Information processing apparatus, information processing method, vehicle, and device

Abstract

The embodiment of the invention discloses information processing equipment, an information processing method, a vehicle and a device. A camera module and a first controller are arranged in an information processing device, a pre-trained detection model is deployed in the first controller, video information of a vehicle is obtained through the camera module, then the first controller conducts image recognition on the video information through the detection model to determine event information, and if the event information represents that the vehicle is in danger, the vehicle is not in danger. And if yes, executing corresponding security maintenance operation according to the event information. Therefore, dangerous events can be effectively detected, and the corresponding safety maintenance operation can be executed, so that the safety is improved, and the method has relatively high applicability.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicles, and in particular, to an information processing device, an information processing method, a vehicle, and a device. Background Art

[0002] With the development and progress of technology and the improvement of people's living standards, traveling by shared vehicles (such as shared bicycles, shared electric vehicles, etc.) has become a new emerging travel mode in the city, which effectively solves the travel needs of urban populations. However, with the large-scale deployment and use of shared vehicles, there are unsafe events such as reverse driving, occupying motor vehicle lanes, and even collisions and falls during the use of shared vehicles by people. Therefore, how to detect unsafe events to improve user safety has become an urgent problem to be solved. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide an information processing device, an information processing method, a vehicle, and a device, which can effectively detect events in danger and perform corresponding safety maintenance operations to improve safety and have high applicability.

[0004] In a first aspect, an embodiment of the present invention provides an information processing device, which includes:

[0005] A camera module for acquiring video information of the vehicle and transmitting it to a first controller;

[0006] A first controller deployed with a pre-trained detection model for performing image recognition on the video information to determine event information, and in response to the event information indicating that the vehicle is in danger, performing corresponding safety maintenance operations according to the event information.

[0007] In a second aspect, an embodiment of the present invention provides a vehicle, which includes:

[0008] A vehicle body;

[0009] The information processing device as described in the first aspect.

[0010] In a third aspect, an embodiment of the present invention provides an information processing method, which includes:

[0011] Receiving video information of the vehicle transmitted by the camera module;

[0012] Performing image recognition on the video information through a pre-trained detection model to determine event information;

[0013] In response to the event information indicating that the vehicle is in danger, performing corresponding safety maintenance operations according to the event information.

[0014] Fourthly, an embodiment of the present invention provides an information processing device, which includes:

[0015] A receiving unit, configured to receive video information of a vehicle transmitted by a camera module;

[0016] An image recognition unit, configured to perform image recognition on the video information through a pre-trained detection model to determine event information;

[0017] An operation execution unit, configured to, in response to the event information indicating that the vehicle is in danger, perform corresponding safety maintenance operations according to the event information.

[0018] Fifthly, an embodiment of the present invention provides a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method described in the third aspect is implemented.

[0019] In the embodiment of the present invention, by setting a camera module and a first controller in the information processing device, the first controller is deployed with a pre-trained detection model, the video information of the vehicle is obtained through the camera module, and then the first controller performs image recognition on the video information through the detection model to determine event information. If the event information indicates that the vehicle is in danger, corresponding safety maintenance operations are performed according to the event information. Thus, dangerous events can be effectively detected and corresponding safety maintenance operations can be performed to improve safety and have high applicability. Description of the Drawings

[0020] Through the following description of the embodiments of the present invention with reference to the drawings, the above and other objects, features, and advantages of the present invention will become clearer. In the drawings:

[0021] Figure 1 is a schematic diagram of the information processing device according to the embodiment of the present invention;

[0022] Figure 2 is a schematic diagram of the vehicle according to the embodiment of the present invention;

[0023] Figure 3 is a flowchart of the information processing method according to the embodiment of the present invention;

[0024] Figure 4 is a flowchart of performing image recognition on video information through a pre-trained detection model to determine event information according to the embodiment of the present invention;

[0025] Figure 5 is a flowchart of performing image recognition on video information through a pre-trained detection model to determine event information according to the embodiment of the present invention;

[0026] Figure 6It is a flowchart of identifying event information by performing image recognition on video information through a pre-trained detection model in an embodiment of the present invention;

[0027] Figure 7 It is a flowchart of identifying event information by performing image recognition on video information through a pre-trained detection model in an embodiment of the present invention;

[0028] Figure 8 It is a schematic diagram of an information processing device fixed to a vehicle-mounted helmet in an embodiment of the present invention;

[0029] Figure 9 It is a schematic diagram of an information processing device in an embodiment of the present invention;

[0030] Figure 10 It is a schematic diagram of an electronic device in an embodiment of the present invention. Detailed implementation manners

[0031] The following describes the present application based on embodiments, but the present application is not limited to these embodiments. In the following detailed description of the present application, some specific detail parts are described in detail. Those skilled in the art can fully understand the present application without the description of these detail parts. In order to avoid obscuring the essence of the present application, well-known methods, processes, procedures, elements and circuits are not described in detail.

[0032] In addition, those of ordinary skill in the art should understand that the accompanying drawings provided herein are for illustrative purposes only, and the drawings are not necessarily drawn to scale.

[0033] Unless the context clearly requires otherwise, the words such as "including", "comprising" and the like in the entire application document should be interpreted as the meaning of including rather than exclusive or exhaustive; that is, the meaning of "including but not limited to".

[0034] For ease of description, spatially relative terms such as "inside", "outside", "below", "beneath", "lower", "above", "upper", "front", "rear", etc. are used herein to describe the relationship of one element or feature illustrated in the figure with another element or feature. It will be understood that the spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figure is flipped, the element described as "below" or "beneath" another element or feature will then be positioned "above" that other element or feature. Thus, the exemplary term "below" can encompass both the orientation of above and below. The device may be otherwise oriented (rotated 90 degrees or at other orientations), and the spatially relative descriptive words used herein should be interpreted accordingly.

[0035] In the description of the present application, it should be understood that terms such as "first", "second", etc. are only used for descriptive purposes and cannot be construed as indicating or implying relative importance. In addition, in the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0036] For the solutions described in this specification and the embodiments, if they involve personal information processing, they will be processed on the premise of having a legal basis (such as obtaining the consent of the personal information subject, or being necessary for performing a contract, etc.), and will only be processed within the specified or agreed scope. If a user refuses to process personal information other than the necessary information required for basic functions, it will not affect the user's use of basic functions.

[0037] In the following description, an information processing device applied to a two-wheeled vehicle (such as a shared bicycle) scenario, where the vehicle is a two-wheeled vehicle, is used as an example for illustration. It should be understood that the information processing device involved in the embodiments of the present invention can be applied to various scenarios that require detection of dangerous behaviors. For example, the information processing device can be applied to scenarios such as detecting dangerous behaviors during the driving of an automobile.

[0038] Figure 1 is a schematic diagram of the information processing device according to the embodiments of the present invention. As Figure 1 shown, the information processing device of this embodiment includes a camera module 11 and a first controller 12. The camera module 11 is communicatively connected to the first controller 12.

[0039] In this embodiment, since there are many unsafe events such as reverse driving, occupying the motor vehicle lane, and even collisions and falls during the use of shared vehicles in the prior art. In view of this situation, in this embodiment, a pre-trained detection model is deployed in the first controller 12. The camera module 11 collects video information of the vehicle and transmits it to the first controller 12. The detection model performs image recognition on the video information to determine event information. If it is detected that the event information indicates that the vehicle is in danger, the first controller 12 performs corresponding safety maintenance operations according to the event information.

[0040] In this embodiment, training can be performed according to training data and a preset training model (e.g., a deep learning model (Deep Learning, DL), etc.) to obtain a detection model. The detection model can be an artificial intelligence (Artificial Intelligence, AI) model. Among them, the training data is, for example, video information (the video information can be image information of multiple frames, image angles), positioning information (e.g., geographical location coordinates and driving trajectories associated with the video information), traffic sign line data (e.g., lane line data, guiding line data, etc. associated with the video information), vehicle information (e.g., vehicle speed, acceleration, vehicle type, etc. associated with the video information), traffic indication device data (e.g., traffic signs, traffic lights associated with the video information, and information such as indicated turning, prohibited passage, suspended passage, permitted passage, etc.). In the following description, the detection model being an artificial intelligence model is taken as an example for illustration.

[0041] Specifically, the training data is used as input data for model training, enabling the training model to output corresponding event information. For example, by detecting the lane lines and vehicle type, the event information is determined as incorrect lane usage behavior or correct lane usage behavior. Another example is by detecting the status information of traffic lights and the distance between the vehicle and traffic indication devices to determine the event information as correct passing behavior or incorrect passing behavior.

[0042] Optionally, the training model can be set with a machine learning program (Machine Learning Program, MLP) (including a deep learning program, which can also be called a machine learning algorithm or machine learning tool). The machine learning tool is used to analyze and process the training data to construct a detection model. Among them, the machine learning tool is, for example, logistic regression, naive Bayes, random forest, neural network, matrix factorization, and support vector machine tools, etc.

[0043] Optionally, the trained artificial intelligence model includes a multi-modal model. The multi-modal model can integrate multiple models to process different types of input data, such as different types of input data like images, videos, positioning information, acceleration, etc. For example, the multi-modal model integrates a vision and language model (vision and language model, VLM). Since video images usually include various semantic information (such as traffic sign colors, lane text (e.g., bus-only lane), etc.), the vision and language model can effectively identify vehicles, traffic lights, lane lines, and traffic indication devices in the video information.

[0044] Optionally, the trained artificial intelligence model further includes a neural network model. The neural network model is, for example, a recurrent neural network (RNN) or a convolutional neural network (CNN). For example, the pedestrian movement posture features of a certain frame of image in the video are extracted through a convolutional neural network, and the temporal relationship of each frame is determined through a recurrent neural network to determine the movement direction of the pedestrian and whether the vehicle is moving in the wrong direction based on the pedestrian movement posture of each frame of image. Thus, the video information can be subjected to image recognition through the trained detection model to determine the event information.

[0045] In an embodiment of the present invention, a camera module and a first controller are provided in an information processing device. The first controller is deployed with a pre-trained detection model. The video information of the vehicle is acquired through the camera module, and then the first controller performs image recognition on the video information through the detection model to determine the event information. If the event information indicates that the vehicle is in danger, corresponding safety maintenance operations are performed according to the event information. Thus, events in danger can be effectively detected and corresponding safety maintenance operations are performed to improve safety, which has high applicability.

[0046] Figure 2 is a schematic diagram of a vehicle according to an embodiment of the present invention. As Figure 2 shown, the vehicle in this embodiment includes a vehicle body 100, an information processing device 1, and a main control device 2. The information processing device 1 includes a camera module 11, a first controller 12, a positioning module 13, a Bluetooth module 14, and a sensor 15. The main control device 2 includes a speaker 21 and a second controller 22.

[0047] Optionally, the vehicle body 100 includes a seat, a seat support, wheels, a handlebar, etc. The main control device 2 can be fixed to the seat support. The main control device 2 can also be fixed at a position such as a bearing assembly above the wheel.

[0048] In this embodiment, the camera module 11, the first controller 12, the positioning module 13, the Bluetooth module 14, and the sensor 15 can communicate in a wired manner, and this wired connection is implemented, for example, through bus interfaces such as CAN (Controller Area Network), LIN (Local Interconnect Network), RS-485, and UART (Universal Asynchronous Receiver / Transmitter).

[0049] In this embodiment, the master device 2 can communicate with the server wirelessly. For example, it can communicate through GSM (Global System for Mobile Communications), GPRS (General Packet Radio Service), eMTC (LTE enhanced MTO), NB-IoT (Narrow Band Internet of Things), etc., to enable data interaction between the vehicle and the server.

[0050] In this embodiment, the information processing device 1 can communicate with the master device 2 through a communication device. The communication device can be a Bluetooth module 14, and the Bluetooth module 14 can include a Bluetooth (BT) chip, an antenna, etc. It should be understood that this embodiment takes the communication device as the Bluetooth module 14 as an example for illustration. It should be understood that the communication device in the information processing device 1 can also be implemented through a Wi-Fi module, a 4G communication module, a 5G communication module, etc.

[0051] In this embodiment, the camera module 11 is used to collect video information of the vehicle.

[0052] Optionally, the camera module 11 can include a lens, an image sensor, a serializer, etc. The vehicle image information is collected through the lens and transmitted to the image sensor. Then the image sensor performs photoelectric conversion to convert the image information into digital information. Furthermore, the serializer performs serial processing on the digital information to obtain the video information of the vehicle, and transmits it to the first controller 12.

[0053] In this embodiment, the vehicle can also include a basket and / or a vehicle helmet. The information processing device 1 can be fixed to the basket or the vehicle helmet.

[0054] Optionally, the camera module 11 can collect video information in front of the vehicle. That is to say, the camera module 11 collects video information in the traveling direction of the vehicle. Correspondingly, the camera module 11 can be fixed in front of the basket or in front of the vehicle helmet.

[0055] Optionally, the camera module 11 can also be arranged inside the vehicle helmet. Specifically, the vehicle helmet can include a helmet shell, and the helmet shell has a hollow part. The lens of the camera module 11 is fixed to the hollow part to collect video information in front of the vehicle. This embodiment takes the information processing device 1 can be fixed to the basket or the vehicle helmet as an example for illustration. It should be understood that the information processing device 1 can be customarily set at different positions of the vehicle according to requirements. For example, the information processing device 1 can be fixed to the handlebar, the front fork, etc.

[0056] In this embodiment, the video information in front of the vehicle collected by the camera module 11 is taken as an example for illustration. It should be understood that the camera module 11 can collect the vehicle environment video information. For example, the camera module 11 can also collect the video information behind, on the side of the vehicle, and during the user's riding process. Thus, the event information can be comprehensively detected to improve safety.

[0057] In this embodiment, the positioning module 13 is used to obtain the positioning information of the vehicle and transmit it to the first controller 12. Among them, the positioning module 13 is, for example, a GPS chip (Global Positioning System), a BDS chip (BeiDou Navigation Satellite System), etc.

[0058] Optionally, the positioning information includes position coordinates (such as longitude and latitude) and the driving trajectory. The positioning module 13 collects multiple position coordinates of the vehicle at different times according to a preset positioning frequency (for example, collects the vehicle position coordinates once every 2 seconds), and then determines the driving trajectory of the vehicle based on the multiple position coordinates.

[0059] Optionally, the positioning information includes position coordinates. The positioning module 13 can transmit multiple position coordinates to the first controller 12, and the driving trajectory of the vehicle is determined based on the multiple position coordinates through a detection model (for example, the detection model can determine the driving trajectory of the vehicle according to the multiple position coordinates and preset map data).

[0060] In this embodiment, the sensor 15 is used to obtain the acceleration information of the vehicle and transmit it to the first controller 12. Among them, the sensor 15 can be implemented, for example, through a six-axis sensor, an accelerometer, etc. Among them, the six-axis sensor includes a three-axis gyroscope and a three-axis acceleration sensor, and the acceleration information of the vehicle can be determined according to the angular velocity detected by the three-axis gyroscope and the acceleration detected by the three-axis acceleration sensor.

[0061] In this embodiment, the first controller 12 and the second controller 22 can be electronic devices with functions such as data transmission, data processing, and data storage. The first controller 12 and the second controller 22 can be implemented through an MCU (Microcontroller Unit), a PLC (Programmable Logic Controller), an FPGA (Field-Programmable Gate Array), a DSP (Digital Signal Processor), etc.

[0062] Optionally, the first controller 12 may be an NPU chip (Neural network Processing Unit), and a pre-trained detection model may be deployed in the NPU chip.

[0063] In this embodiment, an example is given in which the detection model in the first controller 12 processes the received video information, positioning information, and acceleration information to determine event information. It should be understood that the detection model may also determine corresponding event information based on the received target information. For example, the information processing device 1 further includes a vibration sensor, and the vibration sensor is used to obtain vibration information and transmit it to the first controller 12. Then, the detection model of the first controller 12 may determine the vibration change amount based on the vibration information (for example, the detection model determines the amplitude change amount based on the vibration amplitudes at different times included in the vibration information). If the vibration change amount within a target time (for example, within two seconds) is greater than the vibration threshold, the detection model may determine that the event information is a collision accident. In the following description, the first controller 12 determines the event information, that is, the detection model determines the event information. Specifically, the information processing method of this embodiment may refer to Figure 3 。

[0064] Figure 3 is a flowchart of the information processing method according to an embodiment of the present invention. As Figure 3 shown, the information processing process of this embodiment includes the following steps:

[0065] Step S100: Receive the video information of the vehicle transmitted by the camera module

[0066] In this embodiment, the first controller 12 may send a control signal to the camera module 11, and after receiving the control signal, the camera module 11 obtains the video information of the vehicle in real time.

[0067] Optionally, after receiving the unlocking instruction, the first controller 12 sends a control signal to the camera module 11 to obtain the video information of the vehicle.

[0068] In an alternative embodiment, a two-dimensional code is pre-set on the vehicle. When the user needs to borrow the vehicle, the user opens the application pre-installed on the user terminal, scans the two-dimensional code of the vehicle through the application to obtain the Bluetooth communication address of the vehicle, and then establishes a Bluetooth communication connection with the vehicle. The application creates an order and sends an unlocking instruction to the vehicle through Bluetooth communication to achieve borrowing the vehicle.

[0069] In another alternative embodiment, when the user needs to borrow a vehicle, the user opens the pre-installed application on the user terminal, scans the vehicle QR code through the application to obtain the unique identifier of the vehicle, and sends a vehicle borrowing request to the server. The vehicle borrowing request carries the unique identifier of the vehicle. After receiving the vehicle borrowing request, the server creates an order, obtains the communication address of the vehicle through the unique identifier, and sends an unlocking instruction to the vehicle to achieve vehicle borrowing.

[0070] Optionally, when the first controller 12 receives the unlocking instruction and detects the vehicle speed change information, it sends a control signal to the camera module 11 to obtain the video information of the vehicle. For example, the information processing device 1 further includes a speedometer for detecting the vehicle speed and transmitting it to the first controller 12. After receiving the unlocking instruction, the first controller 12 controls the speedometer to detect the vehicle speed in real time to determine the vehicle speed change information. After the vehicle speed change information is detected for the first time, it indicates that the user starts riding, and then the first controller 12 sends a control signal to the camera module 11 to obtain the video information of the vehicle.

[0071] In this embodiment, while the first controller 12 sends a control signal to the camera module 11, the first controller 12 respectively sends control instructions to the positioning module 13 and the sensor 15, so that the positioning module 13 obtains the positioning information of the vehicle in real time, and the sensor 15 obtains the acceleration information of the vehicle in real time.

[0072] Step S200: Perform image recognition on the video information through a pre-trained detection model to determine the event information.

[0073] In this embodiment, the detection model in the first controller 12 can process the received video information, positioning information, and acceleration information to determine the corresponding event information.

[0074] Optionally, step S200 may include steps S211 - S214. Specifically, reference can be made to Figure 4 .

[0075] Figure 4 is the flowchart of performing image recognition on the video information through a pre-trained detection model to determine the event information in the embodiment of the present invention. As Figure 4 shown, the process of performing image recognition on the video information through a pre-trained detection model to determine the event information in this embodiment includes the following steps:

[0076] Step S211: Identify the status information of the traffic indication device in the video information.

[0077] In this embodiment, the detection model can perform image recognition on video information (for example, by extracting feature vectors, that is, extracting the feature vectors of the traffic indication device image) to determine the image information of the traffic indication device and the corresponding status information. Among them, the image information of the traffic indication device includes color, size, text, logo image, image position, etc. The status information includes a passable status and a non-passable status. For example, if the traffic indication device is a traffic light and the image information is red, the corresponding status information is the non-passable status. Another example, if the traffic indication device is a traffic light and the image information is green, the corresponding status information is the passable status. Still another example, if the traffic indication device is a traffic sign and the image information includes text / logo image indicating "no entry", the corresponding status information is the non-passable status.

[0078] In this embodiment, the detection model can perform image recognition on video information in various ways to determine the distance between the traffic indication device and the vehicle.

[0079] Optionally, the detection model can determine the distance between the vehicle and the traffic indication device according to the image position of the traffic indication device and the positioning information of the vehicle. Specifically, the image position of the target object (the target object is, for example, a traffic indication device) in a certain frame of the video information can be determined by a target tracking technology, and the positioning information of the vehicle (such as longitude and latitude) and the parameter information of the camera module 11 can be obtained (such as the internal parameters of the lens (such as lens focal length, principal point and other parameters) and the external parameters (such as the position and direction of the lens fixed to the vehicle). Then, the detection model can perform calculations according to the image position of the traffic indication device, the positioning information of the vehicle and the parameter information of the camera module 11 to determine the distance between the vehicle and the traffic indication device in this frame of image. Similarly, by performing calculations according to the image position of the traffic indication device, the corresponding positioning information of the vehicle and the parameter information of the camera module 11 in each frame of the video information, the detection model can determine the distance between the vehicle and the traffic indication device in real time and determine the change amount of the distance between the vehicle and the traffic indication device. Among them, the image position of the traffic indication device in a certain frame of image represents the position information of the traffic indication device in this frame of image.

[0080] Optionally, the target tracking technology can be implemented by a target tracking algorithm. The target tracking algorithm is, for example, SORT (Simple Online and Realtime Tracking), SSD (Single ShotMultibox Detector), etc. Thus, the detection model can determine the distance between the vehicle and the traffic indication device through the target tracking algorithm.

[0081] Optionally, the video information of the vehicle includes multiple video data from different angles. The detection model can also identify traffic indication devices in different video data to determine the distance between the vehicle and the traffic indication devices. Specifically, the camera module 11 can include multiple lenses, and the vehicle video data from different perspectives can be obtained through the multiple lenses respectively. It is easy to understand that the video data acquisition times of the respective lenses need to be synchronized. The detection model can calculate the parallax information based on the positions of the traffic indication devices in different video data. The parallax information represents the displacement amount of the traffic indication devices in the images of different video data. Then, the detection model can perform calculations based on the parallax information and the parameter information of the camera module 11 (such as the internal parameters and external parameters of each lens), and can determine the distance between the vehicle and the traffic indication devices in real time, and determine the change amount of the distance between the vehicle and the traffic indication devices. Among them, the change amount of the distance can represent that the distance between the vehicle and the traffic indication devices increases, decreases, or remains unchanged.

[0082] It should be noted that, in this embodiment, the detection model performs image recognition on the video information to determine the distance between the traffic indication device and the vehicle as an example for description. It should be understood that the distance between the traffic indication device and the vehicle can also be determined by other means according to requirements. For example, a ranging device (such as a ranging radar, etc.) is set on the vehicle, and the distance between the traffic indication device and the vehicle is determined through the ranging device.

[0083] Step S212: Determine whether the status information is a non-passable status and the distance between the vehicle and the traffic indication device decreases.

[0084] In this embodiment, if the detection model determines that the status information of the traffic indication device is a non-passable status and the distance between the vehicle and the traffic indication device decreases, which indicates that the vehicle is passing dangerously, then step S213 is executed. If the detection model determines that the status information is a passable status, or the status information is a non-passable status and the distance between the vehicle and the traffic indication device increases or remains unchanged, which indicates that the vehicle is passing correctly, then step S214 is executed.

[0085] Step S213: Determine that the event information is an incorrect passing behavior.

[0086] In this embodiment, the detection model determines that the event information is an incorrect passing behavior.

[0087] Step S214: Determine that the event information is a correct passing behavior.

[0088] In this embodiment, the detection model determines that the event information is a correct passing behavior. Thus, the detection model can accurately determine whether the user's passing behavior is correct.

[0089] Optionally, step S200 may include steps S221 - S224. Specifically, reference can be made to Figure 5 .

[0090] Figure 5 This is a flowchart for image recognition of video information by a pre-trained detection model to determine event information in an embodiment of the present invention. As Figure 5 shown, the process of image recognition of video information by a pre-trained detection model to determine event information in this embodiment includes:

[0091] Step S221: Identify traffic marking lines in the video information.

[0092] In this embodiment, the detection model can perform image recognition on the video information to determine the type of traffic marking lines. That is, identify the road surface information in the video information to determine the road surface traffic marking lines, such as lane lines, guiding lines, etc.

[0093] In this embodiment, the detection model can use various methods to identify traffic marking lines in the video information.

[0094] Optionally, the detection model can extract the feature vectors of traffic marking lines, and based on the pre-set correspondence between the feature vectors of marking lines and the types of marking lines, determine the corresponding type of traffic marking lines according to the feature vectors of traffic marking lines. Among them, the types of traffic marking lines are, for example, sidewalks, bicycle lanes, motor vehicle lanes, non-motor vehicle lanes, etc.

[0095] Optionally, the detection model can identify traffic marking lines in the video information through a preset detection algorithm. The detection algorithms are, for example, the Hough algorithm (i.e., the Hough Transform algorithm) and the Canny edge detection algorithm, etc. Among them, the Hough algorithm can extract traffic marking lines from the video image (i.e., video information) by means of feature extraction. It is easy to understand that since many traffic marking lines are located at the edges of lanes, the Canny edge detection algorithm can detect traffic marking lines in the video image.

[0096] It should be noted that in this embodiment, the detection model identifies traffic marking lines in the video information by extracting the feature vectors of traffic marking lines and a preset detection algorithm. It should be understood that the first controller 12 can also identify traffic marking lines in the video information by other means. For example, the first controller 12 identifies traffic marking lines in the video information by means of image segmentation technology, etc. Taking the image segmentation technology as an example, the detection model can extract multiple pixel points of traffic marking lines from a certain frame of the video information, and then perform clustering processing on the multiple pixel points to identify traffic marking lines.

[0097] In this embodiment, the detection model can also obtain the type information of the vehicle to determine whether the type information of the vehicle matches the type of the traffic sign line. It is easy to understand that the first controller 12 can pre-store the type information of the vehicle. The type information of the vehicle is, for example, a bicycle, an electric bicycle, a motor vehicle, etc.

[0098] Step S222: Determine whether the traffic sign line matches the type of the vehicle.

[0099] In this embodiment, if the detection model determines that the traffic sign line does not match the type of the vehicle (for example, the type of the traffic sign line is a sidewalk and the type of the vehicle is a bicycle, then they do not match), step S223 is executed. If the detection model determines that the traffic sign line matches the type of the vehicle (for example, the type of the traffic sign line is a bicycle lane and the type of the vehicle is a bicycle, then they match), step S224 is executed.

[0100] Step S223: Determine that the event information is an act of misusing the lane.

[0101] In this embodiment, the detection model determines that the event information is an act of misusing the lane.

[0102] Step S224: Determine that the event information is an act of misusing the lane.

[0103] In this embodiment, the detection model determines that the event information is an act of misusing the lane.

[0104] Optionally, step S200 may include steps S231 - S237. Specifically, reference can be made to Figure 6 .

[0105] Figure 6 is a flowchart of determining event information by performing image recognition on video information through a pre-trained detection model in an embodiment of the present invention. As Figure 6 shown, the process of performing image recognition on video information through a pre-trained detection model to determine event information in this embodiment includes:

[0106] Step S231: Identify the traffic sign line and road condition information in the video information, where the road condition information includes other vehicles and / or pedestrians.

[0107] In this embodiment, the detection model can identify the traffic sign line in the video information. The specific implementation manner is similar to step S221, and will not be elaborated herein in this embodiment.

[0108] In this embodiment, the detection model can identify the road conditions information in the video information, and the road conditions information includes other vehicles and / or pedestrians. This embodiment takes the road conditions information including other vehicles and / or pedestrians as an example for illustration. It should be understood that the road conditions information may also include target objects, such as traffic guiding devices (e.g., guiding signs, guiding lights, etc.), pets of pedestrians (e.g., puppies, kittens), etc. That is to say, considering that pedestrians may walk with their pets, etc., the moving direction of the pet can also be used as a reference direction for judging whether a vehicle is going in the wrong direction.

[0109] In this embodiment, the detection model can identify the road conditions information in the video information in various ways.

[0110] Optionally, the detection model can perform vehicle recognition on other vehicles in the video information and / or perform face recognition on pedestrians in the video information. That is to say, the video information may include vehicles, or the video information may include pedestrians, or the video information may include both vehicles and pedestrians. Among them, the detection model can perform vehicle recognition and face recognition through models such as YOLO (You Only Look Once) and Faster R-CNN (Faster Region based Convolutional Neural Networks).

[0111] Optionally, the detection model can identify the road conditions information in the video information through a target tracking algorithm.

[0112] Step S232: Determine the lane direction and the moving direction respectively according to the traffic sign lines and the road conditions information.

[0113] In this embodiment, the detection model can determine the lane direction according to the recognized traffic sign lines (such as guiding lines, lane lines, etc.) and determine the moving direction of other vehicles and / or pedestrians according to the road conditions information.

[0114] Optionally, the detection model can perform motion vector analysis (such as analyzing motion speed, acceleration, direction, etc.) on the traffic sign lines, other vehicles and / or pedestrians in the consecutive frame images of the video information to determine the lane direction and the moving direction respectively.

[0115] Optionally, the detection model can determine the lane direction and the moving direction respectively through a target tracking algorithm.

[0116] It should be noted that in this embodiment, the detection model determines the lane direction and the traveling direction through motion vector analysis and target tracking algorithms respectively. It should be understood that the first controller 12 can also determine the lane direction and the traveling direction in other ways. For example, it can perform character recognition on the text information of traffic signs to determine the lane direction, or it can determine the traveling direction by extracting the vector of the target part of pedestrians (such as the vector between the head and the feet, the vector between the waist and the feet, etc.) and performing comparative analysis.

[0117] Step S233: Determine whether the lane direction is consistent with the traveling direction.

[0118] In this embodiment, if the detection model detects that the lane direction is inconsistent with the traveling directions of a predetermined number of other vehicles and / or pedestrians, and the predetermined number is greater than the third threshold, then step S234 is executed. If the detection model detects that the lane direction is inconsistent with the traveling directions of a predetermined number of vehicles and / or pedestrians, and the predetermined number is less than or equal to the third threshold, or the detection model detects that the lane direction is consistent with the traveling directions of all vehicles and / or pedestrians, then step S237 is executed. That is to say, considering the possible reverse traveling of other vehicles and / or pedestrians and other situations, in this case, this embodiment compares the number of other vehicles and pedestrians whose detected traveling directions are inconsistent with the lane direction with the third threshold to accurately determine whether the user is traveling in reverse.

[0119] Optionally, the detection model can calculate the angular deviation value between the lane direction and the traveling direction. If the angular deviation value is greater than the deviation threshold, it can be determined that the lane direction is inconsistent with the traveling direction.

[0120] For example, the third threshold is 2. The video information collected by the camera module 11 of the current user riding the two-wheeler A includes pedestrian B, other vehicles C and D. The traveling directions of pedestrian B, other vehicles C and D are opposite to the lane direction E in the video information. That is, the number of other vehicles and pedestrians whose traveling directions are inconsistent with the lane direction is 3, which is greater than the third threshold. Then step S234 is executed to further determine whether the two-wheeler A is traveling in reverse.

[0121] Step S234: Determine the trajectory direction according to the driving trajectory.

[0122] In this embodiment, the detection model can determine the trajectory direction according to the driving trajectory in various ways. That is to say, the detection model determines whether it is traveling in reverse again according to the trajectory direction of the vehicle to improve the detection accuracy.

[0123] Optionally, the detection model can determine the trajectory direction based on preset road network information and driving trajectory. Specifically, the road network information is used to characterize a road system composed of various roads that are interconnected and interwoven into a network distribution within a certain geographical area. The road network information can include multiple roads and corresponding road directions. The detection model can determine the road section corresponding to the driving trajectory in the road network information, and determine the direction corresponding to the road section as the trajectory direction of the driving trajectory. In other words, considering that the driving trajectory determined by the positioning module 13 may have errors / offsets, the vehicle may have a driving trajectory that turns back multiple times, etc., resulting in the trajectory direction constantly changing, while the direction of the road section is accurate and unchanged, so the direction corresponding to the road section can be determined as the trajectory direction of the driving trajectory.

[0124] Optionally, the detection model can determine multiple trajectory feature vectors based on the driving trajectory, and determine the trajectory direction based on the multiple trajectory feature vectors. For example, the driving trajectory is divided into multiple trajectory feature vectors, and then the multiple trajectory feature vectors are filtered according to a filtering condition to obtain one or more target trajectory feature vectors (the filtering condition may be to filter out driving trajectories that have turned back multiple times, etc.), so as to filter out, and then determine the trajectory direction of the driving trajectory based on the one or more target trajectory feature vectors.

[0125] It should be noted that this embodiment is described by taking the detection model determining the trajectory direction through preset road network information and driving trajectory, and determining the trajectory direction according to multiple trajectory feature vectors corresponding to the driving trajectory as an example. It should be understood that the first controller 12 can also determine the trajectory direction in other ways. For example, the detection model can determine the trajectory direction according to the driving trajectory through a recurrent neural network, a long short-term memory network (LSTM, Long Short-Term Memory), etc.

[0126] Step S235: determine whether the trajectory direction is consistent with the lane direction.

[0127] In this embodiment, if the detection model determines that the trajectory direction is inconsistent with the lane direction, step S236 is executed. If the trajectory direction is consistent with the lane direction, step S237 is executed.

[0128] Step S236: Determine that the event information is a retrograde behavior.

[0129] In this embodiment, the detection model determines that the event information is retrograde behavior.

[0130] Step S237: Determine that the event information is correct driving behavior.

[0131] In this embodiment, the detection model determines that the event information is a correct driving behavior.

[0132] Optionally, the detection model can also determine the acceleration change amount based on the acceleration information (such as moving acceleration, gravitational acceleration, etc.) transmitted in real time by the sensor 15 (for example, calculate the corresponding acceleration change amount according to the acceleration information at different times). If the acceleration change amount within a predetermined time is greater than the first threshold, the detection model determines that the event information is a collision accident. If the acceleration change amount within a predetermined time is less than or equal to the first threshold, the detection model determines that the event information is that no collision accident has occurred. That is to say, when a collision occurs, the acceleration changes greatly. Therefore, by comparing the acceleration change amount with the first threshold, it can be determined whether a collision accident has occurred.

[0133] Optionally, the detection model can determine the acceleration change amount based on the acceleration information through methods such as recurrent neural networks and long short-term memory networks.

[0134] Optionally, the detection model can also determine whether a vehicle has a collision accident based on the vibration change amount. Specifically, if the acceleration change amount within a predetermined time is greater than the first threshold and the vibration change amount within a predetermined time is greater than the vibration threshold, the detection model can determine that the event information is a collision accident.

[0135] Optionally, step S200 may include steps S241 - S245. Specifically, reference can be made to Figure 7 .

[0136] Figure 7 This is a flowchart of image recognition of video information by a pre-trained detection model in an embodiment of the present invention to determine event information. As Figure 7 shown, the process of image recognition of video information by a pre-trained detection model in this embodiment to determine event information includes:

[0137] Step S241: Identify the image angle information of multiple frames in the video information.

[0138] In this embodiment, the detection model can extract the feature information (such as feature points, feature vectors) of the same target object (the target object is, for example, a vehicle, a pedestrian, vegetation, a traffic indication device, etc. in the video information) in consecutive multiple frames of the video information, and then calculate the feature information corresponding to the adjacent frame images of the target object to obtain the image angle information of the adjacent frames. Similarly, by calculating the feature information of each adjacent frame image for the same target object in consecutive multiple frames of the detection model, the image angle information of each frame image can be determined.

[0139] Step S242: Determine the angle change amount according to each image angle information.

[0140] In this embodiment, the detection model may calculate the angle variation of different frame images according to a preset frame interval, for example, calculating the angle variation once every other frame, or once every two frames.

[0141] Step S243: determine whether the angle change is greater than a second threshold.

[0142] In this embodiment, the detection model determines whether the angle change of different frame images in the video information is greater than a second threshold. If the angle change is greater than the second threshold (e.g., 90 degrees, 100 degrees, etc.), step S244 is executed. If the angle change is less than or equal to the second threshold, step S245 is executed.

[0143] Optionally, the sensor 15 can also detect the gravity acceleration of the vehicle. The detection model can also calculate the change in gravity acceleration at different times. If the angle change is greater than the second threshold and the gravity acceleration change is greater than the fourth threshold, it is determined that a fall accident has occurred. In this way, the accuracy of event information detection can be improved.

[0144] Step S244: Determine the event information as a fall accident.

[0145] In this embodiment, the detection model determines that the event information is a fall accident.

[0146] Step S245: Determine the event information that no fall accident occurred.

[0147] In this embodiment, the detection model determines the event information as no fall accident occurred.

[0148] In this embodiment, the detection model can determine that the event information is one or more of wrong traffic behavior, wrong lane use behavior, wrong-way behavior, fall-down accident, and collision accident. For example, if a user rides a bicycle wrongly on a motor vehicle lane, the detection model detects that the event information is wrong lane use behavior and wrong-way behavior.

[0149] Step S300: In response to the event information indicating that the vehicle is in danger, a corresponding safety maintenance operation is performed according to the event information.

[0150] In this embodiment, the first controller 12 determines whether the event information output by the detection model indicates that the vehicle is in danger. If the event information is an incorrect passing behavior, an incorrect lane use behavior, a reverse driving behavior, a falling accident, or a collision accident, the first controller 12 determines that the vehicle is in danger. If there is no incorrect passing behavior, incorrect lane use behavior, reverse driving behavior, falling accident, and collision accident, the process can return to step S100 to continuously detect during the user's riding process. Further, after the user's riding ends, after the first controller 12 receives the lock-up instruction / order end instruction sent by the server / user terminal, it controls the camera module 11 to stop acquiring the video information of the vehicle, controls the positioning module 13 to stop acquiring the positioning information of the vehicle, and controls the sensor 15 to stop acquiring the acceleration information of the vehicle, and the vehicle borrowing process ends.

[0151] In this embodiment, if the event information indicates that the vehicle is in danger, the first controller 12 sends control information to the main control device 2 of the vehicle through the Bluetooth module 14 according to the event information, so that the main control device 2 performs corresponding safety maintenance operations.

[0152] Optionally, if the event information is an incorrect passing behavior, an incorrect lane use behavior, or a reverse driving behavior, the first controller 12 sends first control information to the second controller 22 of the main control device 2 of the vehicle according to the event information. The second controller 22 controls the speaker 21 to play the corresponding warning information according to the first control information. It is easy to understand that the first control information sent corresponding to different event information is different. For example, if the event information is an incorrect lane use behavior, the second controller 22 controls the speaker 21 to play the corresponding warning information according to the first control information: "You are currently using the lane incorrectly. Please use the correct lane." For another example, if the event information is a reverse driving behavior, the second controller 22 controls the speaker 21 to play the corresponding warning information according to the first control information: "You are driving in reverse and are in danger. Please avoid the risk in time."

[0153] Optionally, if the event information is a falling accident or a collision accident, the first controller 12 acquires the video image corresponding to the event information for storage, and sends second control information to the main control device 2 of the vehicle according to the event information. The second controller 22 controls the speaker 21 to play the corresponding help information (such as a help voice, etc.) according to the second control information, and / or sends a corresponding help message to the server. The help message can carry the unique identifier of the vehicle. Then, when the server receives the help message, it can notify the operation personnel to handle it immediately (such as calling the police for help).

[0154] Optionally, after the detection model detects that the event information is a falling accident or a collision accident, the first controller 12 continuously stores the subsequently acquired video information.

[0155] Optionally, the help request message carries a video image corresponding to the event information, and the server can store the video image corresponding to the event information to achieve evidence preservation.

[0156] For example, reference can be made to Figure 8 . Figure 8 It is a schematic diagram of the information processing device fixed to the vehicle-mounted helmet in the embodiment of the present invention. As Figure 8 shown, vehicle b includes a vehicle body 100 and a vehicle-mounted helmet 3, and the information processing device 1 is fixed to the vehicle-mounted helmet 3. User c rides vehicle b, and at the same time, user c wears the vehicle-mounted helmet 3. After starting to ride, the camera module 11 in the information processing device 1 collects video information in front of vehicle b, the positioning module 13 obtains the positioning information of the vehicle, and the sensor 15 obtains the acceleration information of the vehicle. The video information collected by the camera module 11 in front of vehicle b includes traffic lights f, other vehicles a1 and a2, and traffic marking lines d1, d2, and d3. When the first controller 12 detects that the status information of the traffic light f indicates that it is possible to pass, the first controller 12 determines that the event information is a correct passing behavior. When the first controller 12 detects that the lane directions represented by the traffic marking lines d1, d2, and d3 are consistent with the traveling directions of other vehicles a1 and a2, it determines that the event information is a correct driving behavior. When the first controller 12 detects that the angular change amount of multiple frame images in the video information is less than the second threshold, it determines that the event information is that no falling accident has occurred. When the first controller 12 detects that the acceleration change amount within a predetermined time is less than the first threshold, it determines that the event information is that no collision accident has occurred. When the first controller 12 detects that the traffic marking lines d1, d2, and d3 are motor vehicle lane lines and do not match the type of vehicle b, which is a shared bicycle, it determines that the event information is an incorrect lane usage behavior. Then, the first controller 12 sends the first control information to the second controller 22 of the main control device 2 through the Bluetooth module 14. After receiving the first control information, the second controller 22 controls the speaker 21 to play a warning message: "You are currently using the wrong lane. Please use the correct lane." Further, user c rides vehicle b to the bicycle lane. Thus, the safety maintenance operation is completed.

[0157] In the embodiment of the present invention, by setting a camera module and a first controller in the information processing device, the first controller is deployed with a pre-trained detection model. The video information of the vehicle is obtained through the camera module, and then the first controller performs image recognition on the video information through the detection model to determine the event information. If the event information indicates that the vehicle is in danger, the corresponding safety maintenance operation is performed according to the event information. Thus, dangerous events can be effectively detected and the corresponding safety maintenance operations can be performed to improve safety and have high applicability.

[0158] Figure 9 It is a schematic diagram of the information processing device in the embodiment of the present invention. As Figure 9As shown in the figure, the information processing device of this embodiment includes a receiving unit 410, an image recognition unit 420, and an operation execution unit 430. Among them, the receiving unit 410 is used to receive the video information of the vehicle transmitted by the camera module 11. The image recognition unit 420 is used to perform image recognition on the video information through a pre-trained detection model to determine event information. The operation execution unit 430 is used to execute corresponding safety maintenance operations according to the event information in response to the event information indicating that the vehicle is in danger.

[0159] In the embodiment of the present invention, by setting a camera module and a first controller in the information processing device, the first controller is deployed with a pre-trained detection model, the video information of the vehicle is obtained through the camera module, and then the first controller performs image recognition on the video information through the detection model to determine event information. If the event information indicates that the vehicle is in danger, corresponding safety maintenance operations are performed according to the event information. Thus, dangerous events can be effectively detected and corresponding safety maintenance operations can be performed to improve safety and have high applicability.

[0160] Figure 10 is a schematic diagram of the electronic device according to the embodiment of the present invention. As Figure 10 shown, Figure 10 The electronic device shown is a general information processing device, which includes a general computer hardware structure and at least includes a processor 510 and a memory 520. The processor 510 and the memory 520 are connected through a bus 530. The memory 520 is suitable for storing instructions or programs executable by the processor 510. The processor 510 can be an independent microprocessor or a set of one or more microprocessors. Thus, the processor 510 executes the instructions stored in the memory 520 to execute the method flow of the embodiment of the present invention as described above to realize the processing of data and the control of other devices. The bus 530 connects the above-mentioned multiple components together and at the same time connects the above-mentioned components to the display first controller 540, the display device, and the input / output (I / O) device 550. The input / output (I / O) device 550 can be a mouse, a keyboard, a modem, a network interface, a touch input device, a body sensing input device, a printer, and other devices well known in the art. Typically, the input / output (I / O) device 550 is connected to the system through the input / output (I / O) first controller 560.

[0161] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, devices (apparatuses), or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be implemented as a computer program product on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0162] The present application is described with reference to the flowcharts of methods, devices (apparatuses), and computer program products according to the embodiments of the present application. It should be understood that each process in the flowchart can be implemented by computer program instructions.

[0163] These computer program instructions can be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device that implements the process Figure 1 specified functions in one or more of the processes.

[0164] These computer program instructions can also be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the Figure 1 specified functions in one or more of the processes.

[0165] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program, and the computer-readable program is used for a computer to execute the above-mentioned partial or all method embodiments.

[0166] That is, those skilled in the art can understand that all or part of the steps in implementing the above-mentioned embodiment methods can be completed by specifying relevant hardware through a program, and the program is stored in a storage medium, including several instructions for causing a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks, etc., which can store program codes.

[0167] The above are only the preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various modifications and variations can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. An information processing device, characterized in that, The information processing device includes: a camera module for acquiring video information of a vehicle and transmitting it to a first controller; the first controller deployed with a pre-trained detection model for performing image recognition on the video information to determine event information, and in response to the event information indicating that the vehicle is in danger, performing corresponding safety maintenance operations according to the event information.

2. The information processing device according to claim 1, wherein The first controller is further configured to identify status information of a traffic indication device in the video information, and in response to the status information being a non-passable state and detecting that the distance between the vehicle and the traffic indication device decreases, determine the event information as an incorrect passing behavior.

3. The information processing apparatus according to claim 1, wherein The first controller is further configured to identify traffic marking lines in the video information, and in response to the traffic marking lines not matching the type of the vehicle, determine the event information as an incorrect lane usage behavior.

4. The information processing device according to claim 1, wherein The information processing device further includes: a positioning module for acquiring positioning information of the vehicle and transmitting it to the first controller, where the positioning information includes a driving trajectory; wherein, the first controller is further configured to identify traffic marking lines and road condition information in the video information, the road condition information includes other vehicles and / or pedestrians, determine a lane direction and a traveling direction according to the traffic marking lines and the road condition information respectively, in response to the lane direction not being consistent with the traveling direction, determine a trajectory direction according to the driving trajectory, and in response to the trajectory direction not being consistent with the lane direction, determine the event information as a reverse driving behavior.

5. The information processing device according to any one of claims 2-4, characterized in that, The first controller is further configured to send first control information to a main control device of the vehicle according to the event information, so that the main control device plays a corresponding warning message.

6. The information processing apparatus according to claim 1, wherein The information processing device further includes: a sensor for acquiring acceleration information of the vehicle and transmitting it to the first controller; wherein, the first controller is further configured to determine an acceleration change amount according to the acceleration information, and in response to the acceleration change amount within a predetermined time being greater than a first threshold, determine the event information as a collision accident.

7. The information processing apparatus according to claim 1, wherein The first controller is further configured to identify image angle information of multiple frames in the video information, determine an angle change amount according to each of the image angle information, and in response to the angle change amount being greater than a second threshold, determine the event information as a falling accident.

8. The information processing apparatus according to any one of claims 6-7, characterized in that, The first controller is further configured to acquire a video image corresponding to the event information for storage, send second control information to a main control device of the vehicle according to the event information, so that the main control device plays a corresponding help message, and / or send a corresponding help message.

9. The information processing apparatus according to claim 1, wherein The information processing device further includes: a Bluetooth module; wherein, the first controller is further configured to send control information to a main control device of the vehicle through the Bluetooth module according to the event information, so that the main control device performs corresponding safety maintenance operations.

10. A vehicle, characterized in that, The vehicle includes: a vehicle body; the information processing device according to any one of claims 1-9.

11. The vehicle according to claim 10, characterized in that, The vehicle is a two-wheeler, and the vehicle further includes a basket and / or a vehicle helmet, and the information processing device is fixed to the basket or the vehicle helmet.

12. The vehicle according to claim 10, characterized in that, The vehicle further includes: The master control device includes a second controller and a speaker; wherein, the second controller is configured to play a corresponding warning message through the speaker in response to receiving the first control information sent by the information processing device; or the second controller is configured to play a corresponding help message through the speaker and / or send a corresponding help message to the server in response to receiving the second control information sent by the information processing device.

13. An information processing method, characterized in that, The method includes: Receiving video information of the vehicle transmitted by the camera module; Performing image recognition on the video information through a pre-trained detection model to determine event information; In response to the event information indicating that the vehicle is in danger, performing a corresponding safety maintenance operation according to the event information.

14. An information processing apparatus, characterized in that, The device includes: A receiving unit, configured to receive video information of the vehicle transmitted by the camera module; An image recognition unit, configured to perform image recognition on the video information through a pre-trained detection model to determine event information; An operation execution unit, configured to perform a corresponding safety maintenance operation according to the event information in response to the event information indicating that the vehicle is in danger.

15. A computer-readable storage medium storing computer program instructions thereon, characterized in that, The computer program instructions, when executed by a processor, implement the method according to claim 13.

Images