Vehicle monitoring method, electronic device, and storage medium

Abstract

The present application relates to the technical field of vehicle safety, and specifically provides a vehicle monitoring method, an electronic device, and a storage medium, which aim to solve the problem of a surrounding vehicle causing damage to a present vehicle due to sudden door opening. For this purpose, the vehicle monitoring method of the present application comprises: acquiring an image of the surroundings of a present vehicle; on the basis of the image, determining a vehicle around the present vehicle; monitoring a door state of the vehicle, and, on the basis of the door state, determining whether the vehicle is performing a door opening action; and, on the basis of a determination result, determining whether to issue warning information. When the present vehicle is in a parked state, according to the present application, the state of the surrounding vehicle is monitored in real time, so as to determine whether the surrounding vehicle poses a potential risk to the present vehicle during the process of opening a door, so that an owner of the present vehicle can take corresponding countermeasures when receiving the warning information, thereby reducing the property loss of a user.

Description

Vehicle monitoring method, electronic device and storage medium

[0001] This application claims priority to Chinese patent application No. 202311793381.9, filed on December 22, 2023, with the invention name “Vehicle Monitoring Method, Electronic Device and Storage Medium”. The entire contents of the above Chinese patent application are incorporated into this application by reference. Technical Field

[0002] The present application relates to the field of vehicle safety technology, and specifically provides a vehicle monitoring method, electronic equipment, and storage medium. Background Art

[0003] In order to improve vehicle safety, most smart cars on the market today are equipped with a guard function. That is, when the vehicle is stationary, it uses cameras and sensors around the body to monitor the vehicle's surroundings in real time and issue early warning information based on the monitoring information.

[0004] Current vehicle security technologies typically use surveillance footage to determine if a vehicle has been stolen, or detect the movement of pedestrians and other vehicles to determine if they pose a potential threat to the vehicle or its passengers. However, nearby vehicles could potentially collide with or scratch the vehicle if their doors suddenly open, causing damage.

[0005] Accordingly, this field requires a new technical solution to solve the above problems. Summary of the Invention

[0006] The present application aims to solve the above technical problem, that is, to solve the problem of surrounding vehicles causing damage to the vehicle due to sudden door opening.

[0007] In a first aspect, the present application provides a vehicle monitoring method, comprising:

[0008] Acquire images around the vehicle;

[0009] determining vehicles around the vehicle based on the image;

[0010] Monitoring the door status of the vehicle and determining whether the vehicle is in the process of opening the door according to the door status; and

[0011] Determine whether to issue an early warning message based on the judgment results.

[0012] When using the above technical solution, the door status of surrounding vehicles is detected through image recognition, and whether a surrounding vehicle is in the process of opening its door is further determined. Then, based on the judgment result, the vehicle selectively issues a warning message. In this way, when the vehicle is parked, by monitoring the status of surrounding vehicles in real time, it can be determined whether the vehicle's door opening process poses a potential risk to the vehicle. When the warning message is received, the owner can take appropriate countermeasures, thereby reducing the user's property loss.

[0013] In one technical solution of the above monitoring method, the method further includes:

[0014] Determining a position of each of the surrounding vehicles relative to the host vehicle;

[0015] According to the position, the vehicle entering the preset range of the vehicle is screened and marked as a target vehicle;

[0016] The step of “monitoring the door status of the vehicle and determining whether the vehicle is performing a door opening action based on the door status” includes:

[0017] Monitor the door status of the target vehicle and determine whether the surrounding vehicles are performing door opening actions based on the door status.

[0018] When adopting the above technical solution, the surrounding vehicles are screened according to the distance between the surrounding vehicles and the vehicle, and the vehicles that pose a higher potential risk to the vehicle are determined as target vehicles. This not only reduces the processing volume, but also reduces the misjudgment rate and improves the judgment accuracy.

[0019] In one technical solution of the above monitoring method, the step of "screening vehicles that enter the preset range of the vehicle according to the position" includes:

[0020] Inputting the image of each vehicle among the surrounding vehicles into the first object detection model to obtain the area corresponding to each vehicle;

[0021] Determining the distance of each vehicle relative to the vehicle based on the pixel points at the outer edge of the area, and filtering vehicles that enter the preset range of the vehicle based on the distance;

[0022] The first target detection model reflects the mapping relationship between the first feature quantity representing the vehicle appearance and the region.

[0023] In one technical solution of the above monitoring method, the step of “monitoring the door status of the target vehicle” includes:

[0024] Acquiring a current image frame of the target vehicle; and

[0025] Inputting the current image frame into a second target detection model to determine whether the door state of the target vehicle is open or closed;

[0026] The second target detection model reflects the mapping relationship between the second feature quantity representing the vehicle appearance and the door opening state.

[0027] In one technical solution of the above monitoring method, the step of “determining whether the target vehicle is performing a door opening action based on the door state” includes:

[0028] It is determined whether the target vehicle is performing a door opening action according to the door states corresponding to the multiple image frames of the target vehicle.

[0029] In one technical solution of the above monitoring method, the step of “determining whether the target vehicle is performing a door opening action based on the door states corresponding to the multiple image frames of the target vehicle” includes:

[0030] When the door states corresponding to a plurality of consecutive image frames are all open, it is determined that the target vehicle is performing a door opening action.

[0031] In one technical solution of the above monitoring method, the step of “determining whether the target vehicle is performing a door opening action based on the door states corresponding to the multiple image frames of the target vehicle” includes:

[0032] When, in a plurality of consecutive image frames, a ratio of the number of frames in which the door is in the open state divided by the total number of frames is greater than a preset value, it is determined that the target vehicle is performing a door opening action.

[0033] When the above technical solution is adopted, when judging whether the target vehicle is performing a door opening action, the determination is made in combination with multiple continuous image frames, thereby further improving the accuracy and sensitivity of the judgment result.

[0034] In one technical solution of the above monitoring method, the step of "determining whether to issue an early warning message based on the judgment result" includes:

[0035] When it is determined that at least one of the target vehicles is performing a door opening action, a warning message is issued.

[0036] In a second aspect, the present application provides an electronic device comprising a processor and a storage device, wherein the storage device is suitable for storing multiple program codes, and the program codes are suitable for being loaded and run by the processor to execute the monitoring method described in any technical solution in the first aspect.

[0037] In a third aspect, the present application provides a computer-readable storage medium storing a plurality of program codes, wherein the program codes are suitable for being loaded and run by a processor to execute the monitoring method described in any one of the technical solutions in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The preferred embodiments of the present application are described below with reference to the accompanying drawings, in which:

[0039] FIG1 is a flow chart of the main steps of a vehicle monitoring method according to one embodiment of the present application;

[0040] FIG2 is a flowchart of a training procedure for a first target detection model according to one embodiment of the present application;

[0041] FIG3 is a flowchart showing detailed steps of a vehicle monitoring method according to an embodiment of the present application.

[0042] FIG4 is a flowchart showing detailed steps of a vehicle monitoring method according to an embodiment of the present application.

[0043] FIG5 is a schematic diagram of the internal structure of an electronic device according to an embodiment of the present application.

[0044] In the figure, the reference numerals refer to the following: 110, processor; 120, storage device; 130, communication interface; 140, communication bus. DETAILED DESCRIPTION

[0045] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely intended to illustrate the technical principles of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art may adjust these embodiments as needed to suit specific applications.

[0046] In the description of this application, "processor" may include hardware, software or a combination of the two. The processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor or any other suitable processor. The processor has data and / or signal processing functions. The processor may be implemented in software, hardware or a combination of the two. Non-transitory computer-readable storage media include any suitable media that can store program code, such as a magnetic disk, a hard disk, an optical disk, a flash memory, a read-only memory, a random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as just A, just B or A and B. The term "at least one A or B" or "at least one of A and B" has a similar meaning to "A and / or B" and may include just A, just B or A and B.

[0047] The relevant user personal information that may be involved in the various embodiments of this application is strictly in accordance with the requirements of laws and regulations, following the principles of legality, legitimacy and necessity, and based on the reasonable purposes of business scenarios, to process the personal information that users actively provide during the use of products / services or generated due to the use of products / services, as well as the personal information obtained with the user's authorization.

[0048] The user personal information processed in this application will vary depending on the specific product / service scenario and will be based on the specific scenario in which the user uses the product / service. This may involve the user's account information, device information, driving information, vehicle information, or other related information. The applicant will treat the user's personal information and its processing with a high degree of diligence.

[0049] This application attaches great importance to the security of user personal information and has taken reasonable and feasible security protection measures that comply with industry standards to protect user information and prevent personal information from being accessed, disclosed, used, modified, damaged or lost without authorization.

[0050] In some related technologies, the vehicle's guard mode, in addition to providing anti-theft functionality, can also be used in other application scenarios to determine whether pedestrians, vehicles, or any other moving objects are posing a potential threat to the vehicle based on their motion relative to the vehicle, thereby issuing an alarm to remind the vehicle owner to take the next step. Alternatively, in another scenario, when a passenger is about to open the door, the vehicle's body cameras and / or sensors can determine whether there are obstacles around the vehicle or pedestrians or other objects approaching the vehicle, thereby reminding the passenger whether it is safe to exit the vehicle to avoid collisions with pedestrians or objects during the door opening process, thereby ensuring the safety of both the passenger and the vehicle.

[0051] However, in one application scenario, the vehicle is stationary, and the vehicles parked around it are also stationary. However, due to the influence of the distance between the vehicles, when the doors of the surrounding vehicles suddenly open, they may also cause collisions or scratches on the vehicle. Although in some related technologies, sensors are also installed around the vehicle body, and the sensors can sense vibration information to determine whether external objects have caused damage to the vehicle, the sensitivity of the sensors is affected by their set values, resulting in an inability to accurately determine the collisions caused by the opening of doors by surrounding vehicles. When the doors of surrounding vehicles open and collide with the vehicle, although the collision force is not enough to reach the sensitivity range of the sensor, it may still cause scratches on the vehicle. Therefore, the above method has major drawbacks.

[0052] 1 is a flowchart showing the main steps of a vehicle monitoring method according to one embodiment of the present application. The vehicle monitoring method includes the following steps:

[0053] S101: Acquire images around the vehicle.

[0054] In one embodiment of the present application, at least four cameras can be installed around the vehicle. These cameras can be fisheye cameras to prevent blind spots. Specifically, because each camera has its own independent camera coordinate system, after each camera captures image frames from different locations around the vehicle, an image processing module combines the camera coordinate system with the vehicle coordinate system to stitch these frames together, forming a comprehensive image of the vehicle's surroundings.

[0055] S102: Determine vehicles around the vehicle based on the image.

[0056] Step S102 involves determining which image regions are vehicle regions based on the acquired image. Specifically, based on the image acquired in step S101, the image processing module extracts feature information from the image. The feature information reflects the vehicle's external structure, and the image processing module determines which image regions are vehicle regions based on the feature information.

[0057] S103: Monitor the door status of the vehicle and determine whether the vehicle is in the process of opening the door according to the door status.

[0058] It should be noted that the "door status" mentioned above includes both open and closed states. The door status specifically refers to a static result, namely a single-frame image of the vehicle door captured by a camera. Whether the vehicle is currently opening the door specifically refers to a dynamic process. In one embodiment of the present application, whether the vehicle is currently opening the door is determined by combining multiple single-frame images reflecting the door status.

[0059] S104: Determine whether to issue a warning message based on the judgment result.

[0060] In one embodiment of the present application, in step S104, when it is determined that at least one vehicle is performing a door opening action, a warning message is issued.

[0061] It should be noted that in some embodiments, the warning message may be sent by the vehicle to the owner upon detecting a door opening, alerting the owner that the vehicle may be experiencing an abnormal condition that requires attention. Specifically, a communication connection may be established between the vehicle and a user's portable terminal device (e.g., a mobile phone) so that the vehicle can send a message to the user's mobile phone or other terminal device upon detecting an abnormal condition. Simultaneously, while issuing the warning message, the vehicle may also record the current period using cameras located around the vehicle to preserve video evidence. For example, the vehicle may capture the license plates of nearby vehicles with open doors, or the faces of the vehicle's owner or passengers. This allows the vehicle owner to take further action based on this video evidence if the situation occurs but the owner fails to arrive at the scene in time. In other embodiments, the warning message may also be sent by voice notification upon detecting a nearby vehicle with a door opening, reminding the vehicle to be mindful of the door opening to prevent a collision. The voice notification may include, for example, "Please be mindful of the door opening range," "Please open the door slowly," "Please be mindful of the door opening to avoid collisions with nearby vehicles," and so on. Of course, the method of early warning information is not limited to this. Those skilled in the art can adaptively adjust the method of early warning information, and any equivalent or similar replacements made thereto are within the scope of protection of this application.

[0062] This application uses image recognition to detect the door status of surrounding vehicles, further determine whether surrounding vehicles are in the process of opening their doors, and then selectively issues a warning message based on the judgment result. In this way, when the vehicle is parked, by monitoring the status of surrounding vehicles in real time, it can be determined whether they pose a potential risk to the vehicle during the door opening process. When the vehicle owner receives the warning message, he or she can take appropriate countermeasures, thereby reducing the user's property loss.

[0063] Based on the above steps S101 to S104, referring to FIG2 , a detailed flowchart of the vehicle monitoring method disclosed in one embodiment of the present application is shown. Specifically, after step S102, the vehicle monitoring method further includes:

[0064] S1021: Determine the position of each of the surrounding vehicles relative to the vehicle.

[0065] S1022: Filter vehicles that enter the preset range of the vehicle based on the above position and mark them as target vehicles.

[0066] It should be understood that, among the vehicles around this vehicle, the distance between different vehicles and this vehicle may be different. Then, among these vehicles, the possibility of vehicles closer to this vehicle causing a collision with this vehicle when opening the door is definitely greater than that of vehicles farther away from this vehicle. Moreover, when some vehicles are too far away from this vehicle, they will definitely not cause a collision with this vehicle when opening the door. In step S102, the image of the vehicle has been collected. Therefore, through steps S1021 and S102, the surrounding vehicles are screened according to the distance between the surrounding vehicles and this vehicle, and the vehicles that pose a higher potential risk to this vehicle are determined as target vehicles. This can not only reduce the processing volume, but also reduce the misjudgment rate and improve the judgment accuracy.

[0067] In one embodiment of the present application, in step S1021, the position of each surrounding vehicle relative to the vehicle is determined based on the principle of monocular ranging. Specifically, the camera's internal and external parameters are first adjusted. After the camera acquires an image of each vehicle, conversions are performed between the pixel coordinate system, the image coordinate system, the camera coordinate system, and the world coordinate system to obtain a mapping relationship from image pixels to the world coordinate system. Based on this mapping relationship, the position of each vehicle relative to the vehicle is determined. The above-mentioned monocular ranging principle and the method of determining position through conversion between coordinate systems are well known in the art and will not be elaborated upon in this application.

[0068] As an embodiment of the present application, in step S1022, the area corresponding to each vehicle is set as a rectangular area, that is, the projection of each vehicle on the ground corresponds to a rectangular box. Based on this, the step of "screening vehicles that enter the preset range of the vehicle based on the above position" specifically includes:

[0069] The image of each vehicle among the surrounding vehicles is input into the first target detection model to obtain the area corresponding to each vehicle.

[0070] The distance of each vehicle relative to the vehicle is determined based on the pixel points at the outer edge of the area, and vehicles that enter the preset range of the vehicle are filtered based on this distance.

[0071] The first object detection model is pre-trained and maps a first feature representing the vehicle's appearance to a rectangular area. The feature represents the dimensions of the vehicle's outer border. For example, the feature can be the pixels of the vehicle's outer border. When an image of a surrounding vehicle is input into the first object detection model, the model outputs the rectangular area corresponding to that vehicle. The distance between each vehicle and the target vehicle is then determined using the pixels at the outer edge of the rectangular area. This can be achieved by combining the aforementioned monocular ranging method.

[0072] Among them, the above-mentioned preset range can be determined in combination with actual applications and test data, and the numerical value of the preset range reflects the judgment accuracy. In combination with the application scenario of the present application, it can be seen that when the preset range is too large, more target vehicles are screened out, which may cause some target vehicles to actually not cause a collision with the vehicle when opening the door, but issue a warning message, resulting in a misjudgment. When the preset range is too small, fewer target vehicles are screened out, which may cause some vehicles around the vehicle that are not determined as target vehicles to still cause a collision with the vehicle when opening the door. At this time, no warning message is issued, but a collision accident has actually occurred. Therefore, the specific value of the preset range is determined in combination with actual applications and repeated tests, so that it is within a reasonable numerical range, which is conducive to improving the judgment accuracy.

[0073] As shown in FIG2 , it is a flowchart of a program for training a first target detection model according to one embodiment of the present application. As can be seen from the figure, the training process of the first target detection model includes the following steps:

[0074] S201: Obtain a data set.

[0075] In this step, as many data sets as possible can be collected for different scenarios. For example, the scenario may specifically include different parking configurations of vehicles, vehicles of different sizes and shapes, etc. The more types of scenarios are set, the more different application scenarios can be covered.

[0076] S202: Preprocess the data set.

[0077] Dataset preprocessing specifically includes data labeling and cleaning. Data labeling involves manually marking the vehicle's bounding box, followed by cleaning of dirty, invalid, and data that is unfavorable for model training (e.g., imaging issues such as blurred targets and noise contamination). In one embodiment of the present application, data can also be enhanced online, for example, using methods such as mirror inversion, color fine-tuning, exposure fine-tuning, contrast fine-tuning, brightness fine-tuning, random cropping, random rotation, transmission changes, and image shearing to improve the robustness of model training.

[0078] S203: Training a preset first target detection model on the preprocessed data set.

[0079] The entire dataset was split into a test set and a training set at a ratio of 1:9. The model could use any mainstream object detection network (such as YOLO or SSD) as the framework. The training set was fed into the network for 200 training cycles. The center coordinates, width, and height of the box output by the model were compared with the center coordinates, width, and height of the manually annotated boxes, and the L2 loss and IoU loss were calculated. The training optimizer used was SGD (Stochastic Gradient Descent), with a momentum factor of 0.9 and an initial learning rate of 0.0001.

[0080] S204: When it is determined that the number of training times reaches a preset number or the accuracy reaches a preset value, the training may be stopped.

[0081] In one embodiment of the present application, the judgment method uses the PR curve and the area below the line to measure. The horizontal axis of the PR curve is the recall rate, and the vertical axis is the precision rate. The higher the PR curve and the higher the area below the line, the better the model. After adjusting the learning rate, number of training times, model hyperparameters, dataset distribution, etc., training is repeated until the indicators meet the expected requirements.

[0082] 3 , as an embodiment of the present application, the step of “monitoring the door status of the target vehicle” specifically includes:

[0083] S1031: Acquire the current image frame of the target vehicle.

[0084] S1032: Input the current image frame into the second target detection model to determine whether the door state of the target vehicle is open or closed.

[0085] The second object detection model reflects the mapping relationship between a second feature representing the vehicle's exterior and the door's open state. Specifically, the second feature may reflect the overall shape formed between the door and the side surface of the vehicle body. It is understood that the actual value of the second feature varies depending on the door's opening angle. When the current image frame in step S1032 is input to the second object detection model, the second object detection model extracts the second feature in the current image frame and outputs the door's open state as either open or closed. In some embodiments, the model may output a probability of the door being open, and if the probability is greater than a preset value, the door is considered open in that image frame.

[0086] When training the second object detection model, during the data collection phase, as many actual values ​​reflecting the second feature quantity as possible should be collected. For example, the specific values ​​of the door opening angle can be divided as finely as possible to improve the accuracy of model training. The second object detection model can use a lightweight convolutional network model. The training process of the second object detection model also requires data collection, cleaning, and repeated training, which is not detailed in this application.

[0087] In some embodiments of the present application, the specific method of the step of "determining whether the target vehicle is performing a door opening action based on the door status" is: determining whether it is performing a door opening action based on the door status corresponding to multiple image frames detected of the target vehicle.

[0088] 2 , as an embodiment of the present application, after step S1032 , “determining whether a door opening action is being performed based on the door states corresponding to the plurality of detected image frames of the target vehicle” specifically includes:

[0089] S1033: Determine whether the door status detection results in multiple consecutive image frames are all open. When the door status corresponding to the multiple consecutive image frames is all open, it is determined that the target vehicle is performing a door opening action.

[0090] It can be understood that when the vehicle opens the door, the opening angle of the door gradually increases. Therefore, in the case of multiple consecutive image frames detected, the judgment result of each image frame must be that the door is in the open state (if the model outputs the probability of door opening, the probability value gradually increases in the consecutive image frames). Therefore, when it is detected that the door status in multiple consecutive image frames is open, it means that the vehicle is performing the door opening action.

[0091] It should also be noted that in step S1033, the specific number of image frames in the "continuous multiple image frames" can be set according to actual applications and experimental tests. The specific number of image frames determines the accuracy and sensitivity of the judgment result.

[0092] 3 , as one embodiment of the present application, after step S1032 , “determining whether a door opening action is being performed based on the door states corresponding to the plurality of detected image frames of the target vehicle” specifically includes:

[0093] S1034: Determine whether a ratio of the number of frames in which the door is open divided by the total number of frames in the plurality of consecutive image frames is greater than a preset value. If the ratio of the number of frames in which the door is open divided by the total number of frames is greater than the preset value, it is determined that the target vehicle is performing a door opening action.

[0094] It should be noted that when determining the door status based on image frames, factors such as model accuracy and the actual shooting scene may lead to misjudgments, i.e., the door status may actually be open, but the judgment result may be closed. This may result in a situation where the actual door status is open in multiple consecutive image frames, but the judgment result does not confirm that the door is opening, thereby misjudging the door opening action. Therefore, based on this, in step S1044, the ratio of the number of frames in which the door status is open divided by the total number of frames in multiple consecutive image frames is used as the judgment basis. When this ratio is greater than a preset value, it is also determined that the target vehicle is in the process of opening the door.

[0095] For example, if nine or more of the ten consecutive image frames indicate a door opening, the door opening action is determined to be in progress, and the preset value is 0.9. Therefore, the specific value of the preset value also affects the accuracy and sensitivity of the judgment result.

[0096] Furthermore, the present application also discloses an electronic device, which includes a processor 110, a storage device 120, a communication interface 130 and a communication bus 140. The storage device 120 is suitable for storing multiple program codes, and the program codes are suitable for being loaded and run by the processor 110 to execute any of the monitoring methods in the above method embodiments. The above-mentioned communication interface 130 can be one or more, and the communication interface 130 can use any transceiver-like device for communicating with other devices or communication networks. The communication bus 140 may include a path, and the communication bus 140 includes but is not limited to a data bus, a power bus, a control bus and a status signal bus, etc. The processor 110, the storage device 120 and the communication interface 130 are coupled together through the communication bus 140. For ease of explanation, only the parts related to the embodiments of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiments of the present application.

[0097] Furthermore, the present application also discloses a computer-readable storage medium. In a computer-readable storage medium embodiment according to the present application, the computer-readable storage medium can be configured to store a program for executing the monitoring method of the above-mentioned method embodiment, and the program can be loaded and run by the processor to implement the above-mentioned monitoring method. For ease of explanation, only the parts related to the embodiment of the present application are shown. For specific technical details not disclosed, please refer to the method embodiment section of the present application. The computer-readable storage medium can be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiment of the present application is a non-transitory computer-readable storage medium.

[0098] It will be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment of the present application can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer-readable storage medium can include: any entity or device, medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal and software distribution medium that can carry the computer program code. It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electric carrier signals and telecommunication signals.

[0099] Thus far, the technical solutions of the present application have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of the present application is obviously not limited to these specific embodiments. Without departing from the principles of the present application, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present application.

Claims

1. A vehicle monitoring method, characterized in that: include: Acquire images around the vehicle; determining vehicles around the vehicle based on the image; Monitoring the door status of the vehicle, and determining whether the vehicle is in the process of opening the door according to the door status; as well as Determine whether to issue a warning message based on the judgment result.

2. The method according to claim 1, characterized in that The method further comprises: Determine the position of each of the surrounding vehicles relative to the host vehicle; According to the position, the vehicle entering the preset range of the vehicle is screened and marked as a target vehicle; The step of "monitoring the door status of the vehicle and determining whether the vehicle is performing a door opening action according to the door status" includes: The door status of the target vehicle is monitored, and whether the surrounding vehicles are performing door opening actions is determined according to the door status.

3. The method according to claim 2, characterized in that The step of "screening vehicles that enter the preset range of the vehicle according to the position" includes: Inputting the image of each vehicle among the surrounding vehicles into the first target detection model to obtain the area corresponding to each vehicle; Determine the distance of each vehicle relative to the vehicle according to the pixel points at the outer edge of the area, and filter the vehicles that enter the preset range of the vehicle based on the distance; The first target detection model reflects the mapping relationship between the first feature quantity representing the vehicle shape and the region.

4. The method according to claim 2, characterized in that: The step of "monitoring the door status of the target vehicle" includes: Acquire a current image frame of the target vehicle; and Inputting the current image frame into a second target detection model to determine whether the door state of the target vehicle is open or closed; The second target detection model reflects the mapping relationship between the second feature quantity representing the vehicle appearance and the door opening state.

5. The method according to claim 4, characterized in that The step of "determining whether the target vehicle is performing a door opening action according to the door state" includes: It is determined whether the target vehicle is in the process of opening the door according to the door states corresponding to the plurality of image frames of the target vehicle.

6. The method according to claim 5, characterized in that The step of "determining whether a door opening action is being performed according to the door states corresponding to the plurality of image frames of the target vehicle" includes: When the door states corresponding to a plurality of consecutive image frames are all open, it is determined that the target vehicle is performing a door opening action.

7. The method according to claim 6, characterized in that The step of "determining whether a door opening action is being performed according to the door states corresponding to the plurality of image frames of the target vehicle" includes: When, in a plurality of consecutive image frames, a ratio of the number of frames in which the door is in an open state divided by the total number of frames is greater than a preset value, it is determined that the target vehicle is performing a door opening action.

8. The method according to any one of claims 2 to 7, characterized in that The steps of "determining whether to issue a warning message based on the judgment result" include: When it is determined that at least one of the target vehicles is performing a door opening action, a warning message is issued.

9. An electronic device comprising a processor and a storage device, wherein the storage device is suitable for storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and executed by the processor to execute the monitoring method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and run by a processor to execute the monitoring method according to any one of claims 1 to 8.

Images