In-vehicle retention alarm method, device, equipment, medium and product

Abstract

The invention relates to the technical field of automobile safety prevention and control, and discloses an in-vehicle retention alarm method, device, equipment, medium and product, the method comprises the following steps: responding to vehicle flameout and all door closing signals, after setting delay time, calling getting-on and getting-off records of a vehicle, and judging whether a retention object with life exists or not; if yes, determining the type of the detention object, identifying the state of the detention object, and determining the risk level of the detention object; performing risk monitoring according to the in-vehicle environment data, and identifying an environment risk level; and according to the retention object risk level, in combination with the environmental risk level, carrying out differential alarm. According to the invention, differential alarm is carried out by comprehensively considering the risk of the detained object and the in-vehicle environment risk, so that the risk assessment can better fit the actual endurance capability and safety requirements of different objects, and excessive alarm or insufficient alarm is avoided.

Description

Technical Field

[0001] This invention belongs to the field of automotive safety and control technology, and particularly relates to a method, device, equipment, medium, and product for alarming in-vehicle loitering. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] In-vehicle loitering alarm technology is a key technology for ensuring the safety of life inside the vehicle after it is parked. It collects and analyzes data from inside the cabin through various sensors to identify and warn of potential dangers lurking in the vehicle. It is widely used in the safety systems of various motor vehicles, including passenger cars and commercial vehicles. Existing in-vehicle loitering alarm technologies mainly focus on monitoring children and pets. Their detection principles mostly rely on single infrared sensing or visual recognition methods to capture targets. However, when the target is small, located under the seat, or in an unconventional position such as a seat gap, it is easily obstructed by the seat body, interior parts, or other objects, making it impossible for the sensors to effectively collect the target's thermal characteristics or visual outline, resulting in missed detections and false detections. In addition, existing technologies generally neglect the monitoring and warning of adult loitering scenarios. When occupants are resting in the vehicle, the temperature inside the enclosed cabin can rise rapidly and the oxygen concentration can gradually decrease, which may cause occupants to fall asleep or even become unconscious. Without timely intervention, this can easily lead to safety accidents. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the present invention provides a method, device, equipment, medium and product for vehicle occupancy alarm, which improves the accuracy of occupancy detection and provides graded alarm based on the occupancy status.

[0005] To achieve the above objectives, a first aspect of the present invention provides a vehicle in-vehicle loitering alarm method, comprising the following steps: In response to the vehicle being turned off and all doors being closed, after a set delay time, the vehicle's entry and exit records are retrieved to determine if there are any living individuals remaining on board. If present, determine the type of the detained object, identify the status of the detained object, and determine the risk level of the detained object; Risk monitoring is conducted based on in-vehicle environmental data to identify environmental risk levels. Differential alarms are issued based on the risk level of the stranded individuals and the environmental risk level. The method for updating the boarding and alighting records includes: in response to a signal that any door switches from a closed state to an open state, performing boarding / alighting trajectory detection for a living object on that door, and updating the boarding and alighting records.

[0006] In some embodiments, the method for detecting the boarding trajectory is as follows: After receiving the state switching signal of any door from closed to open, it receives the thermal data of the seat area associated with the door as the initial reference thermal data; It continuously receives radar echo signals from a set range outside the door and the door passage area to identify and track targets. When it detects a continuous change in the trajectory of a target moving from a set area outside the vehicle to the door passage, and the moving speed is stable within a set speed range, the target is marked as a suspected vehicle boarding target. It receives image data of the door area, identifies the number of targets based on a pre-trained target type recognition model, and classifies the targets into adults, children, infants, and pets; Receive real-time infrared thermal data of the associated seat area and compare it with the initial baseline thermal value. When it is detected that the thermal value of the area exceeds the baseline thermal value and reaches the set temperature difference, it is considered that the target has entered the vehicle, and the type of the target and the time of entry are recorded.

[0007] In some embodiments, before receiving the door closing signal, the vehicle entry trajectory detection is continuously performed. After receiving the door closing signal, once the infrared thermal data of the associated seat area has stabilized, the seat position occupied by each target is recorded based on the change between the current thermal data and the reference thermal data.

[0008] In some embodiments, the method for detecting the exit trajectory is as follows: Receive the state switching signal of any door from closed to open, retrieve the entry record of the seat area corresponding to that door, and lock the target who may get off the vehicle; It continuously receives infrared thermal data of the associated seat area, identifies one or more target centroids on the seat based on thermal distribution, and when it detects that a target centroid is moving towards the currently open door and the distance to the door is less than a set threshold, it acquires image data of the door area, identifies the number of targets based on a pre-trained target type recognition model, and classifies the targets into adults, children, infants and pets. It continuously receives radar echo signals from the door passage area and outside the door. When a new moving target is detected, it tracks the trajectory of the moving target. When it detects continuous trajectory changes of the target moving from the door passage to a set area outside the vehicle, and the moving speed is stable within the set speed range, it is considered that the target has disembarked.

[0009] In some embodiments, determining the risk level of a stranded individual includes: The vehicle's entry and exit records are used to determine the number, type, and location of stranded objects, and to determine whether the stranded objects include the driver. If not, infrared thermal data and image data of the corresponding area are obtained. The body temperature and respiratory fluctuations of the detained subjects are extracted based on thermal data; the range of limb movements and eye status of the detained subjects are captured based on image data to identify the state of the detained subjects. The risk level of the detained objects is determined based on their type and status.

[0010] In some embodiments, differentiated alarms include: retrieving data on the location of the object where it is staying; determining an alarm strategy based on the risk level of the object and the environmental risk level; sending linkage instructions to the environmental control module, the in-vehicle alarm module, and / or the external alarm module; and simultaneously sending alarm information to the vehicle's associated contacts. The alarm information includes the identification result of the object where it is staying, as well as the detected images of the object getting on and off the vehicle.

[0011] A second aspect of the present invention provides a vehicle occupancy alarm device, comprising: The loitering object determination module is configured to: in response to the vehicle engine being turned off and all doors being closed, retrieve the vehicle's entry and exit records after a set delay time to determine whether there is a surviving object loitering; the method for updating the entry and exit records includes: in response to a signal that any door is switched from a closed state to an open state, perform entry / exit trajectory detection for the door for which a surviving object is located, and update the entry and exit records accordingly. The detention risk assessment module is configured to: if there is a detained object, determine the type of the detained object, identify the status of the detained object, and determine the risk level of the detained object; The environmental risk assessment module is configured to: monitor risks based on in-vehicle environmental data and identify environmental risk levels; The differentiated alarm module is configured to generate differentiated alarms based on the risk level of the detained object and the environmental risk level.

[0012] A third aspect of the present invention provides an electronic device including a processor and a memory, wherein the memory stores computer instructions that, when executed by the processor, cause the electronic device to perform the method described thereon.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method.

[0014] A fifth aspect of the present invention provides a computer program product comprising a computer program that, when executed by a processor, implements the method described herein.

[0015] One or more of the above technical solutions assess the risk of a loitering object by associating the type and status of the loitering object with the risk level of the in-vehicle environment. This approach overcomes the limitations of existing technologies that only focus on a single object type. Furthermore, it makes risk assessment more aligned with the actual tolerance and safety needs of different objects, avoiding interference caused by excessive alarms in low-risk scenarios and preventing safety hazards caused by insufficient alarms in high-risk scenarios. In particular, the real-time detection and recording of entry / exit trajectories triggered by the door opening signal, and the loitering determination based on these records, effectively avoids missed detections due to obstructions and environmental interference, compared to a full-cabin scan only after the vehicle is turned off. Simultaneously, the determination can be completed quickly by retrieving records, ensuring the timeliness of the alarm. Attached Figure Description

[0016] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0017] Figure 1 This is an example application scenario of the in-vehicle loitering alarm method in this invention embodiment; Figure 2 This is a flowchart of the in-vehicle loitering alarm method in an embodiment of the present invention; Figure 3 This is an overall flowchart of the in-vehicle loitering alarm method in an embodiment of the present invention; Figure 4 This is a diagram of the program module architecture of the vehicle occupancy alarm device in an embodiment of the present invention. Detailed Implementation

[0018] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0019] In the description of the embodiments of this application, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on".

[0020] Figure 1An example environment 100 for the application of the in-vehicle loitering alarm method in one or more embodiments of the present invention is shown. Example environment 100 includes a sensor module, a cockpit domain controller, an environmental control module, an in-vehicle alarm module, and an external alarm module. The sensor module is connected to the cockpit domain controller, which is connected to the environmental control module, the in-vehicle alarm module, and the external alarm module, respectively. The cockpit domain controller receives and analyzes the data collected by the sensor module. Based on the detection data from the sensor module and the analysis results from the cockpit domain controller, it determines the loitering risk level in real time and triggers corresponding alarms and protective operations. After successful condition recognition, it uses a multimodal fusion algorithm based on the input data to derive the target type, status, and environmental risk parameters, thereby driving the environmental control module, the in-vehicle alarm module, and the external alarm module to execute corresponding outputs in real time.

[0021] For example, the sensor module includes an infrared thermal imaging sensor, an environmental perception module, a millimeter-wave radar, an in-door camera, and a door status sensor. The infrared thermal imaging sensor collects thermal infrared data inside the vehicle to obtain the outline, temperature distribution, and dynamic characteristics (such as breathing fluctuations and limb movements) of living organisms within the cabin. It is preferably installed on top of the seats, aligned with the seat surface area to avoid obstruction; it is also positioned directly above the B-pillar to cover key areas such as the front seatbacks, rear seats, and trunk entrance. The environmental perception module collects in-vehicle environmental data, including oxygen concentration and temperature data. Oxygen concentration is collected by an oxygen concentration module located below the center console; temperature is collected by multiple temperature sensors in the front and rear rows, and the average value is taken after removing abnormal measurement points to improve data accuracy. The millimeter-wave radar is used for target trajectory tracking and is installed in the front and rear door areas, both covering a set range outside the vehicle and the door passage. The in-door camera is used for visually identifying target types and is installed inside the vehicle, with a shooting angle covering the door passage and associated seat area. The door status sensor is installed inside the door lock body. It uses a Hall sensor and can directly output a closed or open status signal.

[0022] Figure 2 The diagram illustrates a flowchart of one or more embodiments of the present invention providing an in-vehicle loitering alarm method 200, applied to a cockpit domain controller.

[0023] The method 200 is described in detail below. Specifically, the method 200 includes the following steps: S201: In response to the vehicle engine being turned off and all doors being closed, after a set delay time, determine whether there is a living object remaining. If so, proceed to step S202. S202: Determine the type of the detained object, identify the status of the detained object, and determine the risk level of the detained object; S203: Conduct risk monitoring based on in-vehicle environmental data to identify environmental risk levels; S204: Differentiated alarms will be issued based on the risk level of the detained object and the environmental risk level.

[0024] By associating the type of stranded object with its status, this technology overcomes the limitation of focusing only on a single object type in existing technologies. It also makes risk assessment more aligned with the actual tolerance and safety needs of different objects, avoiding interference caused by excessive alarms in low-risk scenarios and preventing safety hazards caused by insufficient alarms in high-risk scenarios.

[0025] In step S201, the vehicle status is monitored in real time. Upon simultaneously receiving a vehicle ignition shutdown signal and all door closure signals, the system automatically enters standby mode and starts a delay timer (default delay of 3 minutes, adjustable as needed). This delay mechanism effectively prevents false triggering caused by the driver briefly leaving the vehicle (e.g., to retrieve an item). If the vehicle battery voltage drops below 12V during monitoring, the system automatically switches to emergency mode, retaining only core sensing and communication functions to ensure basic early warning capabilities. This delay determination mechanism eliminates non-stationary scenarios such as the driver briefly leaving the vehicle to retrieve an item, reducing the probability of false triggering.

[0026] For example, the method for determining whether there is a living object loitering is as follows: based on the current entry and exit records of the vehicle, it is determined whether there is a living object loitering. The method for updating the entry and exit records includes: in response to a signal that any door switches from a closed state to an open state, performing entry / exit trajectory detection for the door and updating the entry and exit records; in response to a signal that the vehicle is turned off and all doors are closed, after a set delay time, retrieving the current entry and exit records and determining whether there is a living object loitering. The vehicle trajectory detection specifically includes steps S201A - S201F: S201A: After receiving a state transition signal from closed to open from any door, it sequentially sends activation commands to the infrared sensor, millimeter-wave radar, and door-in-cabin camera according to a preset sensor activation sequence. Simultaneously, it receives continuous thermal data from the infrared sensor, representing the seat area pre-associated with that door, over a set time period, calculates the average value of this data, and stores it as the initial baseline thermal value. Specifically, the driver's door corresponds to the driver's seat, the passenger's door corresponds to the passenger's seat, and the right and left rear doors correspond to the entire row of rear seats. For example, the sensor activation sequence is set as follows: infrared sensor activates in 10ms, millimeter-wave radar in 20ms, and door-in-cabin camera in 30ms; the baseline thermal data acquisition time is 5 seconds.

[0027] S201B: Continuously receives echo signals from the millimeter-wave radar over a set range outside the door and in the door passage area. It extracts target feature data such as distance, speed, and contour using a Doppler effect algorithm and uses a Kalman filter algorithm to estimate the target's trajectory in real time. When a target's movement is detected to match continuous coordinate changes from the set area outside the vehicle to the door passage, and its speed remains stable within a set speed range (matching the normal speed of a pedestrian boarding the vehicle), the target is marked as a suspected boarding target, assigned a unique identifier, and its trajectory is continuously tracked. For example, the radar's external monitoring range is 0.5-3m, and the target's stable movement speed range is set to 0.3-1.5m / s.

[0028] S201C: After marking a suspected target boarding the vehicle, it sends a capture command to the camera and receives target frame image data of the door area. The image is processed using a target type recognition model to extract texture features such as the target's head-to-body ratio, limb contours, and clothing or hair, thereby determining whether the target is an adult, child, or pet. This differentiation provides a basis for subsequent differentiated risk assessment.

[0029] S201D: Receives real-time infrared thermal data of the associated seat area and compares it with the initial baseline thermal value. When it detects an area in the thermal data where the thermal value rises above the baseline by a set temperature difference, it determines that the target has boarded the vehicle and records the target's type and the current boarding time. It should be noted that for scenarios such as rear doors where multiple people may board continuously, the thermal value of the associated seat area may be unstable for a considerable period during the alternating boarding process. At this stage, only the number of boarding targets is counted; individual target seat positions are not yet bound to avoid position determination errors due to thermal superposition. For example, the set temperature difference for the thermal value rise is 3°C, and the minimum thermal change area standard for determining a target boarding is set to ≥0.05㎡ (covering the minimum thermal radiation range from small pets to adults), ensuring that different types of targets boarding can be effectively identified.

[0030] S201E: Before receiving a current door closing signal, continuously loops the boarding trajectory detection process, updating the number and type of boarding targets in real time; upon receiving a door closing signal from the door status sensor, pauses the boarding count statistics and enters the seat position determination stage. After the infrared thermal data of the associated seat area stabilizes (the determination criterion is that the thermal value fluctuation range is ≤0.5℃ within 5 consecutive seconds and the thermal area outline does not change significantly), based on the difference between the current stable thermal data and the initial baseline thermal data, associates and records each boarding target type and its seat position, for example, locating two adults on the left and right sides of the rear seat, with the adult on the left holding a child.

[0031] S201F: Integrates the number, type, boarding time, and seat position information determined after the door closes for targets boarding the vehicle, synchronously updates the in-vehicle occupant record, and completes the boarding trajectory detection within the current door opening cycle. If multiple targets board the vehicle consecutively through the same door, the correlation between the trajectory tracking time sequence and the boarding timestamp ensures the correspondence between quantity statistics and type identification; the thermal stability determination mechanism after the door closes effectively solves the problem of thermal superposition interference in continuous boarding scenarios, improving the accuracy of seat position binding.

[0032] Of course, when a child is being held by an adult, the adult may obstruct the child's view, preventing the in-vehicle infrared sensors from recognizing the child. In this case, the image captured by the camera will be used as the standard, and only the type of child and the time of boarding will be recorded to retain the tracking record of the target.

[0033] The specific steps for exit trajectory detection include S201G - S201K: S201G: Receives the state switching signal of any door from closed to open, and sequentially activates the infrared sensor, millimeter-wave radar, and door-mounted camera in the corresponding area according to a preset timing sequence; simultaneously, it retrieves the entry record of the corresponding seat area of ​​that door to identify potential exit targets. Understandably, since there is some space in the rear seats, regardless of whether the left or right rear door is opened, the entry record of the entire rear seat is retrieved.

[0034] S201H: Continuously receives infrared thermal data of the associated seat area, identifies the centroid of one or more targets on the seat based on the thermal distribution, and when it detects that the centroid of a target is moving towards the currently open door and the distance to the door is less than a set threshold, such as 0.2m, it is considered that the target intends to get off the vehicle, and a capture command is sent to the camera inside the door. The camera focuses on the door passage and the area between the seat and the door, captures a clear frame image of the target, and identifies the target type. S201I: Simultaneously, it continuously receives echo signals from the door passage area and the 0.5-3m range outside the door transmitted by the millimeter-wave radar. It extracts target distance, speed, and contour data through the Doppler effect algorithm and calls the Kalman filter algorithm to track the target's trajectory in the door passage area in real time. When the radar detects a new moving target in the door passage area, and the target's trajectory conforms to the coordinate change of continuous movement from the inside of the door to the outside, and the moving speed is stable in the normal exit speed range of 0.3-1.5m / s, it is considered that the target has exited the vehicle.

[0035] S201J: Upon receiving a door closing signal, it acquires stable thermal data between the current door opening and the previous door closing as initial baseline thermal data. Once the infrared thermal data of the associated seat area stabilizes, the two sets of infrared thermal data are compared to verify whether the changes in thermal data are consistent with the identification result of the exiting target. The condition for judging the stability of thermal data is that the fluctuation amplitude is ≤0.5℃ for 5 consecutive seconds and the outline of the thermal area no longer changes significantly.

[0036] S201K: If multiple targets get off the vehicle at the same door in succession, the controller distinguishes the trajectory of each target by the unique identifier of the target. It combines the body shape characteristics of visual recognition with the location range division of radar monitoring (for example, using the differences in the outline and movement speed between adults and children) to ensure that the trajectory of each target is accurately bound to itself, avoiding the problem of missed or miscounted targets.

[0037] The aforementioned vehicle entry and exit trajectory detection integrates multimodal data from millimeter-wave radar, infrared sensors, and cameras. By tracking the external trajectory and observing changes in the infrared thermal map of the seats inside the vehicle, it judges the behavior of the target getting on and off the vehicle. During the process of getting on and off the vehicle, machine vision is used to distinguish between different types of objects such as adults, children, and pets. Through the aforementioned vehicle entry and exit trajectory detection, it is possible to more accurately identify whether there are people lingering, effectively solving the problem that existing lingering monitoring may miss small targets such as infants and pets due to obstruction.

[0038] The target type recognition model uses a lightweight YOLOv8 architecture, adding an attention mechanism module after the backbone network module to strengthen the extraction weights of key features such as head-to-body ratio and hair texture. The classification branch for detected heads is changed from a general classification to a specific four-category classification (adult / child / infant / pet), and the loss function is optimized. To improve the model's generalization ability, images of adults, children, and common pets getting in and out of vehicles under different weather and lighting conditions (i.e., images of the door passage area) are collected from cameras inside the car doors. Supplementary datasets include human samples from MSCOCO and pet samples from PASCAL VOC. The samples are divided into training, validation, and test sets in an 8:1:1 ratio and labeled with target category labels.

[0039] In step S202, the following steps are specifically performed: S202A: Retrieve the vehicle entry / exit trajectory data recorded in step S201 to determine the number, type, and location of stranded objects. Determine if the stranded objects include the driver. If not, proceed to step 2. By reusing existing vehicle entry / exit trajectory data, the core information of stranded objects can be quickly locked, avoiding resource waste caused by repeated detection. Identifying whether it is the driver is to exclude scenarios where the driver actively stays, such as when the driver turns off the engine and takes a short rest in the driver's seat. The driver has the ability to autonomously adjust the in-vehicle environment (such as turning on the air conditioner and lowering the windows for ventilation) and cope with risks, thus avoiding false alarms.

[0040] S202B: Acquires infrared thermal and image data of the corresponding area. Based on the thermal data, it extracts physiological characteristics such as body temperature and respiratory fluctuations of the detained object. Based on the image data, it captures the range of limb movement and eye status of the detained object to identify its state. Specifically, this state includes conscious, drowsy, or comatose states. If the vehicle battery voltage is found to be too low during monitoring, the system will automatically simplify the data processing flow, retaining only the core infrared data for analysis, prioritizing the basic capabilities of personnel state identification and driver judgment. For example, in respiratory rate monitoring, the fluctuation range for conscious state can be set to ±2 breaths / minute, for drowsy state to ±5 breaths / minute, and for comatose state to ≤±1 breaths / minute or respiratory arrest exceeding 10 seconds; regarding limb movement, drowsy state can be defined as a reduction of more than 50% in limb movement amplitude and prolonged eye closure time. By employing the fusion of infrared and image multimodal data, the system achieves health status identification of the detained object, improving the accuracy of identification.

[0041] S202C: Determine the risk level of the stranded individuals based on their type and condition. Risk level classification is based on the target's tolerance and the degree of health urgency, combined with a context-defined priority gradient and physiological characteristic standards to establish a multi-level risk level system, and clarifies the priority determination rules when multiple targets coexist. For example, a basic risk level is set based on the type of stranded individual, with infants, children, pets, and adults assigned priority gradients from high to low as highest level, high level, upper-middle level, and basic level, respectively, to accommodate different individuals' tolerance to the harsh in-vehicle environment. Then, the risk level is adjusted based on the stranded individual's condition. If the individual is found to be unconscious, the risk level is directly increased by two levels; if the individual is drowsy, the risk level is increased by one level; only conscious individuals maintain the basic risk level. In the case of multiple stranded individuals, the individual with the highest risk level is used as the benchmark for determining the overall risk level of the stranded individuals, ensuring that high-risk targets receive priority intervention. By establishing a multi-level risk level system, priority intervention is ensured for individuals with weak tolerance and critical health conditions, maximizing the safety of high-risk targets.

[0042] Table 1 Mapping between Detainee Type, Status and Detainee Risk Level

[0043] In step S203, the system receives real-time in-vehicle environmental data uploaded by the environmental perception module and performs quantitative analysis on temperature and oxygen concentration data to identify the environmental risk level. Temperature data is collected from multiple temperature sensors in the front and rear rows; after removing outliers using the Grubbs criterion, the average value is taken as the current in-vehicle temperature. Oxygen concentration data is collected continuously from the oxygen concentration module below the center console, and the average value over 10 seconds is taken to improve accuracy. The risk level is then determined in conjunction with a preset four-level risk standard. The controller updates the environmental data and risk level every 30 seconds. If the risk level is detected to have increased by two levels consecutively within 5 minutes (e.g., from low risk to high risk), it is automatically marked as a rapidly deteriorating environment, and a warning is sent to subsequent alarm steps. If the battery voltage is below 12V, the data processing flow is simplified, and only data from the main temperature sensor and oxygen concentration sensor are used for risk assessment.

[0044] Table 2 Environmental Risk Level Mapping Standards

[0045] In step S204, the location data of the object being stranded is retrieved. Based on the risk level of the stranded object and the environmental risk level, the alarm level is determined, and a linkage command is sent to the environmental control module, the in-vehicle alarm module, and / or the external alarm module. The environmental control module controls ventilation, including air conditioning and windows. The in-vehicle alarm module alerts the target inside the vehicle through sounds, lights, and the central control screen. The external alarm module includes hazard lights and speakers to call for help from the surrounding area.

[0046] At the same time, alarm information is sent to the contacts associated with the vehicle. The alarm information includes the identification results of the stranded object, as well as the images of the detected object getting on and off the vehicle, to help determine whether it is a false alarm.

[0047] For example, alarm intensity is matched based on the risk level of the stranded individual and the environmental risk level. Low-intensity alarm strategies include information, telephone alarms, and / or in-vehicle alerts. Medium and high-intensity alarm strategies implement ventilation control and / or external alarms in addition to the low-intensity alarm strategies. Extreme-intensity alarm strategies implement at least one of the following in addition to the high-intensity alarm strategies: unlocking vehicle doors and contacting emergency medical services. It can be understood that the higher the alarm intensity, the more people and the more frequently information and telephone alarms are associated, the greater the ventilation volume controlled, and the louder the alarm sound. Differentiated alarm strategies are shown in Table 3. In particular, the specific configurations of ventilation control and in-vehicle and external alarms can be differentiated for different types of stranded individuals.

[0048] Table 3 Mapping Relationship between Risk Level of Detained Entities, Environmental Risk Level, and Alarm Strategy

[0049] Figure 4 The diagram shows a program module diagram of an in-vehicle loitering alarm device 300, including: a loitering object judgment module 301, configured to: in response to a vehicle engine shutdown and all door closure signals, retrieve the vehicle's entry and exit records after a set delay time to determine whether a surviving object is loitering; the method for updating the entry and exit records includes: in response to a signal that any door switches from a closed state to an open state, perform entry / exit trajectory detection for a surviving object at that door, and update the entry and exit records; a loitering risk judgment module 302, configured to: if a loitering object exists, determine the type of loitering object, identify the state of the loitering object, and determine the risk level of the loitering object; an environmental risk judgment module 303, configured to: perform risk monitoring based on in-vehicle environmental data and identify the environmental risk level; and a differentiated alarm module 304, configured to: perform differentiated alarms based on the risk level of the loitering object and the environmental risk level.

[0050] Furthermore, one or more embodiments of the present invention also provide an electronic device that can be used to implement the in-vehicle loitering alarm method described in the above embodiments. The electronic device includes one or more processors, one or more memories coupled to the processors, and a communication module coupled to the processors.

[0051] The memory may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, at least one of the following: read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, hard disk, compact disc (CD), digital video disc (DVD), or other magnetic and / or optical storage. Examples of volatile memories include, but are not limited to, at least one of the following: random access memory (RAM), or other volatile memories that do not persist during the power outage duration. The computer program may be stored in the ROM. When the processor executes the computer program, it implements the above-described in-vehicle loitering alarm method.

[0052] In some embodiments, the program may be tangibly contained in a computer-readable medium, which may include a device (such as in memory) or other storage device accessible by the device. The program may be loaded from the computer-readable medium into RAM for execution. The computer-readable medium may include any type of tangible non-volatile memory, such as ROM, EPROM, flash memory, hard disk, whereby the computer-readable storage medium stores a computer program that, when executed by a processor, implements the aforementioned in-vehicle loitering alarm method.

[0053] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a server or terminal, they generate, in whole or in part, the processes or functions described in the embodiments of this application. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic cable, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to the server or terminal, or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, and magnetic tape), an optical medium (e.g., digital video disk (DVD), etc.), or a semiconductor medium (e.g., solid-state drive).

[0054] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0055] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for alarming in-vehicle occupancy, characterized in that, The method comprises the following steps: In response to a vehicle engine-off and all doors closed signal, after a set delay time, the vehicle's boarding and alighting records are called to determine whether there is a living remaining object; If there is, the type of the remaining object is determined, and the state of the remaining object is identified to determine the risk level of the remaining object; Risk monitoring is performed according to the in-vehicle environment data to identify the environmental risk level; According to the risk level of the remaining object, combined with the environmental risk level, a differentiated alarm is performed. The updating method of the boarding and alighting records comprises: in response to a signal that any door switches from a closed state to an open state, performing a living object boarding track / alighting track detection for the door, and updating the boarding and alighting records.

2. The in-vehicle occupancy notification method according to claim 1, wherein The boarding track detection method is: After receiving a signal that any door switches from a closed state to an open state, receive the thermal data of the seat area associated with the door as the initial baseline thermal data; Continuously receive radar echo signals in a set range outside the door body and in the door body passage area to identify and track the target; when detecting a continuous track change of the target moving from the set area outside the vehicle to the door body passage, and the moving speed is stable in the set speed interval, mark the target as a suspected boarding target; Receive image data of the door body area, identify the number of targets based on a pre-trained target type identification model, and distinguish the targets into adults, children, infants, and pets; Receive real-time infrared thermal data of the associated seat area, compare it with the initial baseline thermal data; when detecting that the thermal data in the area exceeds the set area by a set temperature difference, it is considered that the target has boarded, and the type and boarding time of the target are recorded.

3. The in-vehicle occupancy notification method according to claim 2, wherein Before receiving the door closed signal, continuously perform the boarding track detection; after receiving the door closed signal, when the infrared thermal data of the associated seat area is stable, record the seat positions of each target based on the changes between the current thermal data and the baseline thermal data.

4. The in-vehicle occupancy notification method according to claim 2, wherein The alighting track detection method is: Receive a signal that any door switches from a closed state to an open state, call the boarding records of the seat area corresponding to the door, and lock the target that may alight; Continuously receive infrared thermal data of the associated seat area, identify the center of mass of one or more targets on the seat based on the thermal distribution, and when monitoring that the center of mass of the target moves towards the currently open door and the distance from the door is less than a set threshold, acquire image data of the door body area, identify the number of targets based on a pre-trained target type identification model, and distinguish the targets into adults, children, infants, and pets; Continuously receive radar echo signals in the door body passage area and outside the door; when a new moving target is identified, track the track of the moving target; when detecting a continuous track change of the target moving from the door body passage to the set area outside the vehicle, and the moving speed is stable in the set speed interval, it is considered that the target has alighted.

5. The in-vehicle occupancy notification method according to claim 1, wherein Determining the risk level of the remaining object comprises: The boarding and alighting records of the vehicle determine the number, type, and position of the remaining object, determine whether the remaining object includes the driver, if not, acquire the infrared thermal data and image data of the corresponding area; The body temperature and respiratory fluctuation of the stranded object are extracted based on thermal data, and the limb activity amplitude and eye state of the stranded object are captured based on image data to identify the state of the stranded object. The risk level of the stranded object is determined according to the type and state of the stranded object.

6. The in-vehicle occupancy notification method according to claim 5, wherein The differentiated alarm includes: calling the stranded object position data, determining the alarm strategy according to the stranded object risk level and the environmental risk level, sending the linkage instruction to the environmental control module, the in-vehicle alarm module and / or the out-of-vehicle alarm module, and sending the alarm information to the contact associated with the vehicle, wherein the alarm information includes the stranded object identification result and the detected boarding object and alighting object images.

7. An in-vehicle dwell alarm apparatus characterized by comprising: The method comprises: The stranded object judgment module is configured to, in response to the vehicle engine off and all doors closed signal, call the boarding and alighting record of the vehicle after a set delay time, and judge whether there is a living stranded object; the updating method of the boarding and alighting record comprises: in response to the signal that any door is switched from the closed state to the open state, performing the boarding track / alighting track detection of the living object for the door, and updating the boarding and alighting record; The stranded risk judgment module is configured to, if there is a stranded object, determine the stranded object type, identify the state of the stranded object, and determine the stranded object risk level; The environmental risk judgment module is configured to monitor the risk according to the in-vehicle environmental data and identify the environmental risk level; The differentiated alarm module is configured to perform differentiated alarm according to the stranded object risk level and the environmental risk level.

8. An electronic device comprising a processor and a memory, said memory having stored thereon computer instructions, characterized in that, When the computer instructions are executed by the processor, the electronic device executes the method of any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1 to 6.

Images