A dynamic measurement method and system

Abstract

This application relates to the field of data processing technology, and more particularly to a dynamic measurement method and system. It is used to measure the weight of vehicles passing through a monitored area. A multi-line quartz piezoelectric sensor is installed on the ground of the monitored area. The method includes: acquiring pressure information, speed information, and direction of travel information of the vehicle as it passes through the monitored area using the multi-line quartz piezoelectric sensor; inputting the pressure, speed, and direction of travel information at different times as the vehicle passes through the monitored area into a pre-trained weight prediction model to obtain a predicted weight value of the vehicle output by the weight prediction model, and using the predicted weight value as the result of the vehicle weight measurement. This technical solution can effectively achieve dynamic measurement of the weight of vehicles passing through a monitored area.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a dynamic measurement method and system based on a multi-line quartz sensor. Background Technology

[0002] Dynamic measurement is a non-invasive vehicle measurement technology that allows for the measurement of vehicle weight while the vehicle is in motion. Compared to weighing a vehicle while it is parked, dynamic weight measurement can shorten the weighing time and reduce the number of times the vehicle needs to be stopped.

[0003] To achieve dynamic measurement of vehicle weight, Chinese patent application CN114936343A discloses a method for dynamically calculating the mass of an electric vehicle. The method includes: acquiring state data of the vehicle during its acceleration period, including time, speed, altitude, gravitational acceleration, effective driving power, and drag power; calculating the total output work, total drag work, total change in potential energy, and change in kinetic energy of the vehicle during the acceleration period based on the state data; and calculating the mass of the vehicle based on the law of conservation of energy, which states that the total output work of the vehicle's drive motor equals the sum of the total drag work, the change in potential energy at altitude, and the change in kinetic energy.

[0004] The relevant technologies mainly determine the weight of a vehicle by its driving power and total output power. However, as a regulator that dynamically measures the weight of passing vehicles, it is difficult to obtain the driving power and total output power of the vehicles. Therefore, it is difficult to effectively achieve dynamic measurement of the weight of passing vehicles. Summary of the Invention

[0005] To overcome the problem of difficulty in effectively measuring the dynamic weight of passing vehicles in related technologies, this application provides a dynamic measurement method and system based on a multi-line quartz sensor.

[0006] According to a first aspect of the embodiments of this application, a dynamic measurement method is provided for measuring the weight of a vehicle passing through a monitoring area. The monitoring area is equipped with a multi-line quartz piezoelectric sensor on the ground. The method includes: acquiring pressure information and speed and direction information of the vehicle passing through the monitoring area using the quartz piezoelectric sensor when the vehicle to be weighed passes through the monitoring area; inputting the pressure, speed, and direction information at different times when the vehicle passes through the monitoring area into a pre-trained weight prediction model to obtain a predicted weight value of the vehicle output by the weight prediction model, and using the predicted weight value as the weight measurement result of the vehicle; wherein the weight prediction model is obtained by training a pre-built network model using multiple training samples under the supervision of an objective function; the training samples include the actual weight information of the test vehicle, the speed and direction information at different times when passing through the monitoring area, and the pressure information acquired by the quartz piezoelectric sensor; the objective function is used to minimize the difference between the actual weight information and the predicted weight information determined based on the training samples during training, and to maximize the reliability of the training samples.

[0007] In this way, when the vehicle to be weighed passes through the monitoring area, pressure information is obtained through a multi-line quartz piezoelectric sensor, as well as speed and direction information. The pressure, speed, and direction information are then input into a pre-trained weight prediction model to obtain the predicted weight of the vehicle, thus avoiding the need for stopping the vehicle for weighing.

[0008] Optionally, the confidence level of the training samples is obtained in the following way: G i Q represents the confidence level of the i-th training sample. i Let A be the number of sampling times in the i-th training sample, and norm be the normalization function. i,j H represents the first evaluation value of the i-th training sample at the j-th sampling time. This first evaluation value characterizes the degree of difference in parameter values ​​compared to other sampling times. i,j D is the second evaluation value at the j-th sampling time in the i-th training sample. The second evaluation value is used to characterize the degree of difference in parameter values ​​compared with adjacent sampling times; i,l denoted as DTW distance between the i-th training sample and the a-th training sample at multiple sampling times, where N is the number of training samples. The parameter values ​​include the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor of the test vehicle at the sampling time.

[0009] In this way, by comparing the training samples with other training samples and comparing the parameter values ​​in the training samples at different times, the obtained confidence level can better characterize the reliability of the training samples.

[0010] Optionally, the first evaluation value is obtained in the following way: A i,j Let be the first evaluation value of the i-th training sample at the j-th sampling time, norm be the normalization function, W be the number of parameter values, and s be the first evaluation value of the i-th training sample at the j-th sampling time. i,j,k Let Q be the value of the k-th parameter of the i-th training sample at the j-th sampling time. i denoted as the number of sampling times in the i-th training sample.

[0011] In this way, by comparing the parameter values ​​at different times in the training samples, the obtained first evaluation value can better characterize the degree of difference in parameter values ​​with other sampling times.

[0012] Optionally, the second evaluation value is obtained in the following way: Among them, H i,j Here, is the second evaluation value at the j-th sampling time in the i-th training sample, norm is the normalization function, W is the number of parameter values, and s i,j,k Let s be the value of the k-th parameter of the i-th training sample at the j-th sampling time. i,j+f,k Let be the value of the k-th parameter of the i-th training sample at the (j+f)-th sampling time.

[0013] In this way, by determining the difference between the parameter value of the training sample at a certain moment and the parameter value at adjacent moments, the second evaluation value can characterize the probability that the parameter value at a certain moment in the training sample is abnormal, so as to obtain a more accurate quality measurement result for the vehicle.

[0014] Optionally, the method further includes: using the vehicle weight prediction value output by the weight prediction model as the first weight; using the weight determined based on the pressure information obtained by the multi-line quartz piezoelectric sensor at different times as the second weight of the vehicle to be weighed; determining the difference information between the first weight and the second weight, and outputting a first prompt information when the difference information is greater than a preset difference threshold, wherein the first prompt information is used to prompt the vehicle to be weighed to pass through the monitoring area again for weighing.

[0015] In this way, when the difference information is greater than the preset difference threshold, the first prompt information is output, which can prompt the vehicle to be weighed to pass through the monitoring area again for weighing, so as to obtain a more accurate weighing result for the vehicle.

[0016] Optionally, the method further includes: when the vehicle to be weighed passes through the monitoring area again for weighing, taking the weight prediction value output by the weight prediction model after the vehicle passes through the monitoring area again as the third weight; and when the third weight is greater than a preset weight threshold, outputting a second prompt message, the second prompt message being used to indicate that the weight of the vehicle is greater than the preset weight threshold.

[0017] Optionally, the weight prediction model is trained in the following way: the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor at different times when the test vehicle passes through the monitoring area in the training samples are used as inputs to the pre-built network model; the actual weight information of the test vehicle in the training samples is used as the output of the network model; and the training process of the network model is supervised by the objective function; if the training process meets the preset conditions, the trained network model is used as the weight prediction model.

[0018] In this way, by training a pre-built network model and supervising the training process using an objective function to obtain a quality prediction model, the trained quality prediction model can be deployed in advance to electronic devices, thereby enabling dynamic measurement of the quality of vehicles passing through the monitoring area using electronic devices.

[0019] Optionally, the speed information of the vehicle when passing through the monitoring area is obtained by: acquiring multiple frames of images of the vehicle from a top-down perspective when passing through the monitoring area using an image acquisition device set above the monitoring area; determining the speed information of the vehicle at different times when passing through the monitoring area based on the vehicle's position in the multiple frames of images and a preset correspondence, wherein the preset correspondence is used to characterize the relationship between distance in the image coordinate system and distance in the spatial coordinate system.

[0020] Optionally, the objective function G i Let y be the confidence level of the i-th training sample. i This represents the actual weight information of the vehicle in the i-th training sample. N represents the predicted weight information determined based on the i-th training sample during the training process, where N is the number of training samples.

[0021] According to a second aspect of the present application, a dynamic measurement system is provided, comprising: a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions, when executed by the processor, implement the steps of the dynamic measurement method provided in the first aspect of the present application.

[0022] The technical solution provided by the embodiments of this application may include the following beneficial effects: a multi-line quartz piezoelectric sensor is installed on the ground of the monitoring area. When a vehicle to be weighed passes through the monitoring area, pressure information is obtained through the multi-line quartz piezoelectric sensor, as well as speed information and driving direction information of the vehicle when passing through the monitoring area. The pressure information, speed information, and driving direction information are input into a pre-trained weight prediction model to obtain the predicted weight value of the vehicle. The predicted weight value is then used as the weight measurement result of the vehicle. Since the supervisory party that dynamically measures the weight of the passing vehicle can effectively obtain information such as pressure information from the multi-line quartz piezoelectric sensor, speed information, and driving direction information of the vehicle, the dynamic weight measurement process of the vehicle avoids dependence on operating parameters such as the driving power and total output power of the vehicle, thus enabling effective dynamic measurement of the weight of the passing vehicle.

[0023] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a dynamic measurement method according to an exemplary embodiment;

[0025] Figure 2 This is a schematic diagram of the structure of a dynamic measurement system according to an exemplary embodiment;

[0026] Figure 3 This is a schematic diagram of the structure of a multi-line quartz piezoelectric sensor according to an exemplary embodiment. Detailed Implementation

[0027] First, a brief introduction to the application scenarios of the embodiments of this application will be given. In the application scenarios of this application, in order to measure the weight of a vehicle while it is in normal driving condition, the weight of the vehicle in the driving state can be determined by information such as the vehicle's speed, height, gravitational acceleration, effective driving power, drag power, total output work, total drag work, total change in potential energy and change in kinetic energy during the driving process.

[0028] However, as the regulator that measures the weight of vehicles, it is impossible to directly obtain information such as the vehicle's drag power, total output power, total drag power, total change in potential energy, and change in kinetic energy, making it difficult to achieve dynamic measurement of the weight of vehicles passing through the monitoring area.

[0029] To address the aforementioned technical problems, embodiments of this application provide a multi-line quartz piezoelectric sensor and a dynamic measurement method and system. Figure 1 This is a flowchart illustrating a dynamic measurement method according to an exemplary embodiment, such as... Figure 1 As shown, the method includes the following steps.

[0030] In step S101, when the vehicle to be weighed passes through the monitoring area, pressure information is obtained by a multi-line quartz piezoelectric sensor, and speed information and driving direction information of the vehicle when passing through the monitoring area are also obtained.

[0031] Multi-line quartz piezoelectric sensors can be installed on the ground of the monitoring area. The number of multi-line quartz piezoelectric sensors installed on the ground of the monitoring area can be one or more. This application embodiment does not impose any restrictions on the number or location distribution of multi-line quartz piezoelectric sensors installed on the ground of the monitoring area.

[0032] Multi-line quartz piezoelectric sensors can convert the pressure signal on the ground when a vehicle passes through the monitoring area into an electrical signal, thereby enabling the monitoring of the pressure from the vehicle in the monitoring area.

[0033] When a vehicle to be weighed passes through the monitoring area, it can be determined by sensors such as distance sensors and image sensors. For example, the distance sensor can measure the distance along a specified direction, and the distance measured by the distance sensor will change accordingly when the vehicle to be weighed passes through the monitoring area. Alternatively, the image sensor can collect image data, and the image data collected by the image sensor will change accordingly when the vehicle to be weighed passes through the monitoring area.

[0034] The vehicle to be weighed can leave the monitoring area after passing through it. This can also be determined by sensors such as distance sensors and image sensors, which will not be elaborated here.

[0035] The speed and direction of a vehicle when it passes through the monitoring area can be determined using monitoring data acquired by lidar or image sensors. For example, lidar can be set up in the direction of the vehicle's travel to collect point cloud data. Based on the collected point cloud data, the speed and direction of the vehicle at different points along its path when it passes through the monitoring area can be determined.

[0036] Point cloud data is a dataset composed of points in space. Point cloud data contains the position information of at least multiple points in a spatial coordinate system. The position information of these points in the spatial coordinate system can be used to describe the shape information of an object in the spatial coordinate system. The shape information of the object at different times can be used to determine the object's speed and direction of motion in the spatial coordinate system.

[0037] In one embodiment, the speed information of a vehicle passing through a monitoring area is obtained by: acquiring multiple frames of images of the vehicle passing through the monitoring area from a top-down perspective using an image acquisition device set above the monitoring area; determining the speed information of the vehicle at different times when passing through the monitoring area based on the vehicle's position in the multiple frames of images and a preset correspondence, wherein the preset correspondence is used to characterize the relationship between distance in the image coordinate system and distance in the spatial coordinate system.

[0038] For example, an image acquisition device can be set up above the monitoring area to monitor the movement of vehicles in the monitoring area from a top-down perspective. Since the movement of vehicles in the monitoring area can be viewed more intuitively from a top-down perspective, the speed of the vehicle between adjacent time points can be determined by the position of the vehicle in adjacent frames of images. Furthermore, there is a certain correspondence between the distance between two position points in the image and the actual distance between the two position points in the spatial coordinate system. Therefore, based on the position of the vehicle in multiple frames of images and the preset correspondence, the speed information of the vehicle at different times when passing through the monitoring area can be determined.

[0039] The preset correspondence is used to characterize the relationship between the distance in the image coordinate system and the distance in the spatial coordinate system. The preset correspondence can be obtained by calibrating the relationship between the distance in the image coordinate system and the distance in the spatial coordinate system in advance.

[0040] This makes it easier to determine the vehicle's speed within the monitoring area, so that the vehicle's weight information can be determined based on the vehicle's speed information.

[0041] In step S102, the pressure information, speed information and driving direction information of the vehicle at different times when it passes through the monitoring area are input into the pre-trained weight prediction model to obtain the weight prediction value of the vehicle output by the weight prediction model, so as to use the weight prediction value as the weight measurement result of the vehicle.

[0042] Compared to weighing a vehicle when it is stationary, a moving vehicle will exert a greater pressure on the quartz piezoelectric sensor installed on the ground of the monitoring area than the weight of the vehicle itself when it enters the monitoring area. This makes it impossible to directly take the pressure obtained by the quartz piezoelectric sensor as the pressure corresponding to the weight of the vehicle.

[0043] When a vehicle passes through the monitoring area at different speeds, the pressure exerted by the vehicle on the quartz piezoelectric sensor (other than gravity) varies. This results in different pressures being measured by the quartz piezoelectric sensor when a vehicle of the same weight passes through the monitoring area at different speeds. Therefore, the vehicle's speed affects the measurement of its weight.

[0044] Since there is usually a certain angle between the monitoring area and the horizontal plane, the vehicle is subjected to the supporting force from the quartz piezoelectric sensor, the friction between the vehicle's tires and the quartz piezoelectric sensor, and the vehicle's gravity. When the vehicle travels in the monitoring area at different driving angles, the angular relationship between the supporting force, friction, and gravity is different.

[0045] When vehicles of the same weight travel at different speeds within the monitoring area, or when vehicles of the same weight travel at different speeds at different times within the monitoring area, the angular relationship between the supporting force, friction, and gravity is different. This results in different pressures received by the quartz piezoelectric sensors installed on the ground within the monitoring area. Therefore, the direction of vehicle travel within the monitoring area affects the measurement of vehicle weight.

[0046] By inputting pressure, speed, and direction of travel information into a pre-trained weight prediction model, the various factors that affect the measurement of vehicle weight can be fully considered, so as to obtain a more accurate measurement result of vehicle weight.

[0047] The weight prediction model is obtained by training a pre-built network model under the supervision of an objective function using multiple training samples. The training samples include the actual weight information of the test vehicle, as well as the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor at different times when passing through the monitoring area.

[0048] The objective function is used to minimize the difference between the actual weight information and the predicted weight information determined based on the training samples during training, and to maximize the credibility of the training samples.

[0049] For example, the test vehicle can be weighed while it is parked to obtain its actual weight information, and the test vehicle can be controlled to pass through the monitoring area at different speeds, or test vehicles of different weights can be controlled to pass through the monitoring area, with different weights or loads; or the test vehicles can be controlled to enter the monitoring area in different directions; or the test vehicles can be controlled to travel in different directions at different times while traveling within the monitoring area.

[0050] The number of test vehicles can be one or more, and the test vehicles can be the same or different models as the vehicles passing through the monitoring area.

[0051] Since the weight prediction model is obtained by training a pre-built network model under the supervision of the objective function using multiple training samples, the training samples include the actual weight information of the test vehicle, as well as the speed information, driving direction information, and pressure information obtained by the multi-line quartz piezoelectric sensor at different times when passing through the monitoring area.

[0052] Therefore, the weight prediction model can fully learn the actual weight information of the vehicle, the speed information and driving direction information of the vehicle at different times when passing through the monitoring area, and the pressure information obtained by the quartz piezoelectric sensor. This allows the model to construct the relationship between the speed information, driving direction information, and pressure information obtained by the multi-line quartz piezoelectric sensor at different times when the vehicle passes through the monitoring area and the actual weight information of the vehicle. This enables the model to determine the weight information of the vehicle based on the pressure information, speed information, and driving direction information.

[0053] Objective function of weight prediction model G i Let y be the confidence level of the i-th training sample. i This represents the actual weight information of the vehicle in the i-th training sample. N represents the predicted weight information determined based on the i-th training sample during the training process, where N is the number of training samples.

[0054] The credibility of training samples can characterize the reliability or accuracy of training samples. Here, since the purpose of the loss function is to reduce the difference between the actual output and the expected output of the network model, the corresponding part of the objective function has a larger value for training samples with lower credibility. This can reduce the reference value of training samples with lower credibility to the training process, avoid wasting training samples, and make full use of training samples.

[0055] Alternatively, training samples with higher credibility may have smaller values ​​in the corresponding part of the objective function. This increases the reference value of the training samples for the training process and helps to obtain a more accurate quality prediction model. The output of the quality prediction model can then be replaced with an accurate quality prediction value.

[0056] In one embodiment, the weight prediction model is trained as follows: the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor at different times when the test vehicle passes through the monitoring area in the training samples are used as inputs to a pre-built network model; the actual weight information of the test vehicle in the training samples is used as the output of the network model; and the training process of the network model is supervised using the objective function; if the training process meets the preset conditions, the trained network model is used as the weight prediction model.

[0057] Random forest is a tree-based ensemble learning algorithm. By employing an ensemble approach, the network model can reduce the risk of overfitting during prediction, thereby obtaining more accurate prediction results.

[0058] The training process meets preset conditions, such as the objective function value being less than or equal to a preset threshold, or the number of iterations in the training process being greater than or equal to a preset number.

[0059] In this way, by training a pre-built network model and supervising the training process using an objective function to obtain a trained quality prediction model, the trained quality prediction model can be deployed in advance to electronic devices, thereby enabling the dynamic measurement of the quality of vehicles passing through the monitoring area using electronic devices.

[0060] In one embodiment, the confidence level of the training samples can be obtained in the following way: G i Q represents the confidence level of the i-th training sample. i Let A be the number of sampling times in the i-th training sample, and norm be the normalization function. i,j H represents the first evaluation value of the i-th training sample at the j-th sampling time. This first evaluation value characterizes the degree of difference in parameter values ​​compared to other sampling times. i,j D is the second evaluation value at the j-th sampling time in the i-th training sample. The second evaluation value is used to characterize the degree of difference in parameter values ​​compared with adjacent sampling times; i,l denoted as DTW distance between the i-th training sample and the a-th training sample at multiple sampling times, where N is the number of training samples. The parameter values ​​include the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor of the test vehicle at the sampling time.

[0061] ∏ is the cumulative multiplication symbol. In order to give full play to the role of the parameters involved in the cumulative multiplication, a smaller value relative to the DTW distance can be added to the DTW distance between two training samples at multiple sampling times.

[0062] For example, after the DTW distance is normalized by the normalization function, the value range is between 0 and 1. The sum of the DTW distance between two training samples and the preset value can be used as the DTW distance between two training samples again. The preset value can be a value that is relatively small compared to the normalized DTW distance, such as 0.01, 0.015, and 0.02.

[0063] Alternatively, if the DTW distance between two training samples is 0, a preset value can be used as the DTW distance between the two training samples to give full play to the role of the parameters involved in the multiplication.

[0064] DTW (Dynamic Time Warping) distance can reflect the similarity between two time series; the smaller the DTW distance between two time series, the greater the similarity between them; conversely, the larger the DTW distance between two time series, the smaller the similarity between them.

[0065] Different sampling times within the same training sample correspond to different times when the vehicle passes through the monitoring area. The same training sample includes pressure, speed, and direction information of the test vehicle at different times when it passes through the monitoring area. The DTW distance between two training samples can be determined based on the differences in pressure, speed, and direction information between the two training samples.

[0066] For example, the DTW distance between two training samples can be determined based on multiple sub-DTW distances between the two training samples. These multiple sub-DTW distances correspond to the differences between the two training samples in terms of pressure information, speed information, and driving direction information, respectively.

[0067] The greater the difference between the training samples and other training samples in terms of pressure information, speed information, and driving direction information, the lower the probability of the information, speed information, and driving information included in the training samples appearing. The training samples are more likely to be obtained under abnormal driving conditions of the test vehicle. This part of the training samples has less reference value for the subsequent vehicle quality measurement process. Therefore, this part of the training samples can be given lower confidence.

[0068] For the speed and direction information of the test vehicle at different times in the same training sample, the greater the difference in speed and direction when the test vehicle passes through the monitoring area, the more likely the test vehicle is in an abnormal driving state, such as the test vehicle exhibiting rapid acceleration or deceleration.

[0069] Training samples obtained by test vehicles under abnormal driving conditions have lower reference value for subsequent quality measurement processes. Furthermore, the first evaluation value is used to characterize the degree of difference in parameter values ​​with other sampling times, and the second evaluation value is used to characterize the degree of difference in parameter values ​​with adjacent sampling times. Therefore, training samples with lower first and second evaluation values ​​at multiple times can be assigned lower confidence, so that the quality prediction model obtained under the supervision of the objective function can output more accurate quality measurement results.

[0070] In this way, training samples can be compared with other training samples, and the parameter values ​​in the training samples can be compared at different times, so that the obtained confidence can better characterize the reliability of the training samples, thereby obtaining more accurate quality measurement results.

[0071] In one embodiment, the first evaluation value of a training sample can be obtained in the following way: A i,j Let be the first evaluation value of the i-th training sample at the j-th sampling time, norm be the normalization function, W be the number of parameter values, and s be the first evaluation value of the i-th training sample at the j-th sampling time. i,j,k Let Q be the value of the k-th parameter of the i-th training sample at the j-th sampling time. i denoted as the number of sampling times in the i-th training sample.

[0072] For example, when a test vehicle passes through a monitoring area, the fluctuation of pressure information acquired by the quartz piezoelectric sensor installed on the ground in the monitoring area at different times can be determined; or, the fluctuation of speed information of the test vehicle at different times when it passes through the monitoring area can be determined; or, the fluctuation of driving direction information of the test vehicle at different times when it passes through the monitoring area can be determined.

[0073] By comparing the parameter values ​​at different times in the training samples, the obtained first evaluation value can better represent the degree of difference in parameter values ​​with other sampling times.

[0074] In one embodiment, the second evaluation value of the training sample can be obtained in the following way: Among them, H i,j Here, is the second evaluation value at the j-th sampling time in the i-th training sample, norm is the normalization function, W is the number of parameter values, and s i,j,k Let s be the value of the k-th parameter of the i-th training sample at the j-th sampling time. i,j+f,k Let be the value of the k-th parameter of the i-th training sample at the (j+f)-th sampling time.

[0075] In addition to comparing the parameter value at a certain sampling time in the training sample with the parameter values ​​at all other sampling times, the difference between the parameter value at a certain time and the parameter value at adjacent times can also better characterize the possible anomalies in the parameter value at a certain time of the training sample.

[0076] For example, for a training sample with a duration of 60 seconds, the speed of the test vehicle at the 10th second after passing through the monitoring area can be compared with the speed of the test vehicle at the 9th and 11th seconds after passing through the monitoring area to determine whether there are large fluctuations in the speed of the test vehicle.

[0077] If the speed of the test vehicle at a certain moment while passing through the monitoring area differs significantly from that at an adjacent moment, it indicates that the test vehicle is likely in an abnormal driving situation, such as rapid acceleration or deceleration. The training samples corresponding to the abnormal driving situation of the test vehicle have lower reference value.

[0078] In this way, by determining the difference between the parameter value of the training sample at a certain moment and the parameter value at adjacent moments, the second evaluation value can better characterize the probability that the parameter value at a certain moment in the training sample is abnormal, so as to obtain a more accurate quality measurement result for the vehicle.

[0079] In one embodiment, the predicted weight of the vehicle output by the weight prediction model can be used as the first weight; the weight determined based on the pressure information obtained by the quartz piezoelectric sensor at different times can be used as the second weight of the vehicle to be weighed; the difference information between the first weight and the second weight can be determined, and a first prompt information can be output when the difference information is greater than a preset difference threshold. The first prompt information is used to prompt the vehicle to be weighed to pass through the monitoring area again for weighing.

[0080] When a vehicle passes over the quartz piezoelectric sensors installed on the ground in the monitoring area, the pressure directly acquired by the multi-line quartz piezoelectric sensors differs from the vehicle's weight due to the vehicle's movement. However, because the pressure information acquired by the multi-line quartz piezoelectric sensors when the vehicle passes through the monitoring area is correlated with the vehicle's weight, the predicted vehicle weight output by the weight prediction model is correlated with the weight determined directly based on the pressure information acquired by the multi-line quartz piezoelectric sensors at different times. Therefore, the difference between the predicted vehicle weight output by the weight prediction model and the weight determined directly based on the pressure information acquired by the multi-line quartz piezoelectric sensors at different times will be within a certain range.

[0081] If the predicted vehicle weight output by the weight prediction model differs significantly from the weight determined directly from the pressure information obtained by the multi-line quartz piezoelectric sensor at different times, it indicates that the measurement results may be inaccurate due to abnormal driving behavior when the vehicle passes through the monitoring area. In this case, the vehicle can be prompted to pass through the monitoring area again, so that the vehicle weight can be dynamically measured again.

[0082] The difference information could be, for example, the ratio of the difference between the first weight and the second weight. The preset difference threshold can be determined based on the adaptability of the first weight or the second weight, for example, it could be between 20% and 30% of the first weight.

[0083] In this way, if the difference information is greater than the preset difference threshold, the first prompt information will be output, which can prompt the vehicle to be weighed to pass through the monitoring area again for weighing, so as to obtain a more accurate weighing result for the vehicle.

[0084] In one embodiment, when the vehicle to be weighed passes through the monitoring area again for weighing, the weight prediction value output by the weight prediction model after the vehicle passes through the monitoring area again is taken as the third weight; if the third weight is greater than a preset weight threshold, a second prompt message is output, which is used to indicate that the weight of the vehicle is greater than the preset weight threshold.

[0085] For example, when a vehicle to be weighed passes through the monitoring area again, the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor can be obtained. The pressure information, speed information, and driving direction information at different times when the vehicle passes through the monitoring area again are input into the pre-trained weight prediction model, and the weight prediction value output by the weight prediction model is used as the third weight.

[0086] The vehicle that was about to be weighed passed through the monitoring area again for weighing. The driver of the vehicle was likely to have avoided any abnormal driving behavior that may have occurred during the previous weighing, such as changing from rapid acceleration through the monitoring area to a constant speed. Therefore, the weight measurement result obtained again can better reflect the actual weight of the vehicle. If the third weight is greater than the preset weight threshold, it indicates that the vehicle is overweight, and the system can alert the driver that the vehicle's weight exceeds the preset weight threshold.

[0087] For example, when a vehicle is weighed at a monitoring station, a display device can be installed at the monitoring station. If the third weight is greater than a preset weight threshold, the display device can output a second prompt message indicating that the vehicle's weight is greater than the preset weight threshold.

[0088] The preset weight threshold can be determined based on the vehicle's identification information. For example, it can be determined based on the vehicle's model identification information and the allowable load weight pre-bound to the license plate number. Alternatively, since the vehicle will be weighed after loading cargo at its departure point, the vehicle user can submit the weighing result at the vehicle's departure point to the monitoring station. The monitoring station can determine the vehicle's corresponding preset weight threshold based on the received weighing result and the vehicle's model identification information.

[0089] Figure 2 This is a schematic diagram illustrating the structure of a dynamic measurement system 1000 according to an exemplary embodiment. (Refer to...) Figure 2The dynamic measurement system 1000 includes a processor 1100 and a memory 1200. The memory 1200 stores computer program instructions, which, when executed by the processor 1100, implement all or part of the steps of the dynamic measurement method in this application.

[0090] Figure 3 This is a schematic diagram illustrating the structure of a multi-wire quartz piezoelectric sensor 2000 according to an exemplary embodiment. The multi-wire quartz piezoelectric sensor 2000 includes a weighing base 2100 and a signal output component 2200 located within the internal space of the weighing base 2100. The signal output component 2200 may include multiple signal output sub-components 2300. The signal output component 2200 can output signals to a signal processing device (…). Figure 3 (Not shown in the image) Send the acquired pressure information.

[0091] The multi-line quartz piezoelectric sensor 2000 can be installed on the ground in the monitoring area. When a vehicle passes the location where the multi-line quartz piezoelectric sensor 2000 is installed in the monitoring area, the weighing base 2100 can transmit the pressure signal it receives to the signal output component 2200, so that the signal output component 2200 converts the pressure signal into pressure information and sends it to the signal processing equipment.

[0092] The information processing device can be the processor inside the multi-line quartz piezoelectric sensor 2000. Figure 3 (Not shown in the image), or it can be an external terminal device or server that is communicatively connected to the multi-wire quartz piezoelectric sensor 2000. Figure 3 (Not shown in the image).

[0093] The connection relationship between the multiple signal output sub-components 2300 included in the signal output component 2200 is as follows: Figure 3 As not shown, the multiple signal output sub-components 2300 included in the signal output component 2200 can individually send the monitored pressure information to the information processing device; or, the multiple signal output sub-components 2300 can also be interconnected, so that the pressure information monitored by all the multiple signal output sub-components 2300 can be sent to the information processing device together.

[0094] The multi-line quartz piezoelectric sensor 2000 may integrate the dynamic measurement system described in the embodiments of this application. The dynamic measurement system is used to execute the steps of the dynamic measurement method described in the embodiments of this application, so that the multi-line quartz piezoelectric sensor 2000 executes the steps of the dynamic measurement method described in the embodiments of this application.

[0095] Alternatively, the multi-line quartz piezoelectric sensor 2000 can send the pressure information of vehicles passing through the monitoring area to the information processing device, so that the information processing device can execute the dynamic measurement method provided in the embodiments of this application to measure the weight of vehicles passing through the monitoring area, thereby avoiding the need for stopping to measure the weight of vehicles.

[0096] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0097] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A dynamic measurement method for measuring the weight of vehicles passing through a monitoring area, wherein a multi-line quartz piezoelectric sensor is installed on the ground of the monitoring area, characterized in that, include: When a vehicle to be weighed passes through the monitoring area, pressure information is obtained through a multi-line quartz piezoelectric sensor, as well as speed and direction information of the vehicle when it passes through the monitoring area. The pressure, speed, and direction of travel information of the vehicle at different times when it passes through the monitoring area are input into a pre-trained weight prediction model to obtain the weight prediction value of the vehicle output by the weight prediction model, and the weight prediction value is used as the weight measurement result of the vehicle. Among them, the weight prediction model is obtained by training a pre-built network model using multiple training samples under the supervision of the objective function; The training samples include the actual weight information of the test vehicle, as well as its speed, direction of travel, and pressure information acquired by the quartz piezoelectric sensor at different times while passing through the monitoring area. The objective function minimizes the difference between the actual weight information and the predicted weight information determined based on the training samples during training, and maximizes the reliability of the training samples, satisfying the following: , For the The credibility of each training sample For the The number of sampling times in each training sample, where norm is the normalization function. , The first The training sample has a first evaluation value and a second evaluation value at the j-th sampling time. The first evaluation value is used to characterize the degree of difference in parameter values ​​with other sampling times, and the second evaluation value is used to characterize the degree of difference in parameter values ​​with adjacent sampling times. Let be the DTW distance between the i-th training sample and the a-th training sample at multiple sampling times. The number of training samples; the parameter values ​​include the speed information, driving direction information, and pressure information obtained by the quartz piezoelectric sensor of the test vehicle at the sampling time. ， ， W represents the number of possible parameter values. , The first The values ​​of the k-th parameter of each training sample at the j-th and j+f-th sampling times; objective function , For the The actual weight information of the vehicles in each training sample In order to perform training according to the first The predicted weight information determined from each training sample.

2. The dynamic measurement method according to claim 1, characterized in that, The method further includes: The predicted weight of the vehicle output by the weight prediction model is used as the first weight. The weight determined based on the pressure information obtained by the multi-line quartz piezoelectric sensor at different times will be used as the second weight of the vehicle to be weighed. The difference between the first weight and the second weight is determined, and if the difference is greater than a preset difference threshold, a first prompt message is output. The first prompt message is used to prompt the vehicle to be weighed to pass through the monitoring area again for weighing.

3. The dynamic measurement method according to claim 2, characterized in that, The method further includes: If a vehicle to be weighed passes through the monitoring area again for weighing, the weight prediction value output by the weight prediction model after the vehicle passes through the monitoring area again will be used as the third weight. If the third weight is greater than a preset weight threshold, a second prompt message is output, which is used to indicate that the weight of the vehicle is greater than the preset weight threshold.

4. The dynamic measurement method according to claim 1, characterized in that, The weight prediction model was trained in the following way: The speed information, driving direction information, and pressure information obtained by the multi-line quartz piezoelectric sensor at different times when the test vehicle passes through the monitoring area in the training sample are used as inputs to the pre-built network model. The actual weight information of the test vehicle in the training sample is used as the output of the network model. The training process of the network model is supervised by the objective function. If the training process meets the preset conditions, the trained network model will be used as the weight prediction model.

5. The dynamic measurement method according to claim 1, characterized in that, The speed information of the vehicle when it passes through the monitoring area is obtained through the following methods: The image acquisition device set above the monitoring area acquires multiple frames of images from a top-down perspective as the vehicle passes through the monitoring area; Based on the vehicle's location in the multi-frame images and a preset correspondence, the vehicle's speed information at different times when passing through the monitoring area is determined. The preset correspondence is used to characterize the relationship between distance in the image coordinate system and distance in the spatial coordinate system.

6. A dynamic measurement system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the dynamic measurement method according to any one of claims 1-5.

Images