Road Condition Analysis System and Method Based on Multi-Source Data Fusion

Abstract

The present invention discloses a road performance analysis system and method based on multi-source data fusion, belonging to the field of road performance analysis management technology, and includes: a data visualization module, a data management module, a multi-source data fusion module, and a feature extraction and early warning module. The present invention improves the comprehensiveness, accuracy, and timeliness of road performance analysis, optimizes the utilization efficiency of system resources, integrates various types of road-related data through multi-source data fusion, can comprehensively reflect the actual road conditions, avoid the limitations of a single data source, provide a rich and accurate information basis for analysis, accurately assess road damage risks, predict potential problems in advance, provide a reliable basis for road maintenance decisions, reduce safety hazards caused by road diseases, and dynamically configure computing resources and visualization display resources according to the real-time status of the system. On the basis of ensuring data processing accuracy and real-time performance, it avoids resource waste and improves system operation efficiency to adapt to different road conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of road behavior analysis and management, and in particular to a road behavior analysis system and method based on multi-source data fusion. Background Art

[0002] In modern traffic management and road maintenance, accurate understanding of road conditions is crucial for ensuring traffic safety, improving traffic efficiency, and rationally planning road maintenance. With the rapid development of information technology, data from multiple data sources reflects real-time road conditions from different perspectives, including pavement damage, water accumulation, and icing risks. Data fusion technology is becoming increasingly important in many fields, integrating information from various channels to provide more comprehensive and accurate insights. In road performance analysis, the fusion of multi-source data can fully leverage the strengths of each data source, improve the accuracy and reliability of analysis, and provide more accurate, timely, and comprehensive road condition information.

[0003] For example, the invention patent announcement with announcement number: CN118071114B discloses a condition monitoring system and method for road and bridge anti-collision guardrails. It uses statistical analysis and chi-square test to identify features related to road anti-collision guardrails, and combines the DBSCAN algorithm to perform cluster analysis on the selected features to eliminate irrelevant features.

[0004] For example, the invention patent announcement with announcement number: CN119067511B discloses an intelligent road quality evaluation method and system based on the Smooth Road Index, the method including: obtaining collected data of urban roads to be evaluated; dividing the urban roads to be evaluated into different evaluation units; obtaining health status indicators based on the collected data, and calculating the Smooth Road Index corresponding to different evaluation units based on the health status indicators, wherein the health status indicators include at least road damage parameters and bump impact parameters; dividing the evaluation units according to the Smooth Road Index corresponding to different evaluation units to obtain at least a multi-color evaluation map; and outputting at least a multi-color evaluation map.

[0005] However, in the process of implementing the technical solutions of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0006] Current road performance analysis methods often focus on the collection and analysis of road condition data, ignoring the rational allocation of the system's data processing resources. In practice, it is difficult to flexibly adjust resources according to the real-time status of the system, resulting in low resource utilization efficiency, which may affect the accuracy and real-time performance of data processing, reduce the reliability and efficiency of road performance analysis, and make it difficult to cope with complex and changing actual application scenarios. Summary of the Invention

[0007] A first aspect of the present invention provides a road behavior analysis system based on multi-source data fusion, comprising:

[0008] The data visualization module is used to select a processing mode based on the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode, and balanced mode, and dynamically configure system computing resources and visualization display resources based on the processing mode.

[0009] The data management module is used to collect road average stress and stress interference data, analyze the road average stress correction value, and thus obtain the road condition pre-classification result.

[0010] The multi-source data fusion module is used to obtain multi-source data on road structure based on the pre-classification results of road conditions, dynamically adjust the sampling frequency and computing resources according to the data fluctuation characteristics, analyze the data confidence and set the data fusion weight, thereby completing data fusion and generating a fusion feature set.

[0011] The feature extraction and warning module is used to extract road damage features based on the fused feature set, perform road performance analysis, optimize and adjust visual display resources, and simultaneously display the road performance analysis results and resource allocation status through a visual interface.

[0012] A second aspect of the present invention provides a road behavior analysis method based on multi-source data fusion, comprising the following steps:

[0013] S1, selects the processing mode according to the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode and balance mode, and dynamically configures system computing resources and visualization display resources based on the processing mode.

[0014] S2, collecting road average stress and stress interference data, analyzing the road average stress correction value, and thereby obtaining a road condition pre-classification result.

[0015] S3, based on the pre-classification results of road conditions, obtains multi-source data of road structure, dynamically adjusts the sampling frequency and computing resources according to the data fluctuation characteristics, analyzes the data confidence and sets the data fusion weight, thereby completing data fusion and generating a fusion feature set.

[0016] S4, based on the fusion feature set, extracts road damage features and conducts road performance analysis, thereby optimizing and adjusting visualization resources, and simultaneously displays the road performance analysis results and resource allocation status through a visualization interface.

[0017] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0018] 1. The road performance analysis system and method based on multi-source data fusion provided by the present invention improve the comprehensiveness, accuracy and timeliness of road performance analysis and optimize the utilization efficiency of system resources. By integrating various types of road-related data through multi-source data fusion, it can fully reflect the actual road conditions, avoid the limitations of a single data source, and provide a rich and accurate information basis for analysis. During the analysis process, the risk of road damage is accurately assessed, potential problems are predicted in advance, and a reliable basis is provided for road maintenance decisions, reducing the safety hazards caused by road diseases. At the same time, computing resources and visual display resources are dynamically configured according to the real-time status of the system. On the basis of ensuring data processing accuracy and real-time performance, resource waste is avoided, and the system operation efficiency is improved to adapt to different road conditions.

[0019] 2. The present invention can significantly improve the adaptability and operating efficiency of the system in different work scenarios by dynamically configuring system computing resources and visual display resources based on the processing mode. In the accuracy priority mode, the system will invest more computing resources in deep data processing tasks and switch to high-resolution views to make the road performance analysis results more accurate and detailed, meeting the needs of in-depth research and precise evaluation of road conditions. In the real-time priority mode, the switching of computing resource compression and dynamic summary views ensures that the system can still respond quickly under high load, provide users with key road information in a timely manner, and meet the urgent needs of real-time monitoring. The balance mode reasonably allocates computing resources according to the fluctuation of system data volume, avoids resource waste, and maintains efficient and stable operation of the system.

[0020] 3. This invention extracts road damage features based on a fused feature set and performs road performance analysis, enabling timely and accurate assessment of road safety conditions. This analysis, generated by fusing multi-source data, comprehensively encompasses all aspects of road information, making the extracted road damage features more comprehensive and accurate. Risk prediction based on this foundation effectively identifies potential road damage risks, facilitating the development of appropriate maintenance plans and the implementation of targeted measures to prevent further deterioration of road damage, thereby ensuring normal road use and traffic safety.

[0021] 4. This invention significantly improves the efficiency and quality of road condition information users obtain by optimizing and adjusting visual display resources. When the road behavior analysis indicates a secondary state, the system intelligently reallocates visual display resources to highlight important display tasks. For primary states, the system maintains current resource allocation and displays relevant information, avoiding resource waste while ensuring users can readily access the road's basic status. This ensures that the information displayed on the visual interface consistently matches the actual road risk status, providing intuitive and effective data support for road management. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1A schematic diagram of the structure of a road behavior analysis system based on multi-source data fusion provided in an embodiment of the present application;

[0023] Figure 2 A flowchart of a road behavior analysis method based on multi-source data fusion provided in an embodiment of the present application; DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] Reference Figure 1 As shown, the first aspect of the present invention provides a road behavior analysis system based on multi-source data fusion, comprising:

[0026] The data visualization module is used to select a processing mode based on the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode, and balanced mode, and dynamically configure system computing resources and visualization display resources based on the processing mode.

[0027] In this embodiment, the processing mode is selected according to the length of the system real-time data processing queue. The specific analysis method is as follows:

[0028] Get the length of the system's real-time data processing queue.

[0029] Extract the first queue length threshold and the second queue length threshold preset in the database.

[0030] If the length of the system real-time data processing queue is less than or equal to the first queue length threshold, the processing mode is recorded as the accuracy priority mode.

[0031] If the length of the system real-time data processing queue is greater than or equal to the second queue length threshold, the processing mode is recorded as the real-time priority mode.

[0032] If the length of the system real-time data processing queue is greater than the first queue length threshold and less than the second queue length threshold, the processing mode is recorded as the balanced mode.

[0033] In this embodiment, system computing resources and visualization display resources are dynamically configured based on the processing mode. The specific analysis process is as follows:

[0034] If the processing mode is precision priority mode, the computing resource allocation ratio value is extracted according to the length of the system's real-time data processing queue, and the system computing resources are allocated to the data depth processing tasks based on the computing resource allocation ratio value. The visual display resources are simultaneously switched to a high-resolution view.

[0035] In a specific embodiment, the computing resource allocation ratio value is extracted according to the system real-time data processing queue length. The specific analysis process is as follows:

[0036] The first deviation length of the system real-time data processing queue is obtained by subtracting the first queue length threshold from the system real-time data processing queue length.

[0037] The allocation ratio value corresponding to each queue deviation length interval stored in the database is extracted, and the allocation ratio value corresponding to the interval where the first deviation length of the real-time data processing queue of the extraction system is located is mapped and recorded as the computing resource allocation ratio value.

[0038] It's important to understand that the larger the first deviation length of the system's real-time data processing queue, the fewer data processing queues the system is actively processing. To fully utilize system computing resources, improve the quality and depth of data processing, and allow deep data processing tasks to access more computing resources for more accurate and detailed analysis of road condition data, the corresponding extracted computing resource allocation ratio should be larger. This allows, in precision-priority mode, when the system load is low (i.e., fewer data processing queues), more computing resources can be allocated to deep data processing tasks, while the visualization resources are switched to a high-resolution view, allowing users to more clearly and comprehensively view road condition data, providing stronger support for road behavior analysis.

[0039] The high-resolution view refers to a refined visualization mode enabled when system resources are sufficient. It displays multidimensional data analysis results (such as three-dimensional stress distribution heat maps, historical damage trend curves, and sensor spatial topological relationships) with high rendering accuracy, supports interactive detail zooming and multi-layer overlay, and is suitable for in-depth diagnosis and offline backtracking scenarios.

[0040] If the processing mode is real-time priority mode, the computing resource compression ratio is extracted according to the length of the system's real-time data processing queue, and computing resources are compressed accordingly. The visual display of resources is simultaneously switched to a dynamic summary view.

[0041] In a specific embodiment, the computing resource compression ratio is extracted based on the system real-time data processing queue length. The specific process is as follows:

[0042] The second threshold value of the queue length is subtracted from the length of the system real-time data processing queue to obtain a second deviation length of the system real-time data processing queue.

[0043] The compression ratio corresponding to the deviation length interval of each queue stored in the database is extracted, and the compression ratio corresponding to the interval where the second deviation length of the real-time data processing queue of the extraction system is located is mapped and recorded as the computing resource compression ratio.

[0044] It's important to understand that the larger the second deviation length of the system's real-time data processing queue, the more queues the system is currently processing data, and the greater the processing pressure the system faces. To ensure the system can respond quickly even under high load, prioritize real-time data processing, and avoid processing delays caused by excessive backlogs in data processing queues, it's necessary to optimize resource allocation within limited resources. The larger the corresponding compression ratio of the extracted computing resources, the greater the compression of the computing resources, reducing the resources occupied by a single task. This allows more tasks to be processed in parallel, improving overall processing efficiency.

[0045] The dynamic summary view refers to a lightweight visualization mode enabled in resource-constrained or real-time priority modes. It uses data compression and summary generation technology to present only key indicators (such as maximum stress value, risk level, abnormal area positioning box) and dynamically updated statistical charts (such as rolling curves, threshold alarm marks), ensuring real-time monitoring efficiency with low rendering latency.

[0046] If the processing mode is balanced, computing resources are dynamically allocated according to fluctuations in system data volume.

[0047] It should be added that computing resources are dynamically allocated according to fluctuations in system data volume. The specific process is as follows:

[0048] The system data volume fluctuation value is obtained from the system program log according to the preset monitoring period.

[0049] Extract the preset data volume fluctuation threshold in the database.

[0050] Subtract the data volume fluctuation threshold from the system data volume fluctuation value to obtain the system data volume fluctuation deviation value.

[0051] It should be noted that the system data volume fluctuation deviation value can be greater than zero, less than zero, or equal to zero.

[0052] The weight adjustment value corresponding to each data volume fluctuation deviation value interval stored in the database is extracted, and the weight adjustment value corresponding to the interval in which the system data volume fluctuation deviation value is located is mapped and extracted, and recorded as the second adjustment value of the computing resource allocation weight.

[0053] It should be noted that the larger the system data volume fluctuation deviation value is, the more the data volume fluctuates, which means that the amount of data generated or processed by the system during operation varies greatly. In order to make full use of system resources to effectively process a large amount of fluctuating data and ensure the comprehensiveness and accuracy of data processing, the corresponding extracted first weight adjustment value is positive and the larger it is, so that more computing resources can be allocated to data sources or processing tasks with large data volume fluctuations, so that they can better cope with data changes. The larger the system data volume fluctuation deviation value is, the smaller the data volume fluctuation is and the data is relatively stable. In this case, in order to avoid resource waste and reasonably optimize system resource allocation, the corresponding extracted first weight adjustment value needs to be negative and the larger it is, that is, reduce the resource allocation to data sources or processing tasks with small data volume fluctuations, concentrate more resources on where they are more needed, and improve the overall operating efficiency of the system.

[0054] Get the initial computing resource allocation weights for each data source.

[0055] Dynamically adjust the computing resource allocation of each data source based on the initial computing resource allocation weight of each data source and the first adjustment value of the computing resource allocation weight of each data source

[0056] A data management module is used to collect road average stress and stress interference data, analyze the road average stress correction value, and obtain the road condition pre-classification result;

[0057] In this embodiment, the average stress correction value of the road is analyzed to obtain the road condition pre-classification result. The specific process is as follows:

[0058] During the preset collection period, the average road stress and stress interference data are collected through the sensing device.

[0059] The stress interference data includes temperature extreme value difference, humidity extreme value difference, load extreme value difference and wind speed extreme value difference.

[0060] It should be noted that the sensing device includes a temperature sensor, a humidity sensor, a load sensor, a wind speed sensor and a stress sensor.

[0061] It should be understood that the extreme value difference refers to the difference between the maximum and minimum values of the collected values within a period. For example, the temperature extreme value difference refers to the difference between the maximum and minimum temperature values within a preset collection period.

[0062] The reference stress interference data stored in the database are extracted, including the reference temperature extreme value difference, the reference humidity extreme value difference, the reference load extreme value difference, and the reference wind speed extreme value difference.

[0063] Extract the temperature extreme value difference weights, humidity extreme value difference weights, load extreme value difference weights and wind speed extreme value difference weights preset in the database. It should be added that the value ranges of the temperature extreme value difference weights, humidity extreme value difference weights, load extreme value difference weights and wind speed extreme value difference weights are all between 0 and 1. When used, the preset values can be directly extracted from the database. The specific extraction method is, for example, to construct a mapping set one by one with the temperature extreme value difference, humidity extreme value difference weights, load extreme value difference weights and wind speed extreme value difference weights respectively. When used, the temperature extreme value difference, humidity extreme value difference, load extreme value difference and wind speed extreme value difference obtained in real time are input into the corresponding mapping set, thereby extracting the corresponding temperature extreme value difference weights, humidity extreme value difference weights, load extreme value difference weights and wind speed extreme value difference weights.

[0064] The stress interference index is obtained based on the analysis and processing of the stress interference data.

[0065] The stress interference index represents a quantitative indicator of the degree of interference of the temperature extreme difference, humidity extreme difference, load extreme difference and wind speed extreme difference on stress collection. The specific analysis method is to perform differential analysis on the stress interference data and the corresponding reference value, and couple the differential analysis results with the corresponding weights to obtain the stress interference index.

[0066] It should be noted that there is a mutual influence between the temperature extreme difference and the humidity extreme difference. For example, when the temperature extreme difference in the environment increases, the water evaporation rate will be accelerated, which will lead to changes in the humidity extreme difference. Conversely, changes in the humidity extreme difference will also affect the transfer and storage of heat, causing the temperature extreme difference to be affected.

[0067] In a specific embodiment, the stress interference index is specifically expressed as follows:

[0068] ,

[0069] in, is the stress interference index, is the temperature extreme difference, is the humidity extreme difference, is the load extreme value difference, is the extreme difference in wind speed, is the reference temperature extreme value difference, is the reference humidity extreme difference, is the extreme difference of the reference load, is the reference wind speed extreme difference, is the temperature extreme value difference weight, is the humidity extreme value difference weight, is the extreme value difference weight of the load, is the wind speed extreme value difference weight.

[0070] The interference correction value corresponding to each stress interference index interval stored in the database is extracted, and the interference correction value corresponding to the interval where the stress interference index is located is mapped and extracted, and recorded as the stress interference correction value.

[0071] It's important to understand that a larger stress interference index indicates a greater degree of interference from extreme temperature, humidity, load, and wind speed differences on stress collection. To more accurately reflect the road's true stress conditions, when correcting road stress, for high levels of interference, the extracted stress interference correction value should be smaller. This reduces the impact of interference factors on stress collection results, making the corrected road stress value closer to the actual stress experienced by the road. This improves the accuracy of road condition pre-classification results and provides a more reliable data foundation for subsequent road performance analysis.

[0072] The road average stress is combined with the stress interference correction value to perform correction processing to obtain the road average stress correction value. Specifically, the road average stress is multiplied by the stress interference correction value to obtain the road average stress correction value.

[0073] Extract the preset road average stress verification value in the database.

[0074] If the road average stress correction value is less than the road average stress verification value, the road condition pre-classification result is recorded as a normal stress road.

[0075] It's important to understand that if the Corrected Average Road Stress value is less than the Verified Average Road Stress value, it indicates that after accounting for interference factors and making corrections, the average stress experienced by the road is within the normal range. This means that the stress experienced by the road structure is within acceptable standards, the road is in a relatively stable mechanical state, and its structural integrity and safety show no significant stress anomalies. Based on this, it can be preliminarily determined that the road condition is normal.

[0076] If the road average stress correction value is greater than or equal to the road average stress verification value, the road condition pre-classification result is recorded as an abnormal stress road.

[0077] It should be understood that if the road average stress correction value is greater than the road average stress verification value, it means that after considering the interference of interference factors and making corrections, the stress condition still deviates from the normal level. There may be potential risks such as structural damage and reduced bearing capacity, which may cause cracks, subsidence and other defects in the road, affecting the normal use of the road and traffic safety.

[0078] The multi-source data fusion module is used to obtain multi-source data on road structure based on the road condition pre-classification results, dynamically adjust the sampling frequency and computing resources according to the data fluctuation characteristics, analyze the data confidence and set the data fusion weights, thereby completing the data fusion and generating a fusion feature set;

[0079] In this embodiment, multi-source data of road structure is obtained based on the road condition pre-classification results, and the sampling frequency and computing resources are dynamically adjusted according to the data fluctuation characteristics. The specific analysis process is as follows:

[0080] An initial data collection frequency set corresponding to a preset road condition pre-classification result in a database is extracted, and multi-source data collection of road structure is performed based on the initial data collection frequency set.

[0081] It should be added that if the road condition pre-classification result is a normal stress road, it means that the current road condition is good, and the corresponding initial data collection frequency is low. If the road condition pre-classification result is an abnormal stress road, it means that the current road condition is abnormal, and the corresponding initial data collection frequency is high.

[0082] The initial data collection frequency set is the initial data collection frequency of each data source.

[0083] The data fluctuation characteristics include the standard deviation of the data sequence and the number of data points.

[0084] It should be added that the number of data points refers to the amount of data collected by each data source within the preset second data collection window.

[0085] The standard deviation of the data series of each data source is obtained based on the preset first data collection window.

[0086] The standard deviation of the data sequence of each data source is processed by difference processing with the data sequence standard deviation threshold preset in the database to obtain the data sequence standard deviation deviation factor of each data source.

[0087] It should be noted that the difference processing refers to subtracting the data sequence standard deviation threshold preset in the database from the data sequence standard deviation of each data source, and the result of the difference processing can be greater than zero, less than zero or equal to zero.

[0088] The sampling frequency adjustment values corresponding to each standard deviation deviation factor interval stored in the database are extracted, and the sampling frequency adjustment values corresponding to the interval of the standard deviation deviation factor of the data sequence of each data source are mapped and extracted, and recorded as the sampling frequency adjustment values of each data source.

[0089] It should be noted that a positive data series standard deviation deviation factor indicates that the larger the deviation, the greater the deviation of the data series standard deviation of each data source from the preset data series standard deviation threshold in the database. To more accurately capture data fluctuations, avoid missing key information due to abnormal fluctuations, and obtain more detailed data for accurate road condition analysis, the sampling frequency adjustment value should be positive and larger, meaning the sampling frequency should be increased. A negative data series standard deviation deviation factor indicates that the deviation of the data series standard deviation of each data source from the preset threshold is greater and lower, indicating relatively small data fluctuations. To rationally utilize system resources and avoid waste of resources caused by excessive sampling, the sampling frequency adjustment value should be negative and larger, meaning the sampling frequency should be reduced.

[0090] Dynamic adjustment of the sampling frequency is completed based on the initial data collection frequency set and the sampling frequency adjustment value of each data source.

[0091] In a specific embodiment, for a data source, the sampling frequency adjustment is performed specifically as follows: the initial data acquisition frequency of the data source is added to the sampling frequency adjustment value of the data source to obtain the sampling frequency update value of the data source, thereby completing the sampling frequency adjustment of the data source.

[0092] The number of data points from each data source is obtained based on a preset second data acquisition window.

[0093] Dynamically adjust computing resource allocation based on the number of data points in each data source.

[0094] In this embodiment, the computing resource allocation is dynamically adjusted based on the number of data points of each data source. The specific analysis process is as follows:

[0095] Extract the preset number of validation data points from the database.

[0096] The number of data points of each data source is subtracted from the number of verification data points to obtain the number of data deviation data points of each data source.

[0097] It should be noted that the number of data deviation data points can be greater than zero, less than zero, or equal to zero.

[0098] The weight adjustment value corresponding to each data deviation data point interval preset in the database is extracted, and the weight adjustment value corresponding to the interval in which the data deviation data point number of the extracted data source is located is mapped and recorded as the first adjustment value of the computing resource allocation weight.

[0099] It should be noted that a positive and larger number of data deviation points indicates that the data source has a greater volume of data than the preset number of validation data points, indicating that the data source is rich in data and contains a wealth of potentially valuable information for road performance analysis. To more fully process this data, uncover key insights, and thus more accurately analyze road conditions, a positive and larger first weight adjustment value indicates that more computing resources are allocated to this data source. A negative and larger absolute value of the data deviation points indicates that the data source has a smaller volume of data than the validation data points, indicating that the data source generates less data and contains limited information. To optimize the allocation of computing resources and avoid wasting resources on data sources with smaller volumes, a negative and larger first weight adjustment value indicates that computing resources allocated to this data source are reduced, allowing resources to be focused on processing data sources with larger volumes of data and greater analytical value, thereby improving the efficiency and accuracy of the entire road performance analysis system.

[0100] Get the initial computing resource allocation weights for each data source.

[0101] Dynamic adjustment of the computing resource allocation of each data source is performed based on the initial computing resource allocation weight of each data source and the first adjustment value of the computing resource allocation weight of each data source.

[0102] In a specific embodiment, the dynamic adjustment process of computing resource allocation is as follows: the initial computing resource allocation weight of a data source is added to the first adjustment value of the computing resource allocation weight of the data source to obtain the updated computing resource allocation weight value of the data source, each data source is traversed in turn to obtain the updated computing resource allocation weight value of each data source, the updated computing resource allocation weight value of each data source is normalized to obtain the updated implementation value of the computing resource allocation weight of each data source, and the computing resources of each data source are redistributed based on the updated implementation value of the computing resource allocation weight of each data source, thereby completing the dynamic adjustment of the computing resource allocation of each data source.

[0103] In this embodiment, the data confidence is analyzed and the data fusion weight is set, thereby completing the data fusion and generating a fusion feature set. The specific analysis process is as follows:

[0104] The basic parameters of the sensor devices corresponding to each data source are obtained in the preset third data acquisition window and recorded as the basic parameters of each sensor device.

[0105] The basic parameters of each sensor device include the average signal amplitude, average signal drift, power supply voltage fluctuation and average working circuit temperature of each sensor device.

[0106] It should be understood that the power supply voltage fluctuation refers to the difference between the maximum and minimum values of the power supply voltage within the preset third data acquisition window. By connecting the voltage sensor to the power supply circuit of the sensing device, the power supply voltage changes are monitored in real time.

[0107] It should be noted that the average signal amplitude and average signal drift can be obtained using a signal acquisition circuit and signal processing algorithm. When the sensor device collects data, the internal signal conditioning circuit first performs preprocessing on the raw signal, such as filtering and amplification, and then converts the analog signal into a digital signal through analog-to-digital conversion. Using digital signal processing techniques, statistical analysis of the signal over a period of time can be performed to calculate the average signal amplitude and average signal drift. For example, a sliding window algorithm can be used to calculate the signal mean and drift within a continuous time window.

[0108] The average temperature of the working circuit can be collected by a thermistor.

[0109] The basic parameters of the reference sensor device stored in the database are extracted, including the reference average signal amplitude, the reference average signal drift, the reference power supply voltage fluctuation and the reference working circuit average temperature.

[0110] Analyze the confidence of the data of each sensor device based on the basic parameters of each sensor device.

[0111] This embodiment analyzes the confidence level of the data from each sensor device based on its basic parameters, taking into account the mutual influence between these parameters. For example, when the average temperature of the working circuit increases, the performance of the electronic components will change, which may cause the average signal drift to increase. It may also affect the transmission and amplification of the signal, thereby changing the average signal amplitude. Changes in the power supply voltage fluctuation will directly affect the power supply stability of the circuit. Unstable voltage may interfere with signal transmission, causing fluctuations in the average signal amplitude, and also affect the normal operation of the electronic components, causing changes in the average signal drift. Abnormal changes in the average signal amplitude may indicate a change in the circuit's operating state, which in turn will affect the average temperature of the working circuit. It also reflects the instability of the power supply voltage fluctuation or other factors, further affecting the average signal drift.

[0112] The data confidence of each sensor device represents quantitative data on the degree of influence of the average signal amplitude, average signal drift, power supply voltage fluctuation and average working circuit temperature of each sensor device on the credibility of the data source. Specifically, it is expressed as performing differential analysis on the basic parameters of the sensor device and the corresponding reference values, and coupling analysis of the differentiated results with the corresponding weights to obtain the data confidence of each sensor device.

[0113] In a specific embodiment, the confidence level of the data from each sensor device is obtained in the following manner:

[0114] ,

[0115] in, is the confidence level of the data from the i-th sensor device, is the average signal amplitude of the i-th sensor device, is the average signal drift of the i-th sensor device, is the power supply voltage fluctuation of the i-th sensor device, is the average temperature of the working circuit of the i-th sensor device, is the reference average signal amplitude, is the reference average signal drift, is the reference power supply voltage fluctuation, is the average temperature of the reference working circuit, is the average signal amplitude weight, is the average signal drift weight, is the power supply voltage fluctuation weight, is the average temperature weight of the working circuit.

[0116] It should be noted that the value ranges of the average signal amplitude weight, average signal drift weight, power supply voltage fluctuation weight and average working circuit temperature weight are all between 0 and 1. When used, the pre-set values can be directly extracted from the database. The specific extraction method is, for example, to construct a one-to-one mapping set of the average signal amplitude, average signal drift weight, power supply voltage fluctuation weight and average working circuit temperature with the corresponding average signal amplitude weight, average signal drift weight, power supply voltage fluctuation weight and average working circuit temperature weight respectively. When used, the real-time obtained average signal amplitude, average signal drift weight, power supply voltage fluctuation weight and average working circuit temperature are respectively input into the corresponding mapping set, so as to extract the corresponding average signal amplitude weight, average signal drift weight, power supply voltage fluctuation weight and average working circuit temperature weight.

[0117] The fusion weight corresponding to each confidence interval stored in the database is extracted, and the fusion weight corresponding to the interval where the confidence of the sensor device data is located is mapped and extracted, and recorded as the data fusion weight.

[0118] It's important to understand that the greater the data confidence, the more strongly factors—such as the average signal amplitude, average signal drift, power supply voltage fluctuation, and average operating circuit temperature—combined influence the trustworthiness of the data source. In other words, the data source is more reliable and trustworthy. To maximize the value of high-confidence data during multi-source data fusion, ensuring that the fused feature set more accurately and reliably reflects road conditions, the corresponding data fusion weight is increased. This results in higher weighting of high-confidence data in the fusion results, thereby improving the accuracy and reliability of the overall road performance analysis.

[0119] The data source data fusion weight corresponding to each data source is counted and normalized to obtain the fusion weight corresponding to each data source, which is recorded as the data fusion weight of each data source.

[0120] The data fusion weight is set based on the data fusion weight of each data source, thereby fusing the data of each data source and generating a fusion feature set.

[0121] It should be added that the purpose of normalizing the data source data fusion weights corresponding to each data source is to make the sum of each weight equal to 1, thereby allocating the data fusion weights of each data source.

[0122] It should be noted that, in a specific embodiment, data fusion may use model-level fusion.

[0123] The feature extraction and warning module is used to extract road damage features based on the fused feature set, perform road performance analysis, optimize and adjust visual display resources, and simultaneously display the road performance analysis results and resource allocation status through a visual interface.

[0124] In this embodiment, road damage features are extracted based on the fused feature set to perform road behavior analysis. The specific analysis steps are as follows:

[0125] Road damage features are extracted based on the fused feature set, including road smoothness, maximum crack width and settlement.

[0126] In a specific embodiment, the road smoothness can be obtained by performing time domain analysis based on the fused feature set, specifically: obtaining a vibration acceleration signal based on the fused feature set, and determining the road smoothness according to the root mean square value of the vibration acceleration signal. The larger the root mean square value of the vibration acceleration signal, the rougher the road.

[0127] The maximum crack width can be obtained by performing edge detection on the image data of the fused feature set to locate the maximum crack and obtain the maximum crack width.

[0128] The settlement amount can be obtained by extracting strain and displacement through fusing feature sets.

[0129] The reference road damage characteristics stored in the database are extracted, including reference road smoothness, reference maximum crack width, and reference settlement.

[0130] The road damage risk assessment value is obtained based on the analysis and processing of road damage characteristics.

[0131] In a specific embodiment, the road damage risk assessment value is derived by analyzing and processing road damage characteristics, taking into account the mutual influence of these parameters. For example, when road settlement occurs, the original road surface smoothness will be disrupted, causing vehicles to be unevenly stressed when traveling on this road section, further exacerbating damage to the road surface. Over time, this unevenness will lead to local stress concentration in the road surface. In these stress concentration areas, the road surface is more likely to crack, and the maximum crack width will continue to expand. The increase in crack width will reduce the overall strength and stability of the road surface, making the road surface more susceptible to new settlement under external forces such as vehicle loads, further deteriorating road smoothness. In addition, poor road smoothness will cause vehicles to exert greater impact on the road surface when driving, which will not only accelerate the development of settlement but also cause cracks to further expand and increase the maximum crack width.

[0132] The road damage risk assessment value is quantitative data on the degree of influence of road flatness, maximum crack width and settlement on the road damage status. It is specifically expressed by differentiating the road damage characteristics with the corresponding reference values, and coupling the differentiation results with the corresponding weights to obtain the road damage risk assessment value.

[0133] In a specific embodiment, the road damage risk assessment value is obtained in the following manner:

[0134] ,

[0135] in, is the road damage risk assessment value, For road smoothness, is the maximum crack width, is the amount of settlement, For reference road smoothness, is the reference maximum crack width, is the reference settlement, is the road smoothness weight, is the maximum crack width weight, is the settlement weight.

[0136] It should be noted that the value range of the road smoothness weight, maximum crack width weight and settlement weight is 0-1. When used, the pre-set values can be directly extracted from the database. The specific extraction method is, for example, to construct a one-to-one mapping set for the road smoothness, maximum crack width and settlement with the corresponding weights. When used, the road smoothness, maximum crack width and settlement obtained in real time are input one by one into the corresponding mapping set, so as to extract the corresponding road smoothness weight, maximum crack width weight and settlement weight.

[0137] Extract the preset road damage risk assessment thresholds in the database.

[0138] If the road damage risk assessment value is less than or equal to the road damage risk assessment threshold, the road state analysis result is recorded as the first state.

[0139] If the road damage risk assessment value is less than or equal to the road damage risk assessment threshold, it means that after a comprehensive assessment of damage characteristics such as road flatness, crack degree and settlement, the damage risk faced by the road is within an acceptable range. The current conditions of the road are relatively good, its structural integrity and performance are not under major threats, the road damage risk is small, and the current road condition is good.

[0140] If the road damage risk assessment value is greater than the road damage risk assessment threshold, the road behavior analysis result is recorded as the second behavior.

[0141] If the road damage risk assessment value is greater than the road damage risk assessment threshold, it means that the damage risk of the road based on a comprehensive analysis of flatness, crack degree and settlement has exceeded the normal safety standard. The road has a high damage risk and may have more serious problems, such as large areas of unevenness, excessive crack width or severe settlement. The road damage risk is high and the current road condition is low.

[0142] In this embodiment, the optimization adjustment of the visual display resources is performed, and the road performance analysis results and resource allocation status are simultaneously displayed through the visual interface. The specific analysis steps are as follows:

[0143] If the road behavior analysis result is the secondary behavior, then optimize and adjust the visualization resources. The specific optimization and adjustment steps are as follows:

[0144] Extract the importance ratio and importance ratio threshold of each road condition display task preset in the database.

[0145] The road condition display tasks whose importance ratio is less than the importance ratio threshold are recorded as non-important display tasks.

[0146] A road condition display task whose importance ratio is greater than or equal to the importance ratio threshold is recorded as an important display task.

[0147] Each road condition display task is traversed in sequence, thereby obtaining each important display task and each non-important display task by statistics.

[0148] Based on the importance ratio of each road condition display task, the importance ratios of each important display task and each non-important display task are extracted, and normalized respectively to obtain the importance update ratio of each important display task and the importance update ratio of each non-important task.

[0149] The resource increase corresponding to each risk assessment value interval stored in the database is extracted, and the resource increase corresponding to the interval in which the road damage risk assessment value is located is mapped and extracted, and recorded as the visual display resource demand increase.

[0150] According to the updated importance ratio of each non-important task and the increase in the visual display resource demand, the release amount of the visual display resource demand of each non-important task is obtained.

[0151] The visual display resources of each important display task are increased according to the updated importance ratio of each important display task and the increase in the demand for visual display resources, thereby completing the dynamic adjustment of the visual display resources.

[0152] If the road behavior analysis result is the primary behavior, no optimization adjustment of computing resources is performed, and the road behavior analysis result and current resource allocation status are displayed through a visual interface.

[0153] See Figure 2 As shown, the second aspect of the present invention provides a road behavior analysis method based on multi-source data fusion, comprising the following steps:

[0154] S1, selects the processing mode according to the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode and balance mode, and dynamically configures system computing resources and visualization display resources based on the processing mode.

[0155] S2, collecting road average stress and stress interference data, analyzing the road average stress correction value, and thereby obtaining a road condition pre-classification result.

[0156] S3, based on the pre-classification results of road conditions, obtains multi-source data of road structure, dynamically adjusts the sampling frequency and computing resources according to the data fluctuation characteristics, analyzes the data confidence and sets the data fusion weight, thereby completing data fusion and generating a fusion feature set.

[0157] S4, based on the fusion feature set, extracts road damage features and conducts road performance analysis, thereby optimizing and adjusting visualization resources, and simultaneously displays the road performance analysis results and resource allocation status through a visualization interface.

[0158] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0160] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0162] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0163] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. The road behavior analysis system based on multi-source data fusion is characterized by: The following steps are involved: The data visualization module is used to select a processing mode based on the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode, and balanced mode, and dynamically configure system computing resources and visualization display resources based on the processing mode; A data management module is used to collect road average stress and stress interference data, analyze the road average stress correction value, and obtain the road condition pre-classification result; The multi-source data fusion module is used to obtain multi-source data on road structure based on the road condition pre-classification results, dynamically adjust the sampling frequency and computing resources according to the data fluctuation characteristics, analyze the data confidence and set the data fusion weights, thereby completing the data fusion and generating a fusion feature set; The feature extraction and early warning module is used to extract road damage features based on the fused feature set, perform road behavior analysis, optimize and adjust visualization resources, and simultaneously display the road behavior analysis results and resource allocation status through a visualization interface; The specific analysis process of dynamically configuring system computing resources and visual display resources based on processing mode is as follows: If the processing mode is precision-first mode, the computing resource allocation ratio value is extracted according to the system's real-time data processing queue length, and the system computing resources are allocated to the data depth processing tasks based on the computing resource allocation ratio value. The visual display resources are simultaneously switched to a high-resolution view. If the processing mode is real-time priority mode, the computing resource compression ratio is extracted according to the length of the system's real-time data processing queue, and computing resources are compressed accordingly. The visual display of resources is simultaneously switched to a dynamic summary view. If the processing mode is balanced, computing resources are dynamically allocated based on fluctuations in system data volume; The multi-source data of road structure is obtained based on the road condition pre-classification results, and the sampling frequency and computing resources are dynamically adjusted according to the data fluctuation characteristics. The specific analysis process is as follows: Extracting an initial data collection frequency set corresponding to a road condition pre-classification result preset in a database, and performing multi-source data collection of road structure based on the initial data collection frequency set; The initial data collection frequency set is the initial data collection frequency of each data source; The data fluctuation characteristics include the standard deviation of the data series and the number of data points; Obtaining the standard deviation of the data series of each data source based on a preset first data collection window; Perform difference processing on the data sequence standard deviation of each data source and the data sequence standard deviation threshold preset in the database to obtain the data sequence standard deviation deviation factor of each data source; Extract the sampling frequency adjustment value corresponding to each standard deviation deviation factor interval stored in the database, and map and extract the sampling frequency adjustment value corresponding to the interval in which the standard deviation deviation factor of the data sequence of each data source is located, and record it as the sampling frequency adjustment value of each data source; Dynamically adjust the sampling frequency based on the initial data collection frequency set and the sampling frequency adjustment value of each data source; Acquire the number of data points from each data source based on a preset second data collection window; Dynamically adjust computing resource allocation based on the number of data points in each data source; If the road damage risk assessment value is less than or equal to the road damage risk assessment threshold, the road state analysis result is recorded as the first state; If the road damage risk assessment value is greater than the road damage risk assessment threshold, the road state analysis result is recorded as the second state; The optimization adjustment of resources is performed by visually displaying the road behavior analysis results and resource allocation status through the visual interface. The specific analysis steps are as follows: If the road behavior analysis result is the secondary behavior, then optimize and adjust the visualization resources. The specific optimization and adjustment steps are as follows: Extract the importance ratio and importance ratio threshold of each road condition display task preset in the database; The road condition display tasks whose importance ratio is less than the importance ratio threshold are recorded as non-important display tasks; The road condition display tasks whose importance ratio is greater than or equal to the importance ratio threshold are recorded as important display tasks; Traverse each road condition display task in turn, and obtain each important display task and each non-important display task by counting; Based on the importance ratio of each road condition display task, the importance ratios of each important display task and each non-important display task are extracted, and normalized respectively to obtain the importance update ratio of each important display task and the importance update ratio of each non-important task; Extract the resource increase corresponding to each risk assessment value interval stored in the database, and map the extracted resource increase corresponding to the road damage risk assessment value interval, and record it as the visual display resource demand increase; The released amount of visual display resource demand for each non-important task is obtained by analyzing the updated proportion of importance of each non-important task and the increase in visual display resource demand; Increase the visual display resources of each important display task according to the updated importance ratio of each important display task and the increase in visual display resource demand, thereby completing the dynamic adjustment of visual display resources; If the road behavior analysis result is the primary behavior, no optimization adjustment of computing resources is performed, and the road behavior analysis result and current resource allocation status are displayed through a visual interface.

2. The road behavior analysis system based on multi-source data fusion according to claim 1, characterized in that: The processing mode is selected according to the length of the system real-time data processing queue. The specific analysis method is as follows: Get the length of the system's real-time data processing queue; Extracting a first queue length threshold and a second queue length threshold preset in a database; If the length of the system real-time data processing queue is less than or equal to the first queue length threshold, the processing mode is recorded as the accuracy priority mode; If the length of the system real-time data processing queue is greater than or equal to the second queue length threshold, the processing mode is recorded as the real-time priority mode; If the length of the system real-time data processing queue is greater than the first queue length threshold and less than the second queue length threshold, the processing mode is recorded as the balanced mode.

3. The road behavior analysis system based on multi-source data fusion according to claim 1, characterized in that: The analysis of the average road stress correction value to obtain the road condition pre-classification result is as follows: Within a preset collection period, the road average stress and stress interference data are collected through the sensing device; The stress interference data includes temperature extreme value difference, humidity extreme value difference, load extreme value difference and wind speed extreme value difference; According to the stress interference data analysis and processing, the stress interference index is obtained; The stress interference index represents a quantitative indicator of the degree of interference of the temperature extreme value difference, humidity extreme value difference, load extreme value difference and wind speed extreme value difference on stress acquisition. The specific analysis method is to perform differential analysis on the stress interference data and the corresponding reference value, and couple the differential analysis results with the corresponding weights to obtain the stress interference index; Extracting the interference correction value corresponding to each stress interference index interval stored in the database, and mapping the extracted interference correction value corresponding to the interval where the stress interference index is located, and recording it as the stress interference correction value; The road average stress is combined with the stress interference correction value to perform correction processing to obtain the road average stress correction value; Extract the road average stress verification value preset in the database; If the road average stress correction value is less than the road average stress verification value, the road condition pre-classification result is recorded as a normal stress road; If the road average stress correction value is greater than or equal to the road average stress verification value, the road condition pre-classification result is recorded as an abnormal stress road.

4. The road behavior analysis system based on multi-source data fusion according to claim 1, characterized in that: The dynamic adjustment of computing resource allocation based on the number of data points of each data source is analyzed as follows: Extract the preset verification data points in the database; Subtract the number of verification data points from the number of data points of each data source to obtain the number of data deviation data points of each data source; Extract the weight adjustment value corresponding to each data deviation data point interval preset in the database, and map the weight adjustment value corresponding to the interval where the data deviation data point number of the extracted data source is located, and record it as the first adjustment value of the computing resource allocation weight; Obtain the initial computing resource allocation weights for each data source; Dynamic adjustment of the computing resource allocation of each data source is performed based on the initial computing resource allocation weight of each data source and the first adjustment value of the computing resource allocation weight of each data source.

5. The road behavior analysis system based on multi-source data fusion according to claim 1, characterized in that: The data confidence is analyzed and the data fusion weight is set, thereby completing the data fusion and generating a fusion feature set. The specific analysis process is as follows: Acquire basic parameters of the sensor devices corresponding to each data source using a preset third data acquisition window, and record them as basic parameters of each sensor device; The basic parameters of each sensor device include the average signal amplitude, average signal drift, power supply voltage fluctuation and average working circuit temperature of each sensor device; Analyze the confidence level of data from each sensor device based on the basic parameters of each sensor device; The data confidence of each sensor device represents quantitative data on the degree to which the average signal amplitude, average signal drift, power supply voltage fluctuation, and average working circuit temperature of each sensor device jointly influence the credibility of the data source. Specifically, the confidence level of each sensor device data is obtained by performing differential analysis on the basic parameters of each sensor device and the corresponding reference values, and then performing coupled analysis on the differential results in combination with the corresponding weights. Extract the fusion weight corresponding to each confidence interval stored in the database, and map the extracted fusion weight corresponding to the interval where the confidence of the sensor data lies, and record it as the data fusion weight; The data source data fusion weight corresponding to each data source is calculated and normalized to obtain the fusion weight corresponding to each data source, which is recorded as the data fusion weight of each data source; The data fusion weight is set based on the data fusion weight of each data source, thereby fusing the data of each data source and generating a fusion feature set.

6. The road behavior analysis system based on multi-source data fusion according to claim 1, characterized in that: The specific analysis steps for extracting road damage features based on the fusion feature set are as follows: Extract road damage features based on the fused feature set, including road smoothness, maximum crack width, and settlement; Obtain a road damage risk assessment value based on road damage characteristic analysis and processing; The road damage risk assessment value is quantitative data on the degree of influence of road flatness, maximum crack width and settlement on the road damage status. It is specifically expressed by differentiating the road damage characteristics with the corresponding reference values, and coupling the differentiation results with the corresponding weights to obtain the road damage risk assessment value.

7. The method for road behavior analysis system based on multi-source data fusion according to any one of claims 1 to 6, characterized in that: include: S1, selects a processing mode based on the length of the system's real-time data processing queue, including accuracy priority mode, real-time priority mode, and balanced mode, and dynamically configures system computing resources and visualization display resources based on the processing mode; S2, collecting road average stress and stress interference data, analyzing the road average stress correction value, and thereby obtaining a road condition pre-classification result; S3, based on the road condition pre-classification results, obtains multi-source data on road structure, dynamically adjusts sampling frequency and computing resources according to data fluctuation characteristics, analyzes data confidence and sets data fusion weights, thereby completing data fusion and generating a fusion feature set; S4, based on the fusion feature set, extracts road damage features and conducts road performance analysis, thereby optimizing and adjusting visualization resources, and simultaneously displays the road performance analysis results and resource allocation status through a visualization interface.

Images