Abnormality monitoring method for vehicle sentry mode, computer device and storage medium

Abstract

The application relates to the technical field of vehicle sentry, and discloses an abnormality monitoring method for a vehicle sentry mode, a computer device and a storage medium.The method comprises the following steps: acquiring real-time vibration data generated by a vehicle due to external interference, and video stream data of an environment in which the vehicle is located; identifying a target risk event based on the video stream data, and determining a target vibration level corresponding to the real-time vibration data; acquiring a target risk judgment mechanism adapted to the vehicle, wherein the target risk judgment mechanism comprises judgment conditions corresponding to different risk levels, the judgment conditions comprise a risk event and a vibration level, and different judgment conditions differ in at least one of the risk event and the vibration level; and determining whether the vehicle currently has a risk based on the target risk event, the target vibration level and the target risk judgment mechanism.Through the technical scheme, the problem that a risk false positive rate or a risk false negative rate is high in a related art sentry mode is solved, and the risk false positive rate and the risk false negative rate in the sentry mode are reduced.

Description

Technical Field

[0001] This application relates to the field of vehicle sentry technology, specifically to anomaly detection methods, computer equipment, and storage media in vehicle sentry mode. Background Technology

[0002] Sentry mode is a vehicle security monitoring function. When the vehicle is locked and unattended, it uses cameras around the vehicle to continuously monitor the surrounding environment, acting like a "sentinel" to protect the vehicle's safety.

[0003] In related technologies, vehicle sentry mode relies on vision and vibration for independent judgment, typically depending on a fixed set of judgment thresholds or a fusion calculation model with adjustable parameters. This method has significant limitations when facing complex and ever-changing real-world scenarios, making it difficult to balance sensitivity and accuracy, resulting in a high false alarm or false negative rate for sentry mode.

[0004] In other words, the sentinel mode in related technologies suffers from a high false alarm rate or false negative rate. Summary of the Invention

[0005] This application provides an anomaly detection method, computer device, and storage medium for vehicle sentry mode, in order to solve the problem of high false alarm rate or false negative rate in sentry mode in related technologies.

[0006] Firstly, this application provides an anomaly detection method for vehicle sentry mode, the method comprising:

[0007] Acquire real-time vibration data of the vehicle caused by external interference, as well as video stream data of the vehicle's environment; Identify target risk events based on video stream data and determine the target vibration level corresponding to real-time vibration data; Obtain the target risk assessment mechanism for vehicle adaptation. The target risk assessment mechanism includes assessment conditions corresponding to different risk levels. The assessment conditions include risk events and vibration levels. Different assessment conditions include at least one different risk event and vibration level. Based on the target risk event, target vibration level, and target risk assessment mechanism, determine whether the vehicle currently faces any abnormal risks.

[0008] According to the anomaly monitoring method for vehicle sentry mode provided in this application, real-time vibration data of the vehicle caused by external interference and video stream data of the vehicle's environment are acquired, providing heterogeneous and complementary multi-dimensional information for subsequent fusion and judgment. Based on the video stream data, target risk events are identified, and the target vibration level corresponding to the real-time vibration data is determined, realizing the transformation of raw sensor data into structured and quantifiable judgment elements. A vehicle-adapted target risk judgment mechanism is acquired, which includes judgment conditions corresponding to different risk levels. The judgment conditions include risk events and vibration levels, and different judgment conditions include at least one different risk event and vibration level, enabling the system to call the most matching judgment logic according to the scenario or user preference, possessing scenario adaptability. Based on the target risk event, target vibration level, and target risk judgment mechanism, it is determined whether the vehicle currently has an abnormal risk. Through precise logical rule matching, accurate identification and classification of complex risk scenarios are achieved. This reduces the false alarm rate and false negative rate of risk in sentry mode.

[0009] In one optional implementation, the target risk assessment mechanism for vehicle adaptation includes: If the user associated with the vehicle has selected the corresponding preset risk assessment mechanism, then the user's risk assessment mechanism will be used as the risk assessment mechanism adapted to the vehicle; or, if the user has not selected the corresponding preset risk assessment mechanism, then the preset risk assessment mechanism that matches the environmental conditions of the vehicle's environment among multiple preset risk assessment mechanisms will be used as the target risk assessment mechanism.

[0010] This implementation provides a flexible risk assessment mechanism acquisition strategy. By prioritizing user-defined choices, it meets the personalized needs of different users regarding risk sensitivity, thus enhancing the user experience. When the user does not actively select a mechanism, the system can intelligently match the most suitable preset mechanism based on the vehicle's environment, achieving scenario-adaptive risk assessment strategy. This "personalization-first, intelligent backup" mechanism ensures that the system maintains optimal assessment balance in various usage scenarios, making protective measures more aligned with actual needs.

[0011] In one alternative implementation, identifying target risk events based on video stream data includes: The target risk event is obtained by identifying video stream data based on the risk event identification model.

[0012] This implementation method enables the identification of risk events.

[0013] In one optional implementation, determining the target vibration level corresponding to the real-time vibration data includes: The target vibration level is determined based on the range of vibration amplitude corresponding to real-time vibration data.

[0014] In one optional implementation, determining the target vibration level based on the range of vibration amplitudes corresponding to real-time vibration data includes: If the vibration amplitude corresponding to the real-time vibration data is within the first preset vibration range, then the target vibration level is set to the first level. If the vibration amplitude is within the second preset vibration range, the target vibration level is set to the second level, wherein the minimum value of the first preset vibration range is greater than the maximum value of the second preset vibration range, and the first level is higher than the second level. If the vibration amplitude is within the third preset vibration range, the target vibration level is set to the third level, where the minimum value of the second preset vibration range is greater than the maximum value of the third preset vibration range, and the second level is higher than the third level.

[0015] This implementation discretizes continuous vibration amplitude data into clear and ordered level signals (first level, second level, and third level) by defining strictly hierarchical and non-overlapping preset vibration ranges. This quantitative grading method transforms complex physical signals into standardized inputs that the system can directly use for logical judgment. This ensures that the output of vibration sensing is no longer a vague "strong" or "weak," but a level label that can be precisely matched with visual events, laying a solid foundation for subsequent multi-condition joint judgment and ensuring the objectivity and consistency of vibration dimension judgment.

[0016] In one alternative implementation, the target risk event includes at least one of the following: kicking or scratching the vehicle, leaning against or peeping at the vehicle, the vehicle door being opened, suspicious target intrusion, and suspicious target loitering.

[0017] In one optional implementation, the determination of whether a vehicle poses an abnormal risk is based on the target risk event, the target vibration level, and the target risk assessment mechanism, including: Match the target risk event with the risk events in the different judgment conditions included in the target risk judgment mechanism, and determine the first judgment condition for matching the target risk event from the different judgment conditions; and match the target vibration level with the vibration level in the different judgment conditions included in the target risk judgment mechanism, and determine the second judgment condition corresponding to the target vibration level from the different judgment conditions. If the first and second determination conditions are the same, the vehicle is determined to have a risk, and the risk level corresponding to the first or second determination condition is taken as the target risk level of the vehicle; or, if the first and second determination conditions are different, the vehicle is determined not to have any abnormal risk.

[0018] This implementation method defines the core logic of risk assessment: "The event and the vibration must point to the same risk condition." This "AND" logic is crucial; it requires that visual threats and physical impact evidence simultaneously meet the same set of high-level judgment criteria before the system will ultimately confirm the existence of a risk. This greatly enhances the system's anti-interference capability and the rigor of its judgment. For example, detecting only someone approaching (the event) without corresponding vibration, or only slight vibration (such as a truck passing by on the roadside) without threatening behavior, will not trigger a false judgment. This logic effectively eliminates sporadic interference signals from a single dimension, ensuring high confidence in alarms and significantly improving the system's reliability.

[0019] In one alternative implementation, the method further includes: If an abnormal risk is identified in a vehicle, a corresponding target control strategy is matched according to the vehicle's target risk level. According to the target control strategy, the corresponding components of the vehicle perform defensive response operations to obtain the risk disposal results; Security alerts will be issued to users based on the results of risk management.

[0020] This implementation method achieves a closed loop from "risk perception" to "proactive response." Its beneficial effects are that the system goes beyond the judgment stage, executing differentiated and tiered proactive defense measures (target control strategies) based on different risk levels (target risk levels), such as honking the horn, flashing lights, locking doors, and uploading video. Subsequently, the results are simultaneously communicated to the user. This forms a complete security chain of "perception-judgment-response-feedback," transforming the vehicle from a passive monitoring object into an intelligent entity with proactive deterrence and protection capabilities. This significantly improves actual security effectiveness and promptly informs vehicle owners of the situation, enhancing their sense of security and trust in the system.

[0021] In an optional implementation, when the target risk event is determined by a target identification model and the target vibration level is determined based on a preset vibration range, the method further includes: In response to receiving feedback instructions from users after issuing a security alert, and if the feedback instructions indicate a false alarm, the video stream data and real-time vibration data are used as negative samples to adjust the risk event identification model and the preset vibration range, respectively.

[0022] This implementation introduces a crucial "self-evolution" mechanism. By feeding back user-confirmed false alarm data as negative samples to the system, both the visual recognition model and the vibration judgment threshold can be iteratively optimized simultaneously. For the model, this is equivalent to adding "counterexample" training data, helping it better distinguish between threatening and non-threatening scenarios. For the vibration range, the threshold can be calibrated to better adapt to the specific environment in which the vehicle is parked. This closed-loop learning process enables the system to continuously improve during use, constantly reducing the false alarm rate, adapting to different users and environments, and achieving intelligent improvements that become more accurate with use, ensuring the long-term accuracy and user satisfaction of the system.

[0023] Secondly, this application provides a computer device, the device comprising: a controller, the controller including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform an anomaly detection method such as the vehicle sentry mode corresponding to the first aspect and any embodiment thereof.

[0024] Thirdly, this application provides a computer-readable storage medium storing computer instructions for causing a computer to execute an anomaly detection method in vehicle sentinel mode as described in the first aspect and any of its embodiments.

[0025] Fourthly, this application provides a computer program product, including computer instructions for causing a computer to execute an anomaly detection method in vehicle sentry mode as described in the first aspect and any of its embodiments. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram illustrating an application scenario according to an embodiment of this application; Figure 2 This is a flowchart of an anomaly detection method for vehicle sentry mode according to an embodiment of this application; Figure 3 This is a flowchart of another anomaly detection method for vehicle sentry mode according to an embodiment of this application; Figure 4 This is a schematic diagram of a risk event identification process according to an embodiment of this application; Figure 5This is a schematic diagram of a high-sensitivity, high-risk alarm strategy according to an embodiment of this application; Figure 6 This is a schematic diagram of a high-sensitivity, low-risk alarm strategy according to an embodiment of this application; Figure 7 This is a schematic diagram of a medium-sensitivity high-risk alarm strategy according to an embodiment of this application; Figure 8 This is a schematic diagram of a medium-sensitivity, low-risk alarm strategy according to an embodiment of this application; Figure 9 This is a schematic diagram of a low-sensitivity, high-risk alarm strategy according to an embodiment of this application; Figure 10 This is a schematic diagram of a low-sensitivity, low-risk alarm strategy according to an embodiment of this application; Figure 11 This is a structural block diagram of an anomaly detection device in vehicle sentry mode according to an embodiment of this application; Figure 12 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] It is understood that before using the technical solutions disclosed in the various embodiments of this application, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this application in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0030] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0031] As one optional application scenario in the embodiments of this application, such as Figure 1 As shown, the anomaly detection system for the vehicle's sentry mode may include at least one terminal device and at least one server. Figure 1The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.

[0032] Specifically, the terminal device can be an in-vehicle terminal or a vehicle's domain controller. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranets, local area networks, wide area networks, mobile communication networks, and combinations thereof.

[0033] It should be noted that, Figure 1 This is merely an example of an application scenario and does not limit the scope of protection of this application.

[0034] The embodiments of this application will be described below with reference to the accompanying drawings. It should be understood that the pages shown in the drawings are merely examples, and various page designs are possible in practice. The various graphic elements on the page may have different arrangements and different visual representations, one or more elements may be omitted or replaced, and one or more other elements may also be present; no limitations are imposed on the embodiments of this application. Furthermore, the embodiments are primarily described below with reference to terminal device 110. It should be understood that the actions described relative to terminal device 110 can be performed by application 101 on terminal device 110, or can be performed by application 101 in conjunction with its server (e.g., server 120).

[0035] Sentry mode is a vehicle security monitoring function. When the vehicle is locked and unattended, it uses cameras and vibration sensors around the vehicle to continuously monitor the surrounding environment, acting like a "sentinel" to protect the vehicle's safety.

[0036] Sentinel mode defines two risk levels based on visual detection and vibration detection: Low risk: When a person or vehicle is detected approaching by the cameras of the Around View Monitor (AVM) system and lingers for more than 10 seconds, the event resulting from this detection method is considered a low-risk event.

[0037] High-risk: This method involves acquiring the linear acceleration values ​​along the X, Y, and Z axes from vehicle vibration sensors and then using a vibration detection algorithm to determine if the vehicle is experiencing abnormal vibrations. Events detected using this method are considered high-risk.

[0038] Visual inspection and vibration inspection correspond one-to-one with the two risk states, and the two inspection methods are independent of each other.

[0039] In related technologies, visual inspection or vibration detection presents a significant challenge. Even visually detected risks can potentially cause substantial damage to the vehicle body, such as sliding a knife across the car or damaging tires. These actions may be minor and not necessarily cause significant vibrations, and the detection results from these technologies may indicate no high-risk event. However, the actual damage to the vehicle body can be substantial. Failure to promptly notify users and elicit their attention could lead to market or customer complaints.

[0040] Similarly, when a vehicle is parked on the roadside or in a noisy environment such as a construction site, the vehicle body may experience slight vibrations. However, the actual damage to the vehicle is not significant or nonexistent. In the traditional way, the user would be notified of a high-risk event, which would be inconvenient for the user and may also lead to complaints.

[0041] To address the mismatch between detection results and actual risk levels, a new anomaly detection method for vehicle sentry mode is proposed, which integrates visual and vibration information for sentry mode monitoring. This approach no longer defines high or low risk solely based on visual and vibration detection. Instead, it defines high or low risk events based on the user's specific behavior, specifically whether it actually causes damage to the vehicle or involves suspicious actions.

[0042] According to an embodiment of this application, an anomaly detection method for vehicle sentry mode is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0043] This embodiment provides an anomaly detection method for vehicle sentry mode, which can be used in the aforementioned vehicle terminal or vehicle domain controller. Figure 2 This is a flowchart of an anomaly detection method for vehicle sentry mode according to an embodiment of this application, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain real-time vibration data of the vehicle caused by external interference, as well as video stream data of the vehicle's environment.

[0044] Specifically, external interference refers to physical actions outside the vehicle that cause changes in the vehicle's body state, including but not limited to human actions such as patting, scratching, or leaning against the vehicle body; environmental factors such as strong winds, falling objects, or vibrations from nearby construction; and contact with the vehicle by other moving objects, such as collisions from adjacent vehicle doors or bicycle scrapes. Real-time vibration data refers to data collected in real time by onboard vibration sensors that reflects the vehicle's movement caused by external interference, such as acceleration information collected by a three-axis accelerometer. Video stream data refers to a sequence of images continuously collected by one or more cameras deployed around the vehicle, reflecting the dynamics of the vehicle's surrounding environment. This data is used to capture people, vehicles, objects, and their behaviors within the visual range.

[0045] By simultaneously acquiring video stream data and real-time vibration data through image sensors and vibration sensors deployed in the vehicle, raw data from visual and vibration sensors is provided for risk assessment.

[0046] Step S202: Identify the target risk event based on the video stream data and determine the target vibration level corresponding to the real-time vibration data.

[0047] Specifically, a target risk event refers to a specific behavior or state that may pose a safety threat to a vehicle, identified through intelligent analysis of video stream data. This includes at least one of the following: kicking or scratching the vehicle, leaning against or peeping at the vehicle, the vehicle door being opened, suspicious target intrusion, or suspicious target loitering. Examples include direct acts of damage to the vehicle body (such as kicking or scratching), abnormal approach and peeping behavior (such as leaning against or loitering for an extended period), and the vehicle being opened abnormally. The target vibration level refers to the result of a quantitative classification of the intensity of real-time vibration data. For example, by comparing the vibration amplitude with a preset threshold, it can be divided into different levels such as high, medium, and low, corresponding to strong, moderate, and slight vibration intensities, respectively.

[0048] This step performs parallel processing and feature extraction on the synchronously acquired video stream data and real-time vibration data. On one hand, visual analysis maps the continuous video stream into discrete, semantically clear risk events; on the other hand, vibration analysis quantifies the physical signals into graded vibration intensities. These two aspects of processing transform the multi-source heterogeneous raw visual and vibration data into structured judgment elements that can be directly processed by subsequent rules, laying the foundation for accurate risk judgment based on multi-dimensional information fusion.

[0049] Step S203: Obtain the target risk assessment mechanism for vehicle adaptation. The target risk assessment mechanism includes assessment conditions corresponding to different risk levels. The assessment conditions include risk events and vibration levels. Different assessment conditions include at least one different risk event and vibration level.

[0050] Specifically, the target risk assessment mechanism refers to a pre-defined set of core logical rules for comprehensively assessing risks, including judgment conditions corresponding to different risk levels. These judgment conditions are based on risk events and vibration levels, and at least one of these factors must differ between different judgment conditions. The target risk assessment mechanism defines how, under a specific pattern, the identified risk events and vibration levels are combined and correlated, and mapped to specific risk conclusions.

[0051] As an example, one risk assessment mechanism for "high-sensitivity monitoring" might be set to determine high risk when a "suspicious target loitering" event is detected accompanied by "low vibration level." Another mechanism for "routine parking" might require both a "vehicle door being opened" event and "high vibration level" to occur simultaneously for a high-risk assessment. Users can choose to enable different risk assessment mechanisms based on their parking environment (e.g., underground garage, roadside parking) or personal preferences, and the system will adapt the assessment logic to the vehicle.

[0052] Load the pre-stored risk assessment mechanism for analyzing vehicle risks from memory. Alternatively, download the vehicle-specific risk assessment mechanism from the server.

[0053] Step S204: Based on the target risk event, target vibration level, and target risk determination mechanism, determine whether the vehicle has any abnormal risks.

[0054] Specifically, the target risk event and the target vibration level are combined as a group, and this combination is compared and logically matched one by one with the predefined judgment conditions in the target risk judgment mechanism obtained in the above steps.

[0055] As an example, the system will determine whether the current combination fully meets a certain condition, such as "if the risk event is 'suspicious target loitering' and 'vibration level is low,' then a high risk is determined to exist." If the condition is met, the system will output the corresponding risk conclusion, i.e., the existence of risk and its level; if no condition is met, the system will output a conclusion of no risk or the default level.

[0056] The anomaly detection method for vehicle sentry mode provided in this embodiment acquires real-time vibration data of the vehicle caused by external interference, as well as video stream data of the vehicle's environment, providing heterogeneous and complementary multi-dimensional information for subsequent fusion and judgment. Based on the video stream data, it identifies target risk events and determines the target vibration level corresponding to the real-time vibration data, transforming raw sensor data into structured and quantifiable judgment elements. It acquires a vehicle-adapted target risk judgment mechanism, which includes judgment conditions corresponding to different risk levels. These judgment conditions include risk events and vibration levels, with different judgment conditions including at least one different risk event and vibration level. This allows the system to call the most suitable judgment logic based on the scenario or user preference, possessing scenario-adaptive capabilities. Based on the target risk event, target vibration level, and target risk judgment mechanism, it determines whether the vehicle currently faces an abnormal risk. Through precise logical rule matching, it achieves accurate identification and classification of complex risk scenarios, reducing the false alarm rate and false negative rate in sentry mode.

[0057] This embodiment provides an anomaly detection method for vehicle sentry mode, which can be used in the aforementioned vehicle terminal or vehicle domain controller. Figure 3 This is a flowchart of another anomaly detection method for vehicle sentry mode according to an embodiment of this application, such as... Figure 3 As shown, the process includes the following steps: Step S301: Obtain real-time vibration data of the vehicle caused by external interference, as well as video stream data of the vehicle's environment.

[0058] Please see details Figure 2 Step S201 of the illustrated embodiment will not be described again here.

[0059] Step S302: Identify the target risk event based on the video stream data and determine the target vibration level corresponding to the real-time vibration data.

[0060] Specifically, step S302 includes: Step S3021: Identify video stream data based on the risk event identification model to obtain the target risk event.

[0061] Specifically, a risk event identification model refers to a trained machine learning model (such as a convolutional neural network) that can analyze video image sequences and identify specific behavioral patterns related to vehicle safety.

[0062] A pre-trained risk event recognition model is deployed and run in the vehicle's central domain controller to analyze real-time input video stream data. For model input, the model receives continuous video frames or segments from the vehicle's surround-view cameras. For model processing, the model performs multi-label classification analysis on each frame or segment of video based on its internally learned features (such as human posture, motion trajectory, object interaction, dwell time, etc.). It does not simply determine "whether there is risk," but rather identifies the probability of each of several pre-defined risk behavior categories. For the model's output, within the current analysis time window, the model outputs a list of recognition results. Each item in the list corresponds to a candidate risk event it has identified, and each event is accompanied by a confidence score. This indicates which possible risk behaviors the model has identified and how confident it is in each identification result.

[0063] As an example, suppose a vehicle is parked on the side of the road, and its camera captures the following scene: a person approaches the vehicle, stays by its door for about 8 seconds, and touches the door with their finger.

[0064] The risk event identification model analyzes this video. The output is as follows: Candidate risk event 1: Suspicious target loitering, confidence level: 92%.

[0065] Candidate risk event 2: Leaning against or peeping at a vehicle, confidence level: 85%.

[0066] Candidate risk event 3: kicking or scratching the vehicle, confidence level: 15%.

[0067] Candidate risk event 4: The vehicle door is opened, confidence level: 5%.

[0068] This step transforms the unstructured video pixel stream into structured, machine-readable semantic information, namely event categories, providing direct input information for subsequent rule-based fusion judgments.

[0069] In one possible implementation, Figure 4 This is a schematic diagram illustrating a risk event identification process according to an embodiment of this application. For example... Figure 4 As shown, this process begins with the activation of Sentinel Mode. First, Sentinel Mode, running in the vehicle's Domain Control Unit (DCU), transmits configuration parameters to the target risk assessment mechanism and begins detection. Subsequently, the vehicle acquires real-time video streams from the AVM camera via the camera module. The risk event recognition algorithm analyzes the video streams to identify target risk events. Finally, the algorithm returns the recognition results to Sentinel Mode for subsequent risk assessment and decision-making.

[0070] Step S3022: Determine the target vibration level based on the range of vibration amplitude corresponding to the real-time vibration data.

[0071] Specifically, real-time vibration data is compared with multiple preset vibration amplitude ranges. The vibration level corresponding to the vibration amplitude range in which the real-time vibration data falls is used as the target vibration level. This achieves the classification of vibration data, facilitating the matching of judgment conditions with the target determination mechanism.

[0072] Step S303: Obtain the target risk assessment mechanism for vehicle adaptation. The target risk assessment mechanism includes assessment conditions corresponding to different risk levels. The assessment conditions include risk events and vibration levels. Different assessment conditions include at least one different risk event and vibration level.

[0073] Step S304: Based on the target risk event, target vibration level, and target risk determination mechanism, determine whether there is an abnormal risk currently present in the vehicle.

[0074] Specifically, step S304 above includes the following steps: Step S3041: Match the target risk event with the risk events in the different judgment conditions included in the target risk judgment mechanism, and determine the first judgment condition for matching the target risk event from the different judgment conditions; and match the target vibration level with the vibration level in the different judgment conditions included in the target risk judgment mechanism, and determine the second judgment condition corresponding to the target vibration level from the different judgment conditions.

[0075] Specifically, risk event matching refers to comparing the identified target risk event with the risk events defined in the judgment conditions to check whether their event types are consistent. Vibration level matching refers to comparing the determined target vibration level with the vibration level defined in the judgment conditions to check whether their levels are consistent. The first judgment condition is the judgment condition for the risk event that the target risk event matches among the various judgment conditions of the target risk judgment mechanism. The second judgment condition is the judgment condition for the vibration level that the target vibration level matches among the various judgment conditions of the target risk judgment mechanism.

[0076] By matching target risk events and vibration levels with different judgment conditions included in the target risk assessment mechanism, the complex composite condition judgment is decomposed into independent individual conditions for comparison. The system can quickly filter out which rules meet the prerequisites in terms of event type and vibration intensity. This decomposition not only reduces the complexity of subsequent logical processing and prepares for the comprehensive judgment of whether they correspond to the same rule in subsequent steps, but also makes the matching process more modular and efficient.

[0077] Step S3042: If the first judgment condition and the second judgment condition are the same, then it is determined that the vehicle has a risk, and the risk level corresponding to the first judgment condition or the second judgment condition is taken as the target risk level of the vehicle; or, if the first judgment condition and the second judgment condition are different, then it is determined that the vehicle does not have an abnormal risk.

[0078] Specifically, the target risk level refers to the risk level corresponding to when both the vehicle's vibration level and the risk event meet the same judgment criteria.

[0079] If the first and second judgment conditions are the same, meaning the same judgment condition is simultaneously met by the target risk event and the target vibration level, then the judgment condition is deemed valid, thus confirming that the vehicle possesses the risk corresponding to that condition (i.e., the risk level defined by that condition). Conversely, if the first and second judgment conditions are different, it indicates that the currently collected "event-vibration" combination does not meet any preset, complete risk judgment rule, and therefore the vehicle is judged to have no abnormal risk.

[0080] This implementation method achieves accurate risk assessment with scene adaptation, significantly reducing false alarm and false negative rates. By comparing the first and second judgment conditions, this method fundamentally changes the traditional approach of fixed thresholds or single models. The system can invoke the judgment logic most suitable for the current scene (such as "high sensitivity" or "normal mode") based on user selection or environmental perception. Subsequently, by strictly matching the specific combination of "target risk event" and "target vibration level" with preset rules, an abnormal risk is only determined when both conditions simultaneously meet a complete combination condition. This enables the system to intelligently distinguish between complex scenarios such as "malicious minor damage" (e.g., scratching a car with low vibration) and "non-malicious strong interference" (e.g., construction with high vibration), thereby significantly reducing false alarms and false negatives caused by the inability of traditional methods to address both scenarios at the source.

[0081] Furthermore, this implementation provides a strategy configuration method that combines "user-initiated selection" with "system intelligent recommendation." Users can directly select the desired mode based on experience, giving them a strong sense of control; when there are no explicit user instructions, the system can automatically select an appropriate strategy based on environmental information, achieving intelligent assistance. This design enables complex risk assessment systems to work flexibly in a personalized manner, satisfying the customized needs of professional users while ensuring user-friendliness for ordinary users, fundamentally improving the user experience of Sentinel Mode.

[0082] In one possible implementation of this embodiment, Figure 5 This is a schematic diagram of a high-sensitivity, high-risk alarm strategy according to an embodiment of this application. Figure 5As shown, this strategy assesses risk level by monitoring the target vibration level (any one of high, medium, or low in this example) and identifying target risk events (risk events 1, 2, 3, and 4 in this example). Under high sensitivity conditions, when the criteria for a first-level vibration or a second-level vibration are met, or when risk event 1 is detected, the system will trigger a high-risk alarm. Similarly, when any of risk events 2, 3, or 4 are detected, and the criteria for a third-level vibration are also met, the target risk assessment mechanism detects an abnormal risk to the vehicle, and the system will trigger a high-risk alarm.

[0083] Risk event 1 includes at least one of the following: kicking the car, scratching the car, hitting the vehicle, damaging the tires, pulling the door handle, door-opening malfunction, or vehicle collision. Risk event 2 includes at least one of leaning against the vehicle or peeping at the vehicle. Risk event 3 is opening a vehicle door. Risk event 4 is intrusion by a suspicious target, which includes people, non-motorized vehicles, and motorized vehicles. Risk event 5 is loitering by a suspicious target; the higher the sensitivity, the shorter the loitering time of the suspicious target.

[0084] In one possible implementation of this embodiment, Figure 6 This is a schematic diagram of a high-sensitivity, low-risk alarm strategy according to an embodiment of this application. Figure 6 As shown, this strategy assesses the risk level by detecting the target vibration level and identifying target risk events (risk events 2, 3, and 5 in this example). Under high sensitivity conditions, when the third-level vibration is met or at least one of the criteria for risk event 2, risk event 3, or risk event 5 is detected, the target risk assessment mechanism detects an abnormal risk to the vehicle and triggers a low-risk alarm state.

[0085] In one possible implementation of this embodiment, Figure 7 This is a schematic diagram of a medium-sensitivity high-risk alarm strategy according to an embodiment of this application. Figure 7 As shown, this strategy assesses the risk level by detecting the target vibration level and identifying the target risk event (risk event 1 in this example). Under medium sensitivity conditions, when the criteria for determining first-level vibration or second-level vibration are met, or when the criteria for risk event 1 are detected, the target risk determination mechanism detects an abnormal risk to the vehicle and triggers a high-risk alarm state.

[0086] In one possible implementation of this embodiment, Figure 8 This is a schematic diagram of a medium-sensitivity, low-risk alarm strategy according to an embodiment of this application. Figure 8As shown, this strategy assesses risk level by detecting the target vibration level and identifying target risk events (risk events 2, 3, 4, and 5 in this example). Under medium sensitivity conditions, when the criteria for risk event 2 or risk event 5 are met, the target risk assessment mechanism detects an abnormal risk to the vehicle, and the system will trigger a low-risk alarm state. Alternatively, under medium sensitivity conditions, when the criteria for risk event 3 or risk event 4 and level 3 vibration are met, the target risk assessment mechanism detects an abnormal risk to the vehicle, and the system will trigger a low-risk alarm state.

[0087] In one possible implementation of this embodiment, Figure 9 This is a schematic diagram of a low-sensitivity, high-risk alarm strategy according to an embodiment of this application. Figure 9 As shown, this strategy assesses the risk level by detecting the target vibration level and identifying the target risk event (risk event 1 in this example). Under low sensitivity conditions, when the judgment conditions for the first or second level vibration are met, or when the judgment conditions for risk event 1 are met, the target risk judgment mechanism detects an abnormal risk to the vehicle, and the system will trigger a high-risk alarm state.

[0088] In one possible implementation of this embodiment, Figure 10 This is a schematic diagram of a low-sensitivity, low-risk alarm strategy according to an embodiment of this application. Figure 10 As shown, this strategy assesses risk level by identifying target risk events (risk event 2 or risk event 5 in this example). Under low sensitivity conditions, the system will trigger a low-risk alarm state when the judgment conditions for risk event 2 or risk event 5 are met.

[0089] In some optional implementations, step S3022 above includes the following steps: Step a1: If the vibration amplitude corresponding to the real-time vibration data is within the first preset vibration range, then set the target vibration level to the first level.

[0090] Specifically, the first preset vibration range refers to the preset threshold range corresponding to the highest intensity vibration. Its lower limit is usually set at a level that can clearly characterize a vehicle suffering a severe physical impact (such as a violent collision or a violent impact). The first level refers to the vibration intensity level corresponding to this range, which can be called the "high vibration level".

[0091] This step identifies and quantifies severe external impacts by comparing the real-time collected vibration amplitude with a preset maximum intensity threshold range. When the vibration amplitude falls within this range, the system explicitly classifies it as "Level 1" (high vibration). This step clearly marks the most intense interference events at the physical signal level, providing crucial and clear input for subsequent judgment rules (e.g., high vibration directly triggers high risk regardless of visual events), ensuring the system's ability to capture major physical damage events and effectively preventing missed detections.

[0092] Step a2: If the vibration amplitude is within the second preset vibration range, then set the target vibration level to the second level, wherein the minimum value of the first preset vibration range is greater than the maximum value of the second preset vibration range, and the first level is higher than the second level.

[0093] Specifically, the second preset vibration range refers to a preset threshold range corresponding to moderate intensity vibration, used to characterize obvious but not extreme physical actions (such as slamming a door or a moderate impact). The second level refers to the vibration intensity level corresponding to this range, which can be called the "moderate vibration level".

[0094] This step enables fine-grained differentiation of moderate-intensity vibrations. By comparing the vibrations to a threshold range between high and low intensity, the system can classify such events as "Level 2" (moderate vibration). This Level 2 classification is crucial, encompassing numerous scenarios requiring comprehensive judgment incorporating visual information (e.g., potential malicious damage versus unintentional, heavy contact). This step provides critical intermediate input for subsequent fusion judgments, allowing for more refined and accurate risk assessment and avoiding arbitrary "either high or low" treatment of moderate-intensity vibrations.

[0095] Step a3: If the vibration amplitude is within the third preset vibration range, then set the target vibration level to the third level, wherein the minimum value of the second preset vibration range is greater than the maximum value of the third preset vibration range, and the second level is higher than the third level.

[0096] Specifically, the third preset vibration range refers to the preset threshold range corresponding to the lowest intensity vibration. Its upper limit is usually set above the level that can effectively filter environmental background noise (such as slight road vibration transmitted by distant vehicles, wind noise), and is used to sense slight physical contact or disturbance. The third level refers to the vibration intensity level corresponding to this range, which can be called the "low vibration level".

[0097] As an example, a standard vibration threshold is set as a reference value for the preset vibration range, and the preset vibration range corresponding to different vibration levels is determined as follows: The first level of high vibration indicates vibration amplitude >= standard vibration threshold. 110%; The second level, representing moderate intensity, indicates vibration amplitude >= standard vibration value. 100% and vibration amplitude <= standard vibration threshold 110%; The third level, indicating slight intensity: vibration amplitude >= standard vibration value 90% and vibration amplitude <= standard vibration threshold 100%.

[0098] This step aims to identify and quantify physical signals that are weak in intensity but may be significant when combined with specific visual risk events. By setting a low threshold range, the system can detect vibrations generated by actions such as minor scratches or leaning, and classify them as "Level 3" (low vibration). The significance of this step is that it enables the system to capture key scenarios of "strong visual risk accompanied by weak physical signals" (such as scratching a car), thereby significantly reducing the false negative rate for such high-risk malicious behaviors by matching them with rules in the corresponding risk assessment mechanism (e.g., "a scratching event combined with low vibration can be judged as a high-risk level"). At the same time, the clear lower threshold also avoids misjudging background noise as a valid signal.

[0099] In some optional implementations, the anomaly detection method for the above-mentioned vehicle sentry mode includes: Step b1: If it is determined that there is an abnormal risk to the vehicle, then match the corresponding target control strategy according to the target risk level of the vehicle.

[0100] Step b2: Control the corresponding components of the vehicle to perform defensive response operations according to the target control strategy to obtain the risk disposal result.

[0101] Step b3: Issue a security alert to the user based on the risk handling results.

[0102] Specifically, the target control strategy refers to a set of pre-defined vehicle control schemes based on different risk levels (such as high, medium, and low). Each strategy specifies a combination of defensive response actions to be triggered under specific risk conditions, aiming to effectively address risks and reduce losses. Defensive response actions refer to the proactive protective actions performed by vehicle components according to the target control strategy, such as automatically locking doors, activating audible and visual alarms, starting dashcam recording, triggering location reporting, and remotely notifying the owner or security platform. The risk handling result refers to the feedback status obtained by the system after executing the defensive response action, including information such as whether the operation was successful, whether the risk has been resolved, and the current vehicle status, used to evaluate the handling effect and provide a basis for user alerts.

[0103] Based on the real-time risk level assessment (e.g., high risk), the system matches the corresponding target control strategy from the strategy library to ensure that the response measures are appropriate to the level of risk. According to the target control strategy, the system sends instructions to relevant components (such as door locks, alarms, and communication modules) through the vehicle control unit, triggering a series of defensive response operations to form proactive protection. The system collects the risk handling results (e.g., "doors locked" "alarm triggered") and sends safety alarm messages to the user through the vehicle terminal or mobile application. Simultaneously, the handling results can be uploaded to the cloud platform for recording.

[0104] This implementation method establishes a closed-loop handling process from risk identification to proactive response, realizing automated and tiered protection for vehicles when facing threats. Through strategy matching and component control, the system can quickly execute targeted defensive actions, effectively deterring risky behaviors and reducing losses. Real-time alarms allow users to keep abreast of the vehicle's safety status, improving the overall safety protection capability of the vehicle and the user's sense of security.

[0105] In some alternative implementations, the method further includes the following after step b3: Step c1: In response to receiving a user's feedback instruction after issuing a security alert, and the feedback instruction indicates a false alarm, the video stream data and real-time vibration data are used as negative samples to adjust the risk event identification model and the preset vibration range, respectively.

[0106] Specifically, feedback instructions refer to the operational feedback provided by the user through the in-vehicle interface or mobile application after receiving a safety alarm, used to confirm the authenticity (verification) of the alarm or to point out its error (false alarm). This instruction is a key external input for the system to assess the accuracy of its alarms. Preset vibration ranges refer to the pre-set vibration intensity threshold ranges for classifying different risk levels. Specifically, this includes a first preset vibration range corresponding to high risk levels, a second preset vibration range corresponding to medium risk levels, and a third preset vibration range corresponding to low risk levels or the normal range.

[0107] After issuing an alarm, the system proactively provides a concise feedback channel (such as a "Confirmed" or "False Alarm" button) to wait for and receive the user's final judgment instruction on the alarm event. When the feedback instruction determines it is a "false alarm," the system uses the complete data that triggered the alarm as a "negative sample" for learning. Adjust the risk event identification model by adding the video stream data that generated false alarms and their "false alarm" labels to the model's training dataset. Optimize the model parameters through subsequent retraining to reduce the false identification rate for similar non-threat scenarios.

[0108] The system adjusts the preset vibration range by comparing and analyzing real-time vibration data that generates false alarms with the current preset range. For example, if a false alarm is triggered by a non-risk vibration (such as the sound of firecrackers) that enters the "high vibration" range (the first preset range), the system can automatically or after review fine-tune the threshold of that range to make it more accurate in subsequent judgments and reduce false triggers caused by environmental interference.

[0109] This implementation introduces a continuous learning loop of "human-machine collaboration," significantly improving the system's adaptability and long-term reliability. By directly converting user feedback into data fuel for optimizing models and parameters, the system can continuously learn from real-world usage scenarios and dynamically adjust the judgment criteria of the two core modules: risk identification (video analysis) and risk perception (vibration analysis). This not only effectively reduces false alarm rates and alleviates user frustration, but also allows alarm strategies (such as high-sensitivity, high-risk strategies) to become increasingly personalized and precise over time, thereby maintaining high robustness and user trust in complex and ever-changing real-world environments.

[0110] In one possible implementation, step S303 above includes the following steps: Step d1: If the user has already selected a corresponding preset risk assessment mechanism, then the user's risk assessment mechanism will be used as the vehicle's adaptive risk assessment mechanism; or... Step d2: If the user does not select the corresponding preset risk assessment mechanism, then the preset risk assessment mechanism that matches the environmental conditions of the vehicle's environment among the multiple preset risk assessment mechanisms will be used as the vehicle's adaptive risk assessment mechanism.

[0111] Specifically, a preset risk assessment mechanism refers to a complete set of risk assessment logics predefined and stored by the system. Each preset mechanism corresponds to a specific monitoring strategy or application scenario (such as "high-sensitivity monitoring," "routine parking," "noisy environment," etc.), and its core lies in the different judgment conditions it contains. The judgment conditions are rules in the form of "if-then," which clearly define the risk level to be determined when "what kind of risk event" and "what kind of vibration level" are combined in a specific logical relationship. For example, mechanism A might stipulate that "if a 'suspicious target loitering' event occurs and the vibration level is 'low,' then it is judged as high risk"; while mechanism B might stipulate that "only when a 'vehicle door being opened' event occurs and the vibration level is 'high,' it is judged as high risk." This step realizes the final decision-making for dynamically configuring the judgment logic for the vehicle.

[0112] When a user has explicitly selected a preset mechanism through the vehicle system or mobile application, the system will prioritize following the user's instructions and directly load that mechanism as the current risk assessment logic. This reflects the system's respect for and support of the user's personalized settings.

[0113] When the user does not actively select an option, the system initiates an intelligent adaptation process: First, it obtains the current environmental conditions (e.g., parking in an underground garage, on the roadside at night, or in a known high-risk area) through vehicle sensors (such as satellite positioning sensors or light sensors) or network data. Then, it compares the environmental conditions with the recommended applicable scenarios for each preset mechanism, automatically selecting the preset mechanism with the highest matching degree as the current adaptation mechanism. This "environmental condition" may include, but is not limited to, information such as location, time, lighting conditions, and historical event statistics.

[0114] This embodiment also provides an anomaly detection device for vehicle sentry mode, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0115] This embodiment provides an anomaly detection device for vehicle sentry mode, such as... Figure 11 As shown, it includes: The data acquisition module 1101 is used to acquire real-time vibration data of the vehicle caused by external interference, as well as video stream data of the vehicle's environment. The determination module 1102 is used to identify target risk events based on video stream data and determine the target vibration level corresponding to real-time vibration data; The judgment mechanism acquisition module 1103 is used to acquire the target risk judgment mechanism for vehicle adaptation. The target risk judgment mechanism includes judgment conditions corresponding to different risk levels. The judgment conditions include risk events and vibration levels. Different judgment conditions include at least one different risk event and vibration level. The risk determination module 1104 is used to determine whether there is an abnormal risk to the vehicle based on the target risk event, the target vibration level, and the target risk judgment mechanism.

[0116] In some optional implementations, the determination mechanism acquisition module 1103 includes: The determination mechanism unit is used to determine the risk determination mechanism of a vehicle if the user associated with the vehicle has selected the corresponding preset risk determination mechanism; or, if the user has not selected the corresponding preset risk determination mechanism, to select the preset risk determination mechanism that matches the environmental conditions of the vehicle's environment from among multiple preset risk determination mechanisms as the target risk determination mechanism.

[0117] In some alternative implementations, the determining module 1102 includes: The risk event identification unit is used to identify video stream data based on the risk event identification model to obtain the target risk event.

[0118] In some alternative implementations, the determining module 1102 includes: The target vibration level determination unit is used to determine the target vibration level based on the range of vibration amplitude corresponding to real-time vibration data.

[0119] In some optional implementations, the target vibration level determination unit includes: The first-level determination subunit is used to set the target vibration level to the first level if the vibration amplitude corresponding to the real-time vibration data is within the first preset vibration range. The second-level determination subunit is used to set the target vibration level to the second level if the vibration amplitude is within the second preset vibration range, wherein the minimum value of the first preset vibration range is greater than the maximum value of the second preset vibration range, and the first level is higher than the second level. The third-level determination subunit is used to set the target vibration level to the third level if the vibration amplitude is within the third preset vibration range. The minimum value of the second preset vibration range is greater than the maximum value of the third preset vibration range, and the second level is higher than the third level.

[0120] In some alternative implementations, the target risk event includes at least one of the following: kicking or scratching the vehicle, leaning against or peeping at the vehicle, the vehicle door being opened, suspicious target intrusion, and suspicious target loitering.

[0121] In some alternative implementations, the risk determination module 1104 includes: The condition matching unit is used to match the target risk event with the risk events in the different judgment conditions included in the target risk judgment mechanism, and to determine the first judgment condition for matching the target risk event from the different judgment conditions; and to match the target vibration level with the vibration level in the different judgment conditions included in the target risk judgment mechanism, and to determine the second judgment condition corresponding to the target vibration level from the different judgment conditions. The risk level determination unit is used to determine that the vehicle has a risk if the first judgment condition and the second judgment condition are the same, and to take the risk level corresponding to the first judgment condition or the second judgment condition as the target risk level of the vehicle; or, if the first judgment condition and the second judgment condition are different, to determine that the vehicle does not have an abnormal risk.

[0122] In some optional implementations, the risk determination module 1104 further includes: The control strategy matching unit is used to match the corresponding target control strategy according to the target risk level of the vehicle if it is determined that there is an abnormal risk to the vehicle. The risk management unit is used to control the corresponding components of the vehicle to perform defensive response operations according to the target control strategy, and to obtain the risk management results; The alarm unit is used to issue security alerts to users based on the results of risk handling.

[0123] In some optional implementations, the alarm unit further includes: The feedback receiving subunit is used to respond to the user's feedback instruction after a safety alarm is issued to the user. If the feedback instruction indicates a false alarm, the video stream data and real-time vibration data are used as negative samples to adjust the risk event identification model and the preset vibration range, respectively.

[0124] The vehicle sentry mode anomaly detection device provided in this application embodiment can execute the vehicle sentry mode anomaly detection method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects of the method execution. Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0125] Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0126] The following is a detailed reference. Figure 12 The diagram illustrates a structural schematic suitable for implementing the electronic device described in the embodiments of this application. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 1201, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1202 or a program loaded from memory 1208 into random access memory (RAM) 1203. The RAM 1203 also stores various programs and data required for the operation of the electronic device. The processor 1201, ROM 1202, and RAM 1203 are interconnected via a bus 1204. An input / output (I / O) interface 1205 is also connected to the bus 1204.

[0127] Typically, the following devices can be connected to I / O interface 1205: input devices 1206 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 1207 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 1208 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1209. Communication device 1209 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 12 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0128] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 1209, or installed from memory 1208, or installed from ROM 1202. When the computer program is executed by processor 1201, it performs the functions defined in the anomaly detection method of vehicle sentry mode according to embodiments of this application.

[0129] Figure 12 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0130] This application also provides a computer-readable storage medium. The methods described above according to this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the abnormal detection method for vehicle sentry mode shown in the above embodiments is implemented.

[0131] A portion of this application can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0132] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method for anomaly detection in vehicle sentry mode, characterized in that, The method includes: Acquire real-time vibration data of the vehicle caused by external interference, as well as video stream data of the environment in which the vehicle is located; Based on the video stream data, the target risk event is identified, and the target vibration level corresponding to the real-time vibration data is determined; The target risk assessment mechanism for vehicle adaptation is obtained, wherein the target risk assessment mechanism includes assessment conditions corresponding to different risk levels, the assessment conditions include risk events and vibration levels, and different assessment conditions include at least one different risk event and vibration level. Based on the target risk event, the target vibration level, and the target risk determination mechanism, it is determined whether the vehicle has any abnormal risks.

2. The method according to claim 1, characterized in that, The target risk assessment mechanism for obtaining the vehicle adaptation includes: If the user associated with the vehicle has selected a corresponding preset risk assessment mechanism, then the user's risk assessment mechanism will be used as the risk assessment mechanism adapted to the vehicle; or... If the user does not select the corresponding preset risk assessment mechanism, then the preset risk assessment mechanism that matches the environmental conditions of the vehicle's environment among the multiple preset risk assessment mechanisms will be used as the target risk assessment mechanism.

3. The method according to claim 1, characterized in that, The identification of target risk events based on the video stream data includes: The target risk event is obtained by identifying the video stream data based on the risk event identification model.

4. The method according to claim 1, characterized in that, Determining the target vibration level corresponding to the real-time vibration data includes: The target vibration level is determined based on the range of vibration amplitude corresponding to real-time vibration data.

5. The method according to claim 4, characterized in that, Determining the target vibration level based on the range of vibration amplitudes corresponding to real-time vibration data includes: If the vibration amplitude corresponding to the real-time vibration data is within the first preset vibration range, then the target vibration level is set to the first level. If the vibration amplitude is within the second preset vibration range, the target vibration level is set to the second level, wherein the minimum value of the first preset vibration range is greater than the maximum value of the second preset vibration range, and the first level is higher than the second level. If the vibration amplitude is within the third preset vibration range, the target vibration level is set to the third level, wherein the minimum value of the second preset vibration range is greater than the maximum value of the third preset vibration range, and the second level is higher than the third level.

6. The method according to claim 1, characterized in that, The target risk events include at least one of the following: kicking or scratching the vehicle, leaning against or peeping at the vehicle, the vehicle door being opened, suspicious target intrusion, and suspicious target loitering.

7. The method according to claim 1, characterized in that, The determination of whether the vehicle has an abnormal risk based on the target risk event, the target vibration level, and the target risk assessment mechanism includes: The target risk event is matched with risk events in different judgment conditions included in the target risk determination mechanism, and a first judgment condition matching the target risk event is determined from the different judgment conditions; and the target vibration level is matched with vibration levels in different judgment conditions included in the target risk determination mechanism, and a second judgment condition corresponding to the target vibration level is determined from the different judgment conditions. If the first determination condition and the second determination condition are the same, then it is determined that the vehicle poses a risk, and the risk level corresponding to the first determination condition or the second determination condition is taken as the target risk level of the vehicle; or... If the first determination condition and the second determination condition are different, then it is determined that the vehicle does not pose an abnormal risk.

8. The method according to claim 1, characterized in that, The method further includes: If it is determined that the vehicle has an abnormal risk, then a corresponding target control strategy is matched according to the target risk level of the vehicle; According to the target control strategy, the corresponding components of the vehicle are controlled to perform defensive response operations to obtain risk disposal results; Security alerts will be issued to users based on the risk handling results.

9. The method according to claim 7, characterized in that, When the target risk event is determined through a target identification model, and the target vibration level is determined based on a preset vibration range, the method further includes: In response to receiving feedback instructions from the user after issuing a security alert, and the feedback instructions indicating a false alarm, the video stream data and the real-time vibration data are used as negative samples to adjust the risk event identification model and the preset vibration range, respectively.

10. A computer device, characterized in that, The vehicle includes a controller, which includes a memory and a processor, the memory and the processor being communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the anomaly detection method of the vehicle sentry mode according to any one of claims 1 to 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the anomaly detection method of the vehicle sentry mode according to any one of claims 1 to 9.

12. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the anomaly detection method of the vehicle sentry mode as described in any one of claims 1 to 9.

Images