Wheel loosening early warning method and device, medium and vehicle

Abstract

The invention discloses a wheel looseness early warning method and device, a medium and a vehicle, and is applied to the technical field of vehicle control, and the method comprises the steps: obtaining a wheel image, wheel attribute data and driving attribute data of the vehicle; according to the wheel image and the wheel attribute data, target tire pressure detection information and target fastening detection information of the vehicle are obtained; wherein the target tire pressure detection information is used for indicating whether the tire pressure of the wheel is normal or not, and the target fastening detection information is used for indicating whether the mechanical structure of the wheel is fastened or not; if the target tire pressure detection information is used for indicating that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used for indicating that the mechanical structure of the wheel is not fastened, wheel movement track data of the vehicle is obtained according to the wheel image, the wheel attribute data and the driving attribute data; and based on the wheel movement track data, performing wheel looseness early warning processing on the vehicle. According to the invention, the accuracy of wheel loosening early warning can be effectively improved.

Description

Technical Field

[0001] This application relates to the technical field of vehicle control, and in particular, to a method, device, medium, and vehicle for warning of wheel loosening. Background Art

[0002] Vehicles are one of the most common, prevalent, and frequently used means of transportation in daily life. Wheel loosening is a major hidden danger to vehicle safety, which affects the driving stability of the vehicle and increases the risk of traffic accidents. Therefore, warning of wheel loosening is crucial for vehicle safety maintenance. In related technologies, variables such as the tire pressure value or the change value of the tire pressure of the wheel are detected in real time when the vehicle is driving. If the variable reaches a set threshold, a signal for warning of wheel loosening is sent to indicate the risk of wheel loosening. However, the related technologies only judge whether the wheel is loose by the tire pressure state of the wheel, which depends on a single-dimensional factor, resulting in low accuracy of the wheel loosening warning. Summary of the Invention

[0003] Embodiments of this application provide a method, device, medium, and vehicle for warning of wheel loosening, which are used to improve the accuracy of the wheel loosening warning.

[0004] On the one hand, embodiments of this application provide a method for warning of wheel loosening, including the following steps: Obtain the wheel image, wheel attribute data, and driving attribute data of the vehicle; According to the wheel image and the wheel attribute data, obtain the target tire pressure detection information and the target fastening detection information of the vehicle; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; In the case that the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened, obtain the wheel movement trajectory data of the vehicle according to the wheel image, the wheel attribute data, and the driving attribute data; Based on the wheel movement trajectory data, perform wheel loosening warning processing on the vehicle.

[0005] On the other hand, embodiments of this application provide a device for warning of wheel loosening, including: An acquisition module, configured to obtain the wheel image, wheel attribute data, and driving attribute data of the vehicle; The first processing module is configured to obtain the target tire pressure detection information and the target fastening detection information of the vehicle according to the wheel image and the wheel attribute data; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; The second processing module is configured to obtain the wheel movement trajectory data of the vehicle according to the wheel image, the wheel attribute data, and the driving attribute data when the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened; The third processing module is configured to perform a wheel loosening warning process on the vehicle based on the wheel movement trajectory data.

[0006] In another aspect, an embodiment of the present application provides a computer-readable storage medium, and the program executable by the processor is used to implement the above-mentioned wheel loosening warning method when executed by the processor.

[0007] In another aspect, an embodiment of the present application provides a vehicle, including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the above-mentioned wheel loosening warning method.

[0008] According to a wheel loosening warning method, device, medium, and vehicle provided by the present application, first, a wheel image, wheel attribute data, and driving attribute data of a vehicle are obtained; then, target tire pressure detection information and target fastening detection information of the vehicle are obtained according to the wheel image and the wheel attribute data; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; after that, if the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened, then the wheel movement trajectory data of the vehicle is obtained according to the wheel image, the wheel attribute data, and the driving attribute data; finally, a wheel loosening warning process is performed on the vehicle based on the wheel movement trajectory data. According to the technical solution of the present application, multi-dimensional factors such as image data and attribute data associated with the wheels of the vehicle, and attribute data associated with the vehicle driving are fully considered, and wheel loosening warning processing is implemented according to these multi-dimensional factors, so as to exclude the interference of loosening misjudgment caused by single-dimensional factors, thereby effectively improving the accuracy of wheel loosening warning.

[0009] Other features and advantages of the present application will be described in the subsequent specification, and, in part, will be obvious from the specification, or will be understood by implementing the present application. The objectives and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the specification, claims, and drawings. Description of the Drawings

[0010] Figure 1 is a flowchart of a method for warning of wheel looseness provided by the present application; Figure 2 is Figure 1 a specific flowchart of step S102 in Figure 3 is Figure 1 another specific flowchart of step S102 in Figure 4 is Figure 1 a specific flowchart of step S103 in Figure 5 is Figure 1 a specific flowchart of step S104 in Figure 6 is a flowchart of model training provided by the present application; Figure 7 is a specific implementation process diagram of a method for warning of wheel looseness provided by the present application; Figure 8 is a structural diagram of a device for warning of wheel looseness provided by the present application; Figure 9 is an example diagram of a vehicle provided by the present application. Detailed Description of the Embodiments

[0011] In order to make the objectives, technical solutions, and advantages of the present application more clear and understandable, the present application will be further described in detail below in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0012] The present application will be further described below in conjunction with the drawings in the specification and specific embodiments. The described embodiments should not be regarded as limitations on the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0013] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0014] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used herein are for the purpose of describing embodiments of this application only and are not intended to limit this application.

[0015] With the rapid development of technology, vehicles are one of the most common, prevalent, and frequently used means of transportation in daily life and have become important commuting and production tools. When a vehicle is in motion, if there are problems with the wheels, it will affect the driving stability of the vehicle, and situations such as increased noise, vehicle body vibration, and vehicle deviation may occur. In severe cases, it may lead to tire blowout or wheel detachment, thus increasing the risk of traffic accidents. It can be seen that loose wheels are a major hidden danger to vehicle safety, and early warning of wheel looseness is crucial for vehicle safety maintenance.

[0016] In the related art, when a vehicle is in motion, variables such as the tire pressure value or the change value of the tire pressure of the wheels are detected in real time. If the tire pressure value or the change value of the tire pressure reaches a set threshold, a signal for early warning of wheel looseness is sent to indicate the risk of wheel looseness. However, the tire pressure is easily affected by factors such as temperature, load, and altitude. The related art only judges whether the wheels are loose based on the tire pressure state of the wheels, which relies on this single-dimensional factor of tire pressure and is prone to situations such as misjudging the wheels as loose, resulting in low accuracy of the early warning of wheel looseness.

[0017] In view of this, the embodiments of this application provide a method, device, medium, and vehicle for early warning of wheel looseness, aiming to monitor the tire pressure state and fastening state of the wheels and the deviation state of the vehicle in real time, and perform corresponding early warning of wheel looseness according to the monitoring results, thereby effectively improving the accuracy of the early warning of wheel looseness.

[0018] First, the implementation steps of a method for early warning of wheel looseness provided by the embodiments of this application will be elaborated in detail below with reference to the accompanying drawings.

[0019] A method for warning of wheel loosening provided by an embodiment of the present application can be applied to a terminal, or to a server, or can be software running on a terminal or a server, etc. The terminal can be a tablet computer, a notebook computer, a desktop computer, etc., but is not limited thereto. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or can be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms. In addition, the server can also be a node server in a blockchain network, but is not limited thereto. Among them, the blockchain is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithms.

[0020] Referring to Figure 1 , Figure 1 FIG. is a flowchart of a method for warning of wheel loosening provided by the present application. The method for warning of wheel loosening may include the following steps S101-S104: S101, obtaining a wheel image, wheel attribute data, and driving attribute data of a vehicle; S102, obtaining target tire pressure detection information and target fastening detection information of the vehicle according to the wheel image and the wheel attribute data; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; S103, in the case where the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened, obtaining wheel movement trajectory data of the vehicle according to the wheel image, the wheel attribute data, and the driving attribute data; S104, performing a wheel loosening warning process on the vehicle based on the wheel movement trajectory data.

[0021] In the above step S101, image data and attribute data associated with the wheels of the vehicle are obtained before and during the vehicle's driving to serve as the wheel image and wheel attribute data of the vehicle; and, attribute data associated with the vehicle's driving is obtained to serve as the driving attribute data of the vehicle, so as to facilitate detecting and warning of whether the wheels are loose using these data as a basis in subsequent steps.

[0022] Here, the above wheel images may include, but are not limited to, image data associated with the wheel morphology before the vehicle travels, and image data associated with the local temperature of the wheel when the vehicle travels. The above wheel attribute data may include, but are not limited to, attribute data associated with the physical characteristics of the wheel when the vehicle travels, and these attribute data are numerical text data. The above driving attribute data may include, but are not limited to, attribute data associated with the physical characteristics of the vehicle when it travels, and these attribute data are numerical text data.

[0023] In the above step S102, in the wheel loosening warning, first, based on the wheel images and the wheel attribute data, it is detected whether the tire pressure of the wheel is normal and whether the mechanical structure installation of the wheel is tight, so as to obtain the target tire pressure detection information and the target tightness detection information of the vehicle. Among them, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target tightness detection information is used to indicate whether the mechanical structure of the wheel is tight.

[0024] Here, the tire pressure of the wheel is related to the wheel loosening height. When the tire pressure is abnormal because it is lower than the normal value, it means that the tire is leaking air. At this time, there may be situations such as the fixing bolts of the wheel not being tightened, the bearing being damaged, and the wheel hub being deformed, which may cause the wheel to loosen. In other words, the abnormal tire pressure of the wheel can indirectly reflect the wheel loosening. Therefore, detecting whether the tire pressure of the wheel is normal first in the wheel loosening warning helps to quickly identify whether there is a potential risk of wheel loosening in the vehicle.

[0025] Similarly, the tightness of the mechanical structure of the wheel is related to the wheel loosening height. The tightness of the mechanical structure of the wheel generally refers to the tightness of the fixing bolts, wheel hub bolts, wheel hub bearings, etc. of the wheel. When there are situations such as the fixing bolts or bearing bolts of the wheel not being tightened, and the wheel hub bearing being deformed, it means that the tightness of the mechanical structure of the wheel is insufficient. At this time, the wheel may lose its fixed restraint, which may cause the wheel to loosen. In other words, the insufficient tightness of the mechanical structure of the wheel can indirectly reflect the wheel loosening. Therefore, detecting whether the mechanical structure of the wheel is tight first in the wheel loosening warning helps to quickly identify whether there is a potential risk of wheel loosening in the vehicle.

[0026] In the above step S103, if the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal and / or the target tightness detection information is used to indicate that the mechanical structure of the wheel is not tight, it means that the tire pressure of the wheel is abnormal and / or the mechanical structure of the wheel is not tight, and there is a potential risk of wheel loosening in the vehicle. At this time, based on the wheel images, the wheel attribute data, and the driving attribute data, the movement trajectory of the wheel in the future time period is predicted, so as to obtain the wheel movement trajectory data of the vehicle.

[0027] Here, during the driving of the vehicle, if the movement trajectory of the wheel is not a perfect circle, the driving route has a regular bend, or there is a lateral movement trajectory relative to the vehicle body, it indicates that the wheel is out of alignment; otherwise, it indicates that the wheel is not out of alignment. Whether the wheel is out of alignment is related to the looseness height of the wheel. When the wheel is out of alignment, it indicates that there are situations such as insufficient dynamic balance of the wheel and hub displacement. For example, insufficient dynamic balance of the wheel will cause the steering wheel to shake and the wheel to be out of alignment. Another example is that hub displacement of the wheel will cause lateral displacement of the wheel and the wheel to be out of alignment, and at this time, wheel looseness will be triggered. In other words, wheel alignment can directly reflect wheel looseness. Therefore, on the premise of determining that the vehicle has a potential risk of wheel looseness, detecting the wheel movement trajectory data of the vehicle helps to quickly and accurately determine whether the wheel is loose. It should be understood that if the target tire pressure detection information is used to indicate that the tire pressure of the wheel is normal and the target fastening detection information is used to indicate that the mechanical structure of the wheel is fastened, it means that the tire pressure of the wheel is normal and the mechanical structure of the wheel is fastened, and the vehicle does not have a potential risk of wheel looseness. At this time, it will return to the above step S101 to achieve cyclic detection.

[0028] In the above step S104, after determining the wheel movement trajectory data of the vehicle, taking the wheel movement trajectory data as a reference, judge whether the wheels of the vehicle are loose, and accordingly perform a wheel looseness warning process on the vehicle.

[0029] Through the above steps S101 - S104, the embodiment of the present application first detects whether the tire pressure of the wheel is normal and whether the mechanical structure of the wheel is fastened based on the image data and attribute data associated with the wheel, so as to initially identify whether the vehicle has a potential risk of wheel looseness. Then, if the vehicle has a potential risk of wheel looseness, based on the image data and attribute data associated with the wheels of the vehicle, and the attribute data associated with the vehicle driving, predict the wheel movement trajectory data of the vehicle, and accordingly perform a wheel looseness warning process. In this way, the embodiment of the present application does not simply consider single - dimensional factors, but fully considers multi - dimensional factors such as the image data and attribute data associated with the wheels of the vehicle, and the attribute data associated with the vehicle driving, and realizes the wheel looseness warning process according to these multi - dimensional factors. This can eliminate the interference of false looseness judgment caused by single - dimensional factors (tire pressure), thereby effectively improving the accuracy of wheel looseness warning, ensuring that the wheels are always in good condition, guaranteeing the safety of the driver and passengers, reducing the risk of traffic accidents, and helping to improve the safety, intelligence, and comfort experience of users when driving or riding in the vehicle.

[0030] In some embodiments, referring to Figure 2, the above wheel images may include the grounding image of the wheel before the vehicle travels and the infrared image of the wheel when the vehicle travels; the above wheel attribute data may include the first tire pressure data of the wheel before the vehicle travels, and the second tire pressure data, effective radius data, and hub vibration frequency data of the wheel when the vehicle travels; in the above step S102, according to the wheel images and the wheel attribute data, obtaining the target tire pressure detection information and the target fastening detection information of the vehicle may include the following steps S201 - S202: S201, determine the tire pressure risk coefficient of the vehicle before traveling according to the grounding image and the first tire pressure data; S202, perform tire pressure detection processing according to the tire pressure risk coefficient, infrared image, second tire pressure data, effective radius data, and hub vibration frequency data, and obtain the tire pressure detection information of the vehicle when traveling as the target tire pressure detection information.

[0031] Here, in the above wheel images, the grounding image refers to the top-down image of the tire grounding shape of the wheel, which indicates the shape of the wheel before the vehicle travels and is related to the tire pressure height of the wheel. Under normal circumstances, the tire contact area is in a uniform elliptical shape; if the tire contact area becomes wider and the edge deformation is obvious, it indicates that there may be a situation of insufficient tire pressure. The infrared image refers to the infrared thermal imaging image of the wheel, which indicates the change in the thermal distribution of the wheel during rolling when the vehicle is traveling and is related to the tire pressure height of the wheel. Under normal circumstances, the change in the thermal distribution of the wheel during rolling should be uniform; if the local temperature of the wheel increases during rolling, it indicates that there may be a situation where the sidewall is repeatedly bent, resulting in insufficient tire pressure. Therefore, the grounding image and the infrared image are helpful for detecting whether the tire pressure of the vehicle is abnormal.

[0032] Optionally, the grounding image is collected by integrating a normal camera in the wheel arch; the infrared image is collected by integrating an infrared camera in the wheel arch, but it is not limited to this.

[0033] Among the above wheel attribute data, the tire pressure data refers to the tire pressure of the wheel, which directly indicates the tire pressure of the wheel. When the tire pressure data is lower than the normal value, it indicates insufficient tire pressure. The effective radius data refers to the equivalent radius when the tire of the wheel actually rolls under load, which is highly correlated with the tire pressure of the wheel. Under normal circumstances, the effective radius data remains basically unchanged; if the effective radius data decreases, it may indicate insufficient tire pressure. For example, when the tire pressure of the left front wheel is insufficient, the rotational speed of the left front wheel will be slightly higher than that of the right front wheel because the effective radius data of the left front wheel decreases and more rotations are required to maintain the same vehicle speed. The hub vibration frequency data refers to the frequency at which the hub of the wheel vibrates freely without external excitation, which is highly correlated with the tire pressure of the wheel. Under normal circumstances, the hub vibration frequency data remains basically unchanged; if the hub vibration frequency data decreases, it indicates a decrease in tire stiffness and may indicate insufficient tire pressure. Therefore, the tire pressure data, effective radius data, and hub vibration frequency data are helpful for detecting whether the tire pressure of the vehicle is abnormal.

[0034] Optionally, the tire pressure data is collected by a tire pressure sensor, and the hub vibration frequency data is collected by a hub vibration sensor; the vehicle speed data of the vehicle is obtained by a vehicle speed sensor, and the wheel speed data of the vehicle is obtained by a wheel speed sensor. The ratio of the vehicle speed data to the wheel speed data is obtained as the effective radius data, but it is not limited to this.

[0035] In the related art, usually only the tire pressure value or the change value of the tire pressure of the wheel is used to detect whether the tire pressure of the wheel is abnormal. However, the tire pressure is easily affected by factors such as temperature, load, and altitude. The related art is prone to misjudging abnormal tire pressure, and the accuracy of its tire pressure detection needs to be improved. In response to this, this embodiment introduces multi-modal data and adopts a dynamic and static two-stage tire pressure detection method to improve the accuracy of tire pressure detection.

[0036] In the above step S201, the tire pressure risk prediction is carried out based on the grounding image of the wheel and the first tire pressure data before the vehicle travels, and the tire pressure risk coefficient before the vehicle travels is obtained. This tire pressure risk coefficient can reflect the probability of abnormal tire pressure of the wheel before the vehicle travels.

[0037] Here, the vehicle is in a stationary state before driving, and at this time, the wheels are in a cold tire state. The probability of abnormal tire pressure of the vehicle's wheels before driving is determined through the grounding image and tire pressure data of the wheels in the cold tire state, so as to achieve tire pressure detection in the static stage. The data in the image modality can effectively capture the physical deformation characteristics of the wheels in the cold tire state (such as abnormal grounding area, irregular edges, etc.), and the data in the text modality can effectively capture the physical tire pressure characteristics of the wheels in the cold tire state. In this embodiment, the introduction of multi-modal data is used to achieve tire pressure risk prediction. In this way, the accuracy of tire pressure detection in the static stage can be effectively improved, and potential situations suspected of abnormal tire pressure can be quickly and accurately discovered, which helps to improve the efficiency of tire pressure detection.

[0038] In some embodiments, the above-mentioned determining the tire pressure risk coefficient of the vehicle before driving according to the grounding image and the first tire pressure data may include: extracting features from the grounding image to obtain grounding image features; extracting features from the first tire pressure data to obtain first tire pressure features; and determining the tire pressure risk coefficient of the vehicle before driving according to the grounding image features and the first tire pressure features, in combination with a machine learning method.

[0039] Here, in the tire pressure detection in the static stage, first, features are extracted from the grounding image to obtain grounding image features. Among them, the grounding image features can be the width features and edge deformation features of the wheels. The width features reflect the width situation of the wheels, and the edge deformation features reflect the deformation situation of the wheels when they are grounded, so as to more comprehensively reflect the characteristic information of the wheels when they are grounded. At the same time, features are extracted from the first tire pressure data to obtain first tire pressure features. Among them, the specific implementation of feature extraction can be flexibly set according to the actual situation, and this embodiment does not limit this. Then, the grounding image features and the first tire pressure features are input into a pre-trained first tire pressure risk prediction model to obtain the tire pressure risk coefficient of the vehicle before driving, and the risk coefficient is the prediction probability. Among them, the first tire pressure risk prediction model is a machine learning model trained through multiple first feature samples and the label information corresponding to each first feature sample. The first feature samples include grounding image feature samples and tire pressure feature samples, and the label information corresponding to the first feature samples is the tire pressure risk coefficient, which can be determined through expert annotation, historical data statistics, etc. In addition, the type of the first tire pressure risk prediction model can be flexibly set according to the actual situation. For example, it can be models such as support vector machines and logistic regression, but it is not limited thereto. In this way, the tire pressure risk prediction process is performed through a machine learning method. Since the machine learning method has excellent performance, introducing the machine learning method in tire pressure risk prediction can effectively improve the accuracy of tire pressure risk prediction.

[0040] In some embodiments, determining the tire pressure risk coefficient of the vehicle before driving based on the grounding image and the first tire pressure data may include: predicting the tire pressure risk according to the grounding image to obtain a first risk coefficient; predicting the tire pressure risk according to the first tire pressure data to obtain a second risk coefficient; and determining the tire pressure risk coefficient of the vehicle before driving according to the first risk coefficient and the second risk coefficient.

[0041] Here, in the tire pressure detection in the static stage, first, the tire pressure risk is predicted according to the grounding image in combination with a pre-trained second tire pressure prediction model to obtain a first risk coefficient, and the first risk coefficient indicates the probability of abnormal tire pressure of the vehicle's wheels before driving. The risk coefficient is the prediction probability. Among them, the second tire pressure risk prediction model is a model trained by multiple grounding image samples and the label information corresponding to each grounding image sample. The label information corresponding to the grounding image sample is the risk coefficient, which can be determined by means such as expert annotation and historical data statistics. In addition, the type of the second tire pressure risk prediction model can be flexibly set according to the actual situation. For example, it can be a convolutional neural network, but not limited thereto. At the same time, the tire pressure risk is predicted according to the first tire pressure data in combination with a pre-trained third tire pressure prediction model to obtain a second risk coefficient, and the second risk coefficient indicates the probability of abnormal tire pressure of the vehicle's wheels before driving. The risk coefficient is the prediction probability. Among them, the third tire pressure prediction model is a model trained by multiple tire pressure data samples and the label information corresponding to each tire pressure data sample. The label information corresponding to the tire pressure data sample is the risk coefficient, which can be determined by means such as expert annotation and historical data statistics. In addition, the type of the third tire pressure risk prediction model can be flexibly set according to the actual situation. For example, it can be models such as support vector machine and logistic regression, but not limited thereto. Then, based on the first risk coefficient and the second risk coefficient, integration is performed to obtain the tire pressure risk coefficient of the vehicle before driving. For example, calculate the mean of the first risk coefficient and the second risk coefficient as the tire pressure risk coefficient of the vehicle before driving. Another example is to perform weighted summation on the first risk coefficient and the second risk coefficient to obtain the tire pressure risk coefficient of the vehicle before driving. In this way, this embodiment predicts the tire pressure risk from the aspect of image-modal data, and can accurately determine the tire pressure risk based on the physical deformation characteristics, while predicting the tire pressure risk from the aspect of text-modal data, and can accurately determine the tire pressure risk based on the physical tire pressure characteristics. Considering that the tire pressure risks reflected by different-modal data are different, the final tire pressure risk is determined by integrating the tire pressure risks corresponding to multiple modalities. Thus, it helps to further improve the accuracy of tire pressure risk prediction.

[0042] In the above step S202, based on the probability of abnormal tire pressure of the wheels before the vehicle starts moving, target tire pressure detection information for indicating whether the tire pressure of the vehicle's wheels is abnormal during driving is determined through multi-dimensional factors such as the infrared images of the wheels in the driving state, second tire pressure data, effective radius data, and hub vibration frequency data.

[0043] Here, on the one hand, compared with the method of realizing tire pressure detection only through tire pressure values or tire pressure change values, this embodiment fully considers multi-modal data related to the tire pressure height of the vehicle, and realizes tire pressure detection in the dynamic stage through this multi-modal data. On the other hand, compared with the method of realizing tire pressure detection only through real-time data during vehicle driving, in the tire pressure detection in the dynamic stage of this embodiment, not only the real-time data related to the tire pressure height during vehicle driving is considered, but also the tire pressure risk probability before the vehicle starts moving is considered. Thus, this embodiment introduces multi-dimensional data in the tire pressure detection in the dynamic stage, and considers the tire pressure conditions when the vehicle is stationary and when it is not stationary, thereby effectively improving the comprehensiveness, accuracy, and reliability of tire pressure detection and reducing the risk of misjudgment.

[0044] In some embodiments, considering that the tire pressure detection of the vehicle at the current moment is relatively important, the maximum cycle time of the above step S202 is thirty seconds, that is, the above step S202 is executed once every thirty seconds or less than every thirty seconds (for example, every twenty seconds), and the result of the newly executed one covers the result of the previous execution, so as to realize real-time tire pressure detection.

[0045] In some embodiments, the above tire pressure detection process based on the tire pressure risk coefficient, infrared images, second tire pressure data, effective radius data, and hub vibration frequency data to obtain the tire pressure detection information of the vehicle during driving may include: extracting features from the infrared images, second tire pressure data, effective radius data, and hub vibration frequency data to obtain infrared temperature field features, second tire pressure features, effective radius features, and hub vibration frequency features; and obtaining the tire pressure detection information of the vehicle during driving by combining the machine learning method according to the tire pressure risk coefficient, infrared temperature field features, second tire pressure features, effective radius features, and hub vibration frequency features.

[0046] Here, during tire pressure detection, first, feature extraction is performed on the infrared image, second tire pressure data, effective radius data, and wheel hub vibration frequency data respectively to obtain an infrared temperature field feature, a second tire pressure feature, an effective radius feature, and a wheel hub vibration frequency feature. In this way, feature information highly related to the tire pressure during vehicle driving can be extracted. Among them, the specific implementation of feature extraction can be flexibly set according to the actual situation, and this embodiment does not limit it. Then, the tire pressure risk coefficient, infrared temperature field feature, second tire pressure feature, effective radius feature, and wheel hub vibration frequency feature are input into a pre-trained tire pressure detection model, and the tire pressure detection information of the vehicle during driving is obtained through the tire pressure detection model. Among them, the tire pressure detection model is a model trained through multiple second feature samples and the label information corresponding to each second feature sample. The second feature samples include tire pressure risk coefficient samples, infrared temperature field feature samples, tire pressure feature samples, effective radius feature samples, and wheel hub vibration frequency feature samples. The label information corresponding to the second feature samples indicates whether the tire pressure of the wheel is normal, which is usually represented as one or zero, where one indicates normal tire pressure and zero indicates abnormal tire pressure. In addition, the tire pressure detection model can be flexibly set according to the actual situation. For example, it can be a support vector machine, a random forest, etc., but is not limited thereto. In this way, the tire pressure detection process is performed by means of machine learning. Since the method of machine learning has excellent performance, introducing the method of machine learning during tire pressure detection can effectively improve the accuracy of tire pressure detection. In some embodiments, the above tire pressure detection process based on the tire pressure risk coefficient, infrared image, second tire pressure data, effective radius data, and wheel hub vibration frequency data to obtain the tire pressure detection information of the vehicle during driving may include: performing temperature detection according to the infrared image to obtain temperature detection information, and the temperature detection information is used to indicate whether the wheel temperature field of the vehicle is abnormal; if any one of the temperature detection information, tire pressure risk coefficient, second tire pressure data, effective radius data, or wheel hub vibration frequency data meets the first preset condition, then the tire pressure detection information of the vehicle during driving is determined to be used to indicate abnormal vehicle tire pressure, otherwise the tire pressure detection information of the vehicle during driving is determined to be used to indicate normal vehicle tire pressure; where the first preset condition includes: the temperature detection information is used to indicate that the wheel temperature field of the vehicle is abnormal, the tire pressure risk coefficient is higher than the tire pressure risk threshold, the second tire pressure data is outside the tire pressure range, the effective radius data is outside the radius range, and the wheel hub vibration frequency data is outside the first frequency range.

[0047] Here, during tire pressure detection, first, the infrared image is input into a pre-trained temperature detection model, and temperature detection information for indicating whether the wheel temperature field of the vehicle is abnormal is obtained through the temperature detection model. Among them, the temperature detection model is a model trained with multiple infrared image samples and the label information corresponding to each infrared image sample. The label information corresponding to the infrared image sample indicates whether the wheel temperature field is abnormal, which is usually represented as one or zero. One indicates that the wheel temperature field is normal, and zero indicates that the wheel temperature field is abnormal. In addition, the temperature detection model can be flexibly set according to the actual situation. For example, it can be a convolutional neural network, but it is not limited to this. Then, it is determined whether any one of the temperature detection information, the tire pressure risk coefficient, the second tire pressure data, the effective radius data, or the hub vibration frequency data meets the first preset condition. If so, it means that the tire pressure of the vehicle is abnormal at the current moment. At this time, the tire pressure detection information during the vehicle's driving is determined to indicate that the vehicle's tire pressure is abnormal. Otherwise, it means that the tire pressure of the vehicle is normal at the current moment. At this time, the tire pressure detection information during the vehicle's driving is determined to indicate that the vehicle's tire pressure is normal. In this way, the infrared image is converted into feature information that can be used for comparison, and the tire pressure detection in the dynamic stage is realized through a simple comparison operation, which can improve the efficiency of tire pressure detection while ensuring the accuracy of tire pressure detection and meet the real-time requirements. Optionally, the tire pressure risk threshold, the tire pressure range, the radius range, and the first frequency range can all be set according to the actual situation, and specific limitations are not made in this embodiment.

[0048] In some embodiments, referring to Figure 3 , the above-mentioned wheel attribute data includes the fastening torque data, the axial displacement data, and the radial displacement data of the wheel before the vehicle travels, as well as the driving torque data, the wheel speed data, the working sound data, and the hub vibration frequency data of the wheel when the vehicle travels; in the above step S102, according to the wheel image and the wheel attribute data, obtaining the target tire pressure detection information and the target fastening detection information of the vehicle may include the following steps S301-S302: S301, determining the fastening risk coefficient of the vehicle before traveling according to the fastening torque data, the axial displacement data, and the radial displacement data; S302, performing fastening detection processing according to the fastening risk coefficient, the driving torque data, the wheel speed data, the working sound data, and the hub vibration frequency data, and obtaining the fastening detection information during the vehicle's traveling as the target fastening detection information.

[0049] Here, among the above wheel attribute data, the fastening torque data refers to the fastening torque of the wheel nuts and is highly related to the fastening of the wheels. Under normal circumstances, the fastening torque data should be higher than the standard value (for example, usually 80 - 120 N·m for sedans); when the fastening torque data is lower than the standard value, it indicates that the wheel nuts are loose, and at this time, the fastening of the wheels may be insufficient. The axial displacement data refers to the displacement of the wheel in the axial direction and is highly related to the fastening of the wheels. Under normal circumstances, the axial displacement data should be higher than the standard value (for example, higher than 0.5 mm); when the axial displacement data is lower than the standard value, it indicates that the wheel is loose axially, and at this time, the fastening of the wheels may be insufficient. The radial displacement data represents the displacement of the wheel in the radial direction and is highly related to the fastening of the wheels. Under normal circumstances, the radial displacement data should be higher than the standard value (for example, higher than 0.5 mm); when the radial displacement data is lower than the standard value, it indicates that the wheel is loose radially, and at this time, the fastening of the wheels may be insufficient. The driving torque data refers to the driving torque of the wheel and is highly related to the fastening of the wheels. When the driving torque data exceeds the preset range, it indicates that the distribution of the driving torque data is abnormal and the wheel end resistance decreases. At this time, there may be a situation where the wheel is loose and the fastening of the wheel is insufficient. The wheel speed data refers to the rotational speed of the wheel and is highly related to the fastening of the wheels. When the change value of the wheel speed data exceeds the preset range, it indicates that the wheel speed data fluctuates abnormally and the gap between the hub and the brake disc contact surface changes. At this time, there may be a situation where the wheel is loose and the fastening of the wheel is insufficient. The working sound data refers to the frequency data when the wheel is rotating and is highly related to the fastening of the wheels. When the working sound data is abnormal (for example, high-frequency knocking sounds or low-frequency resonance sounds appear), it indicates that there may be a situation where the wheel is loose and the fastening of the wheel is insufficient. The hub vibration frequency data refers to the frequency of the free vibration of the wheel hub without external excitation and is highly related to the fastening of the wheels. Under normal circumstances, the hub vibration frequency data usually concentrates in the low frequency (for example, less than 200 Hz); if the hub vibration frequency data increases, it indicates that the energy of the high-frequency components increases significantly, and there are situations such as bolt knocking and friction resonance. At this time, there may be a situation where the wheel is loose and the fastening of the wheel is insufficient. Therefore, through the fastening torque data, axial displacement data, radial displacement data, driving torque data, wheel speed data, working sound data, and hub vibration frequency data, it is helpful to detect whether the fastening of the vehicle wheels is abnormal.

[0050] Optionally, the fastening torque data is collected by an electronic torque sensor, the axial displacement data and the radial displacement data are collected by a displacement sensor, the driving torque data is collected by a torque sensor, the wheel speed data is collected by a wheel speed sensor, and the working sound data is collected by an acoustic sensor, but it is not limited to this.

[0051] In the related art, usually only the fastening degree of the wheels before the vehicle travels is detected, and there are few methods for detecting the fastening degree of the wheels when the vehicle is traveling. In this regard, this embodiment introduces multi-modal data and adopts a dynamic and static two-stage fastening detection method, aiming to accurately detect the fastening degree of the wheels when the vehicle is traveling.

[0052] In the above step S301, based on the fastening torque data, axial displacement data, and radial displacement data of the wheels before the vehicle travels, a wheel fastening risk prediction is performed to obtain a fastening risk coefficient before the vehicle travels, and this fastening risk coefficient can reflect the probability of insufficient fastening of the wheels of the vehicle before it travels.

[0053] Here, the vehicle is in a stationary state before it travels, and at this time the wheels are in a cold tire state. By using the fastening torque data, axial displacement data, and radial displacement data of the wheels in the cold tire state, the probability of insufficient fastening of the wheels of the vehicle before it travels is determined, thereby realizing the fastening detection of the wheels in the static stage. Text modal data such as fastening torque data, axial displacement data, and radial displacement data can effectively capture the physical characteristics associated with the wheel fastening. This embodiment realizes the wheel fastening risk prediction by introducing multi-dimensional data. In this way, the accuracy of the wheel fastening detection in the static stage can be effectively improved, and potential situations suspected of insufficient wheel fastening can be quickly and accurately discovered, which helps to speed up the efficiency of the wheel fastening detection.

[0054] In some embodiments, the above determining the fastening risk coefficient of the vehicle before it travels according to the fastening torque data, axial displacement data, and radial displacement data may include: extracting features from the fastening torque data, axial displacement data, and radial displacement data to obtain fastening torque features, axial displacement features, and radial displacement features; and determining the fastening risk coefficient of the vehicle before it travels by combining machine learning methods according to the fastening torque features, axial displacement features, and radial displacement features.

[0055] Here, in the wheel tightness detection in the static stage, first, feature extraction is respectively performed on the fastening torque data, axial displacement data, and radial displacement data to obtain the fastening torque feature, axial displacement feature, and radial displacement feature, so that the fastening feature information of the wheel in the cold tire state can be extracted. Among them, the specific implementation of feature extraction can be flexibly set according to the actual situation, and this embodiment does not limit it. Then, the fastening torque feature, axial displacement feature, and radial displacement feature are input into the first fastening risk prediction model to obtain the fastening risk coefficient before the vehicle travels, and the risk coefficient is the prediction probability. Among them, the first fastening risk prediction model is a model trained by multiple third feature samples and the label information corresponding to each third feature sample. The third feature samples include fastening torque feature samples, axial displacement feature samples, and radial displacement feature samples. The label information corresponding to the third feature sample is the fastening risk coefficient, which can be determined by means such as expert annotation and historical data statistics. In addition, the type of the first fastening risk prediction model can be flexibly set according to the actual situation. For example, it can be models such as support vector machines and logistic regression, but it is not limited thereto. In this way, the wheel fastening risk prediction process is performed by the method of machine learning. Since the method of machine learning has excellent performance, introducing the method of machine learning in the wheel fastening risk prediction can effectively improve the accuracy of the wheel fastening risk prediction.

[0056] In some embodiments, determining the fastening risk coefficient of the vehicle before traveling according to the fastening torque data, axial displacement data, and radial displacement data may include: performing wheel fastening risk prediction according to the fastening torque data to obtain a third risk coefficient; performing wheel fastening risk prediction according to the axial displacement data to obtain a fourth risk coefficient; performing wheel fastening risk prediction according to the radial displacement data to obtain a fifth risk coefficient; and determining the fastening risk coefficient of the vehicle before traveling according to the third risk coefficient, the fourth risk coefficient, and the fifth risk coefficient.

[0057] Here, in the wheel tightness detection in the static stage, first, based on the tightening torque data and combined with the pre-trained second tightening risk prediction model, a wheel tightness risk prediction is carried out to obtain the third risk coefficient. Among them, the second tightening risk prediction model is a model trained through multiple tightening torque data samples and the label information corresponding to each tightening torque data sample. At the same time, based on the axial displacement data and combined with the pre-trained third tightening risk prediction model, a wheel tightness risk prediction is carried out to obtain the fourth risk coefficient. Among them, the third tightening risk prediction model is a model trained through multiple axial displacement data samples and the label information corresponding to each axial displacement data sample. And, based on the radial displacement data and combined with the pre-trained fourth tightening risk prediction model, a wheel tightness risk prediction is carried out to obtain the fifth risk data. Among them, the fourth tightening risk prediction model is a model trained through multiple radial displacement data samples and the label information corresponding to each radial displacement data sample. It should be understood that in each of the above tightening risk prediction models, the output risk coefficient is the prediction probability, and the label information corresponding to the data samples during its training is the risk coefficient, which can be determined by means such as expert annotation and historical data statistics. In addition, the types of the above tightening risk prediction models can be flexibly set according to the actual situation. For example, it can be a random forest, logistic regression, etc., but not limited to this. Then, based on the third risk coefficient, the fourth risk coefficient and the fifth risk coefficient, an integration is carried out to obtain the tightening risk coefficient of the vehicle before driving. For example, calculate the mean of the third risk coefficient, the fourth risk coefficient and the fifth risk coefficient as the tightening risk coefficient of the vehicle before driving. Another example is to perform a weighted sum of the third risk coefficient, the fourth risk coefficient and the fifth risk coefficient to obtain the tightening risk coefficient of the vehicle before driving. In this way, the tightening torque data, the axial displacement data and the radial displacement data can respectively reflect the tightening risks in different physical characteristics. Considering that the tightening risks reflected by these three-dimensional data are different, the final tightening risk is determined by integrating the tightening risks corresponding to multiple dimensions. In this way, it helps to further improve the accuracy of the wheel tightness risk prediction.

[0058] In step S302 above, based on the probability of insufficient tightness of the wheels of the vehicle before driving, through multi-dimensional factors such as the driving torque data, wheel speed data, working sound data and hub vibration frequency data of the wheels in the driving state, the target tightening detection information for indicating whether the mechanical structure of the wheels of the vehicle is tight during driving is determined.

[0059] Here, on the one hand, this embodiment fully considers multi-dimensional data highly related to the wheel fastening tightness, and realizes the wheel fastening tightness detection in the dynamic stage through these multi-dimensional data. On the other hand, in the wheel fastening tightness detection in the dynamic stage, this embodiment not only considers the real-time data related to the wheel fastening tightness when the vehicle is driving, but also considers the risk probability of the wheel fastening tightness before the vehicle starts driving. Thus, this embodiment introduces multi-dimensional data in the wheel fastening tightness detection in the dynamic stage, and considers the wheel fastening tightness conditions when the vehicle is stationary and when it is not stationary, thereby effectively improving the comprehensiveness, accuracy and reliability of the wheel fastening tightness detection, and reducing the risk of misjudgment.

[0060] In some embodiments, considering that it is important to detect the wheel fastening tightness of the vehicle at the current moment, the longest cycle time of the above step S302 is thirty seconds, that is, the above step S302 is executed once every thirty seconds or less than every thirty seconds (for example, every twenty seconds), and the result of the new execution overwrites the result of the previous execution, so as to realize the real-time wheel fastening tightness detection.

[0061] In some embodiments, the above-mentioned fastening detection process is performed based on the fastening risk coefficient, driving torque data, wheel speed data, working sound data and hub vibration frequency data to obtain the fastening detection information of the vehicle during driving as the target fastening detection information, which may include: extracting features from the driving torque data, wheel speed data, working sound data and hub vibration frequency data to obtain driving torque features, wheel speed features, working sound features and hub vibration frequency features; according to the fastening risk coefficient, driving torque features, wheel speed features, working sound features and hub vibration frequency features, and combining machine learning methods, the fastening detection information of the vehicle during driving is obtained.

[0062] Here, during the wheel tightness detection, first, feature extraction is performed on the driving torque data, wheel speed data, working sound data, and hub vibration frequency data respectively to obtain the driving torque feature, wheel speed feature, working sound feature, and hub vibration frequency feature. In this way, feature information highly correlated with the wheel tightness during vehicle driving can be extracted. Among them, the specific implementation of feature extraction can be flexibly set according to the actual situation, and this embodiment does not limit it. Then, the tightness risk coefficient, driving torque feature, wheel speed feature, working sound feature, and hub vibration frequency feature are input into a pre-trained tightness detection model to obtain the tightness detection information of the vehicle during driving. Among them, the tightness detection model is a model trained through multiple fourth feature samples and the label information corresponding to each fourth feature sample. The fourth feature sample includes a tightness risk coefficient sample, a driving torque feature sample, a wheel speed feature sample, a working sound feature sample, and a hub vibration frequency feature sample. The label information corresponding to the fourth feature sample indicates whether the wheel tightness is normal, which is usually represented as one or zero. One indicates that the wheel tightness is normal, and zero indicates that the wheel tightness is abnormal. In addition, the tightness detection model can be flexibly set according to the actual situation. For example, it can be a support vector machine, a random forest, etc., but it is not limited to this. In this way, the wheel tightness detection process is performed through machine learning. Since the machine learning method has excellent performance, introducing the machine learning method during wheel tightness detection can effectively improve the accuracy of wheel tightness detection.

[0063] In some embodiments, the above-mentioned tightness detection process based on the tightness risk coefficient, driving torque data, wheel speed data, working sound data, and hub vibration frequency data to obtain the tightness detection information of the vehicle during driving as the target tightness detection information may include: performing sound detection based on the working sound data to obtain sound detection information, and the sound detection information is used to indicate whether the working sound of the wheel is abnormal; if any one of the tightness risk coefficient, driving torque data, wheel speed data, sound detection information, or hub vibration frequency data meets the second preset condition, then the tightness detection information of the vehicle during driving is determined to be used to indicate that the mechanical structure of the wheel is not tightened, otherwise the tightness detection information of the vehicle during driving is determined to be used to indicate that the mechanical structure of the wheel is tightened; where the second preset condition includes: the sound detection information is used to indicate that the working sound of the wheel is abnormal, the tightness risk coefficient is higher than the tightness risk threshold, the driving torque data is outside the driving torque range, the wheel speed data is outside the wheel speed range, and the hub vibration frequency data is outside the second frequency range.

[0064] Here, when detecting the wheel fastening, first, input the working sound data into a pre-trained sound detection model, and obtain sound detection information for indicating whether the working sound of the vehicle's wheels is abnormal through the sound detection model. Among them, the sound detection model is a model trained with multiple working sound samples and the label information corresponding to each working sound sample. The label information corresponding to the working sound sample indicates whether the working sound of the wheel is abnormal, which is usually represented as one or zero. One indicates that the working sound of the wheel is normal, and zero indicates that the working sound of the wheel is abnormal. In addition, the sound detection model can be flexibly set according to the actual situation. For example, it can be a support vector machine, a decision tree, etc., but is not limited thereto. Then, determine whether any one of the fastening risk coefficient, driving torque data, wheel speed data, sound detection information, or hub vibration frequency data meets the second preset condition. If so, it means that the fastening of the wheels of the vehicle at the current moment is insufficient. At this time, determine the fastening detection information during the vehicle's driving as indicating that the mechanical structure of the wheels is not fastened. Otherwise, it means that the fastening of the wheels of the vehicle at the current moment meets the requirements. At this time, determine the fastening detection information during the vehicle's driving as indicating that the mechanical structure of the wheels is fastened. In this way, convert the working sound data into feature information that can be used for comparison, and implement the wheel fastening detection in the dynamic stage through a simple comparison operation, which can improve the efficiency of the wheel fastening detection while ensuring the accuracy of the wheel fastening detection and meet the real-time requirements. Optionally, the fastening risk threshold, driving torque range, wheel speed range, and second frequency range can all be set according to the actual situation, and this embodiment does not make specific limitations on this.

[0065] In some embodiments, referring to Figure 4 , the above-mentioned wheel image may include an infrared image of the wheel when the vehicle is driving; in the above step S103, according to the wheel image, wheel attribute data, and driving attribute data, obtaining the wheel movement trajectory data of the vehicle may include the following steps S401 - S403: S401, perform trajectory prediction processing according to the driving attribute data and wheel attribute data to obtain the initial wheel movement trajectory data of the vehicle; S402, determine the trajectory correction coefficient of the vehicle according to the infrared image; S403, perform correction processing on the initial wheel movement trajectory data according to the trajectory correction coefficient to obtain the wheel movement trajectory data.

[0066] In the above step S401, the driving attribute data and the wheel attribute data are input into a pre-trained first trajectory prediction model to obtain the initial wheel movement trajectory data of the vehicle. The initial wheel movement trajectory data is presented in a line shape, which may include the coordinate positions of the wheels at multiple future moments in the vehicle body coordinate system. Among them, the wheel attribute data input into the first trajectory prediction model may be the second tire pressure data, effective radius data, hub vibration frequency data, driving torque data, wheel speed data, working sound data, and hub vibration frequency data of the wheels when the vehicle is driving. The first trajectory prediction model may be a model trained by a plurality of first data samples and the wheel movement trajectory samples corresponding to each first data sample. The first data samples include driving attribute data samples and wheel attribute data samples. The type of the first trajectory prediction model can be flexibly set according to the actual situation. For example, it can be a support vector machine, logistic regression, etc., but is not limited thereto. Here, the driving attribute data can reflect the real situation of the vehicle when driving, and the wheel attribute data can reflect the real situation of the wheels when the vehicle is driving. The real situation of the vehicle when driving and the real situation of the wheels when the vehicle is driving jointly affect the movement trajectory of the wheels. When it is determined that the vehicle has a potential risk of wheel looseness, in this embodiment, by comprehensively considering these two dimensions of factors for trajectory prediction, the initial wheel movement trajectory data of the vehicle can be obtained, thereby effectively improving the accuracy of wheel movement trajectory prediction.

[0067] Optionally, the driving attribute data can be flexibly set according to the actual situation, and this embodiment does not make specific limitations thereto. For example, the driving attribute data may include vehicle speed data, steering angle data, yaw angular velocity data, lateral acceleration data, longitudinal acceleration data, braking force data, etc., but is not limited thereto.

[0068] In the above step S402, since most of the driving attribute data and the wheel attribute data are data collected by sensors, and the sensors are easily affected by environmental factors, the initial wheel movement trajectory data obtained therefrom may have errors. In this regard, this embodiment determines the trajectory correction coefficient of the vehicle based on the infrared image of the wheels when the vehicle is driving, so as to correct the initial wheel movement trajectory data. Here, since the infrared image of the wheels when the vehicle is driving is data obtained by an infrared camera, it is not easily affected by environmental factors and can accurately reflect the real situation of the wheels. Therefore, determining the trajectory correction coefficient through the infrared image of the wheels when the vehicle is driving helps to accurately correct the initial wheel movement trajectory data.

[0069] In some embodiments, the above determining the trajectory correction coefficient of the vehicle according to the infrared image may include: extracting features from the infrared image to obtain infrared temperature field features; retrieving a preset database according to the infrared temperature field features to obtain the trajectory correction coefficient of the vehicle.

[0070] Here, first, feature extraction is performed on the infrared image to obtain the infrared temperature field feature, so that the feature information characterizing the true condition of the wheel can be extracted. Among them, there is no specific limitation on the specific implementation of feature extraction. In this embodiment, the infrared temperature field feature is manifested as a specific value, which is used to indicate the average temperature in the temperature field of the wheel. Then, the preset database is retrieved according to the infrared temperature field feature. The preset database stores in advance a plurality of infrared temperature field feature samples and the coefficient values corresponding to each infrared temperature field feature sample. Through retrieval, the coefficient value corresponding to the infrared temperature field feature can be obtained and used as the trajectory correction coefficient of the vehicle. Here, the required trajectory correction coefficient can be quickly found through a simple database traversal operation, which can effectively improve the acquisition efficiency of the trajectory correction coefficient.

[0071] In some embodiments, determining the trajectory correction coefficient of the vehicle according to the infrared image as described above may include: performing feature extraction on the infrared image to obtain the infrared temperature field feature; and obtaining the trajectory correction coefficient of the vehicle according to the infrared temperature field feature in combination with the machine learning method.

[0072] Here, first, feature extraction is performed on the infrared image to obtain the infrared temperature field feature, so that the feature information characterizing the true condition of the wheel can be extracted. Among them, there is no specific limitation on the specific implementation of feature extraction. Then, the infrared temperature field feature is input into the coefficient prediction model to obtain the trajectory correction coefficient of the vehicle. Among them, the coefficient prediction model is a machine learning model trained by a plurality of infrared temperature field feature samples and the label information corresponding to each infrared temperature field feature sample, and the label information corresponding to the infrared temperature field feature sample is the trajectory correction coefficient. In addition, the type of the coefficient prediction model can be flexibly set according to the actual situation. For example, it can be a support vector machine, a logistic regression, etc., but it is not limited thereto. Here, the prediction process of the trajectory correction coefficient is performed by the machine learning method. Since the machine learning method has excellent performance, introducing the machine learning method when determining the trajectory correction coefficient can effectively improve the accuracy of the trajectory correction coefficient.

[0073] In the above step S403, after obtaining the trajectory correction coefficient, the initial wheel movement trajectory data is corrected based on it to obtain the wheel movement trajectory data, thereby improving the accuracy of the wheel movement trajectory data. In this way, on the premise of determining that the vehicle has a potential risk of wheel looseness, accurately determining the wheel movement trajectory data of the vehicle helps to quickly and accurately determine whether the wheel is loose.

[0074] In some embodiments, the above-mentioned correction processing of the initial wheel motion trajectory data according to the trajectory correction coefficient to obtain the wheel motion trajectory data may include: obtaining the wheel motion trajectory data by combining the trajectory correction coefficient and the initial wheel motion trajectory data with the method of machine learning.

[0075] Here, by introducing the method of machine learning, the trajectory correction coefficient and the initial wheel motion trajectory data are input into a pre-trained second trajectory prediction model to obtain the wheel motion trajectory data. Among them, the second trajectory prediction model is a model trained by a plurality of second data samples and the wheel motion trajectory samples corresponding to each second data sample. The second data sample may include a wheel motion trajectory sample and a trajectory correction coefficient sample. The type of the second trajectory prediction model can be flexibly set according to the actual situation. For example, it can be a support vector machine, logistic regression, etc., but is not limited thereto. Here, the operation of trajectory correction is performed by the method of machine learning. Since the method of machine learning has excellent performance, introducing the method of machine learning when correcting the trajectory can effectively improve the accuracy of trajectory correction.

[0076] In some embodiments, the above-mentioned correction processing of the initial wheel motion trajectory data according to the trajectory correction coefficient to obtain the wheel motion trajectory data may include: performing a multiplication process on the trajectory correction coefficient and the initial wheel motion trajectory data to obtain the wheel motion trajectory data.

[0077] Here, since the initial wheel motion trajectory data includes the coordinate positions of the wheels at multiple future moments in the vehicle body coordinate system, when correcting the trajectory, for each coordinate position, the trajectory correction coefficient can be multiplied by the coordinate position to obtain the coordinate position with the trajectory correction coefficient as the new coordinate position. By traversing each coordinate position, each new coordinate position can be obtained, thereby obtaining the final wheel motion trajectory data. Here, the operation of trajectory correction is performed by the way of assigning coefficients, which can quickly correct the initial wheel motion trajectory data, accelerate the rate of trajectory correction while ensuring the accuracy of trajectory correction, and meet the real-time requirements.

[0078] In some embodiments, referring to Figure 5 , in the above step S104, based on the wheel motion trajectory data, performing a wheel looseness warning process on the vehicle may include the following steps S501 - S502: S501, performing a deviation detection process according to the wheel motion trajectory data to obtain the deviation detection information of the vehicle; wherein, the deviation detection information is used to indicate whether the wheels deviate in the future time period; S502, performing a wheel looseness warning process on the vehicle according to the deviation detection information.

[0079] In the above step S501, wheel deviation detection processing is performed based on the wheel movement trajectory data to obtain deviation detection information for indicating whether the wheel will deviate in the future time period. Here, based on the movement trajectory data of the wheel in the future time period, it is determined whether the wheel will deviate in the future time period, so that the degree of wheel deviation at the future moment can be accurately determined. At the same time, considering that wheel deviation can directly reflect wheel looseness, determining the degree of deviation from the movement trajectory data can help quickly and accurately determine whether the wheel is loose.

[0080] In some embodiments, the above deviation detection processing based on the wheel movement trajectory data to obtain the deviation detection information of the vehicle may include: inputting the wheel movement trajectory data into a deviation detection model to obtain the deviation detection information.

[0081] Here, the wheel movement trajectory data is input into the deviation detection model, and the deviation detection information is obtained through recognition by the deviation detection model. Among them, the deviation detection model is a model trained by multiple wheel movement trajectory samples and the label information corresponding to each wheel movement trajectory sample. The label information corresponding to the wheel movement trajectory sample indicates whether the wheel deviates, which is usually one or zero. One indicates that the wheel deviates, and zero indicates that the wheel does not deviate. In addition, the type of the deviation detection model can be flexibly set according to the actual situation. For example, it can be a support vector machine, logistic regression, convolutional neural network, etc., but is not limited thereto. In this way, introducing a deviation detection model to realize the automatic detection of deviation, due to the excellent performance of the deviation detection model, the above method can effectively improve the accuracy of deviation detection.

[0082] In the above step S502, when the wheel deviates, it indicates that there are situations such as insufficient dynamic balance and hub displacement of the wheel. For example, insufficient dynamic balance of the wheel will cause the steering wheel to shake and the wheel to deviate. Another example is that the hub displacement of the wheel will cause the wheel to move laterally and the wheel to deviate. At this time, wheel looseness will be caused. In other words, wheel deviation can directly reflect wheel looseness. Therefore, based on the deviation detection information, wheel looseness warning processing is performed on the vehicle.

[0083] In some embodiments, in the above step S501, the deviation detection processing based on the wheel movement trajectory data to obtain the deviation detection information of the vehicle may include: Determine the similarity data between the expected wheel movement trajectory data of the vehicle and the wheel movement trajectory data; If the similarity data is less than or equal to a preset threshold, the deviation detection information is determined to indicate that the wheel will deviate in the future time period, otherwise the deviation detection information is determined to indicate that the wheel will not deviate in the future time period.

[0084] Here, in the deviation detection, first, the expected wheel motion trajectory data of the vehicle is obtained. The expected wheel motion trajectory data includes the expected coordinate positions of the wheels at multiple future moments in the vehicle body coordinate system, which can be pre-stored in a preset database and can be determined by means of expert annotation, historical data statistics, etc. Then, the similarity data between the expected wheel motion trajectory data of the vehicle and the wheel motion trajectory data is calculated, which can effectively measure the difference between the actual motion trajectory of the wheel and the expected motion trajectory. Among them, the type of similarity can be flexibly set according to the actual situation. For example, the similarity can be cosine similarity, Euclidean distance, etc., but it is not limited to this. It should be understood that since both the expected wheel motion trajectory data and the wheel motion trajectory data include multiple coordinate positions, the expected wheel motion trajectory data and the wheel motion trajectory data can actually be represented as a sequence of data, so the similarity between the two can be directly calculated. Next, it is judged whether the similarity data is less than or equal to a preset threshold. If so, it means that the difference between the actual motion trajectory of the wheel and the expected motion trajectory is large, and the wheel is deviated. At this time, the deviation detection information is determined to be used to indicate that the wheel will deviate in the future time period. If not, it means that the difference between the actual motion trajectory of the wheel and the expected motion trajectory is small, and the wheel is not deviated. At this time, the deviation detection information is determined to be used to indicate that the wheel will not deviate in the future time period. It can be seen that in this embodiment, the abstract trajectory difference is converted into a numerical index through similarity calculation, which can accurately measure the difference between the actual motion trajectory of the wheel and the expected motion trajectory. Then, based on the difference between the actual motion trajectory and the expected motion trajectory, it is judged whether the wheel will deviate in the future time period. In this way, the accuracy of deviation detection can be effectively improved.

[0085] In some embodiments, in the above step S502, according to the deviation detection information, performing a wheel looseness warning process on the vehicle may include: If the deviation detection information is used to indicate that the vehicle will deviate in the future time period, then perform a wheel looseness warning process on the vehicle; Or, if the deviation detection information is used to indicate that the vehicle will not deviate in the future time period, then no looseness warning process is performed.

[0086] Here, if the deviation detection information is used to indicate that the vehicle will deviate in the future time period, a wheel loosening warning process is performed on the vehicle. For example, when the body domain controller performs the wheel loosening warning process on the vehicle, it can output a warning display signal to the cockpit domain controller, so that the cockpit domain controller generates a wheel loosening warning icon based on the warning display signal and displays it on the instrument panel to prompt the driver that there is a potential risk of wheel loosening currently. If the deviation detection information is used to indicate that the vehicle will not deviate in the future time period, no loosening warning process is performed. In this way, using the deviation detection information as a benchmark to implement the wheel loosening warning process can further improve the accuracy of the wheel loosening warning.

[0087] In some embodiments, after performing the wheel loosening warning process on the vehicle, return to the above step S101 to achieve cyclic detection until the deviation detection information is used to indicate that the vehicle will not deviate in the future time period, and then the wheel loosening warning stops. Similarly, after not performing the loosening warning process, return to the above step S101 to achieve cyclic detection.

[0088] In some embodiments, the involved models are all integrated in the same artificial intelligence module to form a multi-task learning model.

[0089] In some embodiments, with reference to Figure 6 , the training process of the involved models may include: establishing a training set and a test set; starting to train the model and inputting the training set of the current round into the model; when a round of training ends, using the test set to test the accuracy of the model of the current round to obtain the accuracy of the model of the current round; if the accuracy of the model of the current round is greater than the accuracy threshold, it is determined that the model training is completed, and the model of the current round is output as the trained model, otherwise jump to the next round. Optionally, the accuracy threshold can be flexibly set according to the actual situation. For example, the accuracy threshold can be 98%, but it is not limited thereto.

[0090] To facilitate the understanding of the above wheel loosening warning method of the present application, an actual application scenario of the above wheel loosening warning method of the present application is taken as an example for illustration here.

[0091] In this application scenario, the vehicle is a truck. Due to its use, the truck needs to frequently drive on the highway, and compared with ordinary cars, sports utility vehicles, etc., the probability of the truck having a wheel loosening is higher. With reference to Figure 7 , the method of the embodiment of the present application performs the following loosening warning process on each wheel of the truck: S601, Data source acquisition: Obtain the ground contact image of the wheel before the vehicle travels and the infrared image of the wheel when the vehicle travels as the wheel images of the vehicle. Obtain the first tire pressure data, fastening torque data, axial displacement data, and radial displacement data of the wheel before the vehicle travels, as well as the second tire pressure data, effective radius data, driving torque data, wheel speed data, working sound data, and hub vibration frequency data of the wheel when the vehicle travels as the wheel attribute data of the vehicle. Obtain the vehicle speed data, steering angle data, yaw rate data, lateral acceleration data, longitudinal acceleration data, and braking force data when the vehicle travels as the driving attribute data of the vehicle.

[0092] S602, Tire pressure detection and fastening detection: According to the ground contact image and the first tire pressure data, determine the tire pressure risk coefficient before the vehicle travels. This tire pressure risk coefficient can reflect the probability of abnormal tire pressure of the wheel before the vehicle travels. Subsequently, perform tire pressure detection processing based on the tire pressure risk coefficient, infrared image, second tire pressure data, effective radius data, and hub vibration frequency data to obtain the tire pressure detection information when the vehicle travels as the target tire pressure detection information. The target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal.

[0093] At the same time, according to the fastening torque data, axial displacement data, and radial displacement data, determine the fastening risk coefficient before the vehicle travels. This fastening risk coefficient can reflect the probability of insufficient fastening of the wheel before the vehicle travels. Subsequently, perform fastening detection processing based on the fastening risk coefficient, driving torque data, wheel speed data, working sound data, and hub vibration frequency data to obtain the fastening detection information when the vehicle travels as the target fastening detection information. The target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened.

[0094] S603, Trajectory Detection: If the target tire pressure detection information is used to indicate abnormal tire pressure of the wheel, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened, it indicates that there is a potential risk of wheel looseness in the vehicle. At this time, first, the second tire pressure data, effective radius data, hub vibration frequency data, driving torque data, wheel speed data, working sound data, and hub vibration frequency data of the wheel when the vehicle is running, as well as the vehicle speed data, steering angle data, yaw rate data, lateral acceleration data, longitudinal acceleration data, and braking force data when the vehicle is running, are input into the pre-trained first trajectory prediction model to obtain the initial wheel movement trajectory data of the vehicle. The initial wheel movement trajectory data presents as a line shape, which can include the coordinate positions of the wheel at multiple future moments in the body coordinate system. Then, based on the infrared image of the wheel when the vehicle is running, the trajectory correction coefficient of the vehicle is determined to correct the initial wheel movement trajectory data. Finally, the initial wheel movement trajectory data is corrected based on the trajectory correction coefficient to obtain the wheel movement trajectory data.

[0095] S604, Looseness Warning: First, determine the similarity data between the expected wheel movement trajectory data and the wheel movement trajectory data of the vehicle. Then, if the similarity data is less than or equal to the preset threshold, the deviation detection information is determined to be used to indicate that the wheel will deviate in the future time period, otherwise the deviation detection information is determined to be used to indicate that the wheel will not deviate in the future time period. Finally, if the deviation detection information is used to indicate that the vehicle will deviate in the future time period, the vehicle is subjected to wheel looseness warning processing, otherwise no looseness warning processing is performed.

[0096] Furthermore, referring to Figure 8 , the embodiment of the present application further provides a wheel looseness warning device, which includes: An acquisition module 701, configured to acquire the wheel image, wheel attribute data, and driving attribute data of the vehicle; A first processing module 702, configured to obtain the target tire pressure detection information and the target fastening detection information of the vehicle according to the wheel image and the wheel attribute data; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; A second processing module 703, configured to obtain the wheel movement trajectory data of the vehicle according to the wheel image, the wheel attribute data, and the driving attribute data when the target tire pressure detection information is used to indicate abnormal tire pressure of the wheel, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened; A third processing module 704, configured to perform wheel looseness warning processing on the vehicle based on the wheel movement trajectory data.

[0097] The content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0098] In addition, an embodiment of the present application further provides a computer-readable storage medium, which stores a program executable by a processor. The program executable by the processor is used to implement the above wheel loosening warning method when executed by the processor.

[0099] The content in the above method embodiments is applicable to the medium embodiments of the present application. The functions specifically implemented by the medium embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0100] Finally, with reference to Figure 9 , an embodiment of the present application further provides a vehicle, including: At least one processor 801; At least one memory 802, configured to store at least one program; When the at least one program is executed by the at least one processor 801, the at least one processor 801 implements the above wheel loosening warning method.

[0101] The above vehicle can be a private car, such as a sedan, a sport utility vehicle (SUV), a multi-purpose vehicle (MPV), or a pickup truck, etc., or an operating vehicle, such as a minivan, a bus, a small truck, or a large trailer, etc., and can also be an oil vehicle or a new energy vehicle such as a hybrid or a pure electric vehicle.

[0102] The above memory 802, as a non-transitory network system, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory 802 can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory 802 optionally includes a memory 802 remotely disposed relative to the processor 801, and these remote memories 802 can be connected to the processor 801 through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0103] The above-mentioned memory 802 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 802 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 802, and the processor 801 is used to call and execute the methods of the embodiments of this application.

[0104] The above-mentioned processor 801 can be implemented in the form of a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0105] In some embodiments, the above-mentioned vehicle may further include: An input / output interface for implementing information input and output; A communication interface for implementing communication interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.); A bus for transmitting information between various components of the device (such as the processor 801, the memory 802, the input / output interface, and the communication interface); Among them, the processor 801, the memory 802, the input / output interface, and the communication interface can achieve communication connections with each other inside the device through the bus.

[0106] The content in the above method embodiments is applicable to this vehicle embodiment. The functions specifically implemented by this vehicle embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0107] In some alternative embodiments, the functions / operations recited in the block diagrams may not occur in the order presented in the operational illustrations. For example, depending on the functions / operations involved, two consecutive blocks shown may actually be executed substantially simultaneously or the blocks may sometimes be executed in reverse order. Additionally, the embodiments presented and described in the flowcharts of the present application are provided by way of example for the purpose of providing a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logical flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and in which sub-operations described as part of a larger operation are performed independently.

[0108] Moreover, although the present application has been described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the functions and / or features may be integrated in a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It should also be understood that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present application. Rather, given the attributes, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the modules will be understood within the ordinary skill of an engineer. Thus, those skilled in the art can implement the present application as set forth in the claims without undue experimentation. It should also be understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present application, which is determined by the full scope of the appended claims and their equivalents.

[0109] If the functions are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product stored in a storage medium, including several programs for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0110] The logic and / or steps represented in the flowchart or otherwise described herein can, for example, be considered a sequenced listing of executable instructions for implementing logical functions and can be embodied in any computer-readable medium for use by or in connection with a program execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute the instructions. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport the program for use by or in connection with the program execution system, apparatus, or device.

[0111] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection having one or more wires (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory (EPROM or Flash memory), an optical fiber device, and a portable compact disc read only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or otherwise processing it as appropriate and then storing it in a computer memory.

[0112] It should be understood that the various parts of this application can be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable program execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following technologies well known in the art can be used: discrete logic circuits having logic gates for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0113] In the foregoing description of this specification, the descriptions with reference to the terms "one embodiment / example", "another embodiment / example", or "certain embodiments / examples", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0114] Although the embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present application. The scope of the present application is defined by the claims and their equivalents.

[0115] The above has specifically described the preferred embodiments of the present application, but the present application is not limited to the described embodiments. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the present application, and these equivalent deformations or substitutions are all included within the scope defined by the claims of the present application.

Claims

1. A method for warning of wheel looseness, characterized in that, Including the following steps: Obtain the wheel images, wheel attribute data, and driving attribute data of the vehicle; Based on the wheel images and the wheel attribute data, obtain the target tire pressure detection information and target fastening detection information of the vehicle; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; When the target tire pressure detection information is used to indicate abnormal tire pressure of the wheel, and / or, the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened, based on the wheel images, the wheel attribute data, and the driving attribute data, obtain the wheel movement trajectory data of the vehicle; Based on the wheel movement trajectory data, perform wheel loosening warning processing on the vehicle.

2. The method according to claim 1, wherein The wheel images include the ground contact images of the wheels before the vehicle travels and the infrared images of the wheels when the vehicle travels; the wheel attribute data includes the first tire pressure data of the wheels before the vehicle travels, the second tire pressure data, effective radius data, and hub vibration frequency data of the wheels when the vehicle travels; The obtaining the target tire pressure detection information and target fastening detection information of the vehicle according to the wheel images and the wheel attribute data includes: Based on the ground contact image and the first tire pressure data, determine the tire pressure risk coefficient of the vehicle before traveling; Perform tire pressure detection processing according to the tire pressure risk coefficient, the infrared image, the second tire pressure data, the effective radius data, and the hub vibration frequency data, and obtain the tire pressure detection information of the vehicle when traveling as the target tire pressure detection information.

3. The method according to claim 1, wherein The wheel attribute data includes the fastening torque data, axial displacement data, and radial displacement data of the wheels before the vehicle travels, and the driving torque data, wheel speed data, working sound data, and hub vibration frequency data of the wheels when the vehicle travels; The obtaining the target tire pressure detection information and target fastening detection information of the vehicle according to the wheel images and the wheel attribute data includes: Based on the fastening torque data, the axial displacement data, and the radial displacement data, determine the fastening risk coefficient of the vehicle before traveling; Perform fastening detection processing according to the fastening risk coefficient, the driving torque data, the wheel speed data, the working sound data, and the hub vibration frequency data, and obtain the fastening detection information of the vehicle when traveling as the target fastening detection information.

4. The method according to claim 1, wherein The wheel images include the infrared images of the wheels when the vehicle travels; the obtaining the wheel movement trajectory data of the vehicle according to the wheel images, the wheel attribute data, and the driving attribute data includes: Perform trajectory prediction processing according to the driving attribute data and the wheel attribute data, and obtain the initial wheel movement trajectory data of the vehicle; Based on the infrared image, determine the trajectory correction coefficient of the vehicle; The initial wheel motion trajectory data is corrected according to the trajectory correction coefficient to obtain the wheel motion trajectory data.

5. The method according to claim 1, wherein Based on the wheel motion trajectory data, performing a wheel looseness warning process on the vehicle, including: Performing a deviation detection process according to the wheel motion trajectory data to obtain the deviation detection information of the vehicle; wherein, the deviation detection information is used to indicate whether the wheel will deviate in a future time period; Performing a wheel looseness warning process on the vehicle according to the deviation detection information.

6. The method according to claim 5, wherein The performing a deviation detection process according to the wheel motion trajectory data to obtain the deviation detection information of the vehicle includes: Determining the similarity data between the expected wheel motion trajectory data of the vehicle and the wheel motion trajectory data; If the similarity data is less than or equal to a preset threshold, determining the deviation detection information as being used to indicate that the wheel will deviate in a future time period, otherwise determining the deviation detection information as being used to indicate that the wheel will not deviate in a future time period.

7. The method according to claim 5, characterized in that, The performing a wheel looseness warning process on the vehicle according to the deviation detection information includes: If the deviation detection information is used to indicate that the vehicle will deviate in a future time period, performing a wheel looseness warning process on the vehicle; Or, if the deviation detection information is used to indicate that the vehicle will not deviate in a future time period, no looseness warning process is performed.

8. A wheel loosening warning device, characterized in that, Including: An acquisition module, configured to acquire a wheel image, wheel attribute data, and driving attribute data of a vehicle; A first processing module, configured to obtain target tire pressure detection information and target fastening detection information of the vehicle according to the wheel image and the wheel attribute data; wherein, the target tire pressure detection information is used to indicate whether the tire pressure of the wheel is normal, and the target fastening detection information is used to indicate whether the mechanical structure of the wheel is fastened; A second processing module, configured to obtain the wheel motion trajectory data of the vehicle according to the wheel image, the wheel attribute data, and the driving attribute data when the target tire pressure detection information is used to indicate that the tire pressure of the wheel is abnormal, and / or the target fastening detection information is used to indicate that the mechanical structure of the wheel is not fastened; A third processing module, configured to perform a wheel looseness warning process on the vehicle based on the wheel motion trajectory data.

9. A computer-readable storage medium storing a program executable by a processor, characterized in that, The program executable by the processor, when executed by the processor, is used to implement the wheel looseness warning method according to any one of claims 1-7.

10. A vehicle, characterized in that, Including: At least one processor; At least one memory, configured to store at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the wheel looseness warning method according to any one of claims 1-7.

Images