Vehicle speed adjustment method and related equipment integrating road surface roughness and wheel data

Abstract

The present application provides a vehicle speed adjustment method and related equipment that integrates road surface flatness and wheel data, and relates to the field of intelligent control. The method includes: obtaining road surface images and point cloud data through monitoring equipment, combining the two to determine road surface bump height information and state parameters, and then inputting them into a pre-trained flatness assessment model to obtain a road surface flatness grade. When in an uneven range, obtain the target vehicle wheel data, input the road surface flatness grade and wheel data into a bump impact assessment model to obtain bump data. Determine the vehicle bump impact level in combination with the on-board sensor data, and then determine the standard speed value, so as to control the current speed of the target vehicle, achieve accurate speed adjustment according to the road surface and vehicle conditions, and improve driving safety and comfort.

Description

Technical Field

[0001] The present application relates to the field of intelligent control, and in particular to a vehicle speed adjustment method and related equipment that integrates road surface smoothness and wheel data. Background Art

[0002] With the rapid development of the transportation industry, ensuring the safety and smooth driving of vehicles on the road has become an important issue. This application involves the field of intelligent control, and in particular, it involves a vehicle speed adjustment method and related equipment that integrates road surface flatness and wheel data. When the road condition exceeds this limit value, the on-board computer will automatically decelerate to achieve speed control. The driver will also manually control the speed by using the accelerator, brake, etc. based on his own experience and skills.

[0003] However, existing technologies ignore the importance of road surface flatness and wheel data to vehicle speed control, resulting in the vehicle being jolted due to excessive speed under complex road conditions, affecting the riding experience and even causing damage to the vehicle or posing a driving safety hazard. Summary of the invention

[0004] The present application provides a vehicle speed adjustment method and related equipment that integrates road surface flatness and wheel data, which are used to achieve more precise and intelligent vehicle speed control when the existing technology is difficult to accurately and comprehensively consider the road surface and vehicle conditions to adjust the vehicle speed.

[0005] In a first aspect, the present application provides a vehicle speed adjustment method integrating road surface roughness and wheel data, which is applied to a server, and comprises: obtaining road surface data of a target road through a monitoring device, wherein the road surface data comprises a road surface image and road surface point cloud data; determining road surface unevenness height information in combination with the road surface image and the road surface point cloud data, and obtaining road surface state parameters; inputting the road surface state parameters into a pre-trained roughness evaluation model, and obtaining a road surface roughness grade of the target road, wherein the roughness evaluation model is a neural network model established by training data based on a plurality of real road roughness grades and corresponding road surface image features and three-dimensional point cloud features in advance, and the road surface roughness grades include smooth, slightly uneven, moderately uneven, and severely uneven; When the road surface roughness level is within the unevenness range of the set level, the wheel data of the target vehicle is obtained, and the set level includes slight unevenness, moderate unevenness, and severe unevenness; the road surface roughness level and the wheel data are input into a preset bump impact assessment model to obtain bump data, and the bump impact assessment model is an empirical model established based on the wheel data of different vehicle models and the relationship between the matched road surface roughness level and the actual bump impact degree; the sensor data of the target vehicle is determined based on the on-board sensor; the bump impact level of the target vehicle is determined in combination with the bump data and the sensor data; a matching standard vehicle speed value is determined based on the bump impact level; and the current vehicle speed of the target vehicle is controlled based on the standard vehicle speed value.

[0006] By adopting the above technical solution, the server can obtain comprehensive and accurate road condition data. The monitoring equipment is used to obtain road images and point cloud data, and the road surface flatness level is obtained after a series of processing. The sensor data obtained by the on-board sensor is then combined to determine the vehicle's bumpy condition, and the bumpy condition is combined to achieve precise control of the target vehicle's speed.

[0007] In combination with some embodiments of the first aspect, in some embodiments, the step of determining the road surface bump height information in combination with the road surface image and the road surface point cloud data to obtain the second road surface state parameter specifically includes: after extracting road surface key points from the road surface image, matching the road surface key points with corresponding points in the road surface point cloud data to obtain multiple matching feature point pairs; performing three-dimensional reconstruction based on the feature point pairs to determine the three-dimensional model of the target road; determining the number of extracted road horizontal lines based on the road width data in the three-dimensional model; determining the length of the extracted road horizontal lines based on the road length data in the three-dimensional model; and determining the point cloud coordinates of multiple key points of each road horizontal line based on the number of road horizontal lines and the length of the road horizontal lines, wherein the road horizontal lines are equally spaced and parallel to the edge lines on both sides of the road. a line or lane line, the length of the road horizontal line is the road length; according to the multiple road horizontal lines, the z-axis difference between each key point cloud adjacent to the x-axis and the y-axis is determined; the point cloud coordinate point pair whose z-axis difference is greater than the set threshold is recorded as a convex coordinate pair, and the point cloud coordinate pair whose z-axis difference is less than the set threshold is recorded as a concave coordinate pair; the convex coordinate pair is clustered according to the distance to obtain a convex coordinate set corresponding to the specified road sub-region in the target road; the concave coordinate pair is clustered according to the distance to obtain a concave coordinate set corresponding to the specified road sub-region in the target road; the convexity corresponding to the specified road sub-region is determined according to the convex coordinate set; the concave degree corresponding to the specified road sub-region is determined according to the concave coordinate set; and the road surface concave-convex height information is determined according to the convexity and the concave degree.

[0008] By adopting the above technical solutions, the unevenness of the road surface can be analyzed comprehensively and accurately, providing detailed and reliable data support for accurately evaluating the flatness of the road surface. This helps to more accurately determine the speed adjustment strategy and ensure the stability and safety of vehicle driving.

[0009] In combination with some embodiments of the first aspect, in some embodiments, after the step of determining the z-axis difference between each key point point cloud adjacent to the x and y axes based on multiple road horizontal lines, it also includes: if it is detected that the number of breakpoints of the key point point cloud on the x-axis is greater than the set threshold, then the image coordinates corresponding to the breakpoint are determined to be a longitudinal crack area; if it is detected that the number of breakpoints of the key point point cloud on the y-axis is greater than the set threshold, then the image coordinates corresponding to the breakpoint are determined to be a transverse crack area; if the number of breakpoints of the key point point cloud on the x-axis and the y-axis are both greater than the set threshold, then the image coordinates corresponding to the breakpoint are determined to be a mesh crack area.

[0010] By adopting the above technical solution, different types of crack areas can be identified quickly and accurately, which helps to assess the degree of road damage more specifically and provide strong support for the user's wheel speed.

[0011] In combination with some embodiments of the first aspect, in some embodiments, after the step of acquiring the road surface data of the target road by the monitoring device, the method further includes:

[0012] The pavement image is input into a preset crack recognition model to obtain pavement crack conditions, which at least include crack categories of no cracks, transverse cracks, longitudinal cracks, and mesh cracks, and image coordinates corresponding to each crack category. The crack recognition model is a convolutional neural network model established by training based on crack categories and positions in multiple labeled real pavement images; after feature matching of the pavement image with the pavement point cloud data, the image coordinates corresponding to each crack category are mapped to three-dimensional coordinates in the pavement point cloud data to obtain multiple crack category information corresponding to multiple road sub-areas; after determining the danger level based on the crack category information, the alarm device is controlled to issue an alarm message.

[0013] By adopting the above technical solution, the image coordinates corresponding to the crack category are mapped to the point cloud data, and the crack category information of multiple road sub-areas is further determined. After determining the danger level, the alarm device is controlled to sound an alarm, promptly reminding relevant personnel to pay attention to road safety issues, and providing protection for road maintenance and vehicle driving safety.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after determining the danger level based on the crack category information, after the step of controlling the alarm device to issue an alarm message, it also includes: sending the image coordinates corresponding to the crack category to the receiving end in a visual form through a set channel, so that the receiving end displays the corresponding crack category in different colors at the position corresponding to the image coordinates.

[0015] By adopting the above technical solution, the receiving end displays different crack categories in different colors, and the driver can clearly see the crack situation. During driving, this function can greatly improve the driver's safety awareness. The driver can predict the road conditions in advance according to the crack situation and adjust the driving strategy in time to avoid vehicle loss of control or accidents due to sudden cracks.

[0016] In combination with some embodiments of the first aspect, in some embodiments, when the road surface roughness level is within a set level of unevenness, after the step of obtaining the wheel data of the target vehicle, it also includes: determining the current load based on the current weight of the target vehicle detected by the on-board weight sensor; determining the magnitude of the impact force generated when the target vehicle contacts the road surface based on the road surface roughness level, the wheel data and the current load; comparing the impact force magnitude with the set impact force threshold range to determine the bump impact level.

[0017] By adopting the above technical solution, after obtaining the wheel data of the target vehicle, the impact force is determined by comprehensively considering the road surface flatness level, wheel data and current load, so as to more accurately evaluate the level of bump impact. This comprehensive consideration method can fully consider various actual factors in vehicle driving, making the bump impact assessment more in line with the actual situation.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of controlling the current speed of the target vehicle according to the standard speed value, it also includes: after setting the wheel data update time interval, determining whether there is a difference between the updated new wheel data and the original wheel data; if there is a difference, inputting the new wheel data into the bump impact assessment model, and redetermining the updated bump impact level; redetermining the updated new standard speed value according to the updated bump impact level; if the new standard speed value is different from the original standard speed value, adjusting the current speed of the target vehicle according to the new standard speed value.

[0019] By adopting the above technical solution, after controlling the target vehicle speed according to the standard vehicle speed value, the wheel data is updated at the set time interval and the bump impact level and vehicle speed are re-evaluated, thus realizing dynamic adjustment of the vehicle speed. This dynamic adjustment mechanism can adapt to changes in the road surface and vehicle status in a timely manner, ensuring that the vehicle always travels at a safe and comfortable speed. It helps to maintain the stability of vehicle driving, further reduce the risk of accidents, and at the same time maximize the comfort of the ride and the durability of the vehicle.

[0020] In a second aspect, the present application provides a server, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a server, causes the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a server, enables the server to execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0024] 1. The invention adopts the technical means of obtaining road surface images and point cloud data through monitoring equipment, determining the road surface flatness grade by using the flatness evaluation model, and then determining the level of bump impact in combination with wheel data and sensor data to control the vehicle speed. Therefore, the problem that it is difficult to accurately adjust the vehicle speed according to the actual road surface conditions in the existing technology is effectively solved, thereby achieving the technical effect of improving driving safety and improving road traffic efficiency.

[0025] 2. The method uses a technical method of extracting key points from road surface images and matching them with road surface point cloud data to perform three-dimensional reconstruction to determine the three-dimensional model of the road, and then clustering the convexity and concavity based on the convexity and concavity coordinates to determine the degree of convexity and concavity. Therefore, the problem that it is difficult to comprehensively and accurately analyze the concavity of the road surface in the existing technology is effectively solved, thereby achieving the technical effect of providing detailed and reliable data support for accurately evaluating the flatness of the road surface and ensuring the stability and safety of vehicle driving.

[0026] 3. Since the technical means of determining the impact force by comprehensively considering the road surface flatness grade, wheel data and current load after obtaining the wheel data of the target vehicle, and then evaluating the level of bump impact is adopted, the problem of inaccurate bump impact evaluation in the prior art is effectively solved, thereby achieving the technical effect of making the bump impact evaluation more in line with the actual situation and improving the accuracy of vehicle speed adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a vehicle speed adjustment method integrating road surface flatness and wheel data in an embodiment of the present application;

[0028] Figure 2 is another flow chart of a vehicle speed adjustment method integrating road surface flatness and wheel data in an embodiment of the present application;

[0029] Figure 3 It is a schematic diagram of a physical device structure of a server in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to and includes any or all possible combinations of one or more listed items.

[0031] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0032] By adopting the vehicle speed adjustment method in the embodiment of the present application, road surface data is obtained through monitoring equipment, and the road surface flatness level is determined by combining images and point clouds. The wheel data and sensor data are then combined to determine the level of bump impact to control the vehicle speed, thereby achieving precise adjustment of the target vehicle speed. This not only improves driving safety and comfort, but also reduces vehicle wear and tear and improves road traffic efficiency.

[0033] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above content. Figure 1 , which is a flow chart of a vehicle speed adjustment method integrating road surface roughness and wheel data in an embodiment of the present application.

[0034] S101, obtaining road surface data of a target road through a monitoring device, where the road surface data includes a road surface image and road surface point cloud data;

[0035] The server can obtain road surface data through monitoring equipment installed on both sides of the target road. Among them, the monitoring equipment can be a high-pixel camera device that can capture and obtain high-definition road surface images. The camera is set at a certain height above the road and can capture and record a continuous sequence of road surface images, covering the entire lane range of the target road. The server receives the road surface images from the camera device and stores them. The monitoring equipment can also include a scanning radar device with an integrated laser emission unit, which is used in combination with a camera to obtain three-dimensional point cloud data of the target road. When working, the laser radar device emits a laser to the road surface and changes the emission direction of the laser by mechanical scanning. When the laser hits an object on the road, part of it will be reflected back. The laser radar detects these reflected lasers through the photoelectric receiving unit, and calculates the distance of the laser propagation using the speed of light based on the time interval from the laser emission to the return, thereby judging the distance of the object from the laser radar. At the same time, the camera can capture the visual image of the road. The data of the two are combined to capture the three-dimensional structural information of the road in all directions, providing rich visual details for subsequent flatness evaluation.

[0036] The road surface image acquired by the server may include the overall situation of the target road lane, and may include information that is not so dangerous but may affect driving quality, such as ruts, gravel, loose soil, puddles, mottled and other visual information.

[0037] S102, combining the road surface image and the road surface point cloud data to determine road surface concave and convex height information, and obtain road surface state parameters;

[0038] Specifically, the server will first pre-process the road image, detect and extract some feature points in the road image, such as crack boundaries, markings and other points with obvious edge features. Then these image feature points are matched with the point cloud data to determine the corresponding three-dimensional coordinates in the point cloud. Then, the server can use the three-dimensional reconstruction algorithm to obtain the three-dimensional model of the target road based on the matched image points and point cloud points. In this three-dimensional model, the ups and downs of the road surface can be clearly observed.

[0039] The server will count the road width and length data in the 3D model, and use this to determine the number and length of road horizontal lines that need to be sampled to calculate height information. Then, the 3D coordinates of several representative points on each horizontal line are extracted, and the height difference between these adjacent points on the z-axis is calculated. The height difference of all sampling points can accurately reflect the degree of road surface unevenness, and the road surface state parameters can be determined in combination with the road surface image.

[0040] For example, a linear depression caused by long-term tire wear can be observed from the road image. The area is darker in color and has clear edges. Combined with the point cloud data, the height value of the key points of this linear area can be measured. It is usually lower than the height of the surrounding normal road surface, and the height variation range is concentrated. The linear contour can be determined to be a rut. For another example, a water mark area with a darker local color is observed in the road image, and the boundary is irregular. Water accumulation can be seen in a certain area. The point cloud data height of this area is slightly lower than the surrounding area, but the height variation is not large, and there is no obvious boundary. Therefore, the area can be determined to be a puddle.

[0041] S103, inputting the road surface state parameter into a pre-trained flatness evaluation model to obtain a road surface flatness grade of the target road, wherein the flatness evaluation model is a neural network model established in advance based on data training of a plurality of real road flatness grades and corresponding road surface image features and three-dimensional point cloud features, and the road surface flatness grades include flat, slightly uneven, moderately uneven, and severely uneven;

[0042] For the flatness assessment model, the server needs to collect a large amount of data from real roads with known flatness levels before deploying actual applications to train the model parameters of the assessment model. Specifically, the server will collect road surface image data and three-dimensional laser point cloud data of various roads in different cities. These collected road surface data need to be annotated by professional engineers to give an objective road surface flatness rating for the road section. After collecting a large amount of data with real level annotations, the server will extract the features of the road surface image and point cloud data to construct a training data set. Then, the server uses deep learning models such as convolutional neural networks to train model parameters with the training data set to establish a mapping relationship between road surface features and flatness levels. In order to evaluate the performance of the model, the server also needs to set aside some samples to construct a validation set to test the classification effect of the model on the validation set. Repeat the model training and validation process until the model meets the predetermined accuracy index requirements on the validation set, and the training of the flatness assessment model is completed.

[0043] The trained model can input the road surface characteristics of the new road section and output its road surface smoothness level. The server uses this model to effectively evaluate the smoothness of each target road and provide support for intelligent speed limit.

[0044] S104, when the road surface roughness level is within a set level of unevenness, obtaining wheel data of the target vehicle, where the set level includes slight unevenness, moderate unevenness, and severe unevenness;

[0045] After evaluating the road surface roughness level of the target road, the server will determine whether the level is within the set level of unevenness. The set level can include three levels: slight unevenness, moderate unevenness, and severe unevenness. In actual applications, users can determine the specific requirements of the set level according to actual needs, which is not limited here.

[0046] If it is determined that the road surface roughness level is within the set unevenness range, the server will obtain wheel data of the target vehicle, which includes information such as the wheel speed, wheel direction angle, and wheel height of the current vehicle.

[0047] S105, inputting the road surface flatness grade and the wheel data into a preset bump impact assessment model to obtain bump data, wherein the bump impact assessment model is an empirical model established based on the wheel data of different vehicle models and the relationship between the matched road surface flatness grade and the actual bump impact degree;

[0048] The server will input the obtained road surface roughness grade and wheel data into a pre-established bump impact assessment model. The bump impact assessment model can be a rule-based empirical model, which is constructed by analyzing the relationship between wheel data characteristics and actual bump levels under different vehicle models and different road surface roughness. For example, wheel data of SUVs and sedans under the same road conditions can be collected as samples to establish a correspondence between vehicle models. Then, wheel data of different vehicle models on flat roads and bumpy roads are collected to determine the relationship between wheel data and road surface roughness grades. Combining these corresponding rules, a model for assessing bump impacts is constructed.

[0049] The server can use the bump impact assessment model to input the specific model of a target vehicle and the wheel data obtained in real time, and combine it with the determined flatness of the road section to assess the bump data that the target vehicle will be subjected to.

[0050] S106, determining sensor data of the target vehicle according to the on-board sensor;

[0051] The server needs to collect various sensor data of the target vehicle to determine the specific operating status of the vehicle. The on-board sensors may include tire pressure sensors, suspension sensors and vibration sensors. The tire pressure sensors can be installed in the tire cavity of each wheel to monitor and collect tire pressure data in real time. Too low tire pressure will cause tire deformation and uneven contact between the tire and the road surface, resulting in bumps. Too high tire pressure will reduce the cushioning effect of the tire and make it difficult to effectively absorb road impact. Therefore, abnormal tire pressure may cause the vehicle to be bumpy. The suspension sensor is used to monitor the impact force of the wheel. If the impact force exceeds the normal range, it means that the wheel has suffered too much impact, which will directly cause the vehicle to be bumpy. The vibration sensor can be set at a set position of the vehicle to directly detect the vibration of the vehicle body and frame when the vehicle is driving. It usually contains a three-axis accelerometer that can measure the vibration acceleration of the vehicle in the direction of travel (X axis), vertical direction (Z axis) and lateral direction (Y axis). When the vehicle passes through a bumpy section, the body will vibrate up and down or left and right. At this time, the vibration acceleration in the three-axis direction detected by the vibration sensor may exceed the set normal form level.

[0052] S107, determining the bump impact level of the target vehicle by combining the bump data and the sensor data;

[0053] After the server collects the real-time sensor data of the vehicle, it needs to be integrated with the aforementioned bump prediction data to determine the actual bump impact level of the vehicle. Specifically, the server will analyze the sensor data to detect whether there are abnormal vibrations, acceleration mutations, etc. that exceed the set threshold in the sensor data. For example, if the vehicle body acceleration data suddenly spikes and exceeds the normal driving stability level, it can be determined that strong bumps have occurred. Then, the server will verify these sensor data mutations that exceed the threshold in combination with the bump data predicted by the model in step S105. If the sensor data confirms that the model's bump prediction is accurate, the model's bump level result is directly used. If the two are inconsistent, the server will update the model and retrain to improve the prediction accuracy.

[0054] Through sensor data verification, the server can obtain the actual bump impact level of the vehicle, which can be set according to actual needs, such as mild, medium and severe. For example, if the sensor detects that the vehicle has a large vibration, but the model only predicts mild bumps, the bump impact level of the vehicle is finally determined to be "severe".

[0055] In some embodiments, the server can detect the current weight of the target vehicle in real time through the on-board weight sensor to determine the current load condition of the vehicle. The weight sensor can be set at the chassis of the vehicle. When the vehicle is driving, the sensor can detect the static pressure distribution and various dynamic inertial forces of the entire vehicle, and convert them into electrical signals for output after amplification. The server collects and analyzes these electrical signals to calculate the current overall weight data of the vehicle. The server then compares the weight data with the rated load capacity of the vehicle. If the current weight exceeds the rated load, the vehicle is determined to be fully loaded or overloaded; if it is less than the rated load, it is determined to be unloaded or half-loaded. In this way, the server can determine the current load condition of the target vehicle.

[0056] After obtaining the road surface roughness level, wheel data and current load, the server needs to combine the three to calculate the impact force generated when the target vehicle contacts the road surface. According to the load conditions, a fully loaded vehicle is heavy and generates a large road impact force, while an unloaded vehicle has a small impact force. The above factors are comprehensively analyzed to establish a mathematical model to calculate the impact force of the vehicle in contact with the road surface at a certain moment. For example, a fully loaded heavy truck is traveling at a speed of 60 kilometers per hour on a road with a "severely uneven" level. According to the formula, it can be calculated that its road impact force is 35,000 Newtons. The server compares the calculated impact force data with the set impact force threshold range, and presets an impact force range, which is divided into three levels: low, medium and high. If the impact force is in the low level range, it is judged that the bumpy impact is slight; the medium level is judged as moderate bumps, and the speed needs to be reduced; at the high level, it is judged as severe bumps, and the speed needs to be reduced immediately.

[0057] S108, determining a matching standard vehicle speed value according to the bump impact level;

[0058] After obtaining the target vehicle's bump impact level, the server needs to determine the matching standard speed value based on the level as a basis for controlling the vehicle speed. Specifically, the server will pre-establish a correspondence between different bump impact levels and standard speed values. For example, a mild bump impact level matches a standard speed value of 80-100km / h, a moderate bump impact level matches a standard speed value of 60-80km / h, and a severe bump impact level matches a standard speed value of 30-60km / h. The server also needs to consider factors such as vehicle type and body structure. For open-air sedans, the matching standard speed value will be lower than that of enclosed off-road vehicles. If it is a truck carrying fragile items, the standard speed value also needs to be reduced accordingly.

[0059] By querying the established mapping relationship, the server can find the matching standard speed value based on the specific bump impact assessment results of the target vehicle, providing a reference for subsequent speed control. If the vehicle has special configurations, it can also be adjusted based on the standard speed to formulate a more reasonable and safe speed control strategy.

[0060] S109: Control the current speed of the target vehicle according to the standard speed value.

[0061] After the server determines the standard vehicle speed value, it needs to transmit it to the on-board computer and instruct the on-board computer to control the real-time speed of the vehicle according to the standard value.

[0062] Specifically, the server will send the standard speed command to the vehicle control terminal through the Internet of Vehicles module. The vehicle control terminal will receive the command and compare it with the current vehicle speed. If the current vehicle speed is higher than the standard speed, a control command will be sent to the engine control unit to automatically reduce the throttle output, and a control signal will be sent to the brake control unit to perform slight braking to reduce the vehicle speed to the standard value. If the current vehicle speed is lower than the standard speed, a control command will be sent to increase the throttle output and accelerate the vehicle to the matching standard speed value. While controlling the vehicle speed, the server can also display the speed value and status information to the driver through the display to remind and warn the driver. If the driver actively changes the vehicle speed through operations such as the throttle, the on-board computer will also adjust accordingly to ensure that the vehicle speed remains within the standard range and ensure driving stability and safety.

[0063] The solution in this application is adopted to obtain road data from multiple angles through monitoring equipment, and use image processing and three-dimensional modeling technology to comprehensively and accurately evaluate road conditions, effectively solving the problem that the existing technology is difficult to accurately obtain road information. Based on the road flatness level, the bump level is determined in combination with the on-board sensor, and then the intelligent vehicle speed control strategy is determined. The actual vehicle speed is adjusted in real time through the Internet of Vehicles technology, effectively solving the problem that the existing technology is difficult to actively control the vehicle speed according to the road conditions.

[0064] In some embodiments, the laser radar device may include a laser generator, a transmitting lens, a wavelength selection and modulation component, etc., and may emit laser beams of different wavelengths and operating frequencies to the surface of the road ahead on which the vehicle is traveling. The emitted laser will generate an echo signal after hitting the road surface. The server analyzes the detected echo signal to obtain parameters such as the reflection intensity and waveform characteristics of the signal. Road surfaces of different materials have different reflection mechanisms for lasers. Cement roads have strong specular reflection properties, and the return signal intensity is high and the waveform characteristics are obvious; asphalt roads also have obvious specular reflection, but the intensity is slightly weaker; rough gravel roads will produce strong diffuse reflection, and the return signal intensity is low and the waveform is not obvious; dirt roads will absorb part of the laser energy, and the echo signal is very weak. The server combines the obtained echo signals and realizes the discrimination and classification of different echo signals according to the correspondence between the preset road surface reflection model and the road surface material, thereby inferring which road surface material the road surface belongs to. Among them, the road surface reflection model is a mathematical model used to describe the reflection mechanism and characteristics of different road surface materials to lidar signals. The specific construction process is as follows: in a controlled environment, use a standard lidar to scan roads of different materials and collect reflected echo signals; pre-process and denoise the signals to extract reflection intensity, waveform, spectrum and other characteristics; use a high-speed camera at a set position in the vehicle to capture a two-dimensional image of the road surface reflection to obtain the reflection pattern of the road surface; then collect a large amount of reflection sample data under different angles, distances, wetness and other conditions, and establish a road surface reflection model based on the sample database.

[0065] After determining the road material, since the hardness of different materials varies, the server can add road hardness as an important reference factor for evaluating the level of bump impact. This can be done by consulting relevant industry standards in advance to obtain the mechanical performance parameters of standard road materials such as cement, asphalt, gravel, soil and stone. Then conduct experimental tests to measure the hardness values ​​of various road materials and obtain detailed hardness data. Then simulate these material roads at the test site, let the test vehicle pass on them, collect wheel speed, body vibration, and sensor data, and obtain bump response data under different hardness roads. Then analyze the characteristics of bumps caused by different hardness roads, such as the peak acceleration and duration of bumps, and establish a quantitative correspondence model between road hardness and bump impact. According to safety and comfort requirements, determine the reasonable vehicle speed range under different bump levels, as well as the speed adjustment range, such as reducing 15km / h when bumps intensify. The road hardness data, corresponding bump levels, and vehicle speed adjustment mapping relationships can be integrated into a reference table, with hardness as the reference axis, and the standard bump impact level and speed compensation variables under different hardness roads are specified. A reference table is integrated into the vehicle control system. When the vehicle is driving on different roads and the LiDAR device detects the road material and hardness, the reference table can be queried to obtain the corresponding bump level compensation value and speed compensation variable, and the bump level compensation value and speed can be compensated accordingly. In this way, the road hardness information can be fully utilized to enhance the perception of complex environments, enabling the vehicle to achieve more intelligent speed control.

[0066] In some embodiments, the server may preset a time interval for updating the wheel data, for example, collecting new wheel data every 1 minute. After reaching this time interval, the server will re-acquire the latest wheel data and compare and analyze it with the original wheel data collected at the last moment to determine whether there is a significant difference between the new wheel data.

[0067] For example, the change in wheel speed data in two sets of collected results can be analyzed. If the change exceeds the preset difference threshold, it is determined that there is a significant difference. This indicates that the vehicle or road condition has changed. If the server determines that the new wheel data is significantly different from the original data, the newly collected wheel data is input into the preset bump impact assessment model to re-judge the bump impact data. The model will integrate multiple wheel data, match it with the road parameters, and evaluate the bumpiness of the vehicle. The server inputs new wheel data to re-determine whether the vehicle's bump data has changed under the current actual situation.

[0068] After re-determining the new bump data, the server will combine the on-board sensor to obtain the sensor data again, combine the sensor data to re-determine the bump level, and then query the corresponding relationship between the bump level and the standard speed value to determine the updated new standard speed value. The server will compare the old and new standard speed values. If there is a difference, it is necessary to send a control instruction to the on-board computer to adjust the current actual speed of the vehicle in real time to reach the new standard speed value and complete the dynamic control of the speed. This adjustment can be achieved by manipulating the accelerator and brakes, so that the vehicle can travel at a reasonable speed.

[0069] After combining the above embodiments, the method provided by this embodiment is further described in more detail below. Figure 2 , is another flow chart of the vehicle speed adjustment method integrating road surface roughness and wheel data in an embodiment of the present application.

[0070] S201, after extracting road surface key points from the road surface image, matching the road surface key points with corresponding points in the road surface point cloud data to obtain a plurality of matching feature point pairs;

[0071] After acquiring the road image and point cloud data, the server will first process the road image to detect and extract some feature points in the image. These feature points usually have obvious edges or texture changes in the road image, such as lane line boundaries. Feature point extraction can be achieved through edge detection and other algorithms.

[0072] After extracting the feature point set from the road image, the server will pre-process the application point cloud data, including filtering, denoising, and registration, to obtain clear point cloud data. Then, the server will retrieve the three-dimensional coordinate points corresponding to the image feature points in the point cloud data, and match them by comparing the pixel coordinates, angles, relative position relationships, and other information between the two. Finally, the server can obtain a point pair set of image feature points and matching three-dimensional point clouds.

[0073] The number of matching point pairs will affect the effect of subsequent 3D reconstruction. The server can set a threshold for the number of matching point pairs. If the number of matching point pairs is insufficient, it is necessary to readjust the parameters of image processing and point cloud processing, and repeat the feature point extraction and matching process until a sufficient number of matching point pairs is obtained, which can ensure the accuracy of subsequent 3D reconstruction.

[0074] S202, performing three-dimensional reconstruction according to the feature point pairs to determine a three-dimensional model of the target road;

[0075] After obtaining a large number of accurate matching point pairs between image feature points and point cloud data, the server can use the 3D reconstruction algorithm to restore the 3D model of the target road area. Commonly used 3D reconstruction algorithms include stereo vision method and structured light method. The server can select areas with denser point cloud data as reconstruction templates and set reconstruction accuracy parameters. Then, based on the matching point pairs, the 3D coordinates are filled point by point to gradually form a complete 3D grid model of the target road area.

[0076] After obtaining the preliminary 3D model, the server needs to perform model optimization processing, such as removing outliers and smoothing surfaces. Finally, a detailed 3D model is generated, in which the road contour, lane lines, and road surface bends, slopes, potholes and other ups and downs can be clearly observed. This 3D model can provide a visual and accurate structural reference for subsequent road condition analysis.

[0077] S203, determining the number of extracted horizontal road lines based on the road width data in the three-dimensional model; determining the length of the extracted horizontal road lines based on the road length data in the three-dimensional model;

[0078] After obtaining the three-dimensional model of the target road, the server needs to determine the number and length of the road horizontal lines for sampling and calculating the road surface height information based on the model. First, the server analyzes the road width data contained in the three-dimensional model. For example, the total width of the road in the model and the width of each lane can be measured. Then, the interval distance of the horizontal lines is determined based on empirical values ​​or preset parameters. For example, one horizontal line is extracted every meter. Combined with the total width of the road, the total number of horizontal lines that need to be extracted can be calculated. Secondly, the server analyzes the road length information in the three-dimensional model. The total length of the road section in the model can be directly measured, or the length can be inferred through positioning coordinate data. Then, based on the interval distance, the total length of the horizontal lines that need to be extracted on this road is calculated. Finally, based on the number and length of the horizontal lines, the server can determine how many horizontal lines need to be sampled and extracted in the subsequent process, as well as the number of sampling points on each horizontal line, to fully reflect the height changes of the entire road area.

[0079] S204, determining the point cloud coordinates of multiple key points of each of the horizontal road lines according to the number of horizontal road lines and the length of the horizontal road lines, wherein the horizontal road lines are equally spaced lines or lane lines parallel to the edge lines on both sides of the road, and the length of the horizontal road lines is the length of the road;

[0080] After determining the number and length of horizontal lines to be extracted, the server extracts these horizontal lines in the 3D model and determines the 3D coordinates of the key points on each horizontal line. The server can first generate equally spaced horizontal line frames on the 3D model of the road according to the interval distance. These line segments are parallel to the sidelines on both sides of the road and are distributed on the road lanes to cover the entire road surface. Then, on each horizontal line frame, based on the length of the horizontal line, a number of points on the line segment are reserved as key points using the sampling distance. The number of key points is related to the length of the line segment. Finally, the precise 3D coordinates of each key point in the 3D model are recorded.

[0081] These coordinate points can well reflect the road height information of the area where the horizontal line is located. Among them, the key points of the horizontal line located on the center line of the road can better represent the ups and downs of the vehicle's driving route. Recording these point cloud coordinates is the basis for calculating the road surface roughness.

[0082] The server will adjust the distribution of horizontal lines according to different road types and number of lanes. For example, for a single-lane road, horizontal lines can be evenly set in the middle of the road; for a two-lane road, horizontal lines need to be set on both lanes at the same time. The server can also appropriately increase the density of horizontal lines based on the length and curvature of the road.

[0083] S205, determining the z-axis difference between each key point cloud adjacent to each other on the x-axis and the y-axis according to the plurality of road horizontal lines;

[0084] After obtaining the three-dimensional coordinates of the key points on each road horizontal line, the server can calculate the difference between these points in the height (z-axis) direction, and then analyze the unevenness of the road surface. Specifically, the server will take two adjacent key points on each horizontal line and calculate the difference between the z-axis values ​​in their three-dimensional coordinates, that is, the height difference. By analogy, the height difference of all adjacent key points on the horizontal line can be obtained. Then, the server will continue to take adjacent key points on different horizontal lines and perform similar calculations. For example, the corresponding key points of adjacent horizontal lines in the x-axis direction or the y-axis direction can be taken. Finally, the server can count the height differences of all fixed points in the three-dimensional model.

[0085] These large amounts of height difference data can directly reflect the overall ups and downs and unevenness of the road surface. These data provide basic data support for the subsequent determination of raised and sunken areas on the road surface.

[0086] S206, recording the point cloud coordinate pair whose z-axis difference is greater than the set threshold as a convex coordinate pair, and recording the point cloud coordinate pair whose z-axis difference is less than the set threshold as a concave coordinate pair;

[0087] After calculating the height differences between all key points, the server needs to distinguish between point pairs with larger height differences, which represent raised areas on the road surface, and point pairs with smaller height differences, which represent sunken areas on the road surface.

[0088] Specifically, the server will pre-set a height difference threshold, such as 50 mm. When the z-axis height difference of a point pair is greater than this threshold, it is determined that the point pair corresponds to an area of ​​road protrusion or collision, and the server will record the coordinates of these point pairs and mark them as convex coordinate pairs. On the contrary, when the z-axis height difference of a point pair is less than the threshold, it can be determined that the point pair corresponds to an area of ​​road depression or pothole, and the server will record the coordinates of these point pairs and mark them as concave coordinate pairs.

[0089] Through this comparison and screening, the server can accurately distinguish the coordinates of raised areas and sunken areas with large road undulations from a large number of point pairs, laying the foundation for subsequent road condition analysis.

[0090] S207, clustering the convex coordinate pairs according to the distance to obtain a convex coordinate set corresponding to the specified road sub-region in the target road; clustering the concave coordinate pairs according to the distance to obtain a concave coordinate set corresponding to the specified road sub-region in the target road;

[0091] After obtaining the coordinate pairs of the road surface bulges and depressions, the server needs to perform coordinate clustering analysis to determine the distribution of bulges and depressions in each road section. For bulge coordinate pairs, the server will calculate the distance between each bulge coordinate point pair and determine whether the distance is less than the set distance threshold, such as less than 10 meters. For bulge coordinate point pairs with a distance less than 10 meters, the server will determine that they belong to the same bulge area cluster.

[0092] The server will calculate and judge all the convex coordinate point pairs in turn, and finally determine the coordinate range of each convex area in the target road through repeated distance clustering. Each convex coordinate point pair obtained by clustering constitutes a convex distribution set in a local sub-area of ​​the road. Similarly, for the concave coordinate pairs, the server also uses the same distance density clustering method. Calculate the distance between each concave coordinate point pair. If it is less than 10 meters, they are determined to belong to the same concave area cluster. Repeat clustering until all concave point pairs are clustered. Finally, the in-cluster coordinates of each concave area are a concave distribution set in a local sub-area of ​​the road.

[0093] Through distance clustering, the server can clearly determine whether the road damage is concentrated in a certain range or dispersed. For example, if the convex coordinates are concentrated in a certain section of the road, it can be determined that this may be the trace of uneven repair. If the concave point pairs are scattered, it may be that the road surface is full of potholes due to long-term vehicle driving. Distinguishing the distribution of convex and concave areas in the road surface helps the server to more accurately evaluate the flatness of the road surface, and can also support later road maintenance and improvement.

[0094] S208, determining a convexity corresponding to the designated road sub-region according to the convexity coordinate set; determining a concavity corresponding to the designated road sub-region according to the concavity coordinate set;

[0095] After obtaining the convex coordinate set and concave coordinate set in each road sub-area, the server needs to calculate the convexity and concaveness of the sub-area based on these coordinates. For the convexity, the server will count the number of point pairs in the convex coordinate set, as well as the maximum height difference and average height difference among these point pairs. If the number of convex point pairs is large, and the maximum or average height difference is large, it can be determined that the convexity of the sub-area is high. For example, if a road area with a length of 10 meters contains 20 convex coordinate point pairs, their maximum height difference is 30 mm, and the average height difference is 25 mm. The convexity of this area can be evaluated as medium.

[0096] As for the degree of depression, the server will count the number of point pairs in the depression coordinate set, as well as the minimum height difference and average height difference of these point pairs. If there are a large number of depression point pairs and the minimum or average height difference is small, it can be determined that the depression of the sub-area is high. For example, a road area contains 10 depression point pairs, their minimum height difference is -20 mm, and their average height difference is -15 mm. The depression of this area can be evaluated as a lighter level.

[0097] By integrating information such as the number of bumps and depressions in the area and the height difference, the server can derive the degree of bumps and depressions in the road sub-area, that is, the severity of road surface undulations and damage, providing a basis for subsequent flatness assessment.

[0098] S209: Determine the road surface unevenness height information according to the convexity and the concavity.

[0099] After obtaining the convexity and concavity of each road sub-area, the server can determine the height information of the road surface bumps based on this. Specifically, the server can sort the convexity of each sub-area as a whole, obtain the distribution pattern of the convexity in the complete road area, and determine the location of the most serious convexity. At the same time, the concavity of each sub-area can be counted to determine the location of the most serious concavity in the road area. Then, the server will correspondingly retrieve the original three-dimensional coordinate data of these sub-areas with larger convexity and concavity. By analyzing the distribution of the original height values ​​of these coordinate points, detailed height numerical information of these severely convex and concave areas can be obtained.

[0100] Through comprehensive detection and judgment of road conditions, it is possible to scientifically and rationally control the most reasonable wheel speed and realize intelligent regulation and optimization of wheel speed to adapt to complex and changing road environments.

[0101] In some embodiments, after the server obtains the road surface data in step S101, the road surface image can also be input into a preset crack recognition model to obtain the road surface crack situation. This process is of great significance for timely discovering road safety hazards and ensuring driving safety. The crack recognition model is a convolutional neural network model trained and established based on the crack categories and positions in multiple annotated real road surface images. Through a large amount of training data, the model can accurately identify the categories of road surface cracks, including no cracks, transverse cracks, longitudinal cracks, mesh cracks, etc., and determine the image coordinates of each crack category.

[0102] After feature matching between the road surface image and the road surface point cloud data, the image coordinates corresponding to each crack category are mapped to the three-dimensional coordinates in the road surface point cloud data to obtain multiple crack category information corresponding to multiple road sub-areas. For example, when a transverse crack is detected in the road surface image, the specific position and range of the transverse crack in the road surface point cloud data can be determined through feature matching and coordinate mapping, so as to accurately understand the distribution of the crack on the road.

[0103] After determining the danger level based on the crack category information, the alarm device is controlled to issue an alarm message. This helps to promptly remind relevant personnel to pay attention to road safety issues and take appropriate measures to repair or warn. For example, if the cracks detected are serious mesh cracks with a high danger level, the alarm device will issue a strong alarm signal to attract the attention of relevant personnel. This crack detection and alarm mechanism plays an important role in practical applications. For example, on highways, if road cracks can be discovered and handled in time, traffic accidents caused by crack expansion can be avoided. For another example, on urban roads, timely repair of cracks can improve the smoothness of the road, reduce vehicle bumps, and improve driving comfort.

[0104] The above contents can effectively improve the road safety level, reduce potential safety risks, and provide strong protection for driving safety.

[0105] In some embodiments, the server integrates and processes the crack category with the corresponding image coordinates. Then, through pre-set communication channels, such as wireless networks, wired networks, etc., this information is accurately sent to the receiving end in a visual form. The receiving end can be the monitoring equipment of the road management department, the mobile terminal of the maintenance personnel, etc. When the receiving end receives this information, it will display the corresponding crack category in different colors at the position corresponding to the image coordinates. For example, transverse cracks may be displayed in red, longitudinal cracks in blue, mesh cracks in yellow, and no cracks in green. This kind of visual display method enables users at the receiving end to understand the distribution and type of road cracks very intuitively.

[0106] In this way, road management departments can have a clearer understanding of road conditions. They can see cracks marked with different colors on the monitoring screen and quickly identify areas that need special attention and timely treatment.

[0107] In some embodiments, the server can analyze the continuity of the extracted key point cloud in the x-axis direction. If it is detected that there are a large number of breakpoints in the point cloud on the x-axis, that is, the number of points in this direction is significantly reduced, it can be determined that there is a longitudinal crack here. The server will preset a threshold for the number of breakpoints, for example, more than 10 breakpoints are considered to exist. When the number of breakpoints exceeds the threshold, the server will confirm the coordinate positions of these breakpoints on the image and mark them as representing the longitudinal crack area. Similarly, the server will detect the number of breakpoints in the key point cloud in the y-axis direction. If the number of breakpoints exceeds the preset threshold, it is determined that the image coordinates corresponding to these breakpoints represent the transverse crack area. Because the y-axis represents the transverse direction of the road, the presence of many breakpoints on the y-axis indicates that there may be transverse cracks, such as rutting. The server will also mark the coordinates that meet the breakpoint number conditions as transverse cracks.

[0108] If the server detects that the key point cloud of the same area has a large number of breakpoints on both the x-axis and the y-axis, that is, the continuity of the area in different directions is relatively poor, it can be determined that there are network cracks in the area. The server will identify the image sub-areas where these breakpoints are concentrated, and determine them as network crack areas, marking them as key areas for subsequent repairs.

[0109] In the embodiments of the present application, compared with the prior art that directly determines the flatness of the road surface based on the road surface image, the present application uses the three-dimensional modeling technology that fuses images and point clouds to more accurately and three-dimensionally analyze the undulation height of the road surface, avoid judgment deviations caused by a single data source, and effectively solve the problem in the prior art that it is difficult to accurately obtain road surface bump height information.

[0110] The following describes the server in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of a server in an embodiment of the present application.

[0111] It should be noted that Figure 3 The structure of the server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0112] like Figure 3 As shown, the server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0113] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0114] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0115] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0116] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0117] Specifically, the server of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the vehicle speed adjustment method for integrating road surface roughness and wheel data provided in the above embodiment is implemented.

[0118] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the server described in the above embodiment; or may exist independently without being assembled into the server. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the server, the server implements the vehicle speed adjustment method integrating road surface roughness and wheel data provided in the above embodiment.

[0119] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0120] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0121] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A vehicle speed adjustment method integrating road surface roughness and wheel data, applied to a server, characterized in that: The method comprises: Acquiring road surface data of a target road through a monitoring device, wherein the road surface data includes a road surface image and road surface point cloud data; Determine the road surface bump height information by combining the road surface image and the road surface point cloud data to obtain road surface state parameters; Inputting the road surface state parameters into a pre-trained roughness evaluation model to obtain the road surface roughness grade of the target road, wherein the roughness evaluation model is a neural network model established by training data of a plurality of real road roughness grades and corresponding road surface image features and three-dimensional point cloud features in advance, and the road surface roughness grades include smooth, slightly uneven, moderately uneven, and severely uneven; When the road surface roughness level is within a set level of unevenness, obtaining wheel data of the target vehicle, wherein the set level includes slight unevenness, moderate unevenness, and severe unevenness; Inputting the road surface roughness grade and the wheel data into a preset bump impact assessment model to obtain bump data, wherein the bump impact assessment model is an empirical model established based on the wheel data of different vehicle models and the relationship between the matched road surface roughness grade and the actual bump impact degree; Determine sensor data of a target vehicle based on onboard sensors; Determining a bump impact level of the target vehicle by combining the bump data and the sensor data; Determining a matching standard vehicle speed value according to the bump impact level; Controlling the current speed of the target vehicle according to the standard speed value; The step of determining the road surface unevenness height information by combining the road surface image and the road surface point cloud data to obtain the road surface state parameters specifically includes: After extracting road surface key points from the road surface image, matching the road surface key points with corresponding points in the road surface point cloud data to obtain a plurality of matching feature point pairs; Performing three-dimensional reconstruction based on the feature point pairs to determine a three-dimensional model of the target road; Determine the number of extracted road horizontal lines based on the road width data in the three-dimensional model; determine the length of the extracted road horizontal lines based on the road length data in the three-dimensional model; Determine the point cloud coordinates of multiple key points of each of the horizontal road lines according to the number of horizontal road lines and the length of the horizontal road lines, wherein the horizontal road lines are equally spaced lines or lane lines parallel to the edge lines on both sides of the road, and the length of the horizontal road lines is the length of the road; Determine the z-axis difference between each key point cloud adjacent to each other on the x-axis and the y-axis according to the plurality of road horizontal lines; The point cloud coordinate pairs whose z-axis difference is greater than the set threshold are recorded as convex coordinate pairs, and the point cloud coordinate pairs whose z-axis difference is less than the set threshold are recorded as concave coordinate pairs; Clustering the convex coordinate pairs according to distance to obtain a convex coordinate set corresponding to a designated road sub-region in the target road; clustering the concave coordinate pairs according to distance to obtain a concave coordinate set corresponding to a designated road sub-region in the target road; Determine the convexity corresponding to the designated road sub-region according to the convexity coordinate set; determine the concavity corresponding to the designated road sub-region according to the concavity coordinate set; Determine the road surface unevenness height information according to the convexity and the concavity; After the step of determining the z-axis difference between each key point cloud adjacent to each other on the x-axis and the y-axis according to the plurality of road horizontal lines, the method further includes: If it is detected that the number of breakpoints of the key point cloud on the x-axis is greater than a set threshold, the image coordinates corresponding to the breakpoints are determined to be longitudinal crack areas; If it is detected that the number of breakpoints of the key point cloud on the y-axis is greater than a set threshold, the image coordinates corresponding to the breakpoints are determined to be a transverse crack area; If the number of breakpoints of the key point cloud on the x-axis and the y-axis is greater than the set threshold, the image coordinates corresponding to the breakpoints are determined to be a mesh crack area; After the step of acquiring wheel data of the target vehicle when the road surface flatness level is within the set level of unevenness, the method further includes: determining a current load according to a current weight of the target vehicle detected by an onboard weight sensor; Determining the magnitude of the impact force generated when the target vehicle contacts the road surface according to the road surface flatness grade, the wheel data and the current load; The impact force magnitude is compared with a set impact force threshold range to determine the turbulence impact level.

2. The method according to claim 1, characterized in that After the step of acquiring the road surface data of the target road by the monitoring device, the method further includes: Input the pavement image into a preset crack recognition model to obtain pavement crack conditions, wherein the pavement crack conditions at least include crack categories of no cracks, transverse cracks, longitudinal cracks, and mesh cracks, and image coordinates corresponding to each crack category, wherein the crack recognition model is a convolutional neural network model established by training based on crack categories and positions in a plurality of annotated real pavement images; After feature matching between the road surface image and the road surface point cloud data, the image coordinates corresponding to each of the crack categories are mapped to three-dimensional coordinates in the road surface point cloud data to obtain a plurality of crack category information corresponding to a plurality of road sub-areas; After the danger level is determined according to the crack category information, the alarm device is controlled to issue an alarm message.

3. The method according to claim 2, characterized in that After determining the danger level according to the crack category information, after the step of controlling the alarm device to issue an alarm message, the method further includes: The image coordinates corresponding to the crack category are sent to a receiving end in a visual form through a set channel, so that the receiving end displays the corresponding crack category in different colors at the position corresponding to the image coordinates.

4. The method according to claim 1, characterized in that: After the step of controlling the current speed of the target vehicle according to the standard speed value, the method further includes: After setting the wheel data update time interval, determine whether there is a difference between the updated new wheel data and the original wheel data; If there is a difference, the new wheel data is input into the bump impact assessment model to re-determine the updated bump impact level; According to the updated bump impact level, the updated new standard vehicle speed value is re-determined; If the new standard vehicle speed value is different from the original standard vehicle speed value, the current vehicle speed of the target vehicle is adjusted according to the new standard vehicle speed value.

5. A server, characterized in that: The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the server to execute the method described in any one of claims 1-4.

6. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a server, the server is caused to execute the method according to any one of claims 1 to 4.

7. A computer program product, characterized in that When the computer program product is run on a server, the server is caused to execute the method according to any one of claims 1 to 4.

Images