Road maintenance inspection method and system based on user driving records

Abstract

The invention provides a road maintenance inspection method and system based on user driving records, and belongs to the technical field of electric digital data processing. The method can determine a plurality of maintenance inspection road sections according to user driving records; and determining a maintenance inspection score based on the driving record data and a preset road section map. And generating a preset inspection road image corresponding to the maintenance inspection road section according to the maintenance inspection score and a preset road maintenance simulation model, and sending the preset inspection road image to a corresponding user terminal. And after the simulation feedback information is received, determining a matching index between the user driving record and a preset inspection road image, so as to generate a maintenance inspection sequence in a cascade manner according to the maintenance standard under the condition that the matching index is greater than a preset threshold value. And determining a to-be-maintained defect area and a defect influence value thereof according to the image recognition model and the user driving record, generating a maintenance inspection plan according to the maintenance inspection sequence and the defect influence value, and sending the maintenance inspection plan to a management terminal for display.

Description

Technical Field

[0001] This application relates to the technical field of electronic digital data processing, and particularly to a road maintenance inspection method and system based on user driving records. Background Art

[0002] With the acceleration of urbanization and the continuous increase in traffic flow, highway maintenance work faces unprecedented challenges. Traditional highway maintenance methods mainly rely on manual inspections and regular detections. This method is not only inefficient but also difficult to comprehensively cover and timely detect potential problems. In addition, manual inspections are restricted by various factors, such as weather conditions, traffic conditions, and the experience and subjective judgment of inspectors, which may lead to untimely or omitted maintenance work.

[0003] Although the monitoring based on fixed sensors can continuously obtain some data, there are blind spots in the sensor layout, and it is slow to respond to road problems in some sudden or special locations. At the same time, these traditional methods lack effective interaction with actual users and cannot make full use of the road condition information accumulated by the majority of drivers during their daily driving.

[0004] Based on this, there is an urgent need for a road maintenance inspection method that can combine user driving records to solve the technical problems of low maintenance efficiency, high cost, strong subjectivity, and poor pertinence in existing road maintenance inspection work. Summary of the Invention

[0005] The embodiments of this application provide a road maintenance inspection method and system based on user driving records to solve the technical problems of low maintenance efficiency, high cost, strong subjectivity, and poor pertinence in existing road maintenance inspection work.

[0006] On the one hand, the embodiments of this application provide a road maintenance inspection method based on user driving records. The method includes:

[0007] Determine a number of maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from a number of user terminals; wherein, the maintenance inspection sections are divided based on the driving anomaly feedback density in the user driving records;

[0008] Determine the corresponding maintenance inspection scores based on the driving record data corresponding to each maintenance inspection section and a preset section map;

[0009] Generate a preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and a preset road maintenance simulation model, and send the preset inspection road image to the corresponding user terminal;

[0010] After receiving the analog feedback information from the user terminal, determine the matching index between the user driving record and the preset inspection road image, so as to generate a maintenance inspection sequence corresponding to each maintenance inspection section according to the maintenance standard ladder of the corresponding maintenance inspection score when the matching index is greater than a preset threshold;

[0011] According to the pre-trained image recognition model and the user driving record, determine the area to be maintained and its defect impact value, so as to generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect impact value, and send it to the management terminal for display; wherein, the defect impact value represents the degree of influence of the defect on different types of passing objects.

[0012] In an implementation manner of the present application, determining a plurality of maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from a plurality of user terminals specifically includes:

[0013] According to each user driving record, determine the driving record data set corresponding to the same driving route;

[0014] Match the driving record data set with the pre-clustered abnormal comparison data to determine each driving abnormal data;

[0015] According to each driving abnormal data and its corresponding timestamp, calculate the driving abnormal feedback density corresponding to each preset grid section; wherein, the timestamp is associated with the section position; the driving abnormal feedback density is calculated based on each driving abnormal data in the corresponding preset grid section and the section area of the preset grid section;

[0016] Splice the preset grid sections with the driving abnormal feedback density greater than a predetermined value according to the section position to generate the maintenance inspection section; wherein, the maintenance inspection section includes one or more of the preset grid sections.

[0017] In an implementation manner of the present application, calculating the driving abnormal feedback density corresponding to each preset grid section specifically includes:

[0018] Determine the number of abnormalities and the section area of the driving abnormal data in the same preset grid section;

[0019] Input the number of abnormalities, the section area, and the preset maintenance inspection duration into the density calculation formula to determine the driving abnormal feedback density; the density calculation formula is as follows:

[0020]

[0021] where ρ iis the driving anomaly feedback density of the i-th preset grid road section; F i is the number of anomalies of the i-th preset grid road section; L i is the road section area of the i-th preset grid road section; T gi is the proportion of the traffic peak period corresponding to the i-th preset grid road section; T is the preset maintenance inspection duration corresponding to the i-th preset grid road section; α i is the preset adjustment coefficient corresponding to the i-th preset grid road section.

[0022] In an implementation manner of the present application, based on the driving record data corresponding to each maintenance inspection road section and the preset road section map, the corresponding maintenance inspection score is determined, specifically including:

[0023] According to the driving record data corresponding to the maintenance inspection road section and the preset road section map, the corresponding driving anomaly influencing factors are determined; the driving anomaly influencing factors at least include the following dimensional influencing factors: average speed, speed standard deviation, acceleration change rate, average bump intensity, road grade, traffic flow, and location environment influencing factors;

[0024] According to the driving anomaly influencing factors and the preset influencing factor weights, the maintenance inspection sub-scores of each driving anomaly influencing factor are calculated;

[0025] The arithmetic square root of the sum value of each maintenance inspection sub-score is used as the maintenance inspection score.

[0026] In an implementation manner of the present application, according to the maintenance inspection score and the preset road maintenance simulation model, a preset inspection road image corresponding to the maintenance inspection road section is generated, specifically including:

[0027] According to the location environment influencing factors, the inspection road image sub-model corresponding to the preset road maintenance simulation model is matched; wherein, the preset road maintenance simulation model includes a plurality of pre-trained inspection road image sub-models corresponding to different location environment influencing factors; the inspection road image sub-model is a generative adversarial network GAN model;

[0028] The maintenance inspection scores are normalized, and according to the road section information of the maintenance inspection road section, a maintenance inspection evaluation feature vector corresponding to the maintenance inspection score one by one is generated; the road section information includes location coordinates, road section length, road section width, and road type;

[0029] According to the maintenance inspection evaluation feature vector and the inspection road image sub-model, the preset inspection road image is generated.

[0030] In an implementation manner of the present application, determining the matching index between the user driving record and the preset inspection road image specifically includes:

[0031] When the simulation feedback information is accurate image simulation, according to the user driving record, determine the driving trajectory point vector, speed change curve, and bump intensity curve on the maintenance inspection section;

[0032] According to the preset inspection road image, determine the corresponding defect feature vector and road flatness feature vector;

[0033] Through the preset dynamic time warping (DTW) algorithm, calculate the first matching degree between the driving trajectory point vector and the defect feature vector;

[0034] Calculate the Pearson correlation coefficient between the speed change curve and the defect feature vector as the second matching degree;

[0035] Calculate the mean square error between the bump intensity curve and the road flatness feature vector as the third matching degree;

[0036] According to the first matching degree, the second matching degree, the third matching degree, the preset matching degree weight group, and the preset matching index calculation formula, determine the matching index.

[0037] In an implementation manner of the present application, the preset matching index calculation formula is as follows:

[0038]

[0039] Where M is the matching index; β 1 、β 2 、β 3 are respectively the first matching degree weight, the second matching degree weight, and the third matching degree weight in the preset matching degree weight group, and β 1 +β 2 +β 3 =1; M 1 is the first matching degree; M 2 is the second matching degree; M 3 is the third matching degree.

[0040] In an implementation manner of the present application, according to the maintenance standard ladder corresponding to the maintenance inspection score, generating a maintenance inspection sequence corresponding to each maintenance inspection section specifically includes:

[0041] After determining the maintenance inspection score S, preset the first maintenance inspection score S 1 、the second maintenance inspection score S 2 、the third maintenance inspection score S 3 ;

[0042] According to the relationship between the maintenance inspection score and each maintenance inspection score value, determine the corresponding maintenance standard level, and add each of the maintenance inspection road sections to the maintenance inspection sequence in accordance with the maintenance standard level; the maintenance standard level and the maintenance inspection priority are in a positive correlation relationship;

[0043] When S≥S 1 , determine that the maintenance standard level of the maintenance inspection road section is the first level;

[0044] When S 2 ≤S<S 1 , determine that the maintenance standard level of the maintenance inspection road section is the second level;

[0045] When S 3 ≤S<S 2 , determine that the maintenance standard level of the maintenance inspection road section is the third level;

[0046] When S<S 3 , determine that the maintenance standard level of the maintenance inspection road section is the fourth level.

[0047] In an implementation manner of the present application, according to the pre-trained image recognition model and the user driving record, determine the area of the defect to be maintained and its defect influence value, specifically including:

[0048] Input the preset inspection road image into the image recognition model to determine each area of the defect to be maintained;

[0049] According to the speed change rate, acceleration change rate, and average bump intensity corresponding to the area of the defect to be maintained in the user driving record, determine the initial defect influence value corresponding to the area of the defect to be maintained;

[0050] According to the vehicle type in the user driving record, the initial defect influence value, the area of the defect to be maintained, and the preset defect influence matching list, match the undetermined defect influence values corresponding to each type of passing object; wherein, the preset defect influence matching list includes a number of defect influence values of the same defect area for different types of passing objects under different defect degrees;

[0051] Take the maximum value of the initial defect influence value and each undetermined defect influence value as the defect influence value.

[0052] On the other hand, an embodiment of the present application further provides a road maintenance inspection system based on user driving records, and the system includes:

[0053] The first determination module is used to determine several maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from several user terminals; wherein, the maintenance inspection sections are divided based on the driving anomaly feedback density in the user driving records;

[0054] The second determination module is used to determine the corresponding maintenance inspection scores based on the driving record data corresponding to each maintenance inspection section and the preset section map;

[0055] The generation and sending module is used to generate a preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and the preset road maintenance simulation model, and send the preset inspection road image to the corresponding user terminal;

[0056] The third determination module is used to determine the matching index between the user driving record and the preset inspection road image after receiving the simulation feedback information from the user terminal, and in the case that the matching index is greater than the preset threshold, generate a maintenance inspection sequence corresponding to each maintenance inspection section according to the maintenance standard ladder of the corresponding maintenance inspection score;

[0057] The fourth determination module is used to determine the defect area to be maintained and its defect impact value according to the pre-trained image recognition model and the user driving record, and generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect impact value, and send it to the management terminal for display; wherein, the defect impact value represents the impact degree of the defect on different types of passing objects.

[0058] Compared with the prior art, the present application has the following remarkable effects:

[0059] Through the above technical solutions, by integrating user driving records, the real-time monitoring and comprehensive coverage of road conditions are realized, effectively improving the efficiency and accuracy of maintenance inspections. Compared with traditional manual inspections and fixed sensor monitoring, this method can detect potential problems more timely and avoid omissions in maintenance work. Secondly, the present application combines the actual driving experience of users, reducing the subjectivity and uncertainty of manual inspections. Through the feedback in the user driving records, the road conditions can be evaluated more objectively, and a more reasonable maintenance plan can be formulated. In addition, the present application enhances the interaction with actual users, improves the pertinence and effectiveness of maintenance work, can effectively improve road traffic capacity and safety, and can also meet the needs of different user groups. Furthermore, the present application effectively solves the technical problems of low maintenance efficiency, high cost, strong subjectivity and poor pertinence in the existing road maintenance inspection work. Description of the Drawings

[0060] The accompanying drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0061] Figure 1 It is a schematic flow chart of a road maintenance inspection method based on user driving records in an embodiment of the present application;

[0062] Figure 2 It is a schematic structural diagram of a road maintenance inspection system based on user driving records in an embodiment of the present application. Detailed implementation manners

[0063] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0064] The embodiments of the present application provide a road maintenance inspection method and system based on user driving records to solve the technical problems of low maintenance efficiency, high cost, strong subjectivity, and poor pertinence in existing road maintenance inspection work.

[0065] The following will describe each embodiment of the present application in detail with reference to the drawings.

[0066] The embodiments of the present application provide a road maintenance inspection method based on user driving records. As Figure 1 shown, the method may include steps S101 - S105:

[0067] S101. The server determines several maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from several user terminals.

[0068] Among them, the maintenance inspection sections are divided based on the driving anomaly feedback density in the user driving records. It should be noted that the server, as the execution subject of the road maintenance inspection method based on user driving records, is only an exemplary existence. The execution subject is not limited to the server, and the present application does not make specific limitations on this.

[0069] The user terminal can be an electronic device such as a user's mobile phone, computer, in-vehicle computer, etc. that is network-connected to the vehicle networking platform corresponding to this application. This application does not make specific limitations on this. Among them, the connection and information upload of the user terminal connected to the vehicle networking platform are established after the user's permission. The user driving record can be collected by several sensors pre-set on the user's vehicle, such as a speed sensor, an acceleration sensor, a camera, a Geographic Information System (GIS) sensor, a vibration sensor, etc.

[0070] The user terminal can choose to upload the user driving record to the server in real time, or can selectively upload its own user driving record after the trip. This application does not make specific limitations on this.

[0071] In the embodiment of this application, according to the user driving records from several user terminals, several maintenance inspection sections corresponding to the corresponding driving routes are determined, specifically including:

[0072] According to each user driving record, a driving record data set corresponding to the same driving route is determined. The driving record data set is matched with the pre-clustered abnormal comparison data to determine each driving abnormal data. According to each driving abnormal data and its corresponding timestamp, the driving abnormal feedback density corresponding to each preset grid section is calculated. Among them, the timestamp is associated with the section location. The driving abnormal feedback density is calculated based on each driving abnormal data within the corresponding preset grid section and the section area of the preset grid section. According to the section location, the preset grid sections with a driving abnormal feedback density greater than a predetermined value are spliced to generate a maintenance inspection section. Among them, the maintenance inspection section includes one or more preset grid sections.

[0073] That is to say, the server pre-stores abnormal comparison data obtained by clustering historical driving record data, such as clustering through the Kmeans clustering algorithm. The server extracts the driving records of users on the same driving route and constructs a driving record data set. Subsequently, the driving record data set and the abnormal comparison data are matched, so as to divide the driving record data into the clustering clusters of each abnormal comparison data, and the data belonging to the abnormal clustering cluster is used as driving abnormal data. According to the time stamp of the data, the driving position coordinates corresponding to the driving abnormal data on the driving route can be obtained, so as to calculate the driving abnormal feedback density. Generally speaking, if 100 vehicles pass through the same section of the road and 98 of them have driving abnormal data of sudden braking, then this section of the road has a high driving abnormal feedback density. Further, the server screens out the preset grid sections where the driving abnormal feedback density is greater than the preset value preset by the inspection-related personnel, and splices them into maintenance inspection sections according to the front-back position relationship. It is possible that multiple preset grid sections are not adjacent. This application will calculate the interval distance between two sections, set a blank area between the two sections and mark the interval distance, so as to splice them into a continuous maintenance inspection section. The above-mentioned preset grid sections can be grid areas with preset length and width set by the inspection personnel, such as 10 meters * 50 meters, 20 meters * 50 meters, etc.

[0074] The above method of dividing the maintenance inspection section by calculating the driving abnormal feedback density realizes the refined division of the section, helps to more accurately identify the problem section, and improves the accuracy of the inspection. According to the real-time change of the driving abnormal data, the maintenance inspection section is dynamically adjusted. This flexibility ensures the timeliness and effectiveness of the inspection work.

[0075] Specifically, calculating the driving abnormal feedback density corresponding to each preset grid section includes:

[0076] Determine the number of abnormalities and the section area of the driving abnormal data of the same preset grid section. Input the number of abnormalities, the section area, and the preset maintenance inspection duration into the density calculation formula to determine the driving abnormal feedback density. The density calculation formula is as follows:

[0077]

[0078] Among them, ρ i is the driving abnormal feedback density of the i-th preset grid section. F i is the number of abnormalities of the i-th preset grid section. L i is the section area of the i-th preset grid section. T gi is the proportion of the traffic peak period corresponding to the i-th preset grid section. T is the preset maintenance inspection duration corresponding to the i-th preset grid section. α i is the preset adjustment coefficient corresponding to the i-th preset grid section.

[0079] That is to say, this application can count the abnormal data of driving abnormal data and calculate the driving abnormal feedback density corresponding to the preset grid section through the above density calculation formula. For example, if there are 98 abnormal quantities corresponding to bumps or sudden brakes, the server calculates the driving abnormal feedback density through the road section area corresponding to the abnormal quantity, the proportion of the traffic peak period preset in the database connected to the server, the preset maintenance inspection duration, and the preset adjustment coefficient, and reflects the abnormal influence degree of this road section. The preset adjustment coefficient is a preset value used to adjust the influence degree of the trigonometric function on the density, and this application does not make specific limitations on this.

[0080] Calculating the driving abnormal feedback density by introducing multiple dimensions such as the proportion of the traffic peak period, the preset maintenance inspection duration, and the preset adjustment coefficient improves the accuracy and comprehensiveness of the calculation. And more accurate density calculation helps to optimize the allocation of inspection resources and ensure that key road sections are maintained in a timely and effective manner.

[0081] S102. The server determines the corresponding maintenance inspection score based on the driving record data corresponding to each maintenance inspection road section and the preset road section map.

[0082] In the embodiment of this application, determining the corresponding maintenance inspection score based on the driving record data corresponding to each maintenance inspection road section and the preset road section map specifically includes:

[0083] According to the driving record data corresponding to the maintenance inspection road section and the preset road section map, determine the corresponding driving abnormal influence factors. The driving abnormal influence factors at least include the following dimensional influence factors: average speed, speed standard deviation, acceleration change rate, average bump intensity, road grade, traffic flow, and location environment influence factors. According to the driving abnormal influence factors and the preset influence factor weights, calculate the maintenance inspection sub-scores of each driving abnormal influence factor. Take the arithmetic square root of the sum value of each maintenance inspection sub-score as the maintenance inspection score.

[0084] Among them, the preset road section map is pre-stored in the server. The server matches the location environment influence factors included in the preset road section map through the location coordinates corresponding to the maintenance inspection road section. The location coordinates can be longitude and latitude, and the location environment influence factors can include road geometric features such as road width, curvature, slope, etc.; traffic facilities such as traffic lights, signs and markings, guardrails, etc.; surrounding environment such as building density, greening degree, noise level, etc. At the same time, the server extracts each driving abnormal influence factor in the driving record data and calculates the corresponding maintenance inspection sub-score for each.

[0085] Specifically, the maintenance inspection sub-score corresponding to the average speed is:

[0086]

[0087] where, f 1 (v avg,i ) is the sub - score of maintenance inspection corresponding to the average speed, v avg,i is the average speed on the i - th maintenance inspection section, is the preset maximum recommended speed under the influence of the road grade R i on the i - th maintenance inspection section; is the preset maximum traffic flow; Q i is the traffic flow. Among them, the average speed is an important indicator to measure the road traffic efficiency. In road maintenance inspection, the average speed can reflect the road usage condition and traffic fluency. A higher average speed usually means good road conditions and less traffic congestion. This application can measure the average speed of vehicles under different road grades and traffic flows through on - site measurement, and combine road design standards, traffic flow data, etc., to establish a relationship model between the average speed and road grade, traffic flow, which can be a functional relationship among the three. Further, through experimental verification and adjustment of model parameters, it can accurately reflect the actual situation, and can be specifically calculated by users according to the actual scenario.

[0088] The sub - score of maintenance inspection corresponding to the speed standard deviation is:

[0089]

[0090] where, f 2 (σ v,i ) is the sub - score of maintenance inspection corresponding to the speed standard deviation, σ v,i is the speed standard deviation calculated based on the speeds of each driving record data passing through the maintenance inspection section, σ max is the preset maximum value of the speed standard deviation. The speed standard deviation reflects the dispersion degree of vehicle speeds and is an important indicator to measure road safety. A larger speed standard deviation means a greater difference in vehicle speeds on the road, which may increase the risk of traffic accidents. This application can establish a relationship model between the speed standard deviation and road safety by on - site measurement of the speed standard deviation under different road conditions, combined with traffic accident data, road design standards, etc., and establish it specifically according to the actual measurement data. Similarly, it is necessary to verify and adjust the model parameters through experiments to ensure the accuracy and reliability of the model.

[0091] The sub - score of maintenance inspection corresponding to the acceleration change rate is:

[0092] f 3 (a freq,i ) = log(1 + a freq,i )

[0093] where, f 3 (afreq,i ) is the sub - score of the maintenance inspection for the acceleration change rate; a freq,i is the acceleration change rate. The acceleration change frequency reflects the smoothness of the vehicle during driving. Frequent acceleration changes may lead to problems such as driver fatigue, passenger discomfort, and vehicle wear. This application can simulate or measure the acceleration change frequency under different road conditions in the field, and combine the feedback of drivers and passengers, vehicle wear data, etc., to establish a relationship model between the acceleration change frequency and road smoothness. Using a logarithmic function to reduce the impact of the acceleration change frequency on the score is based on the analysis of experimental data and statistical results, aiming to make the score smoother and more reasonable.

[0094] The average value of the bump intensity b avg,i The corresponding sub - score of the maintenance inspection is:

[0095] f 4 (b avg,i ) = b avg,i

[0096] The bump intensity is an index to measure the flatness of the road surface. A larger bump intensity means the road surface is uneven, which may affect driving comfort and safety. This application measures the bump intensity under different road conditions in the field, and combines the evaluation criteria in aspects such as driving comfort and safety to establish a relationship model between the bump intensity and the road surface flatness. In practical applications, the average value of the bump intensity can be directly used as an input item for scoring, or combined with other factors for comprehensive analysis.

[0097] The sub - score of the maintenance inspection corresponding to the road grade is:

[0098]

[0099] where D represents the highest level of the road grade, and f 5 (R i ) is the sub - score of the maintenance inspection corresponding to the normalized road grade. The road grade reflects the comprehensive level in aspects such as the road design standard, traffic capacity, and safety. Different grades of roads should have different scoring standards and requirements in the maintenance inspection. This application can divide different road grades and establish a relationship model between the road grade and the score according to the road design standard, traffic flow data, traffic accident records, etc. In practical applications, the score can be normalized according to the road grade for comparison and analysis between different roads.

[0100] The sub - score of the maintenance inspection corresponding to the traffic flow is:

[0101]

[0102] where f6 (Q i ) is the sub - score of maintenance inspection corresponding to traffic flow, where Q avg represents the preset average traffic flow value. Traffic flow is an important indicator to measure the road traffic capacity. A larger traffic flow means greater road traffic pressure, and more frequent maintenance inspections and repair work may be required. By measuring the traffic flow under different road conditions on - site and combining road design standards, traffic congestion conditions, traffic accident records, etc., a relationship model between traffic flow and road traffic capacity can be established. In practical applications, the traffic flow can be normalized to the ratio relative to the average value for comparison and analysis between different roads.

[0103] The sub - score of maintenance inspection corresponding to location environment impact factors is as follows:

[0104]

[0105] where, f 7 (E i ) is the sub - score of maintenance inspection corresponding to location environment impact factors, where E i represents the environmental impact factor vector composed of location environment impact factors, m is the number of elements in the environmental impact factor vector, and each vector element represents an impact factor; ω k represents the weight coefficient of the k - th impact factor, and e k,i represents the value of the k - th impact factor of the i - th maintenance inspection section. The above formula can map multiple impact factors into a single value. Environmental impact factors include multiple aspects such as road geometric features, traffic facilities, and climate conditions, which jointly affect the road usage condition and safety. In this application, by on - site investigation and analysis of environmental impact factors under different road conditions and combining traffic accident data, road maintenance records, etc., a relationship model between environmental impact factors and road safety can be established. In practical applications, a single value can be obtained through comprehensive calculation based on the weight coefficient and value of environmental impact factors as an input item for scoring.

[0106] Furthermore, this application calculates the maintenance inspection score through the following formula:

[0107]

[0108] where S represents the maintenance inspection score, γ j represents the weight coefficient of the abnormal driving impact factor of the j - th abnormal driving impact factor, f j(…) represents the sub - score of the maintenance inspection corresponding to the j - th driving anomaly influencing factor. The calculation formula of the maintenance inspection score is the result of comprehensively calculating the above - mentioned various functions and their input parameters. This formula aims to comprehensively reflect the comprehensive level of aspects such as the usage condition, safety, and traffic capacity of the road, providing a scientific basis for road maintenance inspections. Through on - site measurement and investigation of various data under different road conditions, combined with road design standards, traffic flow data, traffic accident records, etc., this application can establish the parameters and weight coefficients of each function in the maintenance inspection score calculation formula. In practical applications, it is necessary to adjust and optimize the parameters and weight coefficients in the formula according to specific data to ensure the accuracy and reliability of the score. At the same time, it is also necessary to verify and update the formula regularly to adapt to the continuous changes and developments of road maintenance inspection work.

[0109] This application determines the maintenance inspection score by comprehensively considering influencing factors from multiple dimensions, improving the accuracy and comprehensiveness of the evaluation.

[0110] S103, the server generates a preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and the preset road maintenance simulation model, and sends the preset inspection road image to the corresponding user terminal.

[0111] In the embodiment of this application, the generation of the preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and the preset road maintenance simulation model specifically includes:

[0112] According to the location environment influencing factors, match the inspection road image sub - model corresponding to the preset road maintenance simulation model. Among them, the preset road maintenance simulation model includes multiple pre - trained inspection road image sub - models corresponding to different location environment influencing factors. The inspection road image sub - model is a Generative Adversarial Networks (GAN) model. Normalize each maintenance inspection score, and generate a maintenance inspection evaluation feature vector corresponding one - to - one with the maintenance inspection score according to the section information of the maintenance inspection section. The section information includes location coordinates, section length, section width, and road type. Generate a preset inspection road image according to the maintenance inspection evaluation feature vector and the inspection road image sub - model.

[0113] That is to say, the preset road maintenance simulation model of the present application is pre-trained with multiple inspection road image sub-models corresponding to different location environmental impact factors. Different location environmental impact factors may cause different diseases and maintenance requirements for the road. Therefore, the present application pre-divides the model according to environmental impact factors and generates corresponding inspection road image sub-models for specific environmental impact factors. For example, the environmental impact factors of large slope and heavy traffic correspond to one inspection road image sub-model, and the environmental impact factors of flat road and small traffic flow correspond to another inspection road image sub-model. The server also normalizes the maintenance inspection scores of each obtained maintenance inspection section so that they are between 0 and 1. The normalization formula is as follows:

[0114]

[0115] Among them, S norm is the normalized maintenance inspection score, S min is the minimum value among the maintenance inspection scores corresponding to the maintenance inspection section, and S max is the maximum value among the maintenance inspection scores.

[0116] The server constructs the normalized maintenance inspection score and the section information of the maintenance inspection section, including but not limited to position coordinates, section length, section width, road type, etc., into a maintenance inspection evaluation feature vector, and inputs it into the matched inspection road image sub-model. The inspection road image sub-model will determine the position, shape, etc. of the defects according to the element values in the maintenance inspection evaluation feature vector and add them to the generated road image. The above inspection road image sub-model is trained with a large amount of road image data in different maintenance conditions and can generate a preset inspection road image corresponding to the maintenance inspection section. The possible cracks, potholes, ruts and other defect positions and approximate shapes can be shown in the image.

[0117] The present application realizes the generation of personalized preset inspection road images by matching the inspection road image sub-models corresponding to different location environmental impact factors, which helps to improve the accuracy and credibility of the user simulation feedback.

[0118] S104. After the server receives the simulation feedback information from the user terminal, it determines the matching index between the user driving record and the preset inspection road image. When the matching index is greater than the preset threshold, it generates a maintenance inspection sequence corresponding to each maintenance inspection section according to the maintenance standard ladder of the corresponding maintenance inspection score.

[0119] In other words, the server can send the above-mentioned preset inspection road image to the user terminal. The user can generate simulated feedback information such as whether the image is accurate or mark the errors in the image by viewing the preset inspection road image. Among them, when the matching index is not greater than the preset threshold, an artificial inspection intervention prompt message can be generated to regenerate the inspection road image or stop further operations on the corresponding maintenance inspection section, such as determining the maintenance standard level and subsequent generation of the maintenance inspection sequence. The preset threshold can be set in actual use, and the present application does not make specific limitations thereon.

[0120] In the embodiment of the present application, the determination of the matching index between the user driving record and the preset inspection road image specifically includes:

[0121] When the simulated feedback information is that the image simulation is accurate, according to the user driving record, determine the driving trajectory point vector, speed change curve, and bump intensity curve on the maintenance inspection section. According to the preset inspection road image, determine the corresponding defect feature vector and road flatness feature vector. Through the preset Dynamic Time Warping (DTW) algorithm, calculate the first matching degree between the driving trajectory point vector and the defect feature vector. Calculate the Pearson correlation coefficient between the speed change curve and the defect feature vector as the second matching degree. Calculate the mean square error between the bump intensity curve and the road flatness feature vector as the third matching degree. According to the first matching degree, the second matching degree, the third matching degree, the preset matching degree weight group, and the preset matching index calculation formula, determine the matching index.

[0122] That is to say, in response to the simulated feedback information that the image simulation is accurate, the server will extract the driving trajectory (which can be longitude and latitude coordinates) of the vehicle in the maintenance inspection section in the user driving record and generate a driving trajectory point vector; extract the speed change in the maintenance inspection section and generate a speed change curve; extract the vibration data in the maintenance inspection section as bump data, and generate a bump intensity curve with time as the abscissa and bump degree as the ordinate. The bump degree can be characterized by measuring the acceleration value through an acceleration sensor installed on the vehicle chassis.

[0123] The server will also extract features from the preset inspection road images to obtain defect feature vectors and road flatness feature vectors in the images. Among them, feature extraction can be achieved by a pre-trained convolutional neural network model or by other means, and this application does not make specific limitations on this. This application uses the DTW algorithm to calculate the first matching degree, calculates the second matching degree between the speed change curve and the defect feature vector through the Pearson correlation coefficient, and calculates the third matching degree between the bump intensity curve and the road flatness feature vector through the mean square error. Among them, for the above speed change curve and bump intensity curve, the curve data can be extracted first, and after constructing one-dimensional arrays with the same length as the defect feature vector and the road flatness feature vector respectively, the above matching degree calculation is performed.

[0124] Among them, the formula for calculating the preset matching index in this application is as follows:

[0125]

[0126] Among them, M is the matching index. β 1 、β 2 、β 3 are respectively the first matching degree weight, the second matching degree weight, and the third matching degree weight in the preset matching degree weight group, and β 1 +β 2 +β 3 =1. M 1 is the first matching degree. M 2 is the second matching degree. M 3 is the third matching degree.

[0127] In road maintenance inspections, the matching degree between the driving trajectory and the road features in the image, the correlation between the speed change curve and the defect features in the image, and the difference between the bump data and the road flatness features in the image have different degrees of importance for evaluating the actual road conditions and maintenance needs. The driving trajectory matching degree can intuitively reflect the consistency between the actual road passage path and the preset inspection road image. If the difference is large, it may mean that the actual road conditions are significantly different from the expected ones; the correlation of the speed change curve reflects the impact of road defects on the vehicle driving speed and is closely related to road safety; the difference degree of the bump data directly reflects the flatness of the road and affects driving comfort and vehicle wear. By setting different weights, the different importance of these factors can be highlighted to accurately evaluate the road conditions. In addition, the weight setting also needs to consider the actual driving experience of users on the road and the actual needs of road maintenance. If the road bumps seriously affect driving comfort and the service life of vehicle components, then the weight of the difference degree between the bump data and the road flatness features should be relatively high; if the road defects have a greater impact on the vehicle speed change and pose a threat to driving safety, the weight of the correlation between the speed change curve and the defect features needs to be increased. This can make the matching index more in line with the actual application scenario and provide a more targeted basis for road maintenance decisions.

[0128] Based on this, the present application can collect a large amount of historical road maintenance data, including driving records of different roads, corresponding preset inspection road images, actual maintenance situations, and user feedback, etc. Deeply analyze these data and statistically analyze the actual roles of different factors in evaluating road conditions and maintenance requirements. For example, analyze the influence degree of driving trajectories, speed changes, and bump data on determining the maintenance plan in the roads that have been maintained, and determine the importance ratio of each factor through data statistics and analysis, so as to initially determine the weight value. The present application can invite experts to evaluate and give suggestions on the weights of different factors according to their professional knowledge and rich experience. Combining the expert opinions, adjust the initially determined weight value to obtain the above-mentioned first matching degree weight, second matching degree weight, and third matching degree weight.

[0129] The exponential function and quadratic operation in the matching index calculation formula of the present application are used to reasonably integrate matching degrees, correlations, and differences of different natures. The exponential function can perform a non-linear transformation on the comprehensive result, making the matching index have better discrimination and stability between 0 and 1. The quadratic operation can amplify the influence of certain factors and highlight the role of important factors. According to the actual needs and data characteristics of road maintenance inspections, construct such a mathematical model so that the matching index can accurately reflect the matching degree between the actual road conditions and the preset inspection road images. Through a large amount of data testing and analysis, it is found that adopting such an operation method can make the matching index more accurately reflect the real situation of the road and provide a reliable basis for maintenance decisions. The present application can experiment and verify the coefficients in the formula when constructing the formula. By collecting data in different road scenarios, calculate the matching index under different coefficient combinations, and conduct a comparative analysis with the actual road conditions. Select the coefficient combination that makes the matching index most consistent with the actual road conditions, determine the values of each coefficient in the formula, so as to determine the specific quantitative relationship between the matching degree and the matching index, and ensure the effectiveness and accuracy of the formula.

[0130] The present application calculates the matching index by introducing multiple dimensions such as driving trajectory point vectors, speed change curves, and bump intensity curves, improving the accuracy and comprehensiveness of the matching. At the same time, it also quantifies the matching index and calculates the matching index through a specific formula, realizing the quantitative evaluation of the matching index. This helps to more intuitively understand the matching degree between the user's driving record and the preset inspection road image, and avoid wasting resources due to analyzing user driving records irrelevant to the road. The quantified matching index helps to optimize the formulation of the maintenance inspection sequence and ensure that key sections are given priority for maintenance.

[0131] In addition, the server can also send reward points to the user terminal that sends simulated feedback information to encourage users to participate in assisting with the maintenance inspection work, which can improve the user's sense of participation and mobilize the user's enthusiasm for participation.

[0132] In another embodiment of the present application, the above-mentioned maintenance standard ladder according to the corresponding maintenance inspection score is used to generate a maintenance inspection sequence corresponding to each maintenance inspection section, specifically including:

[0133] After determining the maintenance inspection score S, a first maintenance inspection score S 1 and a second maintenance inspection score S 2 and a third maintenance inspection score S 3 are preset in advance. According to the relationship between the maintenance inspection score and each maintenance inspection score, the corresponding maintenance standard level is determined, and each maintenance inspection section is added to the maintenance inspection sequence in turn according to the maintenance standard level. The maintenance standard level and the maintenance inspection priority are in a positive correlation. When S≥S 1 , it is determined that the maintenance standard ladder of the maintenance inspection section is the first level. When S 2 ≤S<S 1 , it is determined that the maintenance standard ladder of the maintenance inspection section is the second level. When S 3 ≤S<S 2 , it is determined that the maintenance standard ladder of the maintenance inspection section is the third level. When S<S 3 , it is determined that the maintenance standard ladder of the maintenance inspection section is the fourth level.

[0134] The above-mentioned first maintenance inspection score S 1 , second maintenance inspection score S 2 and third maintenance inspection score S 3 are set by inspection-related workers according to the actual usage scenario, and the present application does not make specific limitations in this regard. The present application can sort the maintenance inspection sections according to the maintenance standard ladder, first number the high-level sections, then number the medium-level sections, and finally number the low-level sections to form a maintenance inspection sequence. For sections of the same level, a secondary sorting is performed according to the product of the length and traffic flow of the section to ensure that sections with longer lengths and larger traffic flows are processed first.

[0135] The present application realizes a gradient maintenance strategy by setting different maintenance inspection scores and corresponding maintenance standard levels, so as to formulate a reasonable maintenance plan according to the actual conditions of different sections. The gradient maintenance strategy helps to optimize the processes of inspection and maintenance work and improve the overall efficiency and quality.

[0136] S105. The server determines the defect area to be maintained and its defect impact value according to the pre-trained image recognition model and the user's driving record, so as to generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect impact value, and send it to the management terminal for display.

[0137] Among them, the defect impact value represents the degree of influence of the defect on different types of passing objects.

[0138] In the embodiment of the present application, determining the defect area to be maintained and its defect influence value according to the pre-trained image recognition model and the user's driving record specifically includes:

[0139] Input the preset inspection road image into the image recognition model to determine each defect area to be maintained. According to the speed change rate, acceleration change rate, and average bump intensity corresponding to the defect area to be maintained in the user's driving record, determine the initial defect influence value corresponding to the defect area to be maintained. According to the vehicle type, initial defect influence value, defect area to be maintained, and preset defect influence matching list in the user's driving record, match the undetermined defect influence values corresponding to each type of passing object. The preset defect influence matching list includes several defect influence values of the same defect area for different types of passing objects under different defect degrees. Take the maximum value of the initial defect influence value and each undetermined defect influence value as the defect influence value.

[0140] The above image recognition model can be a pre-trained convolutional neural network model, which can identify each defect area to be maintained, such as defect areas such as cracks, potholes, and subsidence in the image, and at the same time output information such as the position coordinates, area, and depth corresponding to the defect area. The server further combines the speed change rate, acceleration change rate, and average bump intensity to calculate the initial defect influence value. The formula is as follows:

[0141]

[0142] where, I d is the initial defect influence value of the dth defect area to be maintained, k 1 , k 2 , k 3 are the first weight coefficient, second weight coefficient, and third weight coefficient preset respectively; Δv is the speed change rate of the vehicle in the defect area to be maintained; Δa is the acceleration change rate of the vehicle in the defect area to be maintained; b avg is the average bump intensity of the vehicle in the defect area to be maintained; A d is the area of the defect area to be maintained, h d is the depth of the defect area to be maintained.

[0143] It should be noted that the degrees of influence of the sudden speed drop amplitude, acceleration mutation value, steering wheel vibration intensity, defect area and depth on driving safety are different. A sudden speed drop may lead to accidents such as rear-end collisions. The acceleration mutation affects the vehicle handling stability. The steering wheel vibration intensity reflects the vehicle driving stability and the driving difficulty for the driver. The defect area and depth are directly related to the road structure integrity. The weight setting needs to be determined according to the severity of the influence of these factors on driving safety. The greater the influence, the higher the weight. For example, on a high-speed section, the sudden speed drop amplitude poses a great threat to safety, and the corresponding weight k 1 should be relatively high; while on some sections with lower vehicle speeds, the steering wheel vibration intensity has a more obvious impact on the driver's comfort and controllability, and the corresponding weight k 3 can be appropriately increased.

[0144] For different types of roads, such as highways, urban roads, rural roads, etc., the vehicle driving characteristics and the sensitivity to defects are different. On highways, the vehicle speed is fast, and it is more sensitive to sudden speed drops and acceleration mutations. In urban roads, the steering wheel vibration intensity may have a greater impact on the driver because urban driving requires frequent steering wheel operation. Rural roads may pay more attention to the impact of defect area and depth on the vehicle chassis. Therefore, the weight setting should combine the road type and usage scenario, highlighting the weight of the factors that have a greater impact on driving in that scenario. For roads where large trucks often travel, the defect depth has a more critical impact on the vehicle, and the weight corresponding to the depth can be appropriately increased. The above first weight coefficient, second weight coefficient, and third weight coefficient can refer to the acquisition methods of the above first matching degree weight, second matching degree weight, and third matching degree weight, and are obtained by expert scoring.

[0145] First, the specific values of the weights and each factor in this application are squared or other operations are performed, and then comprehensive calculations are carried out through square roots, arctangent functions, etc. This operation method makes the weights and each factor correlated with each other and jointly determines the defect influence value. For example, the change in weight will change the magnitudes of the terms inside the square root, thereby affecting the value of the entire radical, and ultimately affecting the magnitude of the defect influence value. At the same time, the introduction of the arctangent function also makes the influence of the defect depth h d on the defect influence value show a non-linear change. In this complex operation relationship, the weights interact with other factors to accurately reflect the comprehensive effect of each factor on the road defect influence.

[0146] Subsequently, the server matches the vehicle type, initial defect influence value, and the lane position of the defect area to be maintained in the user's driving record from the preset defect influence matching list to obtain the pending defect influence values of this defect area to be maintained for different types of passing objects.

[0147] The types of passing objects include small cars, large buses, motorcycles, bicycles, pedestrians, etc. The same defect area may have different degrees of impact on different types of passing objects.

[0148] Through the above solution, this application realizes accurate identification of defective areas to be maintained through comprehensive analysis of image recognition models and user driving records, which helps to improve the pertinence and effectiveness of maintenance work. At the same time, the evaluation of the impact value of defects on different types of traffic objects and vehicle types is considered, and personalized defect impact evaluation is realized, so that more detailed and scientific maintenance plans can be formulated to meet the needs of different user groups.

[0149] In an embodiment of the present application, the server will further generate a maintenance inspection plan according to the order in the maintenance inspection sequence, giving priority to the maintenance inspection sections that are at the front of the sequence and have large defect impact values. The maintenance inspection plan includes the type of maintenance operation (such as filling potholes and repairing cracks), the required materials, the equipment list and the estimated maintenance time. Subsequently, the server can send the maintenance inspection plan to the management terminal for display through wireless communication technology. Specifically, the maintenance inspection plan is displayed in a visual manner, including marking the location of the maintenance inspection section in the form of a map, distinguishing the maintenance standard levels with different colors, and displaying the defect impact value and maintenance operation details of each section. Moreover, on the map, the display size of the section is proportional to the defect impact value.

[0150] The management terminal can be understood as the mobile phones, computers and other devices of road maintenance and inspection personnel, and this application does not make specific limitations on this.

[0151] Through the above-mentioned technical solution, the user's driving records are integrated to achieve real-time monitoring and comprehensive coverage of road conditions, effectively improving the efficiency and accuracy of maintenance inspections. Compared with traditional manual inspections and fixed sensor monitoring, this method can detect potential problems more promptly and avoid omissions in maintenance work. Secondly, this application combines the user's actual driving experience to reduce the subjectivity and uncertainty of manual inspections. Through the feedback from the user's driving records, the road conditions can be evaluated more objectively and a more reasonable maintenance plan can be formulated. In addition, this application enhances the interaction with actual users, improves the pertinence and effectiveness of maintenance work, can effectively improve road capacity and safety, and can meet the needs of different user groups. Furthermore, this application effectively solves the technical problems of low maintenance efficiency, high cost, strong subjectivity, and poor pertinence in existing road maintenance inspections.

[0152] Figure 2 A schematic diagram of a road maintenance inspection system based on user driving records provided in an embodiment of the present application is shown in FIG. Figure 2As shown in the figure, the road maintenance inspection system 200 based on user driving records can execute the corresponding steps of the above-mentioned road maintenance inspection method based on user driving records, which specifically includes:

[0153] A first determination module 201, configured to determine several maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from several user terminals. Among them, the maintenance inspection sections are divided based on the driving anomaly feedback density in the user driving records. A second determination module 202, configured to determine the corresponding maintenance inspection scores based on the driving record data corresponding to each maintenance inspection section and a preset section map. A generation and sending module 203, configured to generate a preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and a preset road maintenance simulation model, and send the preset inspection road image to the corresponding user terminal. A third determination module 204, configured to determine the matching index between the user driving record and the preset inspection road image after receiving the simulation feedback information from the user terminal, so as to generate a maintenance inspection sequence corresponding to each maintenance inspection section according to the maintenance standard ladder of the corresponding maintenance inspection score when the matching index is greater than a preset threshold. A fourth determination module 205, configured to determine the to-be-maintained defect area and its defect influence value according to a pre-trained image recognition model and the user driving record, so as to generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect influence value, and send it to the management terminal for display. Among them, the defect influence value represents the influence degree of the defect on different types of passing objects.

[0154] Each embodiment in this application is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment.

[0155] The system provided by the embodiment of this application corresponds to the method one by one. Therefore, the system also has beneficial technical effects similar to those of the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the system will not be elaborated here.

[0156] It should also be noted that the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. Without further limitation, the element defined by the statement "including a..." does not exclude the existence of another identical element in the process, method, commodity or device including the said element.

[0157] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A road maintenance inspection method based on user driving records, characterized in that: The method comprises: Determine, according to user driving records from a plurality of user terminals, a plurality of maintenance inspection sections corresponding to the corresponding driving routes; wherein the maintenance inspection sections are divided based on the density of driving abnormality feedback in the user driving records; Determine a corresponding maintenance inspection score based on the driving record data corresponding to each maintenance inspection road section and a preset road section map; Generate a preset inspection road image corresponding to the maintenance inspection road section according to the maintenance inspection score and the preset road maintenance simulation model, and send the preset inspection road image to the corresponding user terminal; After receiving the simulation feedback information from the user terminal, determining the matching index between the user driving record and the preset inspection road image, and generating a maintenance inspection sequence corresponding to each maintenance inspection road section according to the maintenance standard ladder corresponding to the maintenance inspection score when the matching index is greater than a preset threshold; According to the pre-trained image recognition model and the user's driving record, the defect area to be maintained and its defect impact value are determined, so as to generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect impact value, and send it to the management terminal for display; wherein the defect impact value represents the degree of influence of the defect on different types of passing objects.

2. A road maintenance inspection method based on user driving records according to claim 1, characterized in that: According to the user driving records from several user terminals, several maintenance inspection sections corresponding to the corresponding driving routes are determined, including: Determining a driving record data set corresponding to the same driving route according to the driving records of each user; Matching the driving record data set with the pre-clustered abnormal comparison data to determine each driving abnormality data; According to each of the driving abnormality data and its corresponding timestamp, the driving abnormality feedback density corresponding to each preset grid road section is calculated; wherein the timestamp is associated with the road section position; the driving abnormality feedback density is calculated based on each of the driving abnormality data in the corresponding preset grid road section and the road section area of ​​the preset grid road section; According to the road section location, the preset grid sections with the driving abnormality feedback density greater than a predetermined value are spliced ​​to generate the maintenance inspection section; wherein the maintenance inspection section includes one or more preset grid sections.

3. A road maintenance inspection method based on user driving records according to claim 2, characterized in that: Calculating the driving abnormality feedback density corresponding to the preset grid sections respectively includes: Determining the abnormal number of the abnormal driving data and the area of ​​the road section of the same preset grid road section; The number of abnormalities, the area of ​​the road section and the preset maintenance inspection duration are input into a density calculation formula to determine the driving abnormality feedback density; the density calculation formula is as follows: Among them, ρ i F is the driving abnormality feedback density of the i-th preset grid section; i is the number of anomalies in the i-th preset grid section; L i is the road section area of ​​the i-th preset grid section; T gi is the proportion of peak traffic hours corresponding to the i-th preset grid section; T is the preset maintenance inspection time corresponding to the i-th preset grid section; α i is the preset adjustment coefficient corresponding to the i-th preset grid section.

4. A road maintenance inspection method based on user driving records according to claim 1, characterized in that: Based on the driving record data corresponding to each maintenance inspection section and the preset section map, the corresponding maintenance inspection score is determined, specifically including: Determine corresponding driving abnormality influencing factors according to the driving record data corresponding to the maintenance inspection section and the preset section map; the driving abnormality influencing factors at least include the following dimensional influencing factors: average speed, speed standard deviation, acceleration change rate, bump intensity average value, road grade, traffic flow and location environment influencing factors; Calculate the maintenance inspection sub-score of each driving abnormality influencing factor according to the driving abnormality influencing factor and the preset influencing factor weight; The arithmetic square root of the sum of the maintenance inspection sub-scores is used as the maintenance inspection score.

5. A road maintenance inspection method based on user driving records according to claim 4, characterized in that: Generating a preset inspection road image corresponding to the maintenance inspection road section according to the maintenance inspection score and the preset road maintenance simulation model, specifically comprising: According to the location environment influencing factors, the inspection road image sub-model corresponding to the preset road maintenance simulation model is matched; wherein the preset road maintenance simulation model includes a plurality of pre-trained inspection road image sub-models corresponding to different location environment influencing factors; the inspection road image sub-model is a generative adversarial network GAN model; Normalizing each of the maintenance inspection scores, and generating a maintenance inspection evaluation feature vector corresponding to the maintenance inspection score one by one according to the section information of the maintenance inspection section; the section information includes location coordinates, section length, section width, and road type; The preset inspection road image is generated according to the maintenance inspection evaluation feature vector and the inspection road image sub-model.

6. A road maintenance inspection method based on user driving records according to claim 1, characterized in that: Determining a matching index between the user driving record and the preset inspection road image specifically includes: In the case where the simulation feedback information indicates that the image simulation is accurate, determining the driving trajectory point vector, the speed change curve and the bump intensity curve on the maintenance inspection section according to the user driving record; Determine the corresponding defect feature vector and road smoothness feature vector according to the preset inspection road image; Calculating a first matching degree between the driving trajectory point vector and the defect feature vector by using a preset dynamic time warping (DTW) algorithm; Calculating the Pearson correlation coefficient between the speed change curve and the defect feature vector as a second matching degree; Calculating a mean square error between the bump intensity curve and the road smoothness feature vector as a third matching degree; The matching index is determined according to the first matching degree, the second matching degree, the third matching degree, a preset matching degree weight group and a preset matching index calculation formula.

7. A road maintenance inspection method based on user driving records according to claim 6, characterized in that: The preset matching index calculation formula is as follows: Among them, M is the matching index; β1, β2, and β3 are respectively the first matching weight, the second matching weight, and the third matching weight in the preset matching weight group, β1+β2+β3=1; M1 is the first matching degree; M2 is the second matching degree; and M3 is the third matching degree.

8. A road maintenance inspection method based on user driving records according to claim 1, characterized in that: According to the maintenance standard level corresponding to the maintenance inspection score, a maintenance inspection sequence corresponding to each maintenance inspection section is generated, specifically including: After the maintenance inspection score S is determined, a first maintenance inspection score S1, a second maintenance inspection score S2, and a third maintenance inspection score S3 are preset; According to the relationship between the maintenance inspection score and each maintenance inspection score, a corresponding maintenance standard level is determined, and each maintenance inspection section is sequentially added to the maintenance inspection sequence according to the maintenance standard level; the maintenance standard level is positively correlated with the maintenance inspection priority; When S≥S1, determine that the maintenance standard level of the maintenance inspection section is the first level; When S2≤S<S1, determine that the maintenance standard level of the maintenance inspection section is the second level; When S3≤S<S2, determine that the maintenance standard level of the maintenance inspection section is the third level; When S<S3, determine that the maintenance standard level of the maintenance inspection section is the fourth level.

9. A road maintenance inspection method based on user driving records according to claim 1, characterized in that: According to the pre-trained image recognition model and the user driving record, determine the area with defects to be maintained and its defect impact value, specifically including: Input the preset inspection road image into the image recognition model to determine each area with defects to be maintained; According to the speed change rate, acceleration change rate and average bump intensity corresponding to the area with defects to be maintained in the user driving record, determine the initial defect impact value corresponding to the area with defects to be maintained; According to the vehicle type in the user driving record, the initial defect impact value, the area with defects to be maintained and the preset defect impact matching list, match the pending defect impact values corresponding to each type of passing object; wherein, the preset defect impact matching list includes several defect impact values of the same defect area for different types of passing objects under different defect degrees; Take the maximum value of the initial defect impact value and each pending defect impact value as the defect impact value.

10. A road maintenance inspection system based on user driving records, characterized in that: The system includes: The first determination module is used to determine several maintenance inspection sections corresponding to the corresponding driving routes according to the user driving records from several user terminals; wherein, the maintenance inspection sections are divided based on the driving anomaly feedback density in the user driving records; The second determination module is used to determine the corresponding maintenance inspection score based on the driving record data corresponding to each maintenance inspection section and the preset section map; The generation and sending module is used to generate a preset inspection road image corresponding to the maintenance inspection section according to the maintenance inspection score and the preset road maintenance simulation model, and send the preset inspection road image to the corresponding user terminal; The third determination module is used to determine the matching index between the user driving record and the preset inspection road image after receiving the simulation feedback information from the user terminal, and in the case that the matching index is greater than the preset threshold, generate a maintenance inspection sequence corresponding to each maintenance inspection section according to the maintenance standard level of the corresponding maintenance inspection score; The fourth determination module is used to determine the area with defects to be maintained and its defect impact value according to the pre-trained image recognition model and the user driving record, and generate a maintenance inspection plan for each maintenance inspection section according to the maintenance inspection sequence and the defect impact value, and send it to the management terminal for display; wherein, the defect impact value represents the degree of influence of the defect on different types of passing objects.

Images