Switching method, system, car machine and medium of vehicle sentry mode

Abstract

The application provides a switching method, system, vehicle machine and medium for a vehicle sentry mode. The method comprises: when the sentry mode of the vehicle is activated: acquiring environmental information of the surroundings of the vehicle in real time, and detecting whether there is active personnel in the surroundings of the vehicle based on the environmental information; when it is detected that there is no active personnel in the surroundings of the vehicle, setting the sentry mode to a standby state; when it is detected that there is active personnel in the surroundings of the vehicle, performing identity recognition on the active personnel based on the environmental information, and comparing the identity recognition result with a trusted identity library to determine the identity category of the active personnel; when the active personnel is trusted active personnel, setting the sentry mode to a protection state; when the active personnel is suspicious active personnel, setting the sentry mode to a warning state; wherein the warning level corresponding to the warning state is higher than the warning level corresponding to the protection state. The application realizes differential security control based on the difference in identity categories.

Description

Technical Field

[0001] This application relates to the field of intelligent vehicle safety control technology, and in particular to a method, system, vehicle-mounted unit, and medium for switching vehicle sentry mode. Background Technology

[0002] With the development of intelligent vehicles, some vehicles are now equipped with a parking sentry mode. After the vehicle is locked, the surrounding environment is monitored by devices such as cameras, radar, or vibration sensors installed on the vehicle. When people approach or abnormal behavior is detected, functions such as recording, sound and light alarms, and mobile phone push notifications are automatically triggered to improve the safety of the vehicle while it is parked.

[0003] However, existing sentry modes typically only issue uniform alerts based on target proximity or abnormal behavior, lacking the ability to identify and differentiate the approaching person. Therefore, regardless of whether the person approaching the vehicle is the owner, family, friends, or a stranger, the same alarm process is triggered, resulting in a large number of invalid alarms during daily use. This not only affects user experience but also makes it easy for users to ignore genuine risk warnings, reducing the actual security effectiveness of sentry mode. Therefore, there is a need to provide a method, vehicle-mounted system, device, and medium for switching vehicle sentry modes. Summary of the Invention

[0004] This application provides a method, vehicle-mounted system, device, and medium for switching vehicle sentry mode, in order to improve the technical problem that existing sentry modes are difficult to switch in a differentiated manner based on personnel identity.

[0005] This application provides a method for switching a vehicle's sentry mode. The method includes: when the vehicle's sentry mode is activated: acquiring real-time environmental information around the vehicle and detecting whether there are active personnel around the vehicle based on the environmental information; when no active personnel are detected around the vehicle, setting the sentry mode to standby state; when active personnel are detected around the vehicle, identifying the active personnel based on the environmental information and comparing the identification result with a trusted identity database to determine the active personnel's identity category; when the active personnel are trusted active personnel, setting the sentry mode to protection state; when the active personnel are suspicious active personnel, setting the sentry mode to alert state; wherein, the alert level corresponding to the alert state is higher than the alert level corresponding to the protection state.

[0006] In one embodiment of this application, the step of identifying active personnel based on environmental information and comparing the identification result with a trusted identity database to determine the identity category of the active personnel includes: identifying the active personnel based on the personnel image in the environmental information to obtain the identification result; comparing the identification result with the trusted identity database: if the identification result matches the trusted identity information in the trusted identity database, the active personnel's identity category is determined to be a trusted active personnel; if the identification result does not match the trusted identity information in the trusted identity database, the active personnel's identity category is determined to be a suspicious active personnel.

[0007] In one embodiment of this application, the step of identifying an active person based on a person image in environmental information and obtaining an identification result includes: performing an illumination assessment on the person image to determine whether the image meets preset illumination conditions; if the person image meets the illumination conditions, then identifying the active person based on the person image in environmental information and obtaining an identification result; if the person image does not meet the illumination conditions, then identifying the active person based on gait information in environmental information and determining the identification result based on the gait identification result.

[0008] In one embodiment of this application, the step of identifying an active person based on a person image in environmental information and obtaining an identification result includes: performing facial occlusion detection on the person image; if no facial occlusion is detected, performing face recognition on the person image and using the face recognition result as the identification result; if facial occlusion is detected, performing local feature matching based on the unoccluded local facial region to obtain the corresponding local feature matching result and its local matching confidence; determining whether the local matching confidence is greater than or equal to a preset local confidence threshold; if so, using the local feature matching result as the identification result; otherwise, performing gait recognition on the active person based on gait information in environmental information and determining the identification result based on the gait recognition result.

[0009] In one embodiment of this application, the step of performing gait recognition on an active person based on gait information in environmental information, and determining the identity recognition result based on the gait recognition result includes: performing gait recognition on the active person based on gait information to obtain gait recognition result and its corresponding gait recognition confidence level; if the gait recognition confidence level is greater than or equal to a preset gait confidence level threshold, then the gait recognition result is used as the identity recognition result of the active person; if the gait recognition confidence level is less than the gait confidence level threshold, then performing body contour recognition on the active person based on body contour information in environmental information to determine the identity recognition result of the active person.

[0010] In one embodiment of this application, when the active person is a trusted active person, the method further includes: performing behavior detection on the active person based on environmental information to obtain the behavior category of the active person; determining whether the behavior category matches a preset abnormal behavior category: if it matches, then setting the sentry mode to an alert state; otherwise, keeping the sentry mode in a protected state.

[0011] In one embodiment of this application, the trusted identity database is dynamically updated based on all historical interaction information of the vehicle; wherein, the historical interaction information includes all interaction behavior information between active personnel and the vehicle collected by the vehicle.

[0012] This application also provides a vehicle sentry mode switching system, the system comprising: an acquisition module for acquiring environmental information around the vehicle in real time when the vehicle's sentry mode is activated; a detection module for detecting whether there are active personnel around the vehicle based on the environmental information; an identity recognition module for recognizing the active personnel based on the environmental information when active personnel are detected around the vehicle, and comparing the identity recognition result with a trusted identity database to determine the identity category of the active personnel; and a setting module for setting the sentry mode to a standby state when no active personnel are detected around the vehicle; setting the sentry mode to a protected state when the active personnel are identified as trusted active personnel; and setting the sentry mode to a vigilance state when the active personnel are identified as suspicious active personnel; wherein the vigilance level corresponding to the vigilance state is higher than the vigilance level corresponding to the protected state.

[0013] This application also provides a vehicle infotainment system, including: one or more processors; and a storage device for storing one or more programs, which, when executed by one or more processors, enable the vehicle infotainment system to implement the vehicle sentry mode switching method described above.

[0014] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by the computer's processor, causes the computer to perform any of the above-mentioned methods for switching vehicle sentry modes.

[0015] The beneficial effects of this application are as follows: This application proposes a method, system, vehicle-mounted unit, and medium for switching vehicle sentry modes. After the vehicle's sentry mode is activated, it acquires real-time environmental information about the vehicle's surroundings and detects the presence of active personnel based on this information. When no active personnel are present, the sentry mode is set to standby mode, thus avoiding the continuous activation of high-power monitoring links when no one is approaching, reducing power consumption during vehicle parking. When active personnel are detected, the system further identifies them based on environmental information and determines their identity category using a trusted identity database. This allows the vehicle to not only sense whether someone is approaching but also identify the person approaching the vehicle and switch the corresponding sentry state based on different identity categories. Specifically, for trusted active personnel, the sentry mode is switched to a protected state; for suspicious active personnel, it is switched to a higher alert level alert state. Compared to existing technologies that use a uniform alarm method for all approaching personnel, this application achieves differentiated security responses based on identity categories, reducing invalid alarms and wasted computing power, improving the intelligence level of the vehicle sentry mode and the user experience. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] In the attached diagram: Figure 1 A flowchart illustrating a method for switching vehicle sentry mode according to an embodiment of this application; Figure 2 A flowchart illustrating the state switching process of Sentinel mode provided in an embodiment of this application; Figure 3 This is a schematic diagram of a layered detection area provided in an embodiment of this application; Figure 4 A schematic diagram illustrating the dynamic update process of a trusted identity database provided in an embodiment of this application; Figure 5 A flowchart of the downgraded process for identifying active personnel is provided in one embodiment of this application; Figure 6 This is a structural block diagram of a vehicle sentry mode switching system provided in one embodiment of this application; Figure 7 This is a schematic diagram of the structure of a vehicle system provided in one embodiment of this application. Detailed Implementation

[0018] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0019] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0020] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the present application. However, it will be apparent to those skilled in the art that embodiments of the present application may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the present application.

[0021] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution of this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0022] The inventors discovered that some high-end smart vehicles are already equipped with a sentry mode. After the vehicle is locked, devices such as cameras, vibration sensors, radar, or proximity sensors continuously monitor the vehicle's surroundings and automatically activate security features such as recording and alarm push notifications when abnormal approach, abnormal touch, or collision behavior is detected. However, the existing sentry mode still has the following drawbacks: First, the existing sentry mode usually uses a uniform alarm logic for all people approaching the vehicle, regardless of whether the approacher is the car owner, family member, friend, delivery person, or stranger, all of which may trigger recording, mobile phone push notifications, or alarm processing. This results in the car owner receiving a large number of low-value notifications frequently, making it easy to overlook real security risks after long-term use, leading to a poor user experience. Second, when the existing sentry mode detects the approach of a person, it usually directly wakes up high-power processing links such as high-definition cameras, image recognition, video storage, cloud uploading, and mobile terminal push notifications. Even if the approacher is a trusted person, the vehicle control system still performs a complete security response process, resulting in unnecessary power consumption and computing power occupation during vehicle parking. Furthermore, the sentry mode's status switching is typically based solely on information such as distance changes, collision intensity, or behavioral risk. Therefore, it cannot achieve differentiated security control by maintaining a low-interference response when trusted personnel approach and increasing the alert level when strangers approach. On the other hand, existing sentry modes usually require users to manually enter trusted personnel information or actively authorize the vehicle control system to collect corresponding facial data. This process is cumbersome and easily raises privacy concerns, resulting in low actual usage of the trusted identity management function and making it difficult to achieve automatic maintenance and dynamic updates of the trusted identity database.

[0023] To address the above issues, this application provides a method for switching vehicle sentry mode. After the vehicle's sentry mode is activated, it acquires real-time environmental information about the vehicle's surroundings and detects the presence of any active personnel based on this information. When no active personnel are present, the sentry mode is set to standby mode, thus avoiding the continuous activation of high-power monitoring links when no one is approaching, reducing power consumption during vehicle parking. When active personnel are detected, the system further identifies them based on environmental information and determines their identity category using a trusted identity database. This allows the vehicle to not only sense whether someone is approaching but also identify the person approaching the vehicle and switch the corresponding sentry state based on different identity categories. Specifically, for trusted active personnel, the sentry mode is switched to a protection state to reduce unnecessary alarm triggering. For suspicious active personnel, the system switches to a higher alert level alert state to promptly trigger security measures such as recording, alarms, or remote notifications. Compared to existing technologies that use a uniform alarm method for all approaching personnel, this application can achieve differentiated security responses based on identity categories, reduce invalid alarms and wasted computing power, and improve the intelligence level of vehicle sentry mode and user experience.

[0024] like Figure 1 As shown, the method for switching vehicle sentry mode is executed when the vehicle's sentry mode is activated, and it includes the following steps: S100: Acquire real-time environmental information about the vehicle's surroundings and detect the presence of any people or activities around the vehicle based on this information.

[0025] After the vehicle's sentry mode is activated, the vehicle control system acquires real-time environmental information about the vehicle's surroundings. The sentry mode can be activated in response to an activation command, which can be generated by the vehicle's infotainment interface, mobile application, voice control commands, or preset automatic triggering conditions. It should be noted that after the sentry mode is activated, the vehicle does not directly initiate high-power functions such as high-definition recording, cloud uploading, continuous video analysis, or audible and visual alarms. Instead, it first continuously senses the vehicle's surroundings using low-power monitoring methods to reduce system power consumption and computing power consumption during vehicle parking. This environmental information may include radar detection information, infrared sensing information, and proximity sensor information corresponding to the vehicle's surroundings. Based on the acquired environmental information, the system detects whether there are any active personnel around the vehicle. If active personnel are detected, subsequent identity recognition processing is triggered. Conversely, if no active personnel are detected, the sentry mode is set to standby mode.

[0026] In an optional embodiment of this application, step S100 includes the following process: real-time acquisition of environmental information around the vehicle; detection of the environmental information based on a personnel detection algorithm to obtain target objects corresponding to the vehicle; determination of whether the target objects meet preset human characteristic conditions: if they meet the conditions, it is determined that there are active personnel around the vehicle; if they do not meet the conditions, it is determined that there are no active personnel around the vehicle.

[0027] Once the vehicle's sentry mode is activated, the vehicle control system puts the sentry mode into standby mode, acquiring real-time environmental information about the vehicle's surroundings and using personnel detection algorithms to identify target objects in the vicinity. These personnel detection algorithms include, but are not limited to, those based on convolutional neural networks, object detection models, or human contour analysis. After detecting a target object, the vehicle control system determines whether the target object meets preset human characteristic conditions. These conditions may include whether the target object possesses human contour features, conforms to human movement patterns, and is located within a preset detection area around the vehicle. For example, if the target object possesses human contour features, the distance between the target object and the vehicle continuously decreases, and it enters a preset personnel detection area around the vehicle (e.g., within 1-5 meters of the vehicle), then it is determined that there are active personnel in the vicinity, triggering subsequent identity recognition processing. Conversely, if the target object does not meet the above human characteristic conditions, it is determined that there are no active personnel in the vicinity. Through the above process, this application can not only achieve real-time detection of people moving around the vehicle, but also effectively avoid misidentifying animals, swaying trees or other non-human targets as people, thereby improving the accuracy of people detection and reducing the probability of false triggering of sentry mode.

[0028] S200: When no active personnel are detected around the vehicle, the sentry mode is set to standby mode.

[0029] Once the vehicle's sentry mode is activated, the vehicle control system continuously acquires information about the surrounding environment within a preset first range using low-power detection devices such as low-power radar and proximity sensors. Based on this information, it detects whether any people are present around the vehicle. The first range is a pre-defined personnel detection area centered on the vehicle, used to trigger subsequent identity recognition processing when a person approaches. For example, the first range can be an area within 1-5 meters of the vehicle. It's important to note that during the detection process, the vehicle does not directly activate high-power functions such as high-definition cameras, cloud uploads, or audible and visual alarms. Instead, it first continuously senses the surrounding environment using low-power detection methods to reduce system power consumption and computing power usage while the vehicle is parked. When no people are detected, the sentry mode is set to standby to maintain the vehicle in a low-power monitoring state. Conversely, when people are detected, the camera and identity recognition functions are activated to identify them. After identity recognition, the vehicle control system switches the sentry mode according to the identity category of the person.

[0030] S300: When a person is detected moving around the vehicle, the person is identified based on environmental information, and the identification result is compared with a trusted identity database to determine the person's identity category.

[0031] When a person is detected moving around the vehicle, the vehicle control system activates the vehicle's camera to acquire an image of that person. This camera may include, but is not limited to, in-vehicle cameras, surround-view cameras, or wide-angle cameras, as long as they can capture image information of the person. To improve the accuracy of subsequent identity verification, the vehicle control system can dynamically adjust the camera angle, camera area, or image acquisition parameters based on changes in the person's position, ensuring that the acquired image includes information such as the person's face, body, or movement area. After acquiring the image, the vehicle control system performs identity verification based on the image and compares the verification result with trusted identity information in a trusted identity database.

[0032] When the identity recognition result matches the trusted identity information in the trusted identity database, step S400 is executed, that is, the currently active person is determined to be a trusted active person, and the sentry mode is set to the protection state. In the protection state, the vehicle can reduce the alarm level and only perform low-level environmental monitoring, local recording, or low-frequency image acquisition to avoid frequent false alarms to the vehicle owner, family members, or other trusted persons. Conversely, when the identity recognition result does not match the trusted identity information in the trusted identity database, step S500 is executed, that is, the currently active person is determined to be a suspicious active person, and the sentry mode is set to the alert state. The alert level corresponding to the alert state is higher than the alert level corresponding to the protection state. In the alert state, the vehicle can further activate security protection measures such as high-definition recording, abnormal behavior monitoring, audible and visual alarms, and remote notification, thereby improving the security protection capability of the vehicle when parked. Through the above method, this application can only activate low-power devices such as radar or proximity sensors when no active person is detected to maintain low-power monitoring, and only activate the corresponding image acquisition and identity recognition functions when an active person is detected, thereby avoiding the vehicle being in a high-power operation state for a long time and reducing vehicle energy consumption.

[0033] In an optional embodiment of this application, step S300 includes the following process: S310. Based on the personnel images in the environmental information, identify the active personnel and obtain the identification results.

[0034] The vehicle control system performs facial recognition processing on images of people in the environmental information to obtain corresponding identity recognition results. Specifically, the vehicle control system can first detect facial regions in the personnel image, extract features from the detected facial regions, and determine the identity recognition result corresponding to the currently active person based on the extracted facial features. Facial recognition includes, but is not limited to, facial recognition algorithms based on convolutional neural networks, facial recognition algorithms based on feature vector matching, or facial recognition algorithms based on deep learning. It is understood that the identity recognition result of this application includes, but is not limited to, the identity name, identity number, or preset identity identifier corresponding to the currently active person, as long as it can represent the identity information corresponding to the currently active person.

[0035] In an optional embodiment of this application, step S310 includes the following process: S3101. Perform illumination assessment on personnel images to determine whether the images meet the preset illumination conditions.

[0036] Considering that simple face recognition is easily affected by ambient lighting conditions, which may lead to inaccurate facial feature extraction and thus affect the identity recognition effect, in order to improve the above problem, after acquiring the person image in the environmental information, the person image will first be evaluated for lighting conditions to determine whether the current image meets the preset lighting conditions. The lighting evaluation algorithm includes, but is not limited to, an evaluation algorithm based on average brightness, an evaluation algorithm based on contrast, an evaluation algorithm based on histogram distribution, or an evaluation algorithm based on image sharpness, and may be a combination of one or more of the above, without specific limitation. Lighting conditions are used to characterize whether the current person image meets the requirements for face recognition. Specifically, the vehicle control system can analyze the brightness information, contrast information, and exposure level in the person image to determine whether the current image has insufficient lighting, backlighting, local overexposure, or excessive darkness. For example, when the overall brightness in the person image is within the preset brightness range, and the face region has clearly distinguishable facial features, then the current image is determined to meet the preset lighting conditions. Conversely, if the image has low illumination at night, strong backlight, or excessively dark facial areas, it is determined that the current image does not meet the preset lighting conditions.

[0037] If in step S3101 it is determined that the personnel image meets the lighting conditions, then step S3102 is executed: based on the personnel image in the environmental information, the identity of the active personnel is identified, and the identity identification result is obtained.

[0038] Specifically, the vehicle control system can first perform image quality analysis on the personnel image and, in conjunction with the facial region information in the image, perform identity recognition processing on the active personnel. For example, when the personnel image meets preset recognition conditions, identity recognition can be performed based on the facial features in the image. When the personnel image is partially occluded or the image quality is degraded, identity recognition can be performed based on local features in the image to obtain the corresponding identity recognition result.

[0039] In an optional embodiment of this application, step S3102 includes the following steps: performing facial occlusion detection on the personnel image; if no facial occlusion is detected, performing face recognition on the personnel image and using the face recognition result as the identity recognition result; if facial occlusion is detected, performing local feature matching based on the unoccluded local facial region to obtain the corresponding local feature matching result and its local matching confidence; determining whether the local matching confidence is greater than or equal to a preset local confidence threshold; if so, using the local feature matching result as the identity recognition result; otherwise, performing gait recognition on the active personnel based on the gait information in the environmental information and determining the identity recognition result based on the gait recognition result.

[0040] Specifically, occlusion detection algorithms can be used, such as those based on object detection, keypoint localization, image segmentation, or deep learning, to detect facial occlusion in a person's image. This allows the system to determine if masks, hats, sunglasses, or other obstructions are covering the face. If no occlusion is detected, the image indicates complete facial features, and the vehicle control system can directly perform face recognition, using the result as the identification result. Conversely, if occlusion is detected, local facial features can be extracted from the unoccluded facial region. These extracted features are then matched with pre-stored known local facial features to obtain the matching result and its confidence level. The local facial region can include at least one of the eye, eyebrow, nose, or facial contour regions. If the confidence level is greater than or equal to a preset threshold, the local facial features are considered sufficient to identify the person, and the matching result is used as the identification result. The local feature matching result is used to represent the identity information of the currently active person, including but not limited to the person's name, ID number, or preset identity identifier. Conversely, when the confidence level of the local match is less than a preset threshold, it indicates that the current local facial features are insufficient to accurately determine the identity of the active person. In this case, gait information from the environmental information is further used to perform gait recognition on the active person, and the identity recognition result is determined based on the corresponding gait recognition result. Through the above process, this application can dynamically switch between different identity recognition methods in complex occlusion environments, thereby improving the accuracy of vehicle sentry mode in identifying active persons.

[0041] If, in step S3101, it is determined that the person image does not meet the lighting conditions, then step S3103 is executed: based on the gait information in the environmental information, gait recognition is performed on the active person, and the identity recognition result is determined based on the gait recognition result.

[0042] When the current person image is determined not to meet the preset lighting conditions, facial recognition based on the person image at this time is prone to inaccurate facial feature extraction. Therefore, in this embodiment, the vehicle control system will downgrade to an identity recognition method based on gait information to obtain the corresponding gait recognition result and its corresponding gait recognition confidence level. The gait recognition result is used to characterize the identity information of the currently active person, which includes, but is not limited to, the person's name, ID number, or preset identity identifier.

[0043] In an optional embodiment of this application, the steps of performing gait recognition on an active person based on gait information in environmental information and determining the identity recognition result based on the gait recognition result include: performing gait recognition on the active person based on gait information to obtain the gait recognition result and its corresponding gait recognition confidence level; if the gait recognition confidence level is greater than or equal to a preset gait confidence level threshold, then the gait recognition result is used as the identity recognition result of the active person; if the gait recognition confidence level is less than the gait confidence level threshold, then performing body contour recognition on the active person based on body contour information in environmental information to determine the identity recognition result of the active person.

[0044] Based on the motion image sequence of an active person within a preset time period from environmental information, gait information of the active person is extracted. Gait information includes, but is not limited to, stride length, stride frequency, limb swing, changes in body center of gravity, or human posture information, and may be one or more combinations thereof. The vehicle control system performs gait recognition on the active person based on the gait information, obtaining the corresponding gait recognition result and its corresponding gait recognition confidence level. Gait recognition can employ gait recognition algorithms based on human key points, gait recognition algorithms based on motion contours, or gait recognition algorithms based on deep learning. When the gait recognition confidence level is greater than or equal to a preset gait confidence threshold, the current gait recognition result is deemed reliable, and the active person's identity is determined based on the gait recognition result. If the gait recognition confidence level is less than the gait confidence threshold, the active person's body contour information is further acquired, and body contour recognition is performed on the active person based on the body contour information to improve the accuracy and stability of identity recognition in complex environments.

[0045] Further, in an optional embodiment of this application, the step of performing body contour recognition on an active person based on body contour information in the environmental information to determine the identity of the active person includes: extracting body contour information corresponding to the active person from the motion image sequence in the environmental information; performing body contour recognition on the active person based on the body contour information to obtain the corresponding body contour recognition result and its corresponding body contour recognition confidence level; determining whether the body contour recognition confidence level is greater than or equal to a preset body contour recognition confidence threshold; if the body contour recognition confidence level is greater than or equal to the preset body contour recognition confidence threshold, then determining the identity of the active person based on the body contour recognition result; if the body contour recognition confidence level is less than the body contour recognition confidence threshold, then determining the identity of the active person based on the face recognition result, local feature matching result, gait recognition result, and body contour recognition result.

[0046] Specifically, the vehicle control system can extract the body contour information of the active person from the motion image sequence corresponding to the active person. For example, a human detection algorithm can be used to locate the body region corresponding to the active person, and information such as the external contour of the human body, the coordinates of key human points, or the human posture sequence can be extracted within the body region to generate body contour information for identity recognition. By extracting the body contour information of the active person, the influence of background interference and non-target areas on the identity recognition process can be reduced, thereby improving the accuracy of subsequent recognition processes. Based on the body contour information, body contour recognition is performed on the active person to obtain the corresponding body contour recognition result and its corresponding body contour recognition confidence level. When the body contour recognition confidence level is greater than or equal to a preset body contour confidence level threshold, the current body contour recognition result is determined to be reliable, and the identity of the active person is determined based on the body contour recognition result. Conversely, when the confidence score for body posture recognition is less than the confidence threshold, it indicates that a single body posture recognition method is insufficient to accurately determine the identity of the person. In this case, the candidate identities and recognition confidence scores corresponding to the face recognition results, local feature matching results, gait recognition results, and body posture recognition results are further weighted and accumulated according to preset weights to obtain the comprehensive confidence score for each candidate identity. Here, a candidate identity refers to the identity name, identity number, or preset identity identifier obtained from face recognition, local feature matching, gait recognition, or body posture recognition, used to represent the identity of the current person. The candidate identity with the highest comprehensive confidence score is determined as the target candidate identity. If the comprehensive confidence score of the target candidate identity is greater than or equal to the preset comprehensive confidence threshold, the target candidate identity is used as the identity recognition result for the person. If the comprehensive confidence score of the target candidate identity is less than the preset comprehensive confidence threshold, it is determined that a reliable identity recognition result cannot be obtained at this time. For example, for a given person, if the candidate identity corresponding to the aforementioned face recognition result, local feature matching result, and body contour recognition result is all user A, with corresponding recognition confidence scores of 0.82, 0.75, and 0.71 respectively, and the candidate identity corresponding to the gait recognition result is user B, with a corresponding recognition confidence score of 0.68. Furthermore, if the preset weights for face recognition, local feature matching, gait recognition, and body contour recognition are 0.4, 0.2, 0.2, and 0.2 respectively, then the overall confidence score for user A is 0.62, and the overall confidence score for user B is 0.136. Since user A has the highest overall confidence score and is greater than the preset overall confidence score threshold (e.g., 0.6), user A is determined as the identity recognition result for the person.

[0047] S320. Compare the identity recognition result with the trusted identity database: If the identity recognition result matches the trusted identity information in the trusted identity database, the identity category of the active person is determined as a trusted active person; if the identity recognition result does not match the trusted identity information in the trusted identity database, the identity category of the active person is determined as a suspicious active person.

[0048] After obtaining the identification result for the active person, the system compares the result with trusted identity information in a trusted identity database to determine the identity category of the current active person. This trusted identity database can pre-store trusted identity information for vehicle owners, family members, authorized users, or other trusted individuals. Specifically, the vehicle control system can determine whether the identity name, identity number, or preset identity identifier corresponding to the current identification result matches the trusted identity information in the trusted identity database. For example, if the identification result is user A, and trusted identity information for user A exists in the trusted identity database, then the current active person is determined to be a trusted active person. Conversely, if the identification result cannot match the trusted identity information in the trusted identity database, then the current active person is determined to be a suspicious active person.

[0049] Furthermore, after determining the identity category of the active personnel, differentiated control can be applied to the sentry mode based on the identity category. As a specific example, when the active personnel are suspected, the system is in an alert state, and further security measures such as high-definition video recording, abnormal behavior monitoring, audible and visual alarms, or remote notifications are triggered, thereby improving the security capabilities of the vehicle while it is parked.

[0050] In an optional embodiment of this application, when the active person is a trusted active person, the above-mentioned vehicle sentry mode switching method further includes: based on environmental information, performing behavior detection on the active person to obtain the behavior category of the active person; determining whether the behavior category matches a preset abnormal behavior category: if it matches, then setting the sentry mode to the alert state; otherwise, keeping the sentry mode in the protection state unchanged.

[0051] Specifically, once a person is identified as a trusted individual, this application continues to detect their behavior based on environmental information to determine the corresponding behavior category. It then determines whether the behavior category matches a preset abnormal behavior category. For example, an abnormal behavior category might include, within a preset range (e.g., 1 meter) around the vehicle, pulling the car door more times than a preset threshold, staying for a longer period than a preset threshold, or knocking on the car window more times than a preset threshold. When the behavior category matches a preset abnormal behavior category, it indicates that while the person is a trusted individual, their current behavior may pose an abnormal risk. Therefore, the sentry mode is switched to alert mode, and further security measures such as high-definition recording, abnormal behavior logging, or remote notification are triggered. Conversely, when the behavior category does not belong to a preset abnormal behavior category, the sentry mode remains in a protected state.

[0052] In an optional embodiment of this application, the trusted identity database is dynamically updated based on all historical interaction information of the vehicle; wherein, the historical interaction information includes all interaction behavior information between active personnel and the vehicle collected by the vehicle.

[0053] It should be noted that the trusted identity database in this application is not fixed, but dynamically updated based on all historical interaction information of the vehicle. This historical interaction information includes all interaction behavior information between the vehicle and any active personnel collected during its historical operation. As a specific example, historical interaction information may include the identity recognition results of the active personnel, the frequency of the active personnel approaching the vehicle, the duration of the active personnel's stay around the vehicle, the types of interaction behavior between the active personnel and the vehicle, and corresponding abnormal behavior records. The vehicle control system can continuously record and statistically analyze the above historical interaction information, and dynamically adjust the trusted identity information in the trusted identity database based on the statistical results. For example, when an active person is repeatedly identified as the same person over a long period of time, and there is no abnormal interaction behavior between them and the vehicle, the vehicle control system can gradually increase the trust level of the active person's corresponding identity and add them to the trusted identity database or increase their trust level. Conversely, when an active person exhibits frequent abnormal interaction behavior, does not appear for a long time, or the corresponding abnormal behavior records exceed a preset threshold, the vehicle control system can reduce the trust level of their corresponding identity or remove them from the trusted identity database. Through the above dynamic update process, this application enables the trusted identity database to be continuously and adaptively updated as the vehicle is used, thereby improving the accuracy of the vehicle sentry mode in identifying trusted active personnel.

[0054] As a concrete example, a vehicle control system can score the trust level of individuals based on their historical behavioral information to quantify the likelihood of a stranger gradually becoming a trustworthy individual. This trust level score can be calculated based on multiple behavioral dimensions using a weighted cumulative method. Specific examples are shown in the table below:

[0055] Table 1. Comparison of different trust mechanisms The system can comprehensively score individuals based on behavioral dimensions such as frequency of approach to the vehicle, abnormal behavior, duration of stay, instances of riding with the vehicle owner, and historical trust levels within a preset time period. Specifically, frequency of approach reflects the long-term contact patterns between the individual and the vehicle; abnormal behavior indicates whether the individual exhibits unusual contact with the vehicle; duration of stay reflects the stability of the individual's behavior around the vehicle; instances of riding with the vehicle owner reflect the degree of connection between the individual and the vehicle owner; and historical trust levels reflect changes in the individual's historical trustworthiness.

[0056] Based on the scoring results corresponding to the aforementioned behavioral dimensions, the vehicle control system calculates the total trust score for each active person and determines whether the trust score reaches a preset trust threshold. If the active person is currently in an untrusted identity category and the trust score reaches the preset threshold, their identity is updated to a trusted identity category and added to the trusted identity database. If the active person is already in a trusted identity category, the trusted identity database status remains unchanged. If the trust score does not reach the preset threshold, the active person's current identity category remains unchanged. The vehicle control system can also update the trust score of each active person after detection or according to a preset period to achieve dynamic maintenance of the trusted identity database.

[0057] As a concrete example, the total trust score can be obtained by weighting and accumulating the scores corresponding to each behavioral dimension. For example: Total Trust Score = Approach Frequency Score + No Abnormal Behavior Score + Dwell Time Score + Ride-Share Record Score + Historical Trust Base Score. When the total trust score reaches a preset value (e.g., 70 points out of 100), and the active person has not triggered any preset abnormal behavior, the vehicle control system automatically adds the corresponding active person to the trusted identity database. Furthermore, the trust score of the active person can be asynchronously updated at fixed times each day, after each detected approach event, or according to a preset update cycle.

[0058] To prevent the trusted identity database from becoming outdated due to long-term accumulation, the vehicle control system can also introduce a time-based trust decay mechanism to dynamically adjust the trust level of active personnel. When an active person does not reappear within a preset time period (e.g., 90 consecutive days), the vehicle control system decays the corresponding trust level based on the duration of absence, reflecting changes in the degree of association between the active person and the vehicle. As a specific example, the trust decay mechanism can be shown in Table 2:

[0059] Table 2 Comparison of different attenuation mechanisms In the example above, when the trust level of an active person decays below a preset threshold and a preset period of absence is exceeded, the vehicle control system removes them from the trusted identity database. Similarly, if an active person triggers abnormal behavior while approaching the vehicle and their current trust level is below a preset threshold, they can also be removed from the trusted identity database. Furthermore, the vehicle control system can provide a user interface allowing users to manually delete corresponding members from the trusted identity database. Even after an active person is removed from the trusted identity database, the vehicle control system retains their historical behavior information for subsequent trust level updates. If the active person is detected again, the vehicle control system recalculates their trust level and re-evaluates whether to include them in the trusted identity database based on the updated trust level, thus achieving dynamic maintenance of the trusted identity database.

[0060] Furthermore, the in-vehicle HMI user interface also provides an entry point for managing a trusted identity database, allowing users to manually add, delete, or adjust trust levels. In addition, considering privacy protection, all identity verification and trusted identity processing are handled locally on the in-vehicle chip and are not uploaded to the cloud. Only images of strangers in alarm events are anonymized (e.g., unrelated personnel are obscured) before being pushed to the system.

[0061] like Figure 2As shown, this demonstrates the state switching process of Sentry Mode. After activating Sentry Mode, the vehicle control system first enters a standby state. In standby mode, the vehicle control system continuously monitors for any active personnel entering a preset first area. When active personnel are detected entering the first area, the vehicle control system activates the camera and identity recognition module to identify the active personnel. If the identification result indicates that the active personnel belongs to a trusted identity category, Sentry Mode is switched to Protection mode; if the active personnel belongs to an untrusted identity category, Sentry Mode is switched to Alert mode. In Protection mode, the vehicle control system continuously monitors for any active personnel entering a preset second area and for any abnormal behavior. When active personnel are detected entering the second area and exhibiting abnormal behavior, Sentry Mode is switched from Protection mode to Alert mode. In Alert mode, security protection measures such as recording and alarm push notifications are performed. When active personnel are detected leaving the first area, or after alarm processing is completed, the vehicle control system returns to standby mode, thereby achieving differentiated security control based on the active personnel's identity category and behavioral information.

[0062] like Figure 3 The diagram illustrates the layered detection area in this application. The vehicle control system establishes different levels of detection range centered on the vehicle and performs different levels of security processing based on the location of the active personnel. The area farther from the vehicle corresponds to the standby zone. When an active person is outside the standby zone, the vehicle control system maintains sentry mode in standby state, only maintaining low-power detection link operation. When an active person is detected entering the identification zone, the vehicle control system activates the camera device and identity recognition module to identify the active person and determine their corresponding identity category. Further, when the active person continues to approach the vehicle and enters the upgrade zone, the vehicle control system upgrades the current sentry mode based on the active person's identity category and behavioral information. For example, when the active person belongs to a trusted identity category and does not trigger abnormal behavior, the protection state is maintained. When the active person belongs to an untrusted identity category, or exhibits abnormal contact, abnormal stay, or abnormal operation of the vehicle within the upgrade zone, the sentry mode is switched to the alert state, and security protection functions such as recording and alarm push are activated, thereby realizing hierarchical security control based on distance level and identity information.

[0063] like Figure 4As shown, this illustrates the dynamic update process of the trusted identity database. Upon detecting an approaching person, the vehicle control system first identifies the person, categorizing them as a recorded acquaintance, a stranger, or a failed identification. For recorded acquaintances, the corresponding approach record is updated. For strangers, their characteristic information and approach data are recorded. For failed identifications, the relevant data is temporarily stored and marked as an unknown person. The vehicle control system calculates the corresponding trust level based on the person's historical behavior information and determines whether the trust level reaches a preset threshold. When the trust level reaches the preset threshold, the vehicle control system further determines whether the person already exists in the trusted identity database. If not, the person is automatically added to the database. If already existing, the trusted identity status remains unchanged. If the trust level does not reach the preset threshold, the person remains in the stranger or unknown status, and historical behavior information continues to accumulate for subsequent trust level updates. Furthermore, the vehicle control system can prompt the user via a user interface to add acquaintance information to the trusted identity database and allows the user to manually adjust the information in the database. Furthermore, the vehicle control system can also remove trusted identity members from the trusted identity database that have not appeared for more than a preset number of days (e.g., 90 days) based on a periodic maintenance mechanism, thereby achieving dynamic learning and maintenance of the trusted identity database. Even further, the identity recognition data and identity information in the trusted identity database in this application are processed locally on the vehicle's onboard chip, without being uploaded to the cloud, to reduce the risk of user identity information leakage and improve the data security and privacy protection capabilities of the vehicle's sentry mode. Only when an alarm event is triggered will the vehicle control system perform desensitization processing on the corresponding unfamiliar person image, such as obscuring or blurring areas of unrelated personnel, before pushing an alarm to the user terminal.

[0064] like Figure 5The diagram illustrates the downgraded processing flow for identifying active individuals. After detecting an active individual entering a preset first area, the vehicle control system first acquires the corresponding individual image and determines whether the current image meets preset lighting conditions based on a lighting evaluation algorithm. If the individual image meets the lighting conditions, it further performs facial occlusion detection. When facial occlusion is detected, the vehicle control system performs local feature matching based on the local facial region and determines whether the active individual's identity can be determined based on the confidence level corresponding to the local matching result. If no facial occlusion is detected, face recognition is performed directly to obtain the corresponding face recognition result and face recognition confidence level. If the local matching confidence level is less than a preset feature matching confidence level threshold, or the face recognition confidence level is less than a preset face confidence level threshold, or the aforementioned individual image does not meet the lighting conditions, it downgrades to gait recognition, obtaining the gait recognition result and corresponding gait recognition confidence level. If the gait recognition confidence level is less than a preset gait recognition confidence level threshold, it further downgrades to body posture recognition, obtaining the body posture recognition result and corresponding body posture recognition confidence level. When the confidence score for body posture recognition is less than a preset confidence threshold, the confidence scores for face recognition, occlusion recognition, gait recognition, and body posture recognition are weighted to obtain the identity recognition result. When the confidence score for any of the above recognition methods is greater than or equal to the corresponding confidence threshold, the identity recognition result is obtained directly. Through multi-step downgrade recognition, the accuracy of identity recognition in complex environments is improved.

[0065] After confirming the identity of the person approaching the vehicle, the vehicle control system adaptively switches the Sentry Mode's operating state based on the corresponding identity category. When the vehicle control system identifies a person approaching the vehicle as belonging to a trusted identity category, it switches the Sentry Mode to Protection Mode. In Protection Mode, the vehicle control system maintains a low-interference monitoring mode, continuously monitoring only the behavior of the person around the vehicle without triggering audible and visual alarms or high-level security protection measures. At this time, the display shows the vehicle's current protection state as a protective shield, along with status information such as "Protection in Progress," thereby reducing false alarms caused by trusted personnel approaching the vehicle normally and minimizing unnecessary system power consumption.

[0066] When the vehicle control system detects that an active person belongs to an untrusted identity category, or that a trusted person exhibits abnormal behavior within a preset range, it switches the sentry mode from protection to alert mode. In alert mode, the vehicle control system activates security functions such as recording, abnormal behavior monitoring, cloud synchronization, and alarm push notifications, and displays prompts such as "Sound alarm activated" and "Stranger detected approaching" on the display interface. At this time, the display interface also highlights the vehicle's current high alert status to alert nearby personnel that the vehicle has entered active security mode, thereby enhancing the vehicle's security capabilities while parked.

[0067] like Figure 6 As shown, the vehicle sentry mode switching system includes: an acquisition module 61, a detection module 62, an identity recognition module 63, and a setting module 64. The acquisition module 61 acquires environmental information surrounding the vehicle in real time when the vehicle's sentry mode is activated. The detection module 62 detects the presence of active personnel around the vehicle based on the environmental information. The identity recognition module 63 identifies the active personnel based on the environmental information when active personnel are detected around the vehicle, and compares the identification result with a trusted identity database to determine the active personnel's identity category. The setting module 64 sets the sentry mode to standby mode when no active personnel are detected around the vehicle; sets the sentry mode to protected mode when the active personnel are identified as trusted; and sets the sentry mode to alert mode when the active personnel are identified as suspicious. The alert level corresponding to the alert state is higher than the alert level corresponding to the protected state.

[0068] Specific limitations regarding the vehicle sentry mode switching system can be found in the limitations on the vehicle sentry mode switching method described above, and will not be repeated here. Each module in the aforementioned vehicle sentry mode switching system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware format or independent of it, or stored in the memory of the computer device in software format, so that the processor can call the corresponding operations of each module.

[0069] It should be noted that, in order to highlight the innovative aspects of this application, this embodiment does not include modules that are not closely related to solving the technical problems proposed in this application, but this does not mean that there are no other modules in this embodiment.

[0070] like Figure 7 As shown, the vehicle infotainment system 7 may include a memory 71, a processor 72 and a bus, and may also include a computer program stored in the memory 71 and that can run on the processor 72, such as a vehicle sentry mode switching program.

[0071] The memory 71 includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 71 can be an internal storage unit of the vehicle's infotainment system 7, such as the portable hard drive of the vehicle's infotainment system 7. In other embodiments, the memory 71 can also be an external storage device of the vehicle's infotainment system 7, such as a plug-in portable hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the vehicle's infotainment system 7. Furthermore, the memory 71 can include both internal storage units and external storage devices of the vehicle's infotainment system 7. The memory 71 can be used not only to store application software and various types of data installed on the vehicle's infotainment system 7, such as the code for switching vehicle sentry mode, but also to temporarily store data that has been output or will be output.

[0072] In some embodiments, the processor 72 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits packaged with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 72 is the control unit of the vehicle infotainment system 7, connecting various components of the system via various interfaces and lines. It performs various functions and processes data by running or executing programs or modules stored in the memory 71 (such as the vehicle sentry mode switching program) and calling data stored in the memory 71.

[0073] The processor 72 runs the operating system of the vehicle's infotainment system 7 and installs various applications. The processor 72 executes the applications to implement the steps in the above-described method for switching the vehicle sentry mode.

[0074] For example, a computer program can be divided into one or more modules, one or more of which are stored in memory 71 and executed by processor 72 to complete this application. One or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the process of the computer program in the vehicle system 7. For example, the computer program can be divided into an acquisition module 61, a detection module 62, an identity recognition module 63, and a setting module 64.

[0075] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium, which can be non-volatile or volatile. The software functional module stored in the storage medium includes several instructions to cause a computer device (which may be a personal computer, computer equipment, or network device, etc.) or processor to perform some functions of the vehicle sentry mode switching method in the various embodiments of this application.

[0076] In summary, this application couples the sentry mode state machine with the identity recognition result, enabling the vehicle to differentiate between protection and alert modes based on the identity categories of trusted / suspicious individuals. This achieves an intelligent security strategy that silently protects familiar individuals while proactively deterring strangers. Compared to existing technologies that use a uniform alarm approach for all approaching individuals, this application effectively filters most familiar approach events, significantly reducing false alarm rates and ensuring users only receive alarms indicating genuine security risks. This improves the practicality of sentry mode and increases user trust. Furthermore, this application employs a layered distance detection and recognition wake-up strategy, activating high-definition image acquisition and identity recognition only when a target enters the first range. In standby mode, it maintains a low-power detection state, eliminating the need for prolonged high-power recording, cloud uploading, and continuous video analysis. This effectively reduces system energy consumption and extends the vehicle's standby time in sentry mode.

[0077] Furthermore, this application introduces a dynamic self-learning trusted identity database update mechanism. Based on historical interaction information between active personnel and vehicles, such as proximity frequency, dwell time, and abnormal behavior records, the trusted identity database is dynamically updated without requiring manual user input, thus reducing user operating costs and improving adaptability in long-term use scenarios. In addition, this application employs a multimodal identity recognition degradation and fusion strategy. When facial recognition is affected by factors such as lighting or occlusion, it can automatically switch to local feature matching, gait recognition, or body contour recognition to ensure the stability and all-weather availability of identity recognition in complex environments. Moreover, this application supports dynamic escalation to an alert state based on abnormal behavior of active personnel in a protected state to prevent malicious use of trusted identities, thereby further improving the security protection capabilities of vehicles in parked states. On the other hand, the identity recognition data and trusted identity database data in this application are processed locally on the vehicle's chip, with only necessary alarm information anonymized before being pushed, thus improving security while further meeting user privacy protection needs.

[0078] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for switching vehicle sentry mode, characterized in that, The switching method includes: when the vehicle's sentry mode is activated: The system acquires real-time environmental information about the vehicle's surroundings and detects the presence of any people or activities in the vicinity of the vehicle based on this information. When no active personnel are detected around the vehicle, the sentry mode is set to standby mode. When a person is detected moving around the vehicle, the person is identified based on the environmental information, and the identification result is compared with a trusted identity database to determine the person's identity category. When the active party is a trusted active party, the sentry mode is set to protected state; When the person in question is a suspicious individual, the sentry mode is set to alert status; wherein the alert level corresponding to the alert status is higher than the alert level corresponding to the protection status.

2. The method for switching vehicle sentry mode according to claim 1, characterized in that, The steps of acquiring real-time information about the vehicle's surrounding environment and detecting the presence of any active individuals around the vehicle based on that information include: Real-time acquisition of environmental information surrounding the vehicle; The environmental information is detected based on a personnel detection algorithm to obtain the target objects around the vehicle; Determine whether the target object meets the preset human characteristic conditions: If the conditions are met, it is determined that there are active people around the vehicle; If the condition is not met, it is determined that there are no active personnel around the vehicle.

3. The method for switching vehicle sentry mode according to claim 1, characterized in that, The steps of identifying the active personnel based on the environmental information and comparing the identification results with a trusted identity database to determine the identity category of the active personnel include: Based on the personnel images in the environmental information, the identities of the active personnel are identified to obtain the identification results; The identity verification result is compared with the trusted identity database: If the identity recognition result matches the trusted identity information in the trusted identity database, then the identity category of the active person is determined as a trusted active person; If the identity recognition result does not match the trusted identity information in the trusted identity database, the identity category of the active person is determined to be a suspicious active person.

4. The method for switching vehicle sentry mode according to claim 3, characterized in that, The steps for identifying the active personnel based on the personnel images in the environmental information and obtaining the identification results include: The lighting conditions of the personnel image are evaluated to determine whether the image meets the preset lighting conditions. If the personnel image meets the lighting conditions, then based on the personnel image in the environmental information, the identity of the active personnel is identified to obtain the identity identification result; If the person image does not meet the lighting conditions, then based on the gait information in the environmental information, gait recognition is performed on the active person, and the identity recognition result is determined based on the gait recognition result.

5. The method for switching vehicle sentry mode according to claim 3, characterized in that, The steps for identifying the active personnel based on the personnel images in the environmental information and obtaining the identification results include: Perform facial occlusion detection on the personnel image: If no facial occlusion is detected, facial recognition is performed on the person image, and the facial recognition result is used as the identity recognition result; If facial occlusion is detected, local feature matching is performed based on the unoccluded local facial region to obtain the corresponding local feature matching result and its local matching confidence. Determine whether the local matching confidence level is greater than or equal to a preset local confidence threshold: If so, the local feature matching result shall be used as the identity recognition result; Otherwise, based on the gait information in the environmental information, gait recognition is performed on the active person, and the identity recognition result is determined based on the gait recognition result.

6. The method for switching vehicle sentry mode according to claim 4 or 5, characterized in that, The steps of performing gait recognition on the active person based on the gait information in the environmental information, and determining the identity recognition result based on the gait recognition result, include: Based on the gait information, gait recognition is performed on the active person to obtain the gait recognition result and its corresponding gait recognition confidence level; If the gait recognition confidence level is greater than or equal to a preset gait confidence level threshold, then the gait recognition result is used as the identity recognition result of the active person. If the gait recognition confidence level is less than the gait confidence level threshold, then based on the body contour information in the environmental information, body contour recognition is performed on the active person to determine the identity recognition result of the active person.

7. The method for switching vehicle sentry mode according to claim 1, characterized in that, When the active party is a trusted active party, the method further includes: Based on the environmental information, the behavior of the active personnel is detected to obtain the behavior category of the active personnel; Determine whether the behavior category matches a preset abnormal behavior category: If a match is found, the sentry mode is set to alert status. Otherwise, the sentinel mode remains in a protected state.

8. The method for switching vehicle sentry mode according to claim 1, characterized in that, The trusted identity database is dynamically updated based on all historical interaction information of the vehicle; wherein, the historical interaction information includes all interaction behavior information between the vehicle and all active personnel collected by the vehicle.

9. A vehicle sentry mode switching system, characterized in that, The system includes: The acquisition module is used to acquire environmental information around the vehicle in real time when the vehicle's sentry mode is activated. The detection module is used to detect whether there are any active people around the vehicle based on the environmental information; The identity recognition module is used to identify the active person based on the environmental information when the presence of an active person is detected in the vicinity of the vehicle, and to compare the identity recognition result with a trusted identity database to determine the identity category of the active person. The setting module is used to set the sentry mode to standby state when no active personnel are detected around the vehicle; to protect state when the active personnel are identified as trustworthy; and to alert state when the active personnel are identified as suspicious. The alert level corresponding to the alert state is higher than the alert level corresponding to the protect state.

10. A vehicle infotainment system, characterized in that, The vehicle infotainment system includes: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the vehicle infotainment system to implement the vehicle sentry mode switching method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by the computer's processor, causes the computer to perform the vehicle sentry mode switching method as described in any one of claims 1 to 8.

Images