Abnormality detection method and device

Abstract

The invention discloses an anomaly detection method and device, and relates to the technical field of intelligent monitoring. Wherein when the anomaly detection device detects an abnormal event based on wireless data collected by the wireless sensor, the image sensor is started, and the anomaly detection device monitors the target space through the image sensor. Wherein the wireless sensor and the image sensor are both arranged in a target space. The environment of the target space is detected through the wireless sensor, when the abnormal event is detected, the image sensor is started, and the target space is monitored through the image sensor; when there is no abnormal event, the image sensor does not need to be started. The power consumption of the wireless sensor is very low, so that compared with monitoring by using an image sensor for a long time, the anomaly detection method provided by the invention can greatly reduce the power consumption.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology, and in particular to an anomaly detection method and device. Background Technology

[0002] Currently, in intelligent monitoring scenarios, cameras are typically used to monitor the surrounding environment. For example, in vehicle monitoring scenarios, to ensure vehicle security and prevent vandalism, theft, or other accidental damage after the vehicle is parked, the vehicle's cameras can remain on after the vehicle is locked and parked, recording suspicious events around the vehicle. Typically, four 180° high-definition ultra-wide-angle cameras can be installed around the vehicle to monitor the surrounding environment and vehicle status in real time.

[0003] However, after the vehicle is locked and parked, the vehicle's camera remains on and recording, resulting in high power consumption. This can lead to significant energy drain when the vehicle is parked for extended periods. The issue of high power consumption from cameras operating for extended periods also exists in other monitoring scenarios, such as smart homes or security surveillance. Summary of the Invention

[0004] This application provides an anomaly detection method and apparatus, which can reduce the power consumption of the anomaly detection apparatus.

[0005] Firstly, this application provides an anomaly detection method. This communication method is executed by an anomaly detection device, or by a chip, chip system, or circuit within the anomaly detection device. The anomaly detection method may include: when the anomaly detection device detects an anomaly based on wireless data collected by a wireless sensor, it activates an image sensor, and the anomaly detection device monitors the target space through the image sensor. Both the wireless sensor and the image sensor are located within the target space.

[0006] The anomaly detection method provided in this application detects the environment of a target space using a wireless sensor. When an anomaly is detected, an image sensor is activated to monitor the target space; otherwise, the image sensor does not need to be activated. Because the wireless sensor has very low power consumption, the anomaly detection method provided in this application can significantly reduce power consumption compared to using an image sensor for monitoring for extended periods.

[0007] In one alternative implementation, the anomaly detection device acquires wireless data collected by a wireless sensor and processes the wireless data using a first network model to detect the presence of anomalies.

[0008] In one alternative implementation, after the image sensor is activated, the anomaly detection device can determine an anomaly judgment result based on the video data acquired by the image sensor, and assign a first sample label to the wireless data based on the anomaly judgment result. The wireless data and the first sample label are used to retrain the first network model.

[0009] In the above implementation, the anomaly detection result is determined by video data collected by the image sensor, and the anomaly detection result is used as feedback to retrain the first network model, which can improve the accuracy of detecting abnormal events based on wireless data.

[0010] In one alternative implementation, the anomaly detection device can determine that an anomaly has been detected based on a comparison between wireless data and a data threshold. For example, the wireless data can be used to characterize the phase and amplitude of an electromagnetic wave signal received by a wireless sensor. The anomaly detection device can determine that an anomaly has been detected by a combination of at least one or more of the following methods: if the wireless data is greater than or equal to a first data threshold, then an anomaly has been detected; or, if the amplitude of the electromagnetic wave signal is greater than or equal to a second data threshold, then an anomaly has been detected; or, if the change in the wireless data over a set period of time is greater than or equal to a third data threshold, then an anomaly has been detected; or, if the change in the phase of the electromagnetic wave signal over a set period of time is greater than or equal to a fourth data threshold, then an anomaly has been detected; or, if the change in the amplitude of the electromagnetic wave signal over a set period of time is greater than or equal to a fifth data threshold, then an anomaly has been detected.

[0011] In the above implementation, wireless data is compared with a data threshold to determine whether an abnormal event has been detected. This method requires less computation and can save computing resources.

[0012] In one alternative implementation, the anomaly detection device can determine that an anomaly event has been detected based on a comparison between parameters determined by wireless data and parameter thresholds. For example, the parameters determined by wireless data may include perturbation values ​​at multiple locations in the Doppler spectrum. The anomaly detection device can determine that an anomaly event has been detected in any of the following ways: if there is a first location in the Doppler spectrum with a perturbation value greater than or equal to a first parameter threshold, and the distance between the first location and the wireless sensor is within a set distance range, then an anomaly event is detected; or, if there is a second location in the Doppler spectrum with a perturbation value greater than or equal to a second parameter threshold, and the distance between the second location and the wireless sensor gradually decreases, then an anomaly event is detected; or, if there is a third location in the Doppler spectrum with a perturbation value greater than or equal to a third parameter threshold, and the movement speed of the third location is greater than or equal to a set speed threshold, then an anomaly event is detected.

[0013] In the above implementation, the parameters determined by wireless data are compared with the parameter threshold to determine whether an abnormal event has been detected. The computational load is relatively small, which can save computing resources.

[0014] In one alternative implementation, after the image sensor is turned on, the anomaly detection device can determine an anomaly judgment result based on the video data collected by the image sensor; set a first sample label for the wireless data based on the anomaly judgment result; and use the wireless data and the first sample label to adjust the data threshold or parameter threshold.

[0015] In the above implementation, the anomaly detection result is determined by video data collected by the image sensor. The anomaly detection result is used as feedback to adjust the data threshold or parameter threshold of the wireless sensing module, which can improve the accuracy of detecting abnormal events based on wireless data.

[0016] In one optional implementation, if the number of wireless data with the first sample label being a negative sample label is greater than or equal to a first quantity threshold, or if the ratio of the number of wireless data with the first sample label being a negative sample label to the number of wireless data with the first sample label being a positive sample label is greater than or equal to a first ratio threshold, the data threshold or parameter threshold is increased.

[0017] If, within a first set time period, the number of wireless data with the first sample label being a negative sample label is greater than or equal to a first quantity threshold, or if the ratio of the number of wireless data with the first sample label being a negative sample label to the number of wireless data with the first sample label being a positive sample label is greater than or equal to a first proportion threshold, it indicates that the wireless sensing module is making many misjudgments and is triggering false alarms in many cases. In this case, the data threshold or parameter threshold can be increased to reduce false alarms.

[0018] In one optional implementation, a non-wireless sensor is also provided in the target space. When the non-wireless sensor triggers an abnormal alarm and no abnormal event is detected based on the wireless data, the abnormal detection device can save the wireless data as non-alarm wireless data. If the number of non-alarm wireless data is greater than or equal to a second quantity threshold within a second set time period, or if the ratio of the number of non-alarm wireless data to the number of wireless data with the first sample tag is greater than or equal to a second ratio threshold, the data threshold or parameter threshold is lowered.

[0019] If, within the second set time period, the number of non-alarm wireless data is greater than or equal to the second quantity threshold, or if the ratio of the number of non-alarm wireless data to the number of wireless data with the first sample tag is greater than or equal to the second proportion threshold, it indicates that the wireless sensing module has many missed alarms and has not triggered an alarm when there is an anomaly. In this case, the data threshold or parameter threshold can be lowered to reduce missed alarms.

[0020] In one optional implementation, the anomaly detection device can transmit video data to a terminal device and receive the anomaly determination result returned by the terminal device. Upon receiving the video data from the anomaly detection device, the terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists. User-based judgment reduces the risk of misjudgments caused by instrument-based analysis.

[0021] In one alternative implementation, the anomaly detection device can process the video data using a second network model to obtain anomaly determination results.

[0022] In one alternative implementation, the anomaly detection device can process the video data using a second network model to determine the presence of anomaly targets in the video data; the judgment result output by the second network model is then used as the anomaly judgment result.

[0023] In one alternative implementation, the anomaly detection device can process the video data using a second network model to determine that there are no abnormal targets in the video data; transmit the video data to the terminal device; and receive the anomaly determination result returned by the terminal device.

[0024] In the above implementation, the second network model first makes a judgment. If it is determined that there is no abnormal target in the video data, the video data is then transmitted to the terminal device for user judgment. If it is determined that there is an abnormal target in the video data, the user does not need to make a judgment, which can reduce the user's workload.

[0025] In one alternative implementation, the anomaly detection device can determine a second sample label for the video data based on the anomaly determination result; the video data and the second sample label are used to retrain the second network model.

[0026] In the above implementation, the user's anomaly judgment result is used as feedback to retrain the second network model, which can improve the accuracy of the second network model in judging abnormal events.

[0027] In one alternative implementation, the wireless sensor is a narrowband signal-based wireless sensor. For example, the wireless sensor includes at least one of a Bluetooth sensor or a starburst sensor.

[0028] In one alternative implementation, the anomaly detection device can transmit narrowband signals on multiple channels using wireless sensors and receive echo signals, obtaining wireless data based on the received multiple sets of echo signals.

[0029] In one alternative implementation, the anomaly detection device can perform channel fusion on multiple sets of received echo signals to obtain a frequency domain signal; and convert the frequency domain signal from the frequency domain to the time domain to obtain wireless data.

[0030] Because the electromagnetic waves transmitted and received by wireless sensors based on narrowband signals are narrowband signals, the accuracy of the wireless data obtained using narrowband signals is low, and the resolution of the plotted spectrum is low, generally considered unsuitable for monitoring changes in the surrounding environment. However, this application provides a method for monitoring environmental changes using a wireless sensor based on narrowband signals. The wireless sensor of this application can transmit narrowband signals on multiple channels and receive echo signals. These multiple channels correspond to multiple frequency bands. The wireless sensor can receive multiple sets of echo signals on multiple channels and transmit these multiple sets of echo signals to an anomaly detection device. The anomaly detection device can perform channel fusion on the received multiple sets of echo signals. The signal obtained after channel fusion is equivalent to a broadband signal, which has higher accuracy and higher resolution in the plotted spectrum, effectively reflecting changes in the surrounding environment. Furthermore, the wireless sensor can transmit narrowband signals on multiple channels, resulting in less interference to the surrounding environment.

[0031] In one alternative implementation, wireless sensors and image sensors are mounted on the vehicle.

[0032] Secondly, this application provides an anomaly detection device, which may include:

[0033] The first monitoring unit can be used to activate the image sensor when an abnormal event is detected by wireless data collected by the wireless sensor, wherein both the wireless sensor and the image sensor are located in the target space.

[0034] The second monitoring unit can be used to monitor the target space using an image sensor.

[0035] In one alternative implementation, the first monitoring unit can be used to: acquire wireless data collected by wireless sensors, and process the wireless data through a first network model to detect whether there are any abnormal events.

[0036] In one alternative implementation, the anomaly detection device may further include a parameter adjustment unit for determining anomaly detection results based on video data acquired by an image sensor, and setting a first sample label for the wireless data based on the anomaly detection results. The wireless data and the first sample label are used to retrain the first network model.

[0037] In one alternative implementation, the first monitoring unit may specifically be used to determine that an abnormal event has been detected based on a comparison between wireless data and a data threshold. For example,

[0038] In one alternative implementation, wireless data can be used to characterize the phase and amplitude of electromagnetic wave signals received by the wireless sensor. Specifically, the first monitoring unit can be used to determine that an abnormal event has been detected by a combination of at least one or more of the following methods: if the wireless data is greater than or equal to a first data threshold, then an abnormal event is detected; or, if the amplitude of the electromagnetic wave signal is greater than or equal to a second data threshold, then an abnormal event is detected; or, if the change in the wireless data over a set time period is greater than or equal to a third data threshold, then an abnormal event is detected; or, if the change in the phase of the electromagnetic wave signal over a set time period is greater than or equal to a fourth data threshold, then an abnormal event is detected; or, if the change in the amplitude of the electromagnetic wave signal over a set time period is greater than or equal to a fifth data threshold, then an abnormal event is detected.

[0039] In one alternative implementation, the first monitoring unit may specifically be used to determine that an abnormal event has been detected based on a comparison between parameters determined by wireless data and parameter thresholds.

[0040] In one optional implementation, the parameters determined by the wireless data may include perturbation values ​​at multiple locations in the Doppler spectrum; the first monitoring unit may specifically be used to determine that an abnormal event has been detected by any of the following methods: if there is a first location in the Doppler spectrum with a perturbation value greater than or equal to a first parameter threshold, and the distance between the first location and the wireless sensor is within a set distance range, then an abnormal event is determined to be detected; or, if there is a second location in the Doppler spectrum with a perturbation value greater than or equal to a second parameter threshold, and the distance between the second location and the wireless sensor gradually decreases, then an abnormal event is determined to be detected; or, if there is a third location in the Doppler spectrum with a perturbation value greater than or equal to a third parameter threshold, and the moving speed of the third location is greater than or equal to a set speed threshold, then an abnormal event is determined to be detected.

[0041] In one alternative implementation, the parameter adjustment unit can be specifically used to: determine anomaly detection results based on video data acquired by the image sensor after the image sensor is turned on; set a first sample label for the wireless data based on the anomaly detection results; and use the wireless data and the first sample label to adjust the data threshold or parameter threshold.

[0042] In one optional implementation, the parameter adjustment unit can be specifically used to: increase the data threshold or parameter threshold when the number of wireless data with the first sample label being a negative sample label is greater than or equal to a first quantity threshold, or the ratio of the number of wireless data with the first sample label being a negative sample label to the number of wireless data with the first sample label being a positive sample label is greater than or equal to a first ratio threshold, within a first set time period.

[0043] In one optional implementation, a non-wireless sensor is also provided in the target space. The parameter adjustment unit can be used to: when the non-wireless sensor triggers an abnormal alarm and no abnormal event is detected based on the wireless data, the abnormal detection device can save the wireless data as non-alarm wireless data; if the number of non-alarm wireless data is greater than or equal to a second quantity threshold within a second set time period, or if the ratio of the number of non-alarm wireless data to the number of wireless data with the first sample tag is greater than or equal to a second ratio threshold, the data threshold or parameter threshold is lowered.

[0044] In one optional implementation, the parameter adjustment unit can specifically be used to: transmit video data to the terminal device and receive the anomaly determination result returned by the terminal device. Specifically, upon receiving the video data from the anomaly detection device, the terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists.

[0045] In one alternative implementation, the parameter adjustment unit can be used to process video data through a second network model to obtain anomaly detection results.

[0046] In one alternative implementation, the parameter adjustment unit can be used to: process the video data through the second network model to determine the presence of abnormal targets in the video data; and use the judgment result output by the second network model as the anomaly judgment result.

[0047] In one alternative implementation, the parameter adjustment unit can be used to: process the video data through a second network model to determine that there are no abnormal targets in the video data; transmit the video data to the terminal device and receive the abnormality determination result returned by the terminal device.

[0048] In one alternative implementation, the parameter adjustment unit can be used to: enable the anomaly detection device to determine a second sample label for the video data based on the anomaly determination result; and use the video data and the second sample label to retrain the second network model.

[0049] In one alternative implementation, the wireless sensor is a narrowband signal-based wireless sensor. For example, the wireless sensor includes at least one of a Bluetooth sensor or a starburst sensor.

[0050] In one alternative implementation, the first monitoring unit can be used to: transmit narrowband signals on multiple channels via wireless sensors and receive echo signals, and obtain wireless data based on the received multiple sets of echo signals.

[0051] In one alternative implementation, the first monitoring unit can be used to: perform channel fusion on multiple sets of received echo signals to obtain a frequency domain signal; and convert the frequency domain signal from the frequency domain to the time domain to obtain wireless data.

[0052] In one alternative implementation, wireless sensors and image sensors are mounted on the vehicle.

[0053] Thirdly, this application also provides an anomaly detection device, which may include a processor and a memory; the memory stores a computer program; the processor is used to read the computer program stored in the memory and execute any of the anomaly detection methods provided in the first aspect above.

[0054] Fourthly, this application also provides a chip that may include a processor and a memory; the memory stores a computer program; the processor is used to read the computer program stored in the memory and execute any of the anomaly detection methods provided in the first aspect above.

[0055] Fifthly, this application provides a computer-readable storage medium storing computer-executable instructions, which are used to cause a computer to perform any of the anomaly detection methods provided in the first aspect above.

[0056] Sixthly, this application provides a computer program product comprising computer-executable instructions, the computer-executable instructions being used to cause a computer to execute any of the anomaly detection methods provided in the first aspect above.

[0057] The technical effects that can be achieved by any of the second to sixth aspects mentioned above can be referred to the description of the beneficial effects in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0058] Figure 1 This is a schematic diagram illustrating one application scenario of an embodiment of this application;

[0059] Figure 2 This is a schematic diagram illustrating another application scenario of an embodiment of this application;

[0060] Figure 3 A flowchart illustrating an anomaly detection method provided in this application embodiment;

[0061] Figure 4 An external schematic diagram of a vehicle provided for an embodiment of this application;

[0062] Figure 5 A schematic diagram of an RD spectrum provided for an embodiment of this application;

[0063] Figure 6A flowchart of another anomaly detection method provided in the embodiments of this application;

[0064] Figure 7 An interactive diagram illustrating an anomaly detection method provided in an embodiment of this application;

[0065] Figure 8 A flowchart of another anomaly detection method provided in the embodiments of this application;

[0066] Figure 9 An interaction diagram for another anomaly detection method provided in an embodiment of this application;

[0067] Figure 10 This is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0068] Figure 11 This is a schematic diagram of another communication device provided in an embodiment of this application;

[0069] Figure 12 This is a schematic diagram of another communication device provided in an embodiment of this application. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the embodiments of this application will be described in detail below with reference to the accompanying drawings. The terminology used in the implementation section of this application is only for explaining specific embodiments of this application and is not intended to limit this application.

[0071] Before introducing the specific solutions provided in the embodiments of this application, some terms used in this application will be explained to facilitate understanding by those skilled in the art, but the terms used in this application are not limited.

[0072] (1) Wireless Sensor Based on Narrowband Signals: A wireless sensor is a device with wireless local area network (WLAN) sensing capabilities, which can use received electromagnetic wave signals to infer information such as the position and movement status of objects or people in the environment. In the embodiments of this application, the wireless sensor based on narrowband signals can transmit and receive electromagnetic wave signals in multiple frequency bands. The range of each frequency band is small. For example, each frequency band can be less than or equal to 1MHz, hence the name narrowband signal. It can even be said that the device can transmit and receive electromagnetic wave signals at multiple frequency points, each frequency point representing a very small frequency band range. Among them, multiple discrete frequency bands or multiple discrete frequency points are located within the operating frequency band of the wireless sensor. For example, the operating frequency band of the wireless sensor is 5150MHz~5350MHz, and the wireless sensor can transmit and receive electromagnetic wave signals in 200 discrete frequency bands such as 5150MHz~5151MHz, 5151MHz~5152MHz, 5152MHz~5153MHz, and 5153MHz~5154MHz.

[0073] (2) Bluetooth Sensor: A sensor based on Bluetooth wireless sensing technology, which, based on Bluetooth technology and possessing WLAN sensing capabilities, can monitor changes in the surrounding environment by detecting changes in received electromagnetic wave signals. Bluetooth technology is a short-range wireless communication technology, including Bluetooth Classic and Bluetooth Low Energy (BLE) technologies. BLE technology is characterized by low power consumption and supports longer standby time. Bluetooth sensors can adopt BLE technology, which not only has low power consumption but also helps to detect changes in electromagnetic wave signals in a shorter time, thus sensing changes in the surrounding environment more promptly.

[0074] (3) SparkLink Sensor: A sensor based on SparkLink wireless sensing technology, a device based on SparkLink technology and possessing WLAN sensing capabilities. SparkLink access technology can include SparkLink Basic (SLB) technology and SparkLink Low Energy (SLE) technology. SLE technology is characterized by low power consumption, and the SparkLink sensor can adopt SLE technology. The SparkLink sensor can also monitor changes in the surrounding environment by detecting changes in the received electromagnetic wave signals.

[0075] In this application embodiment, "multiple" refers to two or more. Therefore, in this application embodiment, "multiple" can also be understood as "at least two". "At least one" can be understood as one or more, such as one, two, or more. For example, "including at least one" means including one, two, or more, and it does not limit which ones are included. For example, including at least one of A, B, and C, then it could include A, B, C, A and B, A and C, B and C, or A and B and C. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / ", unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0076] Unless otherwise stated, the ordinal numbers such as "first" and "second" mentioned in the embodiments of this application are used to distinguish multiple objects, and are not used to limit the order, sequence, priority or importance of multiple objects.

[0077] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0078] The technical solutions in this application embodiment can be applied to intelligent monitoring scenarios, such as vehicle monitoring, smart home, or security monitoring. The following explanation uses a vehicle monitoring scenario as an example. Vehicle monitoring scenarios can be applied to the Internet of Vehicles (IoV), such as vehicle-to-everything (V2X), long-term evolution-vehicle (LTE-V), and vehicle-to-vehicle (V2V). For example, the anomaly detection method provided in this application embodiment can be applied to a vehicle or to an in-vehicle device. The in-vehicle device may include, but is not limited to, an in-vehicle terminal, in-vehicle controller, in-vehicle module, in-vehicle assembly, in-vehicle component, in-vehicle chip, and in-vehicle unit. The vehicle can implement the anomaly detection method provided in this application embodiment through the in-vehicle terminal, in-vehicle controller, in-vehicle module, in-vehicle assembly, in-vehicle component, in-vehicle chip, and in-vehicle unit. The anomaly detection method in this application embodiment can also be used in other intelligent devices with monitoring functions besides vehicles, or installed in intelligent devices with monitoring functions, or installed in components of such intelligent devices. The smart device can be a smart transportation device, a smart home device, a smart monitoring device, etc.

[0079] Figure 1The illustration shows an application scenario to which this application embodiment applies. In this application scenario, a vehicle 100 and a server 200 may be included. The server 200 may be a cloud server, and the vehicle 100 and the server 200 may communicate via a network. In one embodiment, the server 200 may also be implemented using a virtual machine.

[0080] Some or all of the functions of vehicle 100 are controlled by computing platform 150 (or computer system). Computing platform 150 may include at least one processor 151, which can execute instructions 153 stored in a non-transitory computer-readable medium such as memory 152. In some embodiments, computing platform 150 may also be multiple computing devices that control individual components or subsystems of vehicle 100 in a distributed manner. Processor 151 may be any conventional processor, such as a central processing unit (CPU). Alternatively, processor 151 may also include graphics processing unit (GPU), field-programmable gate array (FPGA), system-on-chip (SoC), application-specific integrated circuit (ASIC), or combinations thereof.

[0081] Optionally, the vehicle 100 mentioned above can be a car, truck, motorcycle, bus, ship, airplane, helicopter, lawnmower, recreational vehicle, amusement park vehicle, construction equipment, tram, golf cart, train, etc., and this application embodiment does not impose any special limitations.

[0082] In some embodiments, the application scenario may not include server 200.

[0083] It should be understood that Figure 1 The structure of the vehicle should not be construed as a limitation on the embodiments of this application.

[0084] The method provided in this application embodiment can be implemented by a first device, which can be an independent device or a [missing information - likely a component or component]. Figure 1 The chips or components in the vehicle 100 shown can also be software modules that can be deployed on the relevant on-board equipment of the vehicle 100.

[0085] Figure 2 An exemplary schematic diagram illustrates another application scenario to which the embodiments of this application are applicable. Figure 2 The system architecture shown may include vehicle control devices and anomaly detection devices. Optionally, the system architecture may also include cloud-based servers.

[0086] Anomaly detection devices may include vehicles, or units, modules, chips, chip systems, circuits, etc., installed on vehicles, such as in-vehicle decision centers or vehicle processors. Anomaly detection devices may have network communication capabilities, enabling them to receive commands and control certain vehicle operations accordingly. Vehicles may include, for example, […]. Figure 1 100 vehicles in the middle.

[0087] Vehicle control devices may include vehicle keys, or units, modules, chips, chip systems, circuits, etc., within vehicle keys. Vehicle keys can be physical keys or digital keys; digital keys have no physical form and are software keys. Vehicle control devices may also include terminal equipment or units, modules, chips, chip systems, circuits, etc., installed on terminal equipment. Terminal equipment is a device with wireless transceiver capabilities. Terminal equipment can be user equipment (UE), where UE includes handheld devices, in-vehicle devices, wearable devices, or computing devices with wireless communication capabilities. For example, a UE can be a mobile phone, tablet computer, or computer with wireless transceiver capabilities. Terminal equipment can also be virtual reality (VR) terminal equipment, augmented reality (AR) terminal equipment, wireless terminals in industrial control, wireless terminals in autonomous driving, wireless terminals in telemedicine, wireless terminals in smart grids, wireless terminals in smart cities, wireless terminals in smart homes, etc. Applications for controlling the vehicle, such as vehicle digital keys, can be installed on the terminal equipment. Figure 2 The example illustrates how vehicle control devices can be installed in mobile phones and / or smartwatches. Mobile phones and smartwatches can also connect, for example, using communication protocols such as Wi-Fi, Bluetooth, or Starlink. In practical applications, the terminal device is not limited to mobile phones or smart wearable devices.

[0088] Anomaly detection devices and vehicle control devices can establish connections, such as a direct link. For example, the vehicle control device and the vehicle can establish a connection based on communication protocols such as StarFlash, Bluetooth, or Wi-Fi. Alternatively, anomaly detection devices and vehicle control devices can establish indirect link connections. For example, the vehicle control device and the anomaly detection device can establish a connection through other network devices (such as base stations or Wi-Fi access points), and this connection can be based on protocols such as Long Term Evolution (LTE), New Radio (NR), or Wi-Fi.

[0089] In one possible implementation, the vehicle control device and the anomaly detection device can establish a connection based on a server. For example, the vehicle control device can establish a connection with the server based on protocols such as LTE, NR, or Wi-Fi, or through network devices (such as base stations, Wi-Fi access points, etc.). Similarly, the anomaly detection device can establish a connection with the server based on protocols such as LTE, NR, or Wi-Fi, or through network devices (such as base stations, Wi-Fi access points, etc.). The server and network devices can communicate through the interface between the access network and the core network.

[0090] The network devices involved in the embodiments of this application include, for example, radio access network (RAN) devices. RAN devices can be base stations, evolved NodeBs (eNodeBs), transmission reception points (TRPs), transmission points (TPs), next-generation NodeBs (gNBs) in 5th generation (5G) mobile communication systems, next-generation base stations in next-generation mobile communication systems, base stations in future mobile communication systems, or access nodes in Wi-Fi systems; they can also be modules or units that perform some of the functions of a base station, for example, they can be centralized units (CUs) or distributed units (DUs). The CU here performs the functions of the radio resource control protocol and packet data convergence protocol (PDCP) of the base station, and can also perform the functions of the service data adaptation protocol (SDAP); the DU performs the functions of the radio link control layer and medium access control (MAC) layer of the base station, and can also perform some or all of the physical layer functions. For specific descriptions of the above protocol layers, please refer to the relevant technical specifications of the 3rd Generation Partnership Project (3GPP).

[0091] exist Figure 1 and Figure 2 In the application scenario shown, to ensure vehicle security and prevent vandalism, theft, or other accidental damage after parking, the vehicle's cameras can remain on after the vehicle is locked and parked, recording suspicious events around the vehicle. Typically, four 180° high-definition ultra-wide-angle cameras can be installed around the vehicle to monitor the surrounding environment and vehicle status in real time.

[0092] Figure 1 and Figure 2 The application scenarios shown are merely illustrative examples. The anomaly detection method provided in this application embodiment can also be applied to other application scenarios, such as smart home or security monitoring.

[0093] To address the issue of high power consumption caused by the camera remaining constantly on and in shooting mode in the aforementioned scenarios, this application provides an anomaly detection method. Figure 3 This example illustrates a possible flowchart of an anomaly detection method provided in an embodiment of this application. The anomaly detection method can be executed by a vehicle or an anomaly detection device installed on the vehicle. The anomaly detection device can be understood as a unit, module, chip, chip system, or circuit installed on the vehicle; for example, the anomaly detection device can be a processor on the vehicle. Alternatively, the anomaly detection method can be executed by a monitoring device or an anomaly detection device installed on the monitoring device. The anomaly detection device can be understood as a unit, module, chip, chip system, or circuit installed on the monitoring device; for example, the anomaly detection device can be a processor on the monitoring device.

[0094] like Figure 3 As shown in the embodiments of this application, an anomaly detection method may include the following steps:

[0095] S301: When an abnormal event is detected based on wireless data collected by the wireless sensor, the image sensor is activated.

[0096] In this system, both the wireless sensor and the image sensor are positioned in the target space, which is the space that the wireless sensor and the image sensor need to monitor. For example, for a vehicle, the target space is the space where the vehicle is located, and the wireless sensor and image sensor can be installed on the vehicle to monitor suspicious events around the vehicle. For example, in one embodiment, the wireless sensor may include one or more pairs, such as... Figure 4 As shown, two wireless sensors can be installed on each side of the vehicle, forming a pair. One wireless sensor transmits electromagnetic wave signals, and the other receives them. Alternatively, multiple pairs of wireless sensors can be installed on both sides of the vehicle. In another embodiment, one or more wireless sensors can be installed on the vehicle, each capable of both transmitting and receiving electromagnetic wave signals. One or more image sensors can also be installed on the vehicle to capture images of its surrounding environment. In security monitoring scenarios, the target space is a building, shopping mall, or parking lot that needs to be monitored. The wireless sensors and image sensors can be mounted on supports or walls. The wireless sensors and image sensors can be connected to an anomaly detection device, or they can be part of the anomaly detection device itself.

[0097] Wireless sensors can be ultra-wideband (UWB) sensors, Wi-Fi sensors, or millimeter-wave radar sensors. UWB sensors are based on wireless carrier communication technology and possess WLAN sensing capabilities. UWB sensors operate over a wide frequency range with extremely low radiation spectral density, offering advantages such as strong multipath resolution and low power consumption. Wi-Fi sensors are based on Wi-Fi wireless sensing technology, while millimeter-wave radar sensors are based on millimeter-wave radar technology. Both Wi-Fi and millimeter-wave radar sensors can monitor changes in the surrounding environment by detecting changes in received electromagnetic wave signals.

[0098] In some embodiments, the wireless sensor may be a narrowband signal-based wireless sensor, such as a Bluetooth sensor or a starburst sensor. The wireless sensor may also include a combination of multiple sensors described above.

[0099] Taking vehicle monitoring as an example, wireless sensors and image sensors can be installed on the vehicle. After the vehicle is locked and parked, the image sensor turns off, but the wireless sensor remains on, emitting and receiving electromagnetic wave signals. The wireless sensor converts the received electromagnetic wave signals into wireless data. This wireless data is time-domain data, containing two complex components: phase (I) and amplitude (Q), also known as quadrature amplitude components, hence it can be called I / Q data. In other words, I / Q data can be used to determine the phase and amplitude of the electromagnetic wave signal. Anomaly detection devices include a wireless sensing module. The anomaly detection device acquires the wireless data collected by the wireless sensor and can use the wireless sensing module to detect abnormal events based on this data.

[0100] In one embodiment, if the wireless sensor is a narrowband signal-based wireless sensor, the electromagnetic wave signal received by the wireless sensor, i.e., the echo signal, can be multiple sets of narrowband signals, each with a different signal frequency. For example, the wireless sensor can receive narrowband signals in multiple frequency bands such as 5150MHz~5151MHz, 5151MHz~5152MHz, 5152MHz~5153MHz, and 5153MHz~5154MHz respectively, thus obtaining multiple sets of narrowband signals.

[0101] Generally, UWB sensors, Wi-Fi sensors, or millimeter-wave radar sensors transmit and receive electromagnetic signals with a wide bandwidth, reaching over 20MHz, known as broadband signals. Wireless data obtained using broadband signals has high accuracy, and the plotted spectrum has high resolution, effectively reflecting changes in the surrounding environment. Conversely, wireless sensors based on narrowband signals transmit and receive narrowband electromagnetic signals, resulting in lower accuracy and lower resolution of the plotted spectrum, and are generally considered incapable of monitoring changes in the surrounding environment. However, this application provides a method for monitoring environmental changes using a wireless sensor based on narrowband signals. The wireless sensor in this application can transmit narrowband signals on multiple channels and receive echo signals. These multiple channels correspond to multiple frequency bands, such as 5150MHz–5151MHz, 5151MHz–5152MHz, 5152MHz–5153MHz, and 5153MHz–5154MHz. Wireless sensors can receive multiple sets of echo signals on multiple channels. These echo signals can be transmitted to an anomaly detection device, which can perform channel fusion on the received echo signals to obtain an electromagnetic wave signal. This electromagnetic wave signal is equivalent to a broadband signal and has stronger anti-interference capabilities. For example, assuming the wireless sensor receives 200 sets of echo signals, the anomaly detection device can randomly select one echo signal from the 200 sets, use the phase of the channel statement information (CSI) of that echo signal as a reference phase, and perform phase calibration on the multiple sets of echo signals to eliminate phase errors caused by sensor system parameters and obtain accurate phase information. In some embodiments, the multiple sets of echo signals can be sampled frequency domain signals. The phase-calibrated multiple sets of echo signals can be vector-summed to obtain a frequency domain signal. An inverse Fourier transform can be used to convert the frequency domain signal from the frequency domain to the time domain to obtain the required wireless data. In other embodiments, instead of using inverse Fourier transform to convert the frequency domain signal to the time domain, artificial intelligence or specific algorithms can be used to convert the frequency domain signal to the time domain; alternatively, the frequency domain signal can be used directly for subsequent judgments without conversion to the time domain. In other embodiments, multiple sets of echo signals can be sampled time domain signals. Multiple sets of echo signals after phase calibration can be vector-summed to obtain wireless data representing the time domain signal. The signal obtained after channel fusion of multiple sets of echo signals is equivalent to a broadband signal, which has high accuracy and high resolution of the plotted spectrum, fully reflecting changes in the surrounding environment. Furthermore, the wireless sensor can transmit narrowband signals on multiple channels, resulting in less interference to the surrounding environment.

[0102] After acquiring wireless data collected by the wireless sensor, the wireless sensing module can determine whether there are any abnormal events based on the wireless data.

[0103] In some embodiments, the wireless sensing module can compare wireless data with a set data threshold to obtain a comparison result, and determine whether an abnormal event has been detected based on the comparison result. The data threshold can be set based on the results of pre-conducted test experiments.

[0104] For example, in some embodiments, the wireless sensing module can determine whether an abnormal event exists based on the magnitude of the wireless data. For instance, when a person moves around a vehicle, their movement may cause the surrounding electromagnetic wave signal to strengthen, resulting in a larger wireless data size. Therefore, the wireless sensing module can determine whether an abnormal event exists based on the magnitude of the wireless data. Exemplarily, if the wireless data is greater than or equal to a set data threshold X, the wireless sensing module can consider that an abnormal event has been detected. The magnitude of the wireless data can be the magnitude of the I / Q data, or the magnitude of the phase or amplitude of the electromagnetic wave signal in the I / Q data. The phase of the electromagnetic wave signal can include intermediate values ​​such as the standard deviation, variance, or median of the phases of multiple electromagnetic wave signals received over a period of time. Similarly, the amplitude of the electromagnetic wave signal can include intermediate values ​​such as the standard deviation, variance, or median of the amplitudes of multiple electromagnetic wave signals received over a period of time. For example, in one optional embodiment, if the I / Q data is greater than or equal to a set data threshold x1, the wireless sensing module may consider that an abnormal event has been detected; or, if the phase of the electromagnetic wave signal is greater than or equal to a set data threshold x2, the wireless sensing module may consider that an abnormal event has been detected; or, if the amplitude of the electromagnetic wave signal is greater than or equal to a set data threshold x3, the wireless sensing module may consider that an abnormal event has been detected. In another optional embodiment, if the phase of the electromagnetic wave signal is greater than or equal to a set data threshold x2, and the amplitude of the electromagnetic wave signal is greater than or equal to a set data threshold x3, the wireless sensing module may consider that an abnormal event has been detected; or, if the I / Q data is greater than or equal to a set data threshold x1, and the amplitude of the electromagnetic wave signal is greater than or equal to a set data threshold x3, the wireless sensing module may consider that an abnormal event has been detected; or, if the I / Q data is greater than or equal to a set data threshold x1, and the phase of the electromagnetic wave signal is greater than or equal to a set data threshold x2, the wireless sensing module may consider that an abnormal event has been detected. In another alternative embodiment, if the I / Q data is greater than or equal to a set data threshold x1, and the phase of the electromagnetic wave signal is greater than or equal to a set data threshold x2, and the amplitude of the electromagnetic wave signal is greater than or equal to a set data threshold x3, the wireless sensing module may consider that an abnormal event has been detected.

[0105] In another embodiment, the wireless sensing module can monitor changes in wireless data to determine the presence of abnormal events. For example, when a person moves around a vehicle, their movement causes changes in the surrounding electromagnetic wave signals. These changes in signal intensity can be reflected in the wireless data collected by the wireless sensor. When the wireless data collected by the sensor changes, it indicates that there may be a moving target around the vehicle, which could be a person, an animal, or an object. Therefore, the wireless sensing module can determine the presence of abnormal events based on these changes. For example, the wireless sensing module can determine the change in wireless data within a unit of time, which could be 1 second, 3 seconds, or other values. If the magnitude of the change in wireless data is greater than or equal to a set threshold value Y, the wireless sensing module can consider that an abnormal event has been detected. The magnitude of the change in wireless data can be the magnitude of the I / Q data change, or the magnitude of the phase change or amplitude change of the electromagnetic wave signal within the I / Q data.

[0106] For example, in one optional embodiment, if the amplitude of the I / Q data change is greater than or equal to a set change threshold y1, the wireless sensing module may consider that an abnormal event has been detected; or, if the amplitude of the phase change of the electromagnetic wave signal is greater than or equal to a set change threshold y2, the wireless sensing module may consider that an abnormal event has been detected; or, if the amplitude of the amplitude change of the electromagnetic wave signal is greater than or equal to a set change threshold y3, the wireless sensing module may consider that an abnormal event has been detected. In another optional embodiment, if the amplitude of the phase change of the electromagnetic wave signal is greater than or equal to a set change threshold y2, and the amplitude change of the electromagnetic wave signal is greater than or equal to a set change threshold y3, the wireless sensing module may consider that an abnormal event has been detected; or, if the amplitude of the I / Q data change is greater than or equal to a set change threshold y1, and the amplitude change of the electromagnetic wave signal is greater than or equal to a set change threshold y3, the wireless sensing module may consider that an abnormal event has been detected; or, if the amplitude of the I / Q data change is greater than or equal to a set change threshold y1, and the amplitude of the phase change of the electromagnetic wave signal is greater than or equal to a set change threshold y2, the wireless sensing module may consider that an abnormal event has been detected. In another optional embodiment, if the amplitude of the I / Q data change is greater than or equal to a set change threshold y1, and the amplitude of the phase change of the electromagnetic wave signal is greater than or equal to a set change threshold y2, and the amplitude change of the electromagnetic wave signal is greater than or equal to a set change threshold y3, the wireless sensing module may consider that an abnormal event has been detected.

[0107] In other embodiments, the wireless sensing module can compare spectral parameters obtained based on wireless data with a set threshold to obtain a comparison result, and determine whether an abnormal event has been detected based on the comparison result. For example, the wireless sensing module can process the acquired wireless data to generate a range Doppler (RD) spectrum and determine the spectral parameter values ​​of the RD spectrum, which can be simply referred to as the Doppler spectrum. For example, the RD spectrum generated by the wireless sensing module can be as follows: Figure 5 As shown, Figure 5 In the image, (a) represents the RD spectrum without interference from moving targets. Figure 5 (b) in the image shows the RD spectrum when there is interference from a moving target. Figure 5 In (b) of the diagram, the circle encloses the location of the moving target, which can be a person, animal, or moving object. When there is interference from a moving target at a certain location, the perturbation value at that location in the RD spectrum will increase.

[0108] In one optional embodiment, if there is a first position in the RD spectrum with a disturbance value greater than or equal to threshold z1, and the distance between the first position and the wireless sensor is less than a set distance threshold, the wireless sensing module can consider that an abnormal event has been detected. In another optional embodiment, if there is a second position in the RD spectrum with a disturbance value greater than or equal to threshold z2, and the distance between the second position and the wireless sensor gradually decreases, it indicates that the distance between the moving target and the wireless sensor is continuously decreasing, and the wireless sensing module can consider that an abnormal event has been detected. In another optional embodiment, if there is a third position in the RD spectrum with a disturbance value greater than or equal to threshold z3, and the moving speed of the third position is greater than or equal to a set speed threshold, it indicates that the moving speed of the moving target is greater than or equal to the set speed threshold, and the wireless sensing module can consider that an abnormal event has been detected. The thresholds z1, z2, and z3 mentioned above can be collectively referred to as parameter thresholds, and thresholds z1, z2, and z3 can be the same or different.

[0109] In other embodiments, the wireless sensing module can process wireless data using a first network model to detect the presence of abnormal events. The first network model can be a trained network model. In one optional embodiment, the wireless sensing module can input the acquired wireless data into the first network model to obtain a detection result output by the first network model, which indicates whether an abnormal event has been detected. In another optional embodiment, the wireless sensing module can input the RD spectrum obtained based on the wireless data into the first network model to obtain a detection result output by the first network model, which also indicates whether an abnormal event has been detected.

[0110] In other embodiments, when the vehicle connects to the server, the anomaly detection device can send the acquired wireless data to the server for anomaly detection. In one optional embodiment, the server receives the wireless data transmitted by the vehicle's anomaly detection device, compares the wireless data or spectral parameters obtained based on the wireless data with a threshold, obtains a comparison result, determines whether an anomaly event has been detected based on the comparison result, and returns the detection result to the vehicle's anomaly detection device. In another optional embodiment, the server receives the wireless data transmitted by the vehicle's anomaly detection device, processes the wireless data using a first network model to detect the presence of anomaly events, obtains a detection result, and returns the detection result to the vehicle's anomaly detection device. The vehicle's anomaly detection device, upon receiving the detection result returned by the server, can determine whether an anomaly event has been detected based on the wireless data.

[0111] When an anomaly is detected based on wireless data, the anomaly detection device can control the image sensor to turn on. The image sensor may include one or more cameras.

[0112] S302 monitors the target space using an image sensor.

[0113] After the anomaly detection device controls the image sensor to be activated, the image sensor can collect video data of the target space. The anomaly detection device, upon acquiring the video data collected by the image sensor, can monitor the target space based on the video data to determine whether the target space contains an abnormal target. In some embodiments, the anomaly detection device can process the video data using a second network model to determine whether the target space contains an abnormal target. When it is confirmed that the target space contains an abnormal target, it can issue an alarm message. The second network model is a trained network model.

[0114] The anomaly detection method provided in this application uses a wireless sensor to detect the environment of a target space. When an anomaly is detected, an image sensor is activated to monitor the target space; otherwise, the image sensor does not need to be activated. Since the wireless sensor has very low power consumption, the solution provided in this application significantly reduces power consumption compared to using an image sensor for monitoring over a long period.

[0115] In some embodiments, the wireless sensor can be a narrowband signal-based wireless sensor, such as a Bluetooth sensor or a starburst sensor. Using a narrowband signal-based wireless sensor allows for faster and more rapid detection of changes in the surrounding environment, improves interference resistance, and makes the monitoring results more accurate.

[0116] In some embodiments, to further improve the accuracy of detecting abnormal events based on wireless data, data collected during the use of the anomaly detection device can be used as samples to adjust the parameters of the wireless sensing module. For example... Figure 6 As shown, in some embodiments, the method may include the following steps:

[0117] S601: When an abnormal event is detected based on wireless data collected by the wireless sensor, the image sensor is activated.

[0118] In the process of detecting anomalies based on wireless data collected by wireless sensors, the anomaly detection device can save the wireless data and / or the spectrum parameters obtained based on the wireless data.

[0119] S602 determines the anomaly detection result based on the video data collected by the image sensor, and sets the first sample tag for the wireless data based on the anomaly detection result.

[0120] When an anomaly is detected based on wireless data, the anomaly detection device can activate the image sensor. The image sensor can acquire video data of the target space, and the anomaly detection device acquires the video data acquired by the image sensor.

[0121] In some embodiments, the anomaly detection device can process video data using a second network model, or it can invoke a video recognition algorithm to identify objects in the video data, obtain an anomaly determination result, and set a first sample label for the wireless data based on the anomaly determination result. For example, if the anomaly determination result indicates the presence of an abnormal target in the target space, a positive sample label can be set for the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result indicates the absence of an abnormal target in the target space, a negative sample label can be set for the wireless data, indicating that the wireless data is a negative sample. In other embodiments, the anomaly detection device can also acquire images captured by an image sensor. The anomaly detection device can process the images using a second network model, or it can invoke an image recognition algorithm to identify objects in the images, obtain an anomaly determination result, and set a first sample label for the wireless data based on the anomaly determination result.

[0122] In other embodiments, after acquiring video data collected by the image sensor, the anomaly detection device can transmit the video data to the terminal device of the target account associated with the anomaly detection device. For example, when the anomaly detection device is an in-vehicle device, the target account can be the vehicle owner, and the target account's terminal device can be the vehicle owner's mobile phone. Upon receiving the video data sent by the anomaly detection device, the target account's terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can have "Anomaly" and "Non-Anomaly" buttons on the video data display interface. If the user determines that an anomaly exists, they can click the "Anomaly" button; if the user determines that no anomaly exists, they can click the "Non-Anomaly" button. Upon receiving user input or user actions, the terminal device can generate an anomaly determination result based on the user input and send the result to the anomaly detection device. Upon receiving the anomaly determination result returned by the terminal device, the anomaly detection device can set a first sample label for the wireless data based on the result. If the anomaly determination result is "Anomaly," a positive sample label can be set for the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result is "non-anomaly", then a negative sample label can be set for the wireless data to indicate that the wireless data is a negative sample.

[0123] In other embodiments, the anomaly detection device can process video data using a second network model to obtain an initial judgment result. If the initial judgment result indicates the presence of an abnormal target in the target space, consistent with the detection result of the wireless data, the initial judgment result can be used as the anomaly judgment result. Based on the anomaly judgment result, a positive sample label is set for the wireless data, indicating that the wireless data is a positive sample. If the initial judgment result indicates the absence of an abnormal target in the video data, inconsistent with the detection result of the wireless data, the anomaly detection device can transmit the video data to the target account's terminal device. Upon receiving the video data sent by the anomaly detection device, the target account's terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can have "Abnormal" and "Non-Abnormal" buttons on the video data display interface. If the user determines an anomaly exists, they can click the "Abnormal" button; if the user determines no anomaly exists, they can click the "Non-Abnormal" button. Upon receiving user input or user actions, the terminal device can generate an anomaly judgment result based on the user input and send the anomaly judgment result to the anomaly detection device. The anomaly detection device receives the anomaly determination result returned by the terminal device and can assign a first sample label to the wireless data based on the result. If the anomaly determination result is "anomaly," a positive sample label can be assigned to the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result is "not anomaly," a negative sample label can be assigned to the wireless data, indicating that the wireless data is a negative sample. To avoid misjudgments by the image acquisition device, such as errors caused by fog, correcting the labeling results through user feedback can improve the accuracy of sample labeling, which is beneficial for making the network model trained using the data samples more accurate.

[0124] The anomaly detection device can save the wireless data and the first sample tag of the wireless data, or the anomaly detection device can send the wireless data and the first sample tag of the wireless data to the server.

[0125] S603 adjusts the parameters of the wireless sensing module based on wireless data and the first sample tag.

[0126] In some embodiments, the anomaly detection device can periodically construct a training dataset using stored wireless data and first sample labels, or periodically download a training dataset based on wireless data and first sample labels from a server. The training dataset downloaded from the server may include more wireless data, such as wireless data collected by other vehicles. The anomaly detection device can perform normalization preprocessing on the wireless data in the training dataset, adjusting the wireless data to the specifications required by the first network model described above, and then retrain the first network model using the wireless data and first sample labels to adjust the network parameters of the first network model so that the first network model can more accurately detect abnormal events. The process of retraining the first network model can be called an incremental learning process, in which the network parameters of the first network model can be adjusted once every set period using the latest wireless data and first sample labels.

[0127] In other embodiments, the anomaly detection device can periodically construct a training dataset using saved wireless data and first sample labels, or periodically download a training dataset based on wireless data and first sample labels from a server. The data threshold or parameter threshold used in step S301 can be adjusted using the wireless data and first sample labels in the training dataset, so that the wireless sensing module can more accurately monitor abnormal events. The threshold can be adjusted once every set period using the latest wireless data and first sample labels. For example, if, within a set time period, the number of wireless data with negative first sample labels is greater than or equal to a first quantity threshold, or the ratio of the number of wireless data with negative first sample labels to the number of wireless data with positive first sample labels is greater than or equal to a first proportion threshold, it indicates that the wireless sensing module is making many misjudgments and triggering false alarms in many cases. In this case, the data threshold or parameter threshold can be increased to reduce false alarms.

[0128] In other embodiments, non-wireless sensors, such as vibration sensors, audio sensors, and inertial navigation sensors, can also be installed on the vehicle. The anomaly detection device can also monitor abnormal events through these non-wireless sensors. When a non-wireless sensor triggers an anomaly alarm, if the wireless sensing module does not detect the abnormal event, the anomaly detection device can record the wireless data collected by the wireless sensor at this time and label the wireless data with positive samples. This positive sample label, along with the aforementioned wireless data and the first sample label, forms a training dataset. The parameters of the wireless sensing module are then adjusted using the training dataset.

[0129] In other embodiments, when a non-wireless sensor triggers an alarm and no abnormal event is detected based on the wireless data collected by the wireless sensor, the wireless data can be saved as non-alarm wireless data. If, within a set time period, the number of non-alarm wireless data is greater than or equal to a second quantity threshold, or if the ratio of the number of non-alarm wireless data to the number of wireless data with the first sample tag is greater than or equal to a second proportion threshold, it indicates that the wireless sensing module has many missed alarms, meaning that no alarm was triggered when an anomaly existed. In this case, the data threshold or parameter threshold can be lowered to reduce missed alarms.

[0130] In some embodiments, the anomaly detection device may omit step S403 and instead send the recorded wireless data and first sample labels to a server for storage. In one optional embodiment, the server may periodically construct a training dataset using the stored wireless data and first sample labels, including wireless data and first sample labels reported by multiple vehicles. The server may perform normalization preprocessing on the wireless data in the training dataset, adjusting the wireless data to the specifications required by the aforementioned first network model. Then, it may retrain the first network model using the wireless data and first sample labels, adjusting the network parameters of the first network model to enable it to more accurately detect anomalous events. The retrained first network model is then sent to the vehicle's anomaly detection device. In another optional embodiment, the server may periodically construct a training dataset using the stored wireless data and first sample labels, adjust the threshold used in step S301 using the wireless data and first sample labels in the training dataset, and send the adjusted threshold to the vehicle's anomaly detection device. The anomaly detection device can use the adjusted threshold to update the parameters of the wireless sensing module, enabling the wireless sensing module to more accurately detect anomalous events.

[0131] In some embodiments, when constructing the training dataset, wireless data with negative sample labels can be selectively selected, while wireless data with positive sample labels can be omitted, in order to reduce the number of training samples and speed up the training process of the first network model.

[0132] To make it easier to understand, Figure 5 An exemplary embodiment illustrates the interaction flowchart between a terminal device, an anomaly detection device, and a server. The terminal device is the terminal device of the target account associated with the anomaly detection device. Taking a vehicle monitoring scenario as an example, the anomaly detection device may include a vehicle, or units, modules, chips, chip systems, or circuits installed on the vehicle. The anomaly detection device may have network communication capabilities so that it can receive instructions and control certain operations of the vehicle according to those instructions. The terminal device may be... Figure 2The vehicle control device shown may include the vehicle owner's terminal device, or units, modules, chips, chip systems, or circuits within the terminal device. The server may be a cloud server. For example... Figure 7 As shown, the interaction process may include the following steps:

[0133] S701, the anomaly detection device processes the wireless data collected by the wireless sensor through the first network model to detect whether there is an abnormal event.

[0134] S702, when the anomaly detection device detects an abnormal event based on wireless data, it activates the image sensor and acquires video data through the image sensor.

[0135] S703, the anomaly detection device sends video data to the terminal device.

[0136] The anomaly detection device can directly send the video data collected by the image sensor to the target account's terminal device through the communication channel between the anomaly detection device and the terminal device, or it can send the video data collected by the image sensor to the server, and then the server will send the video data collected by the image sensor to the target account's terminal device.

[0137] S704, the terminal device displays video data to the user and generates anomaly determination results based on the information input by the user.

[0138] When the target account's terminal device receives video data from the anomaly detection device, it can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can display "Anomaly" and "Non-Anomaly" buttons on the video data display interface. The user can click the "Anomaly" button if they determine an anomaly exists, and click the "Non-Anomaly" button if they determine no anomaly exists. The terminal device can also generate an anomaly determination result based on user input or user actions.

[0139] S705, the terminal device sends the anomaly determination result to the anomaly detection device.

[0140] S706, the anomaly detection device sets the first sample tag for wireless data based on the anomaly determination result.

[0141] If the anomaly determination result is "abnormal", the anomaly detection device can set a positive sample label for the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result is "not abnormal", the anomaly detection device can set a negative sample label for the wireless data, indicating that the wireless data is a negative sample.

[0142] S707, the anomaly detection device sends wireless data and the first sample tag to the server.

[0143] S708, the server saves the received wireless data and the first sample tag.

[0144] S709, the server periodically uses the saved wireless data and the first sample label to retrain the first network model to obtain the trained first network model.

[0145] The server stores a trained first network model, which is the first network model currently used by the anomaly detection device. The server can construct a training dataset using the stored wireless data and first sample labels according to a set period, retrain the first network model using the training dataset, obtain a trained first network model, save the trained first network model, and delete the previously saved first network model.

[0146] S710, the server sends the trained first network model to the anomaly detection device.

[0147] The server sends the trained first network model to the anomaly detection device so that the anomaly detection device can update the first network model and use the updated first network model to process the wireless data collected by the wireless sensor, thereby improving the accuracy of the anomaly detection results.

[0148] In some embodiments, considering that misjudgments may occur when detecting abnormal events based on image sensors in foggy or other special scenarios, in order to further improve the accuracy of the second network model in detecting abnormal events, data collected during the use of the anomaly detection device can be used as samples to adjust the network parameters of the second network model. For example... Figure 8 As shown, in some embodiments, the method may include the following steps:

[0149] S801: When an abnormal event is detected based on wireless data collected by the wireless sensor, the image sensor is activated.

[0150] S802 processes the video data acquired by the image sensor through a second network model to obtain the detection results.

[0151] When an anomaly is detected based on wireless data, the anomaly detection device can activate the image sensor, which then acquires video data of the target space. The anomaly detection device can then process this video data using a second network model to obtain a detection result. This result indicates whether an anomalous target is present in the target space.

[0152] S803 transmits video data to the target account's terminal device and receives the anomaly determination result returned by the target account's terminal device.

[0153] In some embodiments, after acquiring video data collected by an image sensor, the anomaly detection device can transmit the video data to the terminal device of the target account associated with the anomaly detection device. For example, when the anomaly detection device is an in-vehicle device, the target account can be the vehicle owner, and the target account's terminal device can be the vehicle owner's mobile phone. Upon receiving the video data sent by the anomaly detection device, the target account's terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can have "Anomaly" and "Non-Anomaly" buttons on the video data display interface. If the user determines that an anomaly exists, they can click the "Anomaly" button; if the user determines that no anomaly exists, they can click the "Non-Anomaly" button. Upon receiving user input or user actions, the terminal device can generate an anomaly determination result based on the user input and send the result to the anomaly detection device.

[0154] In other embodiments, the anomaly detection device can obtain detection results obtained by processing video data through a second network model. If the detection result indicates the presence of an abnormal target in the target space, consistent with the detection result of the wireless data, the video data can be withheld from being transmitted to the target account's terminal device. If the detection result indicates the absence of an abnormal target in the video data, inconsistent with the detection result of the wireless data, the anomaly detection device can transmit the video data to the target account's terminal device. Upon receiving the video data from the anomaly detection device, the target account's terminal device can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can include "Abnormal" and "Non-Abnormal" buttons on the video data display interface. If the user determines an anomaly exists, they can click the "Abnormal" button; if the user determines no anomaly exists, they can click the "Non-Abnormal" button. Upon receiving user input or user actions, the terminal device can generate an anomaly determination result based on the user input and send the result to the anomaly detection device.

[0155] S804 sets a second sample label for the video data based on the anomaly detection result.

[0156] The anomaly detection device receives the anomaly determination result returned by the terminal device and can set a second sample label for the video data based on the anomaly determination result. If the anomaly determination result is "anomaly", a positive sample label can be set for the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result is "not anomaly", a negative sample label can be set for the wireless data, indicating that the wireless data is a negative sample.

[0157] The anomaly detection device can save the video data and a second sample tag for the video data, or it can send the video data and the second sample tag for the video data to the server.

[0158] In some embodiments, when the detection result indicates the presence of an abnormal target in the target space, if the anomaly determination result is "abnormal" and consistent with the detection result output by the second network model, the video data and its second sample label may not be saved. When the detection result indicates the absence of an abnormal target in the target space, if the anomaly determination result is "non-abnormal" and consistent with the detection result output by the second network model, the video data and its second sample label may not be saved. When the detection result indicates the presence of an abnormal target in the target space, if the anomaly determination result is "non-abnormal" and inconsistent with the detection result output by the second network model, the video data and its second sample label may be saved. When the detection result indicates the absence of an abnormal target in the target space, if the anomaly determination result is "abnormal" and inconsistent with the detection result output by the second network model, the video data and its second sample label may be saved. Video data whose anomaly determination result is consistent with the detection result output by the second network model may not be saved and may not be used as training samples to reduce the number of training samples and accelerate the training process of the second network model.

[0159] S805 retrains the second network model based on video data and second sample labels.

[0160] In some embodiments, the anomaly detection device can periodically construct a training dataset using stored video data and second sample labels, or periodically download a training dataset based on video data and second sample labels from a server. The training dataset downloaded from the server may include more video data, such as video data collected from other vehicles. The anomaly detection device can perform normalization preprocessing on the video data in the training dataset, adjusting the video data to the specifications required by the second network model described above, and then retrain the second network model using the video data and second sample labels to adjust the network parameters of the second network model so that the second network model can more accurately detect abnormal events. The process of retraining the second network model can be called an incremental learning process, in which the network parameters of the second network model can be adjusted once every set period using the latest video data and second sample labels.

[0161] In other embodiments, the anomaly detection device may omit step S805 and instead send the recorded video data and second sample labels to a server for storage. In an optional embodiment, the server may periodically construct a training dataset using the stored video data and second sample labels, including video data and second sample labels reported by multiple vehicles. The server may perform normalization preprocessing on the video data in the training dataset, adjusting the video data to the specifications required by the aforementioned second network model. Then, it may retrain the second network model using the video data and second sample labels, adjusting the network parameters of the second network model to enable it to more accurately detect anomalous events. The retrained second network model is then sent to the vehicle's anomaly detection device.

[0162] To make it easier to understand, Figure 9 An exemplary embodiment illustrates the interaction flowchart between a terminal device, an anomaly detection device, and a server. The terminal device is the terminal device of the target account associated with the anomaly detection device. Taking a vehicle monitoring scenario as an example, the anomaly detection device may include a vehicle, or units, modules, chips, chip systems, or circuits installed on the vehicle. The anomaly detection device may have network communication capabilities so that it can receive instructions and control certain operations of the vehicle according to those instructions. The terminal device may be... Figure 2 The vehicle control device shown may include the vehicle owner's terminal device, or units, modules, chips, chip systems, or circuits within the terminal device. The server may be a cloud server. For example... Figure 9 As shown, the interaction process may include the following steps:

[0163] S901, the anomaly detection device processes the wireless data collected by the wireless sensor to detect whether there is an abnormal event.

[0164] S902, when the anomaly detection device detects an abnormal event based on wireless data, it activates the image sensor and acquires video data through the image sensor.

[0165] S903 processes the video data acquired by the image sensor through a second network model to obtain the detection results.

[0166] The anomaly detection device acquires video data collected by the image sensor. It can process the video data collected by the image sensor through a second network model to obtain the detection result. The detection result is used to indicate whether there is an abnormal target in the target space.

[0167] S904, the anomaly detection device sends video data to the terminal device.

[0168] The anomaly detection device can directly send the video data collected by the image sensor to the target account's terminal device through the communication channel between the anomaly detection device and the terminal device, or it can send the video data collected by the image sensor to the server, and then the server will send the video data collected by the image sensor to the target account's terminal device.

[0169] In some embodiments, after acquiring video data collected by an image sensor, the anomaly detection device can transmit the video data to the terminal device of the target account associated with the anomaly detection device.

[0170] In other embodiments, the anomaly detection device can obtain detection results obtained by processing video data through a second network model. If the detection result indicates the presence of an anomalous target in the target space, consistent with the detection result of the wireless data, the video data can be withheld from being transmitted to the target account's terminal device. If the detection result indicates the absence of an anomalous target in the video data, inconsistent with the detection result of the wireless data, the anomaly detection device can transmit the video data to the target account's terminal device.

[0171] S905, the terminal device displays video data to the user and generates anomaly determination results based on the information input by the user.

[0172] When the target account's terminal device receives video data from the anomaly detection device, it can alert the user and display the video data to allow the user to determine if an anomaly exists. The terminal device can display "Anomaly" and "Non-Anomaly" buttons on the video data display interface. The user can click the "Anomaly" button if they determine an anomaly exists, and click the "Non-Anomaly" button if they determine no anomaly exists. The terminal device can also generate an anomaly determination result based on user input or user actions.

[0173] S906, the terminal device sends the anomaly determination result to the anomaly detection device.

[0174] S907, the anomaly detection device sets a second sample label for the video data based on the anomaly determination result.

[0175] If the anomaly determination result is "abnormal", the anomaly detection device can set a positive sample label for the wireless data, indicating that the wireless data is a positive sample. If the anomaly determination result is "not abnormal", the anomaly detection device can set a negative sample label for the wireless data, indicating that the wireless data is a negative sample.

[0176] S908, the anomaly detection device sends video data and second sample tags to the server.

[0177] S909, the server saves the received video data and the second sample tag.

[0178] S910: The server periodically uses the saved video data and second sample labels to retrain the second network model, thus obtaining a trained second network model.

[0179] The server stores a trained second network model, which is the second network model currently used by the anomaly detection device. The server can construct a training dataset using the stored video data and second sample labels according to a set period, retrain the second network model using the training dataset, obtain a trained second network model, save the trained second network model, and delete the previously saved second network model.

[0180] S911, the server sends the trained second network model to the anomaly detection device.

[0181] The server sends the trained second network model to the anomaly detection device so that the anomaly detection device can update the second network model and use the updated second network model to process the video data collected by the image sensor, thereby improving the accuracy of the anomaly detection results.

[0182] Based on the same design concept as the above-described method embodiments, this application also provides an anomaly detection device. This anomaly detection device can be applied to... Figure 1 or Figure 2 The anomaly detection device shown in the vehicle can also be applied to other monitoring scenarios. It can achieve the functions of the above-described method embodiments, thus realizing the beneficial effects of the above-described method embodiments. For example... Figure 10 As shown, the anomaly detection device 1000 may include a first monitoring unit 1001 and a second monitoring unit 1002.

[0183] The first monitoring unit 1001 can be used to activate the image sensor when an abnormal event is detected based on wireless data collected by the wireless sensor. Both the wireless sensor and the image sensor are located in the target space. The first monitoring unit 1001 can be understood to include the wireless sensing module in the above embodiments.

[0184] The second monitoring unit 1002 can be used to monitor the target space through an image sensor.

[0185] In some embodiments, such as Figure 11As shown, the anomaly monitoring device 1000 may further include a parameter adjustment unit 1101. The parameter adjustment unit 1101 can be used to determine anomaly judgment results based on video data acquired by an image sensor, and to set a first sample tag for wireless data based on the anomaly judgment results. The wireless data and the first sample tag are used to adjust the parameters of the first monitoring unit 1001.

[0186] It should be noted that, in some embodiments, the first monitoring unit 1001 can be used to execute any step in the anomaly detection method, the second monitoring unit 1002 can be used to execute any step in the anomaly detection method, and the parameter adjustment unit 1101 can be used to execute any step in the anomaly detection method. The steps implemented by the first monitoring unit 1001, the second monitoring unit 1002, and the parameter adjustment unit 1101 can be specified as needed. The first monitoring unit 1001, the second monitoring unit 1002, and the parameter adjustment unit 1101 respectively implement different steps in the anomaly detection method to achieve all the functions of the anomaly detection device.

[0187] In the embodiments of this application, the functional modules can be integrated into a single processor, or each module can exist physically separately, or two or more modules can be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional units.

[0188] Based on the same technical concept as the above-described method embodiments, this application also provides an anomaly detection device. Exemplarily, this anomaly detection device can be a vehicle or a chip within a vehicle. This anomaly detection device can be used to implement the functions of the above-described method embodiments, thus achieving the beneficial effects of the above-described method embodiments.

[0189] In some embodiments, the structure of the anomaly detection device 1200 can be as follows: Figure 12 As shown, the system includes a processor 1201 and a memory 1202 connected to the processor 1201. The processor 1201 and the memory 1202 can be interconnected via a bus. The processor 1201 can be a general-purpose processor, such as a microprocessor, or other conventional processor. The bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc.

[0190] The memory 1202 can be used to store software programs and modules. The processor 1201 executes various functional applications and data processing of the anomaly detection device 1200 by running the software programs and modules stored in the memory 1202, such as the anomaly detection method provided in the embodiments of this application.

[0191] The memory 1202 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs of at least one application, etc.; the data storage area may be used to store user data, etc. In addition, the memory 1202 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0192] The processor 1201 in the anomaly detection device 1200 is used to run computer instructions or programs stored in the memory 1202 to perform the functions in any of the above method embodiments. For example, the processor can be used to monitor the signal strength of the serving cell of the terminal during emergency bearer existence; when the signal strength of the serving cell is less than or equal to a first threshold, neighbor cell signal measurement is performed to obtain at least one candidate neighbor cell; the signal strength of the candidate neighbor cell is greater than the first threshold. When there is no neighbor cell among the at least one candidate neighbor cell that meets the emergency bearer conditions, the terminal continues to camp on the serving cell. In some embodiments, the processor 1201 may include one or more processing units, which may be independent devices or integrated into one or more processors. The processor 1201 may also include a controller, which can generate operation control signals according to the instruction opcode and timing signals to complete the control of instruction fetching and execution.

[0193] In one embodiment, the anomaly detection device 1200 may further include a communication module for communicating with a server and terminal devices.

[0194] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the anomaly detection device. In other embodiments of this application, the anomaly detection device may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0195] This application also provides a chip, which may be a processor chip, including a processor and a memory. The structure of the chip can be referred to... Figure 12 The structure of the anomaly detection device shown will not be described in detail here.

[0196] This application also provides a computer program product comprising computer-executable instructions. In one embodiment, the computer-executable instructions are used to cause a computer to perform the functions described in the method embodiments above.

[0197] Computer-executable instructions can be stored in a computer-readable storage medium. This application also provides a computer-readable storage medium storing executable instructions. In one embodiment, the computer-executable instructions are used to cause a computer to perform the functions described in the method embodiments above.

[0198] The computer-readable storage medium provided in the embodiments of this application may be random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), register, hard disk, portable hard disk, CD-ROM, or any other form of computer-readable storage medium known in the art.

[0199] Computer-executable instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access, 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, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video disc (DVD); or it can be a semiconductor medium, such as a solid-state drive.

[0200] In the various embodiments of this application, unless otherwise specified or logically conflicting, the terminology and / or descriptions between different embodiments are consistent and can be referenced mutually. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as including a series of steps or units. A method, system, product, or device is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0201] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made therein without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely illustrative examples of the solutions defined by the appended claims and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application.

[0202] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if these modifications and variations of the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. An anomaly detection method, characterized in that, The method includes: When an abnormal event is detected based on wireless data collected by the wireless sensor, the image sensor is activated; both the wireless sensor and the image sensor are located in the target space. The target space is monitored using the image sensor.

2. The method according to claim 1, characterized in that, The method further includes: The wireless data is processed using a first network model to detect any abnormal events.

3. The method according to claim 2, characterized in that, The method further includes: Anomaly detection results are determined based on the video data acquired by the image sensor. Based on the anomaly determination result, a first sample label is set for the wireless data; the wireless data and the first sample label are used to retrain the first network model.

4. The method according to claim 1, characterized in that, The method further includes: An abnormal event is determined to have been detected based on the comparison result between the wireless data and the data threshold; or, An abnormal event was detected based on a comparison between the parameters determined from the wireless data and the parameter thresholds.

5. The method according to claim 4, characterized in that, The wireless data is used to characterize the phase and amplitude of the electromagnetic wave signal received by the wireless sensor. Based on the comparison result between the wireless data and the data threshold, an abnormal event is detected, including: An abnormal event is determined to have been detected by a combination of at least one or more of the following methods: If the wireless data is greater than or equal to a first data threshold, then an abnormal event is determined to have been detected; or, If the amplitude of the electromagnetic wave signal is greater than or equal to the second data threshold, then an abnormal event is determined to have been detected; or, If the change in the wireless data within a set time period is greater than or equal to a third data threshold, then an abnormal event is determined to have been detected; or, If the phase change of the electromagnetic wave signal within a set time period is greater than or equal to the fourth data threshold, then an abnormal event is determined to have been detected; or, If the amplitude of the electromagnetic wave signal changes by more than or equal to the fifth data threshold within a set time period, an abnormal event is detected.

6. The method according to claim 4, characterized in that, The parameters include perturbation values ​​at multiple locations in the Doppler spectrum; the determination of detected abnormal events based on the comparison between the parameters determined by the wireless data and parameter thresholds includes: If there is a first location in the Doppler spectrum with a perturbation value greater than or equal to a first parameter threshold, and the distance between the first location and the wireless sensor is within a set distance range, then an abnormal event is determined to have been detected; or, If there is a second location in the Doppler spectrum with a perturbation value greater than or equal to the second parameter threshold, and the distance between the second location and the wireless sensor gradually decreases, then an abnormal event is determined to have been detected; or, If there is a third position in the Doppler spectrum with a perturbation value greater than or equal to the third parameter threshold, and the movement speed of the third position is greater than or equal to the set speed threshold, then an abnormal event is determined to have been detected.

7. The method according to any one of claims 4 to 6, characterized in that, The method further includes: Anomaly detection results are determined based on the video data acquired by the image sensor. Based on the anomaly determination result, a first sample label is set for the wireless data; the wireless data and the first sample label are used to adjust the data threshold or the parameter threshold.

8. The method according to claim 7, characterized in that, The method further includes: If, within a first set time period, the number of wireless data with the first sample label being a negative sample label is greater than or equal to a first quantity threshold, or if the ratio of the number of wireless data with the first sample label being a negative sample label to the number of wireless data with the first sample label being a positive sample label is greater than or equal to a first ratio threshold, then the data threshold or the parameter threshold is increased.

9. The method according to claim 8, characterized in that, The method further includes: When a non-wireless sensor triggers an abnormal alarm, and no abnormal event is detected based on the wireless data, the wireless data is saved as non-alarm wireless data. If, within a second set time period, the number of non-alarm wireless data is greater than or equal to a second quantity threshold, or if the ratio of the number of non-alarm wireless data to the number of wireless data with a first sample tag is greater than or equal to a second ratio threshold, the data threshold or the parameter threshold is lowered.

10. The method according to claim 3 or 7, characterized in that, The determination of anomaly based on the video data acquired by the image sensor includes: The video data is transmitted to the terminal device; Receive the anomaly determination result returned by the terminal device.

11. The method according to claim 3 or 7, characterized in that, The determination of anomaly based on the video data acquired by the image sensor includes: The video data is processed by a second network model to obtain the anomaly determination result.

12. The method according to claim 3 or 7, characterized in that, The determination of anomaly based on the video data acquired by the image sensor includes: The video data is processed using a second network model to determine the presence of abnormal targets in the video data. The judgment result output by the second network model is used as the anomaly judgment result.

13. The method according to claim 3 or 7, characterized in that, The determination of anomaly based on the video data acquired by the image sensor includes: The video data is processed using a second network model to determine that there are no abnormal targets in the video data; The video data is transmitted to the terminal device, and the anomaly determination result returned by the terminal device is received.

14. The method according to claim 13, characterized in that, The method further includes: Based on the anomaly detection result, a second sample label is determined for the video data; the video data and the second sample label are used to retrain the second network model.

15. The method according to any one of claims 1 to 14, characterized in that, The wireless sensor is a wireless sensor based on narrowband signals.

16. The method according to claim 15, characterized in that, The wireless sensor includes at least one of a Bluetooth sensor or a star-flash sensor.

17. The method according to any one of claims 1 to 16, characterized in that, The method further includes: The wireless sensor transmits narrowband signals on multiple channels and receives echo signals. The wireless data is obtained based on multiple sets of received echo signals.

18. The method according to claim 17, characterized in that, The wireless data obtained based on multiple received echo signals includes: The multiple sets of echo signals are channel-fused to obtain a frequency domain signal; The frequency domain signal is converted from the frequency domain to the time domain to obtain the wireless data.

19. The method according to any one of claims 1 to 18, characterized in that, The wireless sensor and the image sensor are mounted on the vehicle.

20. An anomaly detection device, characterized in that, include: The first monitoring unit is used to activate the image sensor when an abnormal event is detected based on the wireless data collected by the wireless sensor. Both the wireless sensor and the image sensor are located in the target space; The second monitoring unit is used to monitor the target space through the image sensor.

21. An anomaly detection device, characterized in that, Including processor and memory; The processor is configured to execute a computer program or instructions stored in the memory, causing the anomaly detection device to implement the method described in any one of claims 1 to 19.

22. A chip, characterized in that, Including processor and memory; The processor is configured to execute a computer program or instructions stored in the memory, such that the anomaly detection device including the chip implements the method described in any one of claims 1 to 19.

23. A computer-readable storage medium, characterized in that, The computer storage medium stores computer-readable instructions that, when executed on the anomaly detection device, cause the method as described in any one of claims 1 to 19 to be performed.

24. A computer program product, characterized in that, When the computer program product is run on the anomaly detection device, the anomaly detection device performs the method according to any one of claims 1 to 19.

Images