding environmental factors, a road collapse risk assessment model is used to determine the target risk level, enabling a more comprehensive and accurate assessment of road collapse risks, providing a scientific basis for road maintenance and safety management.

[0033] In an optional embodiment, the target risk level of the road to be inspected is determined using a road collapse risk assessment model in combination with the identification results and surrounding environmental factors, including: determining a weight vector corresponding to an evaluation factor set, the evaluation factor set including the type of diseased body, the scale of the diseased body, the depth of the diseased body, the distance to the building, the distribution of underground pipelines, the soil type and the groundwater level, the weight vector including the weight values ​​corresponding to each evaluation factor in the evaluation factor set, wherein the identification results include the type of diseased body, the scale of the diseased body and the depth of the diseased body, and the surrounding environmental factors include the distance to the building, the distribution of underground pipelines, the soil type and the groundwater level; constructing a fuzzy relationship matrix, and obtaining an evaluation result vector based on the weight vector and the fuzzy relationship matrix, wherein the fuzzy relationship matrix is ​​used to represent the fuzzy relationship between each evaluation factor in the evaluation factor set and each evaluation level in the evaluation level set; and determining the target risk level of the road to be inspected based on the evaluation result vector.

[0034] In the above embodiment, the road collapse risk can be assessed using weight vectors and fuzzy relationship matrices in combination with the identification results and surrounding environmental factors, thereby determining the target risk level of the road to be inspected. Specifically, the following methods are used: First, by combining the identification results of the type, scale, depth, and other aspects of the damaged body with surrounding environmental factors such as the distance to buildings and the distribution of underground pipelines, various factors affecting road collapse are comprehensively considered, thereby improving the accuracy of risk assessment; second, by using a fuzzy relationship matrix to describe the fuzzy relationship between evaluation factors and evaluation levels, the problem of inaccurate assessment caused by the complex relationship between factors in traditional methods is effectively solved; third, by calculating the evaluation result vector based on the weight vector and fuzzy relationship matrix, a quantitative analysis of the road collapse risk is achieved, providing a scientific basis for the subsequent implementation of targeted maintenance measures.

[0035] This embodiment evaluates the risk of road collapse based on the principle of Fuzzy Comprehensive Evaluation (FCE). The weight vector reflects the importance of each evaluation factor in the road collapse risk assessment. The evaluation factor set includes the type of diseased body, the scale of the diseased body, the depth of the diseased body, the distance from the building, the distribution of underground pipelines, the type of soil and the groundwater level. These factors have a comprehensive impact on the risk of road collapse. By determining the weight value of each factor, its contribution to the risk can be quantified. In this embodiment, 7 key factors that affect the risk of road collapse (type of diseased body, scale, depth, distance from the building, distribution of underground pipelines, type of soil and groundwater level) are selected and weight values ​​are assigned to each factor. For example, the analytic hierarchy process (AHP) can be used to determine the weight of each evaluation factor. Different factors have different contributions to the risk. For example, voids are more important than loose bodies. Shallow damage is more dangerous than deeper damage, and the weight vector quantifies the difference in their importance. The fuzzy relationship matrix represents the fuzzy relationship between each evaluation factor and the evaluation level, which can be low risk, medium risk, or high risk. For example, a set of evaluation levels is defined, such as low risk, medium risk, and high risk. Membership functions are established between each factor and the risk level based on expert experience or historical data. For example, the membership of a damage depth of "<1m" to high risk is 0.8. Based on the weight vector and the fuzzy relationship matrix, a fuzzy comprehensive evaluation is performed to obtain an evaluation result vector. The evaluation result vector reflects the membership of the road under inspection at each evaluation level, thereby determining its target risk level. The fuzzy matrix addresses boundary ambiguity, for example, a building distance of 10m is considered close or medium distance, making the risk classification more realistic. In practical applications, weights can be adjusted according to the environment, such as increasing the weight of groundwater levels during rainy seasons. The model can also adapt to different urban geological characteristics, such as focusing on damage types in karst areas and soil types in soft soil areas. For example, if a road is detected to have a 50cm diameter void (1m shallow), a gas pipeline within 3m of the surrounding area, and silty clay soil, a "high risk" signal is output after weighted calculation, triggering emergency repairs. Related technologies often rely on single data sources or simple assessment methods for road collapse risk assessment. These methods often fail to fully consider multiple influencing factors, resulting in insufficient accuracy and reliability in the assessment results. This embodiment can more comprehensively assess road collapse risk by comprehensively considering the type, scale, depth of the diseased body, and surrounding environmental factors (such as distance from buildings and distribution of underground pipelines). The fuzzy comprehensive evaluation method can handle the ambiguity and uncertainty of the evaluation factors. By synthesizing the weight vector and fuzzy relationship matrix, a more accurate evaluation result is obtained. The weight vector and fuzzy relationship matrix can also be adjusted according to actual conditions, making the assessment model more adaptable and flexible.

[0036] In an optional embodiment, a fuzzy relationship matrix is ​​constructed, and an evaluation result vector is obtained based on the weight vector and the fuzzy relationship matrix, including: determining the membership of each evaluation factor in the evaluation factor set to each evaluation level in the evaluation level set to obtain the fuzzy relationship matrix; multiplying the weight vector and the fuzzy relationship matrix to obtain an evaluation result vector; and determining the target risk level of the road to be inspected based on the evaluation result vector, including: determining the evaluation level corresponding to the maximum value of the element in the evaluation result vector as the target risk level.

[0037] In the above embodiment, by determining the membership of each evaluation factor in the evaluation factor set to the evaluation level set and constructing a fuzzy relationship matrix, the evaluation process is made more scientific and reasonable, and the impact of multiple factors on the risk level can be fully considered; the weight vector is multiplied by the fuzzy relationship matrix to obtain the evaluation result vector, which realizes quantitative analysis and improves the accuracy and reliability of the evaluation results; the target risk level is determined according to the evaluation level corresponding to the maximum value of the element in the evaluation result vector, which simplifies the decision-making process, ensures that the evaluation results are intuitive and clear, and facilitates the subsequent implementation of targeted measures. Describing the relationship between the evaluation factors and the evaluation level based on the membership can more accurately reflect the impact of each factor on the risk level, avoid the errors caused by simple classification or weighted average in traditional methods, and thus improve the accuracy of road risk level assessment. This embodiment can achieve a more accurate assessment of the road collapse risk level.

[0038] For each factor in the evaluation factor set (such as damage type, damage size, damage depth, and distance to buildings), its membership to each evaluation level in the evaluation level set (e.g., low risk, medium risk, and high risk) is determined. The membership reflects the degree of ambiguity of a factor at a given evaluation level and is typically determined using fuzzy mathematical methods (such as fuzzy statistics and expert scoring). The memberships of all evaluation factors to each evaluation level are combined into a matrix, known as a fuzzy relationship matrix, which reflects the fuzzy relationships between the evaluation factors and the evaluation levels. The weight vector (representing the importance of each evaluation factor) is multiplied by the fuzzy relationship matrix to produce an evaluation result vector. Each element in the evaluation result vector represents the comprehensive membership of the road under inspection at each evaluation level. The evaluation level corresponding to the maximum value of the element in the evaluation result vector is selected as the target risk level. This is because the maximum membership indicates that the evaluation level best reflects the actual situation of the road under inspection after comprehensive consideration of all factors. Related art risk assessment methods often rely on a single indicator or a simple weighted summation method, which struggles to address the ambiguity and uncertainty of the evaluation factors. The concepts of fuzzy relationship matrices and membership can effectively address the ambiguity and uncertainty of evaluation factors, making the assessment results more consistent with actual conditions. This method comprehensively considers multiple evaluation factors and their weights, avoiding the one-sided influence of a single factor on the assessment results, thereby improving the comprehensiveness and accuracy of the assessment. Furthermore, membership and weights can be adjusted based on actual conditions, making the assessment model more adaptable and flexible. For example, for each evaluation factor (such as the depth of the diseased area or soil type), a membership function is assigned to each risk level (low, medium, or high). For example, a depth of 1.5 meters has a membership of 0.6 for "high risk," 0.3 for "medium risk," and 0.1 for "low risk." The weight vector (e.g., a weight of 0.3 for the diseased area type) is then weighted and synthesized with the fuzzy relationship matrix to produce an evaluation result vector. For example, [0.2, 0.5, 0.3] indicates the highest probability of medium risk. It is important to note that membership functions can be trained based on historical data (e.g., the relationship between depth and collapse probability from 10,000 disease case studies) to avoid arbitrary artifacts. For example, if a damaged area is detected at a depth of 1.2 meters (with a high-risk membership of 0.7) and there is a building within 5 meters (with a membership of 0.4), the weight matrix will calculate the evaluation vector [0.1, 0.2, 0.7], which is then determined to be high risk, and the system will immediately issue an alert. In practice, a review mechanism can be triggered for "fuzzy areas" (such as the evaluation vector [0.3, 0.35, 0.35]) to improve reliability.

[0039] In practical applications, the analytic hierarchy process (AHP) can be used to determine the weights of each evaluation factor in the evaluation factor set, and the fuzzy comprehensive evaluation method can be used to assess the road collapse risk. The specific steps include: Constructing a hierarchical model: Decompose the road collapse risk assessment problem into a target layer, a criterion layer, and an indicator layer. The target layer is the road collapse risk assessment; the criterion layer may include aspects such as damage body characteristics and surrounding environmental factors (e.g., surrounding environmental conditions and engineering geological conditions); and the indicator layer contains specific assessment indicators for each criterion layer. For example, the damage body characteristics criterion layer may include indicators such as damage body type, size, and depth; the surrounding environmental conditions criterion layer may include indicators such as distance to surrounding buildings and underground pipeline distribution; and the engineering geological conditions criterion layer may include indicators such as soil type and groundwater level. Constructing a judgment matrix: For each factor in the criterion layer, pairwise comparisons are performed with the corresponding indicators in the indicator layer. A scaled value from 1 to 9 is assigned based on relative importance to construct a judgment matrix. For example, if damage body size is considered more important than damage body type in affecting road collapse risk, a corresponding scaled value can be assigned in the judgment matrix. Calculating the weight vector: The maximum eigenvalue of the judgment matrix and its corresponding eigenvector are calculated and consistency tested. If the test passes, the eigenvector is normalized to obtain the weight vector for each indicator relative to the criterion layer. Based on the importance of each factor in the criterion layer relative to the target layer, a judgment matrix for the criterion layer is constructed. Repeat the above steps to obtain the weight vectors of each factor in the criterion layer relative to the target layer. Finally, a weighted combination is used to obtain the comprehensive weight of each indicator relative to the target layer. For each evaluation factor, its membership function for each evaluation level is determined based on its value range and impact on road collapse risk. Through field measurements or surveys of each evaluation factor, its membership degree for each evaluation level is calculated based on the membership function, thus constructing a fuzzy relationship matrix. The comprehensive weight vectors of each indicator are combined with the fuzzy relationship matrix to obtain a comprehensive evaluation result vector. Based on the principle of maximum membership, the road collapse risk level is determined.

[0040] For example, target layer: road collapse risk assessment; Criteria layer: characteristics of the diseased body (C1), surrounding environmental conditions (C2), engineering geological conditions (C3); Indicator layer: Disease body characteristics (C1): disease body type (X1), disease body size (X2), disease body depth (X3); Surrounding environmental conditions (C2): distance to surrounding buildings (X4), underground pipeline distribution (X5); Engineering geological conditions (C3): soil type (X6), groundwater level (X7); Construct a judgment matrix: Taking the indicators under the disease body characteristic criterion layer as an example, the judgment matrix is ​​constructed as follows. The first column represents the relative importance between factor X1 and each factor X1, X2, and X3. For example, 1 can indicate that the two factors are equally important; 3 indicates that one factor is slightly more important than the other factor; 5 indicates that one factor is significantly more important than the other factor; 7 indicates that one factor is strongly more important than the other factor; 9 indicates that one factor is extremely more important than the other factor; 2, 4, 6, and 8 represent the intermediate values ​​of the above adjacent judgments.

[0041]

[0042] Calculate the weight vector: By solving the maximum eigenvalue of the above judgment matrix and its corresponding eigenvector, and performing a consistency test (assuming the test passes), the weight vector of each indicator under the disease body characteristic criterion layer is obtained, assuming: W C =(0.23, 0.64, 0.13); Similarly, the weight vector of each factor in the criterion layer relative to the target layer is: W=(0.4, 0.3, 0.3); the final comprehensive weight vector of each indicator relative to the target layer is: W X =(0.4×0.23,0.4×0.64,0.4×0.13,0.3×0.3,0.3×0.7,0.3×0.4,0.3×0.6)=(0.092,0.256,0.052,0.09,0.21,0.12,0.18).

[0043] Determine the evaluation factor set and the evaluation level set. For example, the evaluation factor set U={X1,X2,X3,X4,X5,X6,X7}, and the evaluation level set V={high risk, medium risk, low risk}. Of course, you can also set multiple evaluation levels, such as higher risk, lower risk. Then determine the membership function. Taking the scale of the disease body (X2) as an example, assuming that its value range is [0,100] (unit: cubic meters), when the scale is greater than 80 cubic meters, the membership to the high risk level is 1; when the scale is between 60-80 cubic meters, the membership to the high risk level is linearly decreasing, and the membership to the higher risk level is linearly increasing; and so on to determine the membership function of each factor to each comment level; and so on, then construct the fuzzy relationship matrix R, obtain the specific value of each evaluation factor, calculate its membership to each comment level according to the membership function, and construct the fuzzy relationship matrix R; finally, the comprehensive weight vector W X Perform synthesis operation with the fuzzy relationship matrix R, such as using the maximum-minimum synthesis method, to obtain the comprehensive evaluation result vector B, that is, B=W X *R, assuming that the obtained B=(0.45,0.33,0.22), the maximum value of 0.45 corresponds to high risk.

[0044] In an optional embodiment, the road to be inspected is scanned by a ground-penetrating radar to obtain radar image data, including: transmitting a detection signal to the road to be inspected by the ground-penetrating radar; receiving an echo signal, wherein the echo signal is a signal generated by the detection signal being transmitted through the underground medium interface of the road to be inspected; and obtaining radar image data based on the echo signal.

[0045] In the above embodiment, by transmitting a detection signal to the road to be inspected by a ground-penetrating radar, the geological structures of different depths and ranges below the road can be covered, ensuring the comprehensiveness of the detection; receiving the echo signal and generating radar image data based on it can accurately reflect the reflection characteristics of the underground medium interface, thereby providing a reliable data basis for subsequent feature extraction and identification.

[0046] Ground-penetrating radar (GPR) emits high-frequency electromagnetic pulse signals into the underground of a road, typically in the 100MHz-2.5GHz frequency range. Its penetration depth and resolution depend on the frequency: high frequencies offer high resolution but shallow penetration, and vice versa. When electromagnetic waves encounter underground medium interfaces (such as cavities or soil layer boundaries), they are reflected or refracted. The receiving antenna captures the echo signal, whose parameters, such as amplitude and phase, carry information about the medium's characteristics. For example, the amplitude of a cavity echo is strong, while that of an aquifer is reversed. Based on parameters such as the time delay and intensity of the echo signal, combined with the propagation speed of the electromagnetic wave in the medium, the depth and reflection characteristics of the underground medium interface are calculated and then converted into two-dimensional or three-dimensional radar image data, visually presenting the underground structure. The time-domain echo signals are arranged according to the scanning position, and a radar profile image (B-Scan) is formed through time-to-depth conversion (depth is calculated based on the dielectric constant of the medium). The horizontal axis represents horizontal distance, the vertical axis represents depth, and the grayscale / color represents signal strength. Related technologies also include drilling sampling, which is usually time-consuming and destructive, and cannot quickly and effectively assess the underground conditions in large areas. Ground penetrating radar technology provides an efficient and non-destructive detection method that can obtain information on underground structures without causing any damage to the road. It can complete large-area road inspections in a short period of time, significantly improving inspection efficiency.

[0047] In an optional embodiment, preprocessing the radar image data to obtain target image data includes: performing denoising and normalization processing on the radar image data to obtain target image data.

[0048] In the above embodiment, denoising and normalizing the radar image data can effectively reduce noise interference and improve data quality. At the same time, normalization helps to unify the data range, providing more reliable basic data for subsequent feature extraction and identification, thereby improving the accuracy and efficiency of road damage detection.

[0049] Image data acquired by ground-penetrating radar (GPR) is often subject to noise. This noise can originate from environmental interference, electronic noise from the device itself, and attenuation during signal transmission. Denoising is the process of removing this noise and improving image clarity and quality. Common denoising methods include filtering techniques (such as low-pass filtering, median filtering, and wavelet transforms), which can effectively remove high-frequency noise while preserving important image features. Normalization adjusts the pixel values ​​of image data to a uniform range (such as 0 to 1 or -1 to 1) to facilitate more efficient subsequent processing. Normalization eliminates differences between different data sources, making image data more consistent and comparable, and also helps improve the performance of subsequent feature extraction and recognition algorithms. Denoising effectively removes noise from images, improving image clarity and quality, and providing a more reliable data foundation for subsequent feature extraction and recognition. Normalization also unifies image data acquired by different devices into a standard range, facilitating subsequent processing and analysis, and improving data consistency and comparability. Traditional manual interpretation is susceptible to random noise, such as misidentifying noise spikes as defects. This embodiment uses adaptive denoising algorithms (such as wavelet threshold shrinkage) to automatically suppress noise, reducing false defects by over 20%. Normalization balances the dynamic range of shallow and deep layer signals, improving full-depth detection consistency. This preprocessed target image data is of higher quality, reducing misidentifications and missed detections, and improving the accuracy of defect identification.

[0050] In an optional embodiment, a set of features is identified to obtain an identification result, including: using a target recognition model to identify a set of features to obtain an identification result; wherein the target recognition model is trained in the following manner: obtaining a training sample set, wherein each training sample in the training sample set contains a set of sample features and corresponding road disease results; using the training sample set to train a network model to be trained until an end condition is met, and determining the network model obtained after the training as the target recognition model, wherein the end condition includes that the loss value of the network model to be trained meets a preset convergence condition or the number of iterations reaches a preset threshold number.

[0051] In the above-mentioned embodiment, the use of a target recognition model to identify features significantly improves the accuracy and efficiency of road damage identification. By training on a training sample set containing sample features and road damage results, the network model learns the mapping relationship between different features and damage bodies, enabling more accurate determination of the presence and type of underground damage bodies on the road under inspection. Furthermore, by setting appropriate termination conditions, the model ensures sufficient training and avoids overfitting, further enhancing the reliability of the identification results.

[0052] To obtain a training sample set, a large number of training samples containing known road damage information (such as damage type and size) are first collected. Each sample contains a set of features (such as amplitude, frequency, and phase) and the corresponding road damage result. A target recognition model is then trained using this training sample set to train a network model (such as a deep neural network). During training, network parameters are adjusted to minimize the difference between the predicted and actual results (i.e., the loss value). The training process continues until a predetermined termination condition is met, such as when the loss value reaches a preset convergence criterion or the number of iterations reaches a set maximum threshold. Once the model is trained and verified, the target recognition model is used to analyze the feature data of the road to be inspected and automatically output recognition results, including whether a damage has been found and the specific type of the damage. In this way, the target recognition model is fed into the target recognition model. Based on the relationship between the learned features and the damage, the model outputs recognition results such as the damage type, size, and depth. Through massive sample training, the model can learn complex disease characteristic patterns and reduce subjective misjudgments, especially significantly enhancing the ability to identify tiny or hidden disease bodies; the deep learning model has strong generalization capabilities and can adapt to detection data under different geological conditions and radar equipment parameters, reducing the interference of environmental factors on the results.

[0053] The characteristics of a cavity or void include phase: negative, positive, and negative (black, white, and black); amplitude: relatively strong, clearly different from the background field; continuity of the event axis: relatively good continuity; generally, the upper end of the reaction is relatively smooth, with edge reactions at both ends (tailing obliquely downward); multiple reflections: most voids do not have multiple reflections; when the bottom of the cavity is water or a strong reflector, there will be multiple reflections, and based on the bottom reflection signal, the cavity can be inferred to be clear. For example, for a cavity, the frequency characteristic is typically characterized by attenuation of high-frequency signals and enhancement of low-frequency signals. This is because high-frequency signals are more easily absorbed when propagating in air (cavities), while low-frequency signals are relatively stable. The amplitude characteristic is characterized by a higher amplitude of the reflected signal from a cavity, due to the large difference in dielectric constant between air and the surrounding medium, resulting in a high reflectivity. The phase characteristic is characterized by the phase of the reflected signal from a cavity being consistent with the incident wave, as the electromagnetic wave passes from a medium with a high dielectric constant (such as soil) to a medium with a low dielectric constant. For example, assuming a 400MHz ground-penetrating radar antenna, the reflected signal from a cavity in the frequency domain manifests as an enhancement of low-frequency components, with a higher amplitude and a phase consistent with the incident wave. For water-rich areas, the frequency characteristic is characterized by a smaller frequency variation of the reflected signal, but an enhanced amplitude due to the high dielectric constant of water. Water has a high dielectric constant and a strong ability to reflect electromagnetic waves. The amplitude characteristic is that the amplitude of the reflected signal in water-rich areas is usually stronger because water has a high reflectivity. The phase characteristic is that the phase of the reflected signal in water-rich areas is opposite to that of the incident wave because the electromagnetic wave enters a medium with a large dielectric constant (such as water) from a medium with a small dielectric constant (such as air). For example, using a 400MHz ground penetrating radar antenna, the reflected signal in the water-rich area is characterized by a small frequency change, a high amplitude, and a phase opposite to that of the incident wave in the frequency domain.

[0054] In an optional embodiment, before scanning the road to be inspected by the ground penetrating radar, the above method also includes: using a group of sensors to collect the working environment parameters of the ground penetrating radar in real time, wherein the working environment parameters include soil moisture and electromagnetic interference intensity; according to the working environment parameters, automatically adjusting the transmission frequency, power and sampling rate of the ground penetrating radar through a preset algorithm.

[0055] In the above embodiment, sensors can be used to collect real-time environmental parameters, including soil moisture and electromagnetic interference intensity, before the ground-penetrating radar (GPR) scans a road. Based on these parameters, a pre-set algorithm automatically adjusts the GPR's transmission frequency, power, and sampling rate. Dynamically adjusting the GPR's parameters reduces environmental interference with detection signals, improves the GPR's adaptability to various environments, and mitigates the impact of external factors on detection results, thereby enhancing the accuracy and reliability of detection data.

[0056] This embodiment adds adaptive parameter adjustment for ground-penetrating radar (GPR). A set of sensors collects GPR operating environment parameters in real time, including soil moisture and electromagnetic interference intensity. These parameters significantly influence radar signal propagation and reflection. For example, soil moisture affects the propagation speed and attenuation of electromagnetic waves. Based on these collected operating environment parameters, a preset algorithm automatically adjusts the GPR's transmission frequency, power, and sampling rate. For example, when soil moisture is high, the transmission frequency may need to be adjusted to accommodate higher electromagnetic wave attenuation; when electromagnetic interference intensity is high, the transmission power may need to be increased to improve the signal-to-noise ratio. Automatic adjustments are made based on a preset algorithm (such as a lookup table or fuzzy logic controller). For example, for high-moisture soil (high attenuation), the frequency is reduced to increase penetration depth; for high-EMI environments, the power is increased to improve the signal-to-noise ratio; and for shallow-seated disease detection, the sampling rate is increased to enhance resolution.

[0057] In an optional embodiment, after determining the target risk level of the road to be inspected using a road collapse risk assessment model, the above method further includes: when it is determined that the target risk level is a preset risk level, issuing a warning message, the warning message including target position information, identification results and target risk level, wherein the target position is used to indicate the current location of the ground penetrating radar.

[0058] In the above embodiment, when it is determined that the target risk level of the road to be inspected reaches the preset risk level, early warning information including the target location, identification results and target risk level can be issued in a timely manner; this embodiment improves the initiative of road detection and can quickly notify relevant personnel when the potential risk is high, so that effective measures can be taken to prevent accidents; since the early warning information contains the target location and identification results, it helps to quickly locate the problem area and understand the specific situation of the diseased body, thereby improving the efficiency and accuracy of emergency response.

[0059] After using a road collapse risk assessment model to determine the target risk level for the road to be inspected, if the target risk level reaches a preset level, the system automatically issues an early warning message. This warning message includes the target location (the current location of the ground-penetrating radar), identification results (damage type and size, etc.), and the target risk level. For example, the current detection location can be obtained using the ground-penetrating radar's built-in positioning module (such as GPS or Beidou). Identification results include the damage type (such as a cavity), size (such as a diameter of 50 cm), depth (such as 1.2 m), and a high risk level. This information provides road management departments with detailed information on the location and risk of the hazard. This embodiment enables an automatic early warning mechanism, shortening the time between detection and response. In particular, minute-level warnings can be achieved for high-risk sections, reducing the probability of collapse accidents. The warning information, including location, damage type, and risk level, avoids the subjectivity and information loss associated with manual reporting, providing comprehensive data support for emergency decision-making. Optionally, the warning information can be integrated with a geographic information system (GIS) to enable visual management of risky sections, assisting departments in formulating comprehensive maintenance plans and promoting intelligent road management upgrades.

[0060] As an optional implementation, a Geographic Information System (GIS) map can be generated based on the target location information to visualize the detection data. This GIS map can be used to analyze the distribution of collapse risks in different geographic locations and identify high-risk areas. Furthermore, based on historical and real-time data from high-risk areas, targeted early warning information can be generated and distributed to relevant management departments.

[0061] This application also provides a road detection system based on ground penetrating radar, such as Figure 2 As shown, Figure 2 This is a structural block diagram of a road detection system based on ground penetrating radar provided in an embodiment of the present application, which includes: A scanning module 21 is used to scan the road to be inspected using a ground penetrating radar to obtain radar image data; The processing module 22 is used to pre-process the radar image data to obtain target image data; An extraction module 23 is used to extract features from the target image data to obtain a set of features, wherein the set of features includes amplitude, frequency, phase, event continuity and multiple reflection features; The identification module 24 is used to identify a set of features and obtain an identification result, wherein the identification result is used to indicate whether an underground disease body exists on the road to be inspected, and to determine the type of the underground disease body if an underground disease body exists on the road to be inspected; The determination module 25 is used to determine the target risk level of the road to be detected by combining the recognition results and surrounding environmental factors using the road collapse risk assessment model.

[0062] In an optional embodiment, the above-mentioned determination module 25 includes: a first determination unit, used to determine the weight vector corresponding to the evaluation factor set, the evaluation factor set includes the type of diseased body, the scale of the diseased body, the depth of the diseased body, the distance from the building, the distribution of underground pipelines, the soil type and the groundwater level, and the weight vector includes the weight value corresponding to each evaluation factor in the evaluation factor set, wherein the identification result includes the type of diseased body, the scale of the diseased body and the depth of the diseased body, and the surrounding environmental factors include the distance from the building, the distribution of underground pipelines, the soil type and the groundwater level; a construction unit, used to construct a fuzzy relationship matrix, and obtain an evaluation result vector based on the weight vector and the fuzzy relationship matrix, wherein the fuzzy relationship matrix is ​​used to represent the fuzzy relationship between each evaluation factor in the evaluation factor set and each evaluation level in the evaluation level set; a second determination unit, used to determine the target risk level of the road to be inspected based on the evaluation result vector.

[0063] In an optional embodiment, the above-mentioned construction unit includes: a first determination subunit, used to determine the membership of each evaluation factor in the evaluation factor set to each evaluation level in the evaluation level set to obtain a fuzzy relationship matrix; an acquisition subunit, used to multiply the weight vector with the fuzzy relationship matrix to obtain an evaluation result vector; the above-mentioned second determination unit includes: a second determination subunit, used to determine the evaluation level corresponding to the maximum value of the element in the evaluation result vector as the target risk level.

[0064] In an optional embodiment, the above-mentioned scanning module 21 includes: a transmitting unit for transmitting a detection signal to the road to be detected through a ground penetrating radar; a receiving unit for receiving an echo signal, wherein the echo signal is a signal generated by the detection signal being transmitted through the underground medium interface of the road to be detected; and an obtaining unit for obtaining radar image data based on the echo signal.

[0065] In an optional embodiment, the processing module 22 includes: a processing unit configured to perform denoising and normalization processing on the radar image data to obtain target image data.

[0066] In an optional embodiment, the above-mentioned recognition module 24 includes: a recognition unit, which is used to use a target recognition model to identify a set of features and obtain a recognition result; wherein, the target recognition model is trained in the following manner: obtaining a training sample set, wherein each training sample in the training sample set contains a set of sample features and corresponding road disease results; using the training sample set to train the network model to be trained until the end condition is met, and determining the network model obtained after the training as the target recognition model, wherein the end condition includes that the loss value of the network model to be trained meets the preset convergence condition or the number of iterations reaches a preset threshold.

[0067] In an optional embodiment, the above-mentioned system also includes: an acquisition module, which is used to use a group of sensors to collect the working environment parameters of the ground penetrating radar in real time before scanning the road to be inspected by the ground penetrating radar, wherein the working environment parameters include soil moisture and electromagnetic interference intensity; an adjustment module, which is used to automatically adjust the transmission frequency, power and sampling rate of the ground penetrating radar through a preset algorithm according to the working environment parameters.

[0068] In an optional embodiment, the above-mentioned system also includes: an early warning module, which is used to determine the target risk level of the road to be inspected using the road collapse risk assessment model, and when the target risk level is determined to be a preset risk level, issue an early warning message, wherein the early warning message includes the target position, identification result and target risk level, wherein the target position is used to indicate the current location of the ground penetrating radar.

[0069] It should be noted that the above embodiments provide devices that implement their functions using only the division of the above functional modules as examples. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0070] The present application also provides a computer-readable storage medium, which stores instructions. When the instructions are executed, any one of the above-mentioned method steps is executed.

[0071] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0072] This application also discloses an electronic device. Figure 3 As shown, Figure 3 The electronic device 300 may include: at least one processor 301 , at least one communication bus 302 , a user interface 303 , at least one network interface 304 , and a memory 305 .

[0073] The communication bus 302 is used to implement the connection and communication between these components.

[0074] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0075] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0076] The processor 301 may include one or more processing cores. The processor 301 utilizes various interfaces and circuits to connect various components within the electronic device (e.g., a server). It executes instructions, programs, code sets, or instruction sets stored in the memory 305 and accesses data stored in the memory 305 to perform various server functions and process data. Optionally, the processor 301 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 301 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing content displayed on the display; and the modem handles wireless communications. It is understood that the modem may also be implemented as a separate chip, rather than integrated into the processor 301.

[0077] Among them, the memory 305 may include a random access memory (RAM) or a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may also optionally be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 , the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a road detection method based on ground penetrating radar.

[0078] exist Figure 3 In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call an application program of a road detection method based on ground penetrating radar stored in the memory 305. When executed by one or more processors 301, the electronic device 300 executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should know that this application is not limited to the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for this application.

[0079] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0080] In the several embodiments provided in this application, it should be understood that the disclosed device or system can be implemented in other ways. For example, the device or system embodiments described above are merely schematic, such as the division of units, which is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0081] Units described as separate components may or may not be physically separate, and components shown 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 these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0082] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0083] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of this application. The aforementioned memory includes various media that can store program code, such as USB flash drives, mobile hard drives, magnetic disks, or optical disks.

[0084] The foregoing is merely an exemplary embodiment of the present disclosure and is not intended to limit the scope of the present disclosure. In other words, any equivalent variations and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the disclosure herein.

[0085] This application is intended to cover any modifications, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the technical field not described in the present disclosure.

Claims

1. A road detection method based on ground penetrating radar, characterized in that: include: Scan the road to be inspected by ground penetrating radar to obtain radar image data; Preprocessing the radar image data to obtain target image data; Performing feature extraction on the target image data to obtain a set of features, wherein the set of features includes amplitude, frequency, phase, event continuity, and multiple reflection features; Identifying the set of features to obtain an identification result, wherein the identification result is used to indicate whether an underground diseased body exists on the road to be inspected, and determining the type of the underground diseased body if the underground diseased body exists on the road to be inspected; In combination with the identification results and surrounding environmental factors, a road collapse risk assessment model is used to determine the target risk level of the road to be detected.

2. The method according to claim 1, characterized in that Combining the identification results and surrounding environmental factors, a road collapse risk assessment model is used to determine the target risk level of the road to be detected, including: Determining a weight vector corresponding to an evaluation factor set, wherein the evaluation factor set includes the type of the diseased body, the size of the diseased body, the depth of the diseased body, the distance to the building, the distribution of underground pipelines, the soil type, and the groundwater level, and the weight vector includes the weight value corresponding to each evaluation factor in the evaluation factor set, wherein the identification result includes the type of the diseased body, the size of the diseased body, and the depth of the diseased body, and the surrounding environmental factors include the distance to the building, the distribution of underground pipelines, the soil type, and the groundwater level; Constructing a fuzzy relationship matrix, and obtaining an evaluation result vector based on the weight vector and the fuzzy relationship matrix, wherein the fuzzy relationship matrix is ​​used to represent the fuzzy relationship between each evaluation factor in the evaluation factor set and each evaluation grade in the evaluation grade set; The target risk level of the road to be inspected is determined based on the evaluation result vector.

3. The method according to claim 2, characterized in that Constructing a fuzzy relationship matrix and obtaining an evaluation result vector based on the weight vector and the fuzzy relationship matrix, comprising: determining the membership of each evaluation factor in the evaluation factor set to each evaluation level in the evaluation level set to obtain the fuzzy relationship matrix; multiplying the weight vector and the fuzzy relationship matrix to obtain the evaluation result vector; Determining the target risk level of the road to be inspected based on the evaluation result vector includes: determining the evaluation level corresponding to the maximum value of the elements in the evaluation result vector as the target risk level.

4. The method according to claim 1, wherein The ground penetrating radar is used to scan the road to be inspected to obtain radar image data, including: transmitting a detection signal to the road to be detected by the ground penetrating radar; receiving an echo signal, wherein the echo signal is a signal generated when the detection signal is transmitted through an underground medium interface of the road to be detected; The radar image data is obtained according to the echo signal.

5. The method according to claim 1, characterized in that Preprocessing the radar image data to obtain target image data includes: Denoising and normalization processing are performed on the radar image data to obtain the target image data.

6. The method according to claim 1, wherein Identifying the set of features to obtain an identification result includes: Using a target recognition model to identify the set of features to obtain the recognition result; The target recognition model is trained in the following way: Acquire a training sample set, wherein each training sample in the training sample set includes a set of sample features and corresponding road disease results; The network model to be trained is trained using the training sample set until an end condition is met, and the network model obtained after the training is determined as the target recognition model, wherein the end condition includes that the loss value of the network model to be trained meets a preset convergence condition or the number of iterations reaches a preset threshold.

7. The method according to claim 1, characterized in that Before scanning the road to be inspected by the ground penetrating radar, the method further includes: Using a group of sensors to collect working environment parameters of the ground penetrating radar in real time, wherein the working environment parameters include soil moisture and electromagnetic interference intensity; According to the working environment parameters, the transmission frequency, power and sampling rate of the ground penetrating radar are automatically adjusted through a preset algorithm.

8. A road detection system based on ground penetrating radar, characterized in that: include: A scanning module is used to scan the road to be inspected by using a ground-penetrating radar to obtain radar image data; a processing module, configured to pre-process the radar image data to obtain target image data; an extraction module, configured to perform feature extraction on the target image data to obtain a set of features, wherein the set of features includes amplitude, frequency, phase, event continuity, and multiple reflection features; an identification module, configured to identify the set of features and obtain an identification result, wherein the identification result is used to indicate whether an underground diseased body exists on the road to be inspected, and to determine the type of the underground diseased body if the underground diseased body exists on the road to be inspected; The determination module is used to determine the target risk level of the road to be detected by combining the recognition result and surrounding environmental factors and using a road collapse risk assessment model.

9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images