Road disease automatic detection method and system based on image recognition

Abstract

The invention relates to the technical field of road management, in particular to an automatic road disease detection method and system based on image recognition, which improve the accuracy and efficiency of road disease detection and ensure the safety and smoothness of urban traffic. The method comprises the steps of collecting a road disease growth trend sequence; inputting the road disease growth trend sequence into a pre-trained road disease hazard assessment model to obtain a first-order road disease hazard level; in response to the situation that the first-order road disease hazard level is smaller than a preset hazard level threshold, collecting construction feature information of the position where the road disease is located; identifying and analyzing the construction feature information by using a pre-trained construction feature influence analysis model to obtain a road construction evaluation factor; evaluating and correcting the first-order road disease hazard grade by using the road construction evaluation factor to obtain a second-order road disease hazard grade; and comparing the second-order road disease hazard grade with a preset hazard grade threshold.

Description

Technical Field

[0001] The present invention relates to the technical field of road management, and in particular to an automatic road disease detection method and system based on image recognition. Background Art

[0002] With the acceleration of urbanization, roads, as an important part of urban infrastructure, are directly related to the smoothness of urban traffic and the safety of citizens' travel. However, various defects such as cracks, potholes, and peeling will inevitably appear in the process of road use. These defects not only affect the flatness and driving comfort of the road, but also may pose a safety hazard to vehicles and pedestrians. In the past, road defect detection mainly relied on manual inspections, which is not only inefficient, but also easily affected by human factors, resulting in inaccurate or missed detection results. In recent years, automatic road defect detection technology based on image recognition has gradually become a research hotspot.

[0003] However, existing automatic detection methods for road defects based on image recognition mostly remain at the simple identification and classification of defects, lacking the assessment of the growth trend of defects; at the same time, they ignore the impact of road construction, such as road structure, materials, and service life, on road defects. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method and system for automatic detection of road defects based on image recognition, which improves the accuracy and efficiency of road defect detection and ensures the safety and smoothness of urban traffic.

[0005] In a first aspect, the present invention provides a method for automatically detecting road damage based on image recognition, the method comprising:

[0006] Collect road disease growth trend series;

[0007] The road disease growth trend sequence is input into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level;

[0008] In response to the first-order road disease hazard level being less than a preset hazard level threshold, collecting construction characteristic information of a location where the road disease is located;

[0009] Use the pre-trained construction feature impact analysis model to identify and analyze construction feature information to obtain road construction evaluation factors;

[0010] The first-order road hazard level is evaluated and corrected using the road construction evaluation factor to obtain the second-order road hazard level.

[0011] The second-order road disease hazard level is compared with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately.

[0012] Furthermore, the method for collecting the road disease growth trend sequence includes:

[0013] Collect road disease image information when the road disease is first discovered;

[0014] Extract features from road damage image information to obtain a road damage feature set;

[0015] According to the preset acquisition frequency, multiple image acquisition and feature extraction are performed on the road damage at the same location to obtain multiple road damage feature sets;

[0016] Arrange multiple road disease feature sets in chronological order to obtain a road disease growth trend sequence.

[0017] Furthermore, the method for setting the acquisition frequency includes:

[0018] Quantifying the road damage feature set obtained for the first time to obtain a road damage feature index vector;

[0019] The road disease feature index vector is used to traverse the pre-built frequency reference library to extract the acquisition frequency corresponding to the road disease feature index vector.

[0020] Furthermore, the method for constructing the road hazard assessment model includes:

[0021] Collecting series data on the growth trend of road diseases;

[0022] Perform pre-processing operations such as cleaning, denoising and standardization on the collected data;

[0023] Selecting a machine learning model as the infrastructure of the road hazard assessment model, wherein the machine learning model includes a random forest, a gradient boosting tree, a convolutional neural network, and a recurrent neural network;

[0024] Use the preprocessed data to train the model. During the training process, adjust the parameters and structure of the model to optimize its performance.

[0025] Use an independent validation set to validate the trained model and evaluate its performance in real applications;

[0026] Tune the model based on the validation results;

[0027] Deploy the trained and verified models to actual application scenarios.

[0028] Furthermore, the construction characteristic information includes road structure, service life, traffic flow, environmental factors and maintenance history.

[0029] Furthermore, the method for obtaining the second-order road disease hazard level includes:

[0030] Based on the first-order road hazard level and the construction characteristics of the road hazard location, their impact on the growth trend and hazard level of road hazard is evaluated;

[0031] Setting correction thresholds for different construction features;

[0032] comparing the road construction characteristic information with a set correction threshold;

[0033] According to the comparison results, the first-order road disease hazard level is adjusted;

[0034] After applying the correction criteria for all relevant construction characteristics, the second-order road disease hazard level is obtained by integrating all factors.

[0035] Furthermore, the influencing factors for setting the preset hazard level threshold include road type, road material, traffic flow, road age and environmental conditions.

[0036] On the other hand, the present application also provides an automatic road disease detection system based on image recognition, the system comprising:

[0037] Data collection module, collecting road disease growth trend sequences;

[0038] The preliminary disease assessment module inputs the road disease growth trend sequence into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level;

[0039] A construction characteristic information acquisition module, in response to the first-order road disease hazard level being less than a preset hazard level threshold, collects construction characteristic information of the location of the road disease;

[0040] The construction feature impact analysis module uses the pre-trained construction feature impact analysis model to identify and analyze construction feature information and obtain road construction evaluation factors;

[0041] The disease hazard level correction module uses the road construction evaluation factor to evaluate and correct the first-order road disease hazard level to obtain the second-order road disease hazard level;

[0042] The decision output module compares the second-order road disease hazard level with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately.

[0043] In a third aspect, the present application provides an electronic device comprising a bus, a transceiver, a camera, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and the computer program implements the steps of any one of the above methods when executed by the processor.

[0044] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps in any one of the above-mentioned methods when executed by a processor.

[0045] Compared with the prior art, the invention has the following beneficial effects: the method not only focuses on the simple identification and classification of road diseases, but also deeply evaluates the growth trend of diseases; by collecting the disease growth trend sequence, it can more comprehensively understand the development of diseases and provide a more sufficient basis for subsequent maintenance decisions; it introduces the consideration of road construction characteristics, and through the construction characteristics influence analysis model, it can more accurately evaluate the hazard level of road diseases and improve the accuracy and reliability of the evaluation;

[0046] By combining the road construction evaluation factors to correct the first-order disease hazard level, the second-order disease hazard level is obtained. This process realizes the dynamic adjustment and optimization of the evaluation results; it helps to more accurately reflect the actual situation of road diseases and provide a more scientific basis for maintenance decisions;

[0047] Compared with the traditional manual inspection method, this method, based on image recognition technology, can realize automatic detection of road diseases, greatly improving the detection efficiency; at the same time, by introducing advanced data processing and analysis models, it reduces the interference of human factors and improves the accuracy and reliability of the detection results;

[0048] The method can classify and treat road damage according to the evaluation results, continuously observe the damage in a normal development state, and promptly issue an alarm for the damage that needs immediate repair, thus providing strong support for road management departments.

[0049] In summary, this method improves the accuracy and efficiency of road damage detection, and also provides road management departments with more scientific and reasonable decision-making support to ensure the safety and smoothness of urban traffic. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of the present invention;

[0051] Figure 2 It is a flow chart of the method for obtaining the second-order road disease hazard level;

[0052] Figure 3 This is a structural diagram of the automatic road damage detection system based on image recognition. DETAILED DESCRIPTION

[0053] In the description of the present application, those skilled in the art should know that the present application can be implemented as a method, an apparatus, an electronic device, and a computer-readable storage medium. Therefore, the present application can be specifically implemented in the following forms: complete hardware, complete software (including firmware, resident software, microcode, etc.), a combination of hardware and software. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer-readable storage media, and the computer-readable storage medium contains computer program code.

[0054] The above-mentioned computer-readable storage medium may adopt any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, optical disc read-only memories, optical storage devices, magnetic storage devices, or any combination of the above. In the present application, computer-readable storage media can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0055] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws.

[0056] The present application describes the provided methods, devices, and electronic devices through flowcharts and / or block diagrams.

[0057] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine, and these computer-readable program instructions are executed by a computer or other programmable data processing device to produce a device that implements the functions / operations specified by the boxes in the flowchart and / or block diagram.

[0058] These computer-readable program instructions may also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to work in a specific manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction device product including functions / operations specified in the blocks in the flowchart and / or block diagram.

[0059] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby enabling the instructions executed on the computer or other programmable data processing apparatus to provide a process for implementing the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0060] The present application is described below in conjunction with the drawings in the present application.

[0061] Embodiment 1: Figures 1 to 3 As shown, the automatic road damage detection method based on image recognition of the present invention specifically includes the following steps:

[0062] S1. Collect the growth trend sequence of road damage;

[0063] Methods for collecting road disease growth trend series include:

[0064] Use high-resolution camera equipment to conduct initial inspections of roads to ensure that subtle road defects can be captured. The camera equipment includes high-definition cameras on drones and high-definition cameras on vehicles. Choose clear weather and sufficient light to avoid shadows and reflections that affect image quality. Record the location information of the defects, including latitude and longitude coordinates and reference objects, for subsequent follow-up inspections. Take photos of the same defect area from multiple angles to more comprehensively assess the defect status.

[0065] Use image recognition algorithms to automatically extract features from disease images and extract a feature set that can characterize the morphology, size, distribution and other information of the disease; standardize the extracted features to eliminate differences caused by different acquisition conditions and ensure the consistency and comparability of the features;

[0066] According to factors such as road type, traffic flow and disease type, a reasonable collection frequency can be preset; for areas or sections where diseases develop rapidly, the number of collections can be appropriately increased; regular image collection of road diseases at the same location is carried out according to the preset collection frequency; ensure that the parameters of the camera equipment remain consistent during each collection to reduce errors; extract features from each collected disease image and compare and analyze it with the previous feature set; by comparing the feature changes at different time points, the growth trend of the disease can be evaluated;

[0067] Add a timestamp to each feature set to record the collection time; sort the feature sets according to the timestamp to form a time series; analyze the changing trend of the disease over time, display the trend series in the form of a chart, and construct a road disease growth trend series to help intuitively understand the development of the disease.

[0068] In this step, the use of high-resolution camera equipment and multi-angle shooting strategies can capture subtle changes in road diseases, which not only improves the accuracy of disease detection, but also helps to identify and intervene in the early stages of the disease to prevent it from further deteriorating; by presetting a reasonable acquisition frequency and ensuring the consistency of camera equipment parameters, it is possible to regularly obtain image data of road diseases at the same location, and perform continuous feature extraction and comparative analysis; it not only helps to capture the entire process of disease growth, but also makes data at different time points highly comparable, providing a reliable basis for evaluating the growth trend of the disease; by comparing the feature changes at different time points, the growth rate and trend of the disease can be quantitatively evaluated; quantitative evaluation not only helps to more accurately judge the degree of harm of the disease, but also provides a scientific basis for formulating maintenance plans; at the same time, displaying the trend sequence in the form of a chart can intuitively reflect the development of the disease, which is convenient for managers to understand and make decisions; automated feature extraction and standardized data processing procedures greatly reduce the impact of manual participation and subjective judgment, and improve the efficiency and accuracy of disease management; in addition, by constructing a road disease growth trend sequence, potential safety hazards can be discovered and warned in a timely manner to ensure the safety and smoothness of road traffic.

[0069] S2, inputting the road disease growth trend sequence into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level;

[0070] The construction method of the road hazard assessment model includes:

[0071] Collect road damage growth trend series data, covering the changes of different types of road damage in different time periods;

[0072] Clean, denoise and standardize the collected data to ensure data quality and consistency, including removing invalid data, correcting erroneous data, and converting data into a unified format and unit;

[0073] Selecting a machine learning model as the infrastructure of the road hazard assessment model, wherein the machine learning model includes a random forest, a gradient boosting tree, a convolutional neural network, and a recurrent neural network;

[0074] Use the preprocessed data to train the model. During the training process, continuously adjust the model's parameters and structure to optimize its performance. This includes selecting appropriate hyperparameters, applying regularization techniques to prevent overfitting, and using early stopping to avoid long training times. The training process also needs to be monitored and evaluated. The generalization ability of the model is evaluated through methods such as cross-validation to ensure that the model can maintain good performance on unseen data.

[0075] Use an independent validation set to validate the trained model and evaluate its performance in real applications;

[0076] Tune the model based on the verification results; if the model performance is poor, try to adjust the model structure, change the algorithm, and increase the training data to improve its performance;

[0077] Deploy the trained and verified models to actual application scenarios.

[0078] In this step, the automated road hazard assessment model can significantly improve the efficiency of road hazard detection and reduce the time and labor costs of manual inspections. At the same time, the model is based on a large amount of data and advanced algorithms, which can more accurately identify and assess the hazard level of road hazards and avoid errors and omissions caused by human factors. The model comprehensively considers the growth trend of road hazards and road construction characteristics, providing a scientific decision-making basis for road management departments. By assessing the hazard level of road hazards, management departments can more accurately formulate maintenance plans and budgets, reasonably allocate resources, and ensure the safety and durability of roads. The trained model is deployed in actual application scenarios to achieve real-time monitoring and early warning of road hazards. The construction and application process of the road hazard assessment model promotes the innovation and upgrading of related technologies. By continuously collecting new data, optimizing model structure and algorithms, the performance and generalization ability of the model can be continuously improved, promoting technological progress in the field of road management.

[0079] S3, in response to the first-order road disease hazard level being less than a preset hazard level threshold, collecting construction feature information of the location of the road disease;

[0080] The construction feature information includes:

[0081] Road structure: including road type, pavement structure and roadbed conditions;

[0082] Age: how long the road has been built and when it was last overhauled or maintained;

[0083] Traffic volume: average daily traffic volume on the road section, as well as the volume of heavy vehicles;

[0084] Environmental factors: climatic conditions, groundwater levels, and soil types in the area;

[0085] Maintenance history: a record of maintenance activities carried out on the road section in the past few years;

[0086] The factors affecting the setting of the preset hazard level threshold include:

[0087] Road type: Different types and grades of roads differ in design standards, traffic flow, vehicle types, etc., so their tolerance for damage and safety requirements are also different; high-grade roads tend to have a lower tolerance for damage and require stricter hazard level thresholds;

[0088] Road materials: The materials and design of roads directly affect their durability and the probability of disease occurrence; for roads with high-performance materials or excellent designs, the hazard level threshold can be appropriately increased;

[0089] Traffic volume: Roads with heavy traffic and heavy vehicles are more susceptible to wear and tear, so lower hazard level thresholds are needed to detect and treat damage in a timely manner;

[0090] Road age: The longer a road is used, the more its performance deteriorates and the more likely it is to develop problems. At the same time, roads with a good maintenance history can slow down the development of problems. Therefore, the age and maintenance status of roads need to be considered when setting hazard level thresholds.

[0091] Environmental conditions: Climate and environmental conditions have a significant impact on the development of road hazards; in severe climate or unfavorable environment, lower hazard level thresholds need to be set to address potential damage risks.

[0092] In this step, by collecting construction feature information at the location of road defects, the assessment of road defects is no longer limited to simple identification and classification, but goes deep into a comprehensive analysis of the background of the occurrence, development trends and potential risks of the defects; it provides a more scientific and accurate basis for subsequent disease treatment decisions; the setting of the preset hazard level threshold takes into account multiple factors, so that the threshold can be dynamically adjusted and optimized according to the actual conditions of different roads, ensuring the applicability and effectiveness of the disease detection and assessment process; through the judgment of the preset hazard level threshold, only when the first-order road disease hazard level is lower than the threshold, will the construction feature information be further collected for in-depth analysis, effectively avoiding unnecessary waste of resources; this step combines image recognition technology with road management practice, and promotes the development of road management towards intelligence through automated data collection, analysis and decision support; this not only improves the efficiency and accuracy of road management, but also lays the foundation for the construction of future smart cities.

[0093] S4. Using the pre-trained construction feature impact analysis model to identify and analyze the construction feature information, and obtain the road construction evaluation factor;

[0094] Input the construction feature information of the location of the road damage to be analyzed into the trained construction feature impact analysis model;

[0095] The model calculates and outputs road construction evaluation factors based on the input construction feature information, which is used to quantify the impact of construction features on the growth trend of road diseases;

[0096] Based on the road construction evaluation factors output by the model, the construction characteristic information is deeply analyzed to understand which factors are the main driving forces of the current road disease growth trend and how these factors interact with each other;

[0097] Input layer: Input features include various construction feature information of the location of road damage, including road structure type, material properties, service life, traffic flow, design load, drainage system status and maintenance records;

[0098] Embedding layer: If the input features contain categorical data, the embedding layer is used to convert these categorical data into dense vector representations; the embedding layer can learn the potential semantic relationship of each category and help the model better understand and utilize this categorical information;

[0099] Hidden layer: A multi-layer neural network is used to extract the complex relationship between input features. Each layer of the network introduces nonlinearity through the activation function, so that the model can learn and express complex mapping relationships.

[0100] Output layer: used to output road construction evaluation factors; road construction evaluation factors are used to quantify the impact of construction characteristics on the growth trend of road diseases.

[0101] In this step, by using the pre-trained construction feature impact analysis model, this step can accurately quantify the impact of road construction features on the growth trend of disease; the output of road construction evaluation factors provides a scientific basis for subsequent disease assessment, making the assessment results more objective and accurate; the model not only outputs road construction evaluation factors, but also uses these factors to deeply analyze which construction features are the main driving force of the current road disease growth trend and how these factors interact with each other; it helps road managers better understand the causes of diseases and thus formulate more targeted repair and maintenance strategies; compared with traditional manual inspection methods, methods based on image recognition and machine learning have greatly improved the efficiency of road disease detection and further accelerated the process of disease assessment and maintenance decision-making. The output of road construction evaluation factors provides strong decision-making support for road managers. Based on these factors, managers can evaluate the potential effects of different maintenance plans and select the best plan for implementation. This helps optimize the allocation of maintenance resources and improve the efficiency of capital use. By timely and accurately identifying and analyzing road diseases and their causes, road managers can promptly discover potential safety hazards and take measures to resolve them. This helps improve the overall safety and traffic capacity of roads and ensure the safety of citizens' travel. The implementation of this step promotes the intelligentization of road management. By applying machine learning and image recognition technology to the field of road disease detection, real-time monitoring and accurate evaluation of road conditions are achieved.

[0102] S5. Use the road construction evaluation factor to evaluate and correct the first-order road disease hazard level to obtain the second-order road disease hazard level;

[0103] The method for obtaining the second-order road disease hazard level includes:

[0104] Based on the first-order road hazard level and the construction characteristics of the road hazard location, their potential impact on the growth trend and hazard level of road hazard is evaluated;

[0105] Based on the professional knowledge and experience in the field of road management, correction thresholds are set for different construction characteristics;

[0106] comparing the road construction characteristic information with a set correction threshold;

[0107] According to the comparison results, the first-order road disease hazard level is adjusted up or down; if the road service life is far beyond the standard, the hazard level is increased by one level; if the road material has good resistance to the disease, the hazard level is maintained or slightly reduced;

[0108] After applying the correction criteria for all relevant construction characteristics, the final second-order road disease hazard level is obtained by taking all factors into consideration.

[0109] In this step, by introducing road construction characteristic information and correcting the first-order road disease hazard level, it can more comprehensively reflect the actual situation and potential risks of road diseases, thereby improving the accuracy of the assessment; based on the assessment results of the second-order road disease hazard level, the road management department can formulate a more scientific and reasonable maintenance or repair strategy; for sections with higher hazard levels, maintenance resources can be arranged in priority to ensure road safety; for sections with lower hazard levels, a continuous observation strategy can be adopted to reasonably allocate resources; through accurate disease assessment, excessive or insufficient maintenance can be avoided, thereby optimizing the configuration of road maintenance resources; it can not only improve resource utilization efficiency, but also reduce maintenance costs; timely disease detection and accurate hazard assessment help to timely discover and repair potential safety hazards, thereby improving the overall safety and driving comfort of the road; the evaluation and correction process in step S5 can improve the accuracy of road disease assessment, enhance the scientific nature of decision-making, optimize resource allocation, improve road safety, and promote technological innovation and application.

[0110] S6. Compare the second-order road disease hazard level with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately;

[0111] If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state, which means that the current disease level has not yet reached the urgency of requiring immediate repair; the road management department should continue to monitor the road section regularly to track the development of the disease; at the same time, formulate a preventive maintenance plan to delay the further development of the disease;

[0112] If the second-order road disease hazard level is greater than or equal to the preset hazard level threshold, it means that the road disease needs to be repaired immediately; at this time, the disease has had a significant impact on the safety or functionality of the road, and measures must be taken immediately to repair it; the road management department should immediately organize a professional maintenance team to go to the site for assessment and repair work; the repair plan should be formulated according to the specific situation of the disease and the characteristics of road construction to ensure that the repair effect meets safety standards and extends the service life of the road;

[0113] Detailed records are kept for every maintenance decision and execution; based on actual maintenance results and user feedback, the preset hazard level thresholds, assessment models and maintenance strategies are continuously optimized and improved to improve the accuracy and efficiency of the automatic road hazard detection and assessment system.

[0114] In this step, by comparing the second-order road disease hazard level with the preset hazard level threshold, the system can accurately determine the severity of the road disease and make a decision on whether immediate maintenance is needed; this greatly shortens the time from disease discovery to maintenance response, and improves the timeliness and efficiency of road maintenance; when the second-order road disease hazard level is less than the preset threshold, the system recommends continuous observation and the formulation of a preventive maintenance plan, which helps to take intervention measures in the early stages of the disease and delay its further development, thereby avoiding larger-scale maintenance and higher cost investment due to the deterioration of the disease; at the same time, it also avoids unnecessary waste of maintenance resources; for road diseases that require immediate maintenance, the system can quickly trigger maintenance Response mechanism ensures that defects are repaired in a timely manner; it helps to improve road safety, reduce the risk of traffic accidents caused by defects, and improve driving comfort and road service life; by recording each maintenance decision and execution, and collecting user feedback, the system can continuously optimize and improve the preset hazard level thresholds, assessment models, and maintenance strategies, thereby improving the accuracy of detection and the reliability of assessment; the implementation of this step relies on advanced image recognition technology and data analysis methods to realize the intelligent and automated detection of road defects; this not only reduces the workload of manual inspections, but also improves detection efficiency and accuracy, providing road management departments with more scientific and efficient decision-making support.

[0115] Embodiment 2: Figure 3 As shown, the road disease automatic detection system based on image recognition of the present invention specifically includes the following modules:

[0116] Data collection module, collecting road disease growth trend sequences;

[0117] The preliminary disease assessment module inputs the road disease growth trend sequence into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level;

[0118] A construction characteristic information acquisition module, in response to the first-order road disease hazard level being less than a preset hazard level threshold, collects construction characteristic information of the location of the road disease;

[0119] The construction feature impact analysis module uses the pre-trained construction feature impact analysis model to identify and analyze construction feature information and obtain road construction evaluation factors;

[0120] The disease hazard level correction module uses the road construction evaluation factor to evaluate and correct the first-order road disease hazard level to obtain the second-order road disease hazard level;

[0121] The decision output module compares the second-order road disease hazard level with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately.

[0122] The system automatically collects and analyzes road damage data through image recognition technology, significantly improving detection efficiency and reducing reliance on manual inspections; this not only saves human resources, but also speeds up the discovery of damage and helps to take timely response measures;

[0123] The system uses pre-trained models to identify and assess damage, reducing the impact of human factors on the test results and improving the accuracy of the assessment. At the same time, by comprehensively considering the growth trend of damage and road construction characteristics, the system can provide more comprehensive and accurate damage assessment results.

[0124] The system can track the growth trend of road damage, not only identifying the current damage status, but also predicting its future development trend; it provides valuable decision-making basis for road management departments, helps to extend the service life of roads, and reduce traffic interruptions and safety accidents caused by sudden damage;

[0125] The system automatically outputs maintenance decision suggestions based on the comparison results of the second-order road disease hazard level and the preset threshold; it helps the road management department to quickly respond to the disease situation, reasonably arrange maintenance plans, and improve road management efficiency;

[0126] The system fully considers the impact of road construction characteristics on road damage. By collecting and analyzing construction characteristic information, it incorporates factors such as road structure, materials, and service life into the damage assessment system, making the assessment results more comprehensive and scientific, and helping to formulate more accurate maintenance strategies.

[0127] The various variations and specific embodiments of the method for automatic detection of road defects based on image recognition in the aforementioned embodiment 1 are also applicable to the automatic detection system for road defects based on image recognition in the present embodiment. Through the aforementioned detailed description of the method for automatic detection of road defects based on image recognition, those skilled in the art can clearly understand the implementation method of the automatic detection system for road defects based on image recognition in the present embodiment. Therefore, for the sake of brevity of the specification, it will not be described in detail here.

[0128] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected via a bus. When the computer program is executed by the processor, each process of the above-mentioned method for controlling output data is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0129] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for automatic detection of road damage based on image recognition, characterized in that: The method comprises: Collect road disease growth trend series; The road disease growth trend sequence is input into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level; In response to the first-order road disease hazard level being less than a preset hazard level threshold, collecting construction characteristic information of a location where the road disease is located; Use the pre-trained construction feature impact analysis model to identify and analyze construction feature information to obtain road construction evaluation factors; The first-order road hazard level is evaluated and corrected using the road construction evaluation factor to obtain the second-order road hazard level. The second-order road disease hazard level is compared with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately.

2. The method for automatic detection of road damage based on image recognition according to claim 1, characterized in that: Methods for collecting road disease growth trend series include: Collect road disease image information when the road disease is first discovered; Extract features from road damage image information to obtain a road damage feature set; According to the preset acquisition frequency, multiple image acquisition and feature extraction are performed on the road damage at the same location to obtain multiple road damage feature sets; Arrange multiple road disease feature sets in chronological order to obtain a road disease growth trend sequence.

3. The method for automatic detection of road damage based on image recognition as claimed in claim 2, characterized in that: The method for setting the acquisition frequency includes: Quantifying the road damage feature set obtained for the first time to obtain a road damage feature index vector; The road disease feature index vector is used to traverse the pre-built frequency reference library to extract the acquisition frequency corresponding to the road disease feature index vector.

4. The method for automatic detection of road damage based on image recognition according to claim 1, characterized in that: The construction method of the road hazard assessment model includes: Collecting series data on the growth trend of road diseases; Perform pre-processing operations such as cleaning, denoising and standardization on the collected data; Selecting a machine learning model as the infrastructure of the road hazard assessment model, wherein the machine learning model includes a random forest, a gradient boosting tree, a convolutional neural network, and a recurrent neural network; Use the preprocessed data to train the model. During the training process, adjust the parameters and structure of the model to optimize its performance. Use an independent validation set to validate the trained model and evaluate its performance in real applications; Tune the model based on the validation results; Deploy the trained and verified models to actual application scenarios.

5. The method for automatic detection of road damage based on image recognition according to claim 1, characterized in that: The construction characteristic information includes road structure, service life, traffic flow, environmental factors and maintenance history.

6. The method for automatic detection of road damage based on image recognition according to claim 1, characterized in that: The method for obtaining the second-order road disease hazard level includes: Based on the first-order road hazard level and the construction characteristics of the road hazard location, their impact on the growth trend and hazard level of road hazard is evaluated; Setting correction thresholds for different construction features; comparing the road construction characteristic information with a set correction threshold; According to the comparison results, the first-order road disease hazard level is adjusted; After applying the correction criteria for all relevant construction characteristics, the second-order road disease hazard level is obtained by integrating all factors.

7. The method for automatic detection of road damage based on image recognition according to claim 1, characterized in that: The factors affecting the setting of the preset hazard level threshold include road type, road material, traffic flow, road age and environmental conditions.

8. An automatic road disease detection system based on image recognition, characterized in that: The system comprises: Data collection module, collecting road disease growth trend sequences; The preliminary disease assessment module inputs the road disease growth trend sequence into the pre-trained road disease hazard assessment model to obtain the first-order road disease hazard level; A construction characteristic information acquisition module, in response to the first-order road disease hazard level being less than a preset hazard level threshold, collects construction characteristic information of the location of the road disease; The construction feature impact analysis module uses the pre-trained construction feature impact analysis model to identify and analyze construction feature information and obtain road construction evaluation factors; The disease hazard level correction module uses the road construction evaluation factor to evaluate and correct the first-order road disease hazard level to obtain the second-order road disease hazard level; The decision output module compares the second-order road disease hazard level with the preset hazard level threshold. If the second-order road disease hazard level is less than the preset hazard level threshold, it means that the road disease is in a normal development state and needs to be continuously observed; otherwise, it means that the road disease needs to be repaired immediately.

9. An electronic device for automatic detection of road defects based on image recognition, comprising a bus, a transceiver, a camera, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and characterized in that: When the computer program is executed by the processor, the steps in the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 7 are implemented.

Images