Piezoelectric quartz vehicle load dynamic monitoring system and method based on visual recognition

Abstract

The application discloses a piezoelectric quartz vehicle load dynamic monitoring system and method based on visual identification and belongs to the technical field of vehicle dynamic weighing. The system comprises an image information data acquisition module, a charge signal data acquisition module, a database, a model management module, a real-time monitoring module and a warning module. The image information data acquisition module is used for acquiring image information data. The charge signal data acquisition module is used for acquiring charge signal data, calculating vehicle axle load and vehicle estimated total load. The database is used for storing historical data. The model management module is used for establishing a regional error analysis model and analyzing the influence of errors of each monitoring area on vehicle load dynamic monitoring results. The real-time monitoring module is used for outputting current vehicle estimated total load results. The warning module is used for real-time warning of overweight vehicles. The application compensates for errors of each monitoring area, thereby improving the weighing precision of vehicle load dynamic monitoring.

Description

Technical Field

[0001] This invention relates to the field of vehicle dynamic weighing technology, specifically to a piezoelectric quartz vehicle load dynamic monitoring system and method based on visual recognition. Background Technology

[0002] With the rapid development of intelligent transportation and technology-based enforcement of vehicle weight limits, the dynamic weighing technology of piezoelectric quartz sensors has been widely used in non-site enforcement scenarios of road traffic due to its advantages such as high precision and long lifespan. This technology collects the charge signals generated by wheel rolling and combines them with signal processing algorithms to realize the dynamic measurement of vehicle axle weight and total weight, meeting the vehicle monitoring needs under all-weather and high-traffic conditions.

[0003] In existing piezoelectric quartz vehicle load detection systems, one or two sets of quartz sensors are typically deployed along the lane. The system uses the piezoelectric charge generated by wheel rolling to integrate and obtain axle load data. It then calculates the wheelbase and identifies the number of axles based on the time difference between axles, and finally sums up the axle loads to obtain the total vehicle weight. However, existing solutions do not fully consider the accuracy differences between different sensors, which can easily lead to inconsistent weighing results for the same vehicle on different road sections. Furthermore, they are not sufficiently able to suppress external interference, especially on uneven road surfaces where vehicle bumps can introduce dynamic interference components into the pressure signals detected by the sensors, significantly affecting weighing stability. Therefore, current technology still suffers from poor robustness and insufficient detection accuracy in vehicle weighing under complex conditions. Summary of the Invention

[0004] The purpose of this invention is to provide a technical solution to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition, comprising: Smart cameras are deployed above the monitored road to collect and identify image information data of vehicles during their journey; the image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road; The monitored road is divided into several independent monitoring areas. Piezoelectric quartz sensing units are deployed in each monitoring area to collect charge signal data when the vehicle is rolled over. The collected charge signal data is analyzed to calculate the axle weight of the vehicle in each monitoring area at different timestamps, and the estimated total weight of each vehicle is obtained. A database is established to store image information data and charge signal data as historical data. Based on the historical image information data and charge signal data, a regional error analysis model is established. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, the influence of the error in each monitoring area on the dynamic monitoring results of vehicle load is analyzed, and the error compensation coefficient of each monitoring area is obtained. When a vehicle travels on the monitored road, the real-time collected image information data and charge signal data are analyzed. Combined with the established regional error analysis model, the estimated total weight of the vehicle is output, and real-time warnings are issued for overweight vehicles.

[0006] Furthermore, the method for calculating the vehicle axle weight at different timestamps for each monitoring area to obtain the estimated total vehicle weight is as follows: Based on the vehicle's trajectory on the monitoring road, the projection range of the vehicle's left and right wheels on the monitoring road is determined. Based on the projection range, the set of monitoring areas S that each vehicle axle needs to be associated with is determined. The collected charge signal data is analyzed, and based on the timestamp, the charge signal collected by each monitoring area within the set of monitoring areas S is determined to obtain the vehicle axle weight at different timestamps for each monitoring area. Based on the timestamp and the vehicle's unique identifier, all vehicle axle weights obtained by the identified vehicle are summed to obtain the estimated total vehicle weight W. Where m represents the total number of axles of the vehicle; This indicates the axle weight of vehicle axle j within monitoring area k.

[0007] Furthermore, the method and steps for analyzing the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load are as follows: S1. Retrieve historical image information data and historical charge signal data containing the actual total weight of vehicles from the database for analysis, obtain the actual total weight of different vehicles and the estimated total weight of vehicles, determine the driving trajectory and speed of each vehicle on the monitored road under the estimated total weight of each vehicle, and associate the axle weight of each vehicle obtained in each monitoring area under the estimated total weight of each vehicle with the timestamp. S2. Establish a regional error analysis model, using the vehicle's actual total weight and vehicle speed in different monitoring areas as inputs, and the vehicle axle load obtained from the monitoring areas as output. Analyze the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load, based on the following formula: ; in, This indicates the axle weight of vehicle axle j within monitoring area k; This indicates the actual axle load of vehicle axle j within monitoring area k; This indicates the vehicle speed of axle j within monitoring area k; a k b represents the dynamic influence coefficient of vehicle speed in monitoring area k; k c represents the vehicle speed saturation coefficient in monitoring area k; k The piezoelectric quartz sensor influence coefficient represents the monitoring area k; d k The fixed deviation coefficient representing the monitoring area k; W represents the hyperbolic tangent function; zIndicates the vehicle's true gross weight; S3. Substitute the analysis results of historical image information data and historical charge signal data in S1 into the conditional formula in S2 to obtain the error compensation coefficient for each monitoring area.

[0008] Furthermore, the method for outputting the current estimated total weight of the vehicle is as follows: Based on real-time image information data and real-time charge signal data, the current vehicle's trajectory and speed on the monitored road are determined, the axle load and corresponding vehicle speed of the current vehicle in each monitoring area are obtained, and the current estimated total weight of the vehicle is calculated based on the established regional error analysis model. : ; in, This indicates the axle weight of the vehicle currently in monitoring area k, measured for axle j. This indicates the vehicle speed of the current vehicle axle j within the monitoring area k.

[0009] Furthermore, the monitored road is divided into several monitoring areas of equal width along the transverse direction of the road. At least one set of piezoelectric quartz sensing units is deployed in each monitoring area. Each set of piezoelectric quartz sensing units includes two piezoelectric quartz sensors arranged in parallel front and back with a preset distance.

[0010] Furthermore, the overweight threshold of a vehicle is determined based on its unique identifier and the number of axles. The estimated total weight of the current vehicle is then compared with the overweight threshold. If the vehicle is not overweight, monitoring continues. If the vehicle is overweight, an early warning signal is triggered, and the information about the overweight vehicle is sent to the remote law enforcement platform.

[0011] The piezoelectric quartz vehicle load dynamic monitoring system based on visual recognition includes an image information data acquisition module, a charge signal data acquisition module, a database, a model management module, a real-time monitoring module, and an early warning module. The image information data acquisition module is used to deploy smart cameras above the monitored road to collect and identify image information data during vehicle movement; the image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road; the collected image information data is sent to the database and the real-time monitoring module; The charge signal data acquisition module is used to divide the monitored road into several independent monitoring areas, and to deploy piezoelectric quartz sensing units in each monitoring area to collect charge signal data when the vehicle is rolled over; to analyze the collected charge signal data, to calculate the axle weight of the vehicle in each monitoring area at different timestamps, and to obtain the estimated total weight of each vehicle; and to send the collected charge signal data to the database and the real-time monitoring module. The database is used to store image information data and charge signal data as historical data; The model management module is used to establish a regional error analysis model based on historical image information data and charge signal data. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, it analyzes the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load and obtains the error compensation coefficient for each monitoring area. The real-time monitoring module is used to analyze the real-time collected image information data and charge signal data when a vehicle is traveling on the monitored road, and output the estimated total weight of the vehicle in combination with the established regional error analysis model; and send the estimated total weight of the vehicle to the early warning module. The early warning module is used to monitor the estimated total weight of the vehicle and provide real-time warnings for overweight vehicles.

[0012] Furthermore, an interactive display platform is provided, allowing managers to view image information data and charge signal data of different vehicles and identify overweight vehicles.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: 1. By establishing a regional error analysis model, the influence of errors in each monitoring area on the dynamic monitoring results of vehicle load is analyzed. This allows for the precise separation and elimination of dynamic weighing errors caused by sensor characteristics, road surface potholes and settlement, and vehicle speed within a single monitoring area, effectively achieving accurate restoration of the vehicle's true axle load. 2. By compensating for errors in each monitoring area, the system can still measure accurately under complex road conditions, improving the weighing accuracy and robustness of the dynamic monitoring of vehicle load. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] The present invention provides the following technical solution: Please see Figure 1 In this first embodiment: the piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition includes: Smart cameras are deployed above the monitored road to collect and identify image information data of vehicles during their journey; the image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road.

[0017] In this embodiment, one or more smart cameras are deployed above the monitored road. The installation height and angle are designed to clearly cover the entire monitored road and provide a certain front and rear field of view. The smart cameras collect image information data during vehicle movement, and the embedded image processing unit identifies the image information data. The license plate number is extracted using a license plate recognition algorithm, and combined with the timestamp, the position coordinate changes of each identified vehicle are recorded, thereby generating the vehicle's driving trajectory and speed on the monitored road.

[0018] The monitored road is divided into several independent monitoring areas. Piezoelectric quartz sensing units are deployed in each monitoring area to collect charge signal data when the vehicle is rolled over. The collected charge signal data is analyzed to calculate the axle weight of the vehicle in each monitoring area at different timestamps, and the estimated total weight of each vehicle is obtained.

[0019] Furthermore, the monitored road is divided into several monitoring areas of equal width along the transverse direction of the road. At least one set of piezoelectric quartz sensing units is deployed in each monitoring area. Each set of piezoelectric quartz sensing units includes two piezoelectric quartz sensors arranged in parallel front and rear with a preset distance. Specifically, the transverse direction of the road is perpendicular to the direction of vehicle travel.

[0020] In this embodiment, based on the vehicle's unique identifier, driving trajectory, and timestamp, spatiotemporal matching is performed on the image information data and the charge signal data to obtain the timestamps and corresponding axle weights of a specific vehicle when passing through various monitoring areas. The method for calculating the vehicle axle weights acquired at different timestamps in each monitoring area to obtain the estimated total vehicle weight is as follows: Based on the vehicle's driving trajectory on the monitoring road, the projection intervals of the vehicle's left and right wheels on the monitoring road are determined. Based on the projection intervals, the set of monitoring areas S that each vehicle axle needs to be associated with is determined. For example, if vehicle axle i's left and right wheels span monitoring area 1 and monitoring area 2, the collected charge signal data is analyzed. Based on the timestamps, the charge signals collected in each monitoring area within the set of monitoring areas S are determined to obtain the vehicle axle weights acquired at different timestamps in each monitoring area. Based on the timestamps and the vehicle's unique identifier, all vehicle axle weights acquired by the identified vehicle are accumulated to obtain the estimated total vehicle weight W. Where m represents the total number of axles of the vehicle; This represents the axle weight of vehicle axle j in monitoring area k. Since the projection ranges of the left and right wheels of different axles on the monitoring road are different, the monitoring areas in the monitoring area set S corresponding to different axles of the vehicle are different.

[0021] A database is established to store image information data and charge signal data as historical data. Based on the historical image information data and charge signal data, a regional error analysis model is established. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load is analyzed, and the error compensation coefficient of each monitoring area is obtained.

[0022] Specifically, the method and steps for analyzing the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load are as follows: S1. Retrieve historical image information data and historical charge signal data containing the actual total weight of vehicles from the database for analysis, obtain the actual total weight of different vehicles and the estimated total weight of vehicles, determine the driving trajectory and speed of each vehicle on the monitored road under the estimated total weight of each vehicle, and associate the axle weight of each vehicle obtained in each monitoring area under the estimated total weight of each vehicle with the timestamp. S2. Establish a regional error analysis model, using the vehicle's actual total weight and vehicle speed in different monitoring areas as inputs, and the vehicle axle load obtained from the monitoring areas as output. Analyze the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load, based on the following formula: ; in, This indicates the axle weight of vehicle axle j within monitoring area k; This indicates the actual axle load of vehicle axle j within monitoring area k; This indicates the vehicle speed of axle j within monitoring area k; a k b represents the dynamic influence coefficient of vehicle speed in monitoring area k; k c represents the vehicle speed saturation coefficient in monitoring area k; k The piezoelectric quartz sensor influence coefficient represents the monitoring area k; d k The fixed deviation coefficient representing the monitoring area k; W represents the hyperbolic tangent function; z Indicates the vehicle's true gross weight; S3. Substitute the analysis results of historical image information data and historical charge signal data in S1 into the conditional formula in S2 to obtain the error compensation coefficient for each monitoring area.

[0023] It should be noted that the true total weight of the vehicle represents the actual total weight of the vehicle as a baseline value, which is specifically measured by a high-precision static weighing device. In this embodiment, vehicles are guided to voluntarily or according to a sampling ratio to enter the weighing area, and the total weight of the vehicle is recorded by a static weighbridge to obtain the true total weight of the vehicle. At the same time, the estimated total weight of the vehicle traveling on the monitored road is recorded, and the image information data and charge signal data collected by the vehicle are stored in the database as historical data.

[0024] In this embodiment, by establishing a regional error analysis model, the analysis results of historical data in S1 are substituted into the conditional formula of S2, and the actual total weight of different vehicles is used as W. z The training parameters use the vehicle's speed on the monitored road as... Training parameters, using vehicle axle loads acquired from each monitoring area as... The training parameters are fitted to the conditional formula using the least squares method, and the error compensation coefficient 'a' for each monitoring area is determined. k b k c k and d k Among them, the actual axle load of the vehicle Based on the vehicle's actual total weight W z and vehicle axle load It can be determined, and can be based on the error compensation coefficient and vehicle axle load. Directly determine the actual axle load of the vehicle Because piezoelectric quartz sensors in different transverse monitoring areas have individual differences and operating condition attenuation, and the road surface with different potholes in each monitoring area is different, all of these factors will affect the determination of vehicle axle load. By establishing a regional error analysis model, the influence of errors in each monitoring area on the dynamic monitoring results of vehicle load can be analyzed. This model can accurately separate and eliminate dynamic weighing errors caused by the sensor's own characteristics, road surface potholes and settlement, and vehicle speed in a single monitoring area, effectively achieving accurate restoration of the vehicle's true axle load.

[0025] When a vehicle travels on the monitored road, the real-time collected image information data and charge signal data are analyzed. Combined with the established regional error analysis model, the estimated total weight of the vehicle is output, and real-time warnings are issued for overweight vehicles.

[0026] Specifically, the method for outputting the estimated total weight of the current vehicle is as follows: Based on real-time image information data and real-time charge signal data, determine the current vehicle's trajectory and speed on the monitored road, obtain the vehicle's axle load and corresponding vehicle speed in each monitoring area, and calculate the estimated total weight of the current vehicle based on the established regional error analysis model. : ; in, This indicates the axle weight of the vehicle currently in monitoring area k, measured for axle j. This indicates the vehicle speed of the current vehicle axle j within the monitoring area k.

[0027] It should be noted that the regional error analysis model can be updated online or retrained offline based on the continuously accumulated historical data, thereby improving the system's adaptability to changes in the road environment. By compensating for the errors in each monitoring area, the system can still measure accurately under complex road conditions, improving the weighing accuracy and robustness of the dynamic monitoring of vehicle load.

[0028] Furthermore, the overweight threshold of a vehicle is determined based on its unique identifier and the number of axles. The estimated total weight of the current vehicle is then compared with the overweight threshold. If the vehicle is not overweight, monitoring continues. If the vehicle is overweight, an early warning signal is triggered, and the information about the overweight vehicle is sent to the remote law enforcement platform.

[0029] In this second embodiment: a piezoelectric quartz vehicle load dynamic monitoring system based on visual recognition is provided. The system includes an image information data acquisition module, a charge signal data acquisition module, a database, a model management module, a real-time monitoring module, and an early warning module. The image information data acquisition module is used to deploy smart cameras above the monitored road to collect and identify image information data during vehicle movement; the image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road; the collected image information data is sent to the database and the real-time monitoring module; The charge signal data acquisition module is used to divide the monitored road into several independent monitoring areas, and to deploy piezoelectric quartz sensing units in each monitoring area to collect charge signal data when the vehicle is rolled over; to analyze the collected charge signal data, to calculate the axle weight of the vehicle in each monitoring area at different timestamps, and to obtain the estimated total weight of each vehicle; and to send the collected charge signal data to the database and the real-time monitoring module. The database is used to store image information data and charge signal data as historical data; The model management module is used to establish a regional error analysis model based on historical image information data and charge signal data. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, it analyzes the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load and obtains the error compensation coefficient for each monitoring area. The real-time monitoring module is used to analyze the real-time collected image information data and charge signal data when a vehicle is traveling on the monitored road, and output the estimated total weight of the vehicle in combination with the established regional error analysis model; and send the estimated total weight of the vehicle to the early warning module. The early warning module is used to monitor the estimated total weight of the vehicle and provide real-time warnings for overweight vehicles.

[0030] Furthermore, an interactive display platform is provided, allowing managers to view image information data and charge signal data of different vehicles and identify overweight vehicles.

[0031] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for dynamic monitoring of load on piezoelectric quartz vehicles based on visual recognition, characterized in that: include: Smart cameras are deployed above the monitoring road to collect and identify image information data of vehicles during their movement. The image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road; The monitored road is divided into several independent monitoring areas, and piezoelectric quartz sensing units are deployed in each monitoring area to collect charge signal data when the wheels run over it. The collected charge signal data is analyzed to calculate the axle weight of vehicles in each monitoring area at different timestamps, and the estimated total weight of each vehicle is obtained. A database is established to store image information data and charge signal data as historical data. Based on the historical image information data and charge signal data, a regional error analysis model is established. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, the influence of the error in each monitoring area on the dynamic monitoring results of vehicle load is analyzed, and the error compensation coefficient of each monitoring area is obtained. When a vehicle travels on the monitored road, the real-time collected image information data and charge signal data are analyzed. Combined with the established regional error analysis model, the estimated total weight of the vehicle is output, and real-time warnings are issued for overweight vehicles.

2. The piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to claim 1, characterized in that: The method for calculating the vehicle axle weight at different timestamps in each monitoring area and obtaining the estimated total weight of the vehicle is as follows: Based on the vehicle's driving trajectory on the monitoring road, determine the projection range of the left and right wheels of the vehicle on the monitoring road, and determine the set of monitoring areas S that each vehicle axle needs to be associated with based on the projection range; The collected charge signal data is analyzed. Based on the timestamp, the charge signals collected in each monitoring area within the monitoring area set S are determined, and the vehicle axle weights obtained in each monitoring area at different timestamps are obtained. Based on the timestamp and the vehicle's unique identifier, all vehicle axle weights obtained for the identified vehicle are summed to obtain the estimated total vehicle weight W. Where m represents the total number of axles of the vehicle; This indicates the axle weight of vehicle axle j within monitoring area k.

3. The piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to claim 2, characterized in that: The method and steps for analyzing the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load are as follows: S1. Retrieve historical image information data and historical charge signal data containing the actual total weight of vehicles from the database for analysis, obtain the actual total weight of different vehicles and the estimated total weight of vehicles, determine the driving trajectory and speed of each vehicle on the monitored road under the estimated total weight of each vehicle, and associate the axle weight of each vehicle obtained in each monitoring area under the estimated total weight of each vehicle with the timestamp. S2. Establish a regional error analysis model, using the vehicle's actual total weight and vehicle speed in different monitoring areas as inputs, and the vehicle axle load obtained from the monitoring areas as output. Analyze the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load, based on the following formula: ； in, This indicates the axle weight of vehicle axle j within monitoring area k; This indicates the actual axle load of vehicle axle j within monitoring area k; This indicates the vehicle speed of axle j within monitoring area k; a k b represents the dynamic influence coefficient of vehicle speed in monitoring area k; k c represents the vehicle speed saturation coefficient in monitoring area k; k The piezoelectric quartz sensor influence coefficient represents the monitoring area k; d k The fixed deviation coefficient representing the monitoring area k; W represents the hyperbolic tangent function; z Indicates the vehicle's true gross weight; S3. Substitute the analysis results of historical image information data and historical charge signal data in S1 into the conditional formula in S2 to obtain the error compensation coefficient for each monitoring area.

4. The piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to claim 3, characterized in that: The method for outputting the current estimated total weight of the vehicle is as follows: Based on real-time image information data and real-time charge signal data, determine the current vehicle's trajectory and speed on the monitored road, obtain the vehicle's axle load and corresponding vehicle speed in each monitoring area, and calculate the current vehicle's estimated total weight based on the established regional error analysis model. : ； in, This indicates the axle weight of the vehicle currently in monitoring area k, measured for axle j. This indicates the vehicle speed of the current vehicle axle j within the monitoring area k.

5. The piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to claim 1, characterized in that: The monitored road is divided into several monitoring areas of equal width along the transverse direction of the road. At least one set of piezoelectric quartz sensing units is deployed in each monitoring area. Each set of piezoelectric quartz sensing units includes two piezoelectric quartz sensors arranged in parallel front and back with a preset distance.

6. The piezoelectric quartz vehicle load dynamic monitoring method based on visual recognition according to claim 1, characterized in that: The vehicle's overweight threshold is determined based on its unique identifier and number of axles. The estimated total weight of the current vehicle is then compared with the overweight threshold. If the vehicle is not overweight, continue monitoring; If a vehicle is overweight, an early warning signal is triggered, and the information about the overweight vehicle is sent to the remote law enforcement platform.

7. A piezoelectric quartz vehicle load dynamic monitoring system based on visual recognition, characterized in that: The system includes an image information data acquisition module, a charge signal data acquisition module, a database, a model management module, a real-time monitoring module, and an early warning module; The image information data acquisition module is used to deploy smart cameras above the monitored road to collect and identify image information data during vehicle movement; the image information data includes unique vehicle identifiers identified at different timestamps and the driving trajectory and speed of each vehicle on the monitored road; the collected image information data is sent to the database and the real-time monitoring module; The charge signal data acquisition module is used to divide the monitored road into several independent monitoring areas, and to deploy piezoelectric quartz sensing units in each monitoring area to collect charge signal data when the vehicle is rolled over; to analyze the collected charge signal data, to calculate the axle weight of the vehicle in each monitoring area at different timestamps, and to obtain the estimated total weight of each vehicle; and to send the collected charge signal data to the database and the real-time monitoring module. The database is used to store image information data and charge signal data as historical data; The model management module is used to establish a regional error analysis model based on historical image information data and charge signal data. According to the actual total weight of the vehicle and the estimated total weight of the vehicle, it analyzes the impact of errors in each monitoring area on the dynamic monitoring results of vehicle load and obtains the error compensation coefficient for each monitoring area. The real-time monitoring module is used to analyze the real-time collected image information data and charge signal data when a vehicle is traveling on the monitored road, and output the estimated total weight of the vehicle in combination with the established regional error analysis model; and send the estimated total weight of the vehicle to the early warning module. The early warning module is used to monitor the estimated total weight of the vehicle and provide real-time warnings for overweight vehicles.

8. The piezoelectric quartz vehicle load dynamic monitoring system based on visual recognition according to claim 7, characterized in that: An interactive display platform is provided, allowing managers to view image information data and charge signal data of different vehicles and identify overweight vehicles.

Images