Automobile control method and device, storage medium and equipment

Abstract

The invention provides an automobile control method and device, a storage medium and equipment, and the method comprises the steps: obtaining an environment image around an automobile, carrying out the face detection of the environment image, carrying out the recognition of a detected face through a face recognition model, determining whether a person is an automobile owner or not, and when the person approaching the automobile is determined to be the automobile owner, carrying out the recognition of the person. And a door opening signal is sent to a central control system of the vehicle to relieve the vehicle door locking state. Therefore, identity verification is carried out by utilizing a visual method, the automobile owner can automatically open the door only by approaching the automobile, the use convenience and the user experience are improved, compared with a mode of using a mobile terminal as an intelligent key of the automobile, the visual method is more difficult to be invaded or simulated by illegal users, the risk of theft or loss does not exist, and the safety of the automobile is improved. Therefore, the safety of the vehicle can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of automobile control technology, and in particular to an automobile control method, device, storage medium and equipment. Background Art

[0002] With the improvement of the level of automobile industry, people have higher and higher requirements for the intelligence of automobiles. The automatic door opening function of a car is an intelligent automobile safety assistance system, which can help the driver to automatically open the door when parking, so as to facilitate passengers to get on and off the car. The existing automatic door opening method of a car generally uses a mobile terminal such as a mobile phone as the smart key of the car. When the owner brings the mobile terminal close to the vehicle, the vehicle transmits data with the mobile terminal based on communication technologies such as Bluetooth and NFC (Near Field Communication), and verifies the user's identity information through the ID recognition module, and when the verification is passed, the door is automatically opened. Although this method improves the intelligence of the car, it is easy to affect the safety of the car. For example, when the mobile terminal acting as a smart key is stolen or lost, the safety of the car may be threatened, and illegal users may also invade the mobile terminal to obtain access rights to the vehicle. Summary of the invention

[0003] The purpose of the present application is to provide a car control method, device, storage medium and equipment, aiming to solve the problem of the automatic door opening method of the car in the related art that it is easy to affect the safety of the car.

[0004] In a first aspect, the present application provides a car control method, comprising: acquiring an environmental image around the vehicle, and performing face detection on the environmental image; recognizing the detected face through a face recognition model, and determining whether the person corresponding to the face is the owner of the vehicle; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on a feature vector extracted from an input face image; if the judgment result is yes, sending a door opening signal to the central control system of the vehicle, so that the central control system releases the door lock state.

[0005] In the above implementation process, the environmental image around the vehicle is obtained, and after the face detection is performed on the environmental image, the detected face is recognized by the face recognition model to confirm whether it is the owner of the vehicle. When it is confirmed that the person approaching the vehicle is the owner, a door opening signal is sent to the central control system of the vehicle to unlock the door. In this way, the visual method is used for identity authentication, and the owner only needs to approach the car to open the door automatically, which improves the convenience of use and user experience. In addition, compared with the method of using a mobile terminal as a smart key for the car, the visual method is more difficult to be invaded or simulated by illegal users, and there is no risk of theft or loss, so it can effectively improve the safety of the vehicle.

[0006] Furthermore, in some examples, the performing face detection on the environmental image includes: performing face detection on the environmental image through a face positioning model; during the training process, the face positioning model learns to detect the position information of the face in the input image based on the features of the input image.

[0007] In the above implementation process, a face positioning model is constructed to accurately locate the position of the face part in the environmental image, quickly separate the face area from the complex background, and lay a good foundation for subsequent face recognition.

[0008] Furthermore, in some examples, the face positioning model includes a YOLOv8n model; the backbone network of the YOLOv8n model includes four Stage Layers, wherein the first, second and third Stage Layers each include a convolution module and a C2f module, and the fourth Stage Layer includes an SPPF module.

[0009] In the above implementation process, the face localization model uses YOLOv8n as the basic model to improve the recognition speed and efficiency. In addition, in view of the situation in which the user is far away from the car and the target is small in the face localization task, the YOLOv8n model is improved by removing the convolution part in the Stage Layer4 of the Backbone part and retaining only the SPPF structure, so that the size of the feature map is enlarged, thereby enhancing the robustness and applicability of the model.

[0010] Furthermore, in some examples, before the detected face is recognized by the face recognition model, it includes: cropping the environment image according to the position information detected by the face positioning model; adjusting the cropped face image to a target size; processing the adjusted face image using a high-pass filter to generate a target face image; the recognized face by the face recognition model includes: outputting the target face image to the face recognition model so that the face recognition model performs face recognition on the target face image.

[0011] In the above implementation process, through cropping and preprocessing, a clearer and more accurate input is provided to the face recognition model, so that the identity of the person approaching the vehicle can be verified more accurately.

[0012] Furthermore, in some examples, the face recognition model uses a pre-trained FaceNet model to extract feature vectors of the target face image, and then calculates the similarity between the extracted feature vectors and each feature vector in a car owner database, and authenticates the person corresponding to the target face image based on the similarity; each feature vector in the car owner database represents the facial features of each car owner.

[0013] In the above implementation process, when confirming whether the person in the image is the car owner, the pre-trained FaceNet model is first used to extract the feature vector of the face image, and then it is compared with the feature vector of the registered car owners stored in the car owner database, thereby effectively realizing the identity verification of the person approaching the vehicle.

[0014] Furthermore, in some examples, it also includes: detecting whether a person is approaching the vehicle and detecting the distance between the person and the vehicle through ultrasonic sensors installed around the vehicle; the step of recognizing the detected face through a face recognition model is performed when the ultrasonic sensor detects that a person is approaching the vehicle and the distance reaches a preset distance threshold.

[0015] In the above implementation process, the ultrasonic sensor is used to monitor the distance between the person and the vehicle. The face recognition model is triggered only when it is detected that the distance between the person and the vehicle reaches a preset distance threshold, avoiding the uninterrupted operation of the face recognition model, thereby effectively saving computing resources.

[0016] Furthermore, in some examples, it also includes: if the similarity between the extracted feature vector and each feature vector in the vehicle owner database is lower than a preset similarity threshold, an abnormal warning mechanism is activated; the abnormal warning mechanism includes at least one of: sending a reminder message to the vehicle owner, recording abnormal events, and sending an alarm notification.

[0017] In the above implementation process, a user anomaly detection mechanism is introduced. If the similarity between the feature vector extracted from the detected face image and the feature vector in the car owner database is low, the abnormal warning mechanism is activated and corresponding measures are taken to deal with it, including reminding the car owner, recording abnormal events, sending alarm notifications, etc. In this way, the safety of the vehicle can be effectively improved by timely identifying and responding to abnormal situations, and the safety of the vehicle and its internal items can be guaranteed.

[0018] In a second aspect, the present application provides an automobile control device, comprising: a detection module, used to obtain an environmental image around the vehicle and perform face detection on the environmental image; a recognition module, used to recognize the detected face through a face recognition model, and determine whether the person corresponding to the face is the vehicle owner; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on the feature vector extracted from the input face image; and a control module, used to send a door opening signal to the central control system of the vehicle if the judgment result is yes, so that the central control system releases the door lock state.

[0019] In a third aspect, the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in any one of the first aspects when executing the computer program.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium, on which instructions are stored. When the instructions are executed on a computer, the computer executes the method as described in any one of the first aspects.

[0021] In a fifth aspect, the present application provides a computer program product, which, when running on a computer, enables the computer to execute the method as described in any one of the first aspects.

[0022] Other features and advantages disclosed in the present application will be described in the following description, or some features and advantages can be inferred or determined without doubt from the description, or can be learned by implementing the above-mentioned technology disclosed in the present application.

[0023] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A flow chart of a vehicle control method provided in an embodiment of the present application;

[0026] Figure 2A schematic diagram of the structure of the Backbone part in the improved YOLOv8n provided in an embodiment of the present application;

[0027] Figure 3 A block diagram of an automobile control device provided in an embodiment of the present application;

[0028] Figure 4 A structural block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0030] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0031] As described in the background art, the automatic door opening method of a car in the related art has the problem of easily affecting the safety of the car. Based on this, the embodiment of the present application provides a car control solution to solve the above problem.

[0032] Next, the embodiments of the present application are introduced:

[0033] like Figure 1 As shown, Figure 1 This is a flow chart of a vehicle control method provided in an embodiment of the present application. The method can be applied to a processor mounted on a vehicle. The method can be implemented as an automatic door opening control system. For the convenience of discussion, the execution subject is referred to as a system below.

[0034] The method comprises:

[0035] Step 101: Acquire an environment image around the vehicle and perform face detection on the environment image;

[0036] The environmental image mentioned in this step can be acquired by an image acquisition device arranged around the vehicle, and the image acquisition device can be a monocular camera, a binocular camera, etc. The system uses the image acquisition device to capture the environmental image in real time, detects human faces from the real-time video stream, and identifies people approaching the vehicle.

[0037] In some embodiments, the face detection on the environment image mentioned in this step may include: performing face detection on the environment image through a face positioning model; the face positioning model learns to detect the position information of the face in the input image according to the characteristics of the input image during the training process. In other words, a face positioning model is constructed, and the face positioning model is used to determine the position information of the face in the input image. In this way, the environment image collected in real time is input into the face positioning model. If there is a face in the environment image, the face positioning model can accurately locate the position of the face part in the environment image and output the corresponding position information. Among them, the position of the face in the image can be marked with a rectangular frame. When implemented, the position information output by the face positioning model can be represented by the positions of the four vertices of the face rectangular frame relative to the image origin. In this way, the face positioning model is used to quickly separate the face area from the complex background, laying a good foundation for subsequent face recognition.

[0038] Optionally, the aforementioned face location model may include a YOLOv8n model. YOLO (You Only Look Once) is a deep learning algorithm for target detection, and YOLOv8 is a version of YOLO. The network structure of YOLOv8 is mainly composed of three parts: a backbone network (Backbone), a neck network (Neck), and a head network (Head). The Backbone part is responsible for feature extraction, the Neck part is responsible for multi-scale feature fusion, and the Head part is responsible for the final target detection and classification tasks. In this embodiment, the n model in the YOLOv8 series, i.e., the YOLOv8n model, is used to construct a face location model, which can provide high detection accuracy while maintaining high speed.

[0039] In the face positioning task, when the user is far away from the car, the target to be detected can be regarded as a small target. At this time, the small-size feature map extracted by the YOLOv8n model from the input image will cause a large loss of feature information, which is easy to affect the detection accuracy. Therefore, in some embodiments, the backbone network of the YOLOv8n model may include four StageLayers, wherein the first, second and third Stage Layers each include a convolution module and a C2f module, and the fourth Stage Layer includes an SPPF module. That is to say, in the face positioning task, the YOLOv8n model can be improved to address the situation where the target is small due to the long distance between the user and the car. The Backbone part of the improved YOLOv8n model includes four Stage Layers, namely Stage Layer1, Stage Layer2, Stage Layer3 and Stage Layer4, where Stage Layer1, Stage Layer2 and Stage Layer3 each include a convolution module and a C2f module. These layers extract information from different feature scales of the input image through the convolution module and the C2f module, and StageLayer4 includes an SPPF module. SPPF (Spatial Pyramid Pooling Fast) is a variant of spatial pyramid pooling (SPP), which generates a fixed-length feature representation by performing pooling operations on feature maps at different scales. In this way, by removing the convolution part in Stage Layer4 of the Backbone part and retaining only the SPPF structure, the size of the feature map is enlarged, thereby enhancing the robustness and applicability of the model.

[0040] Step 102: Recognize the detected face through a face recognition model to determine whether the person corresponding to the face is the car owner; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on the feature vector extracted from the input face image;

[0041] In this embodiment, a face recognition model is constructed, which is used to extract feature vectors from an input face image, and to authenticate the person corresponding to the face in the face image based on the extracted feature vectors. In this way, after face detection is performed on the environmental image, the detected face is recognized by the face recognition model to confirm whether the person corresponding to the face is the car owner, so as to make accurate door control decisions.

[0042] In some embodiments, before the detected face is identified by the face recognition model mentioned in this step, it may include: cropping the environment image according to the position information detected by the face positioning model; adjusting the cropped face image to the target size; using a high-pass filter to process the adjusted face image to generate a target face image; accordingly, the detected face is identified by the face recognition model mentioned in this step may include: outputting the target face image to the face recognition model so that the face recognition model performs face recognition on the target face image.

[0043] That is to say, after using the face positioning model to determine the position information of the face in the original image, the system can calculate the position and size of the face part in the original image based on the position information, and perform corresponding cropping, so as to accurately extract the face part from the original image. After that, the cropped face image is preprocessed. Specifically, in order to maintain consistency, the system adjusts the cropped face image to a fixed size. At the same time, taking into account that the details of local areas of the face image may be unclear or blurred, the system uses a high-pass filter to process the adjusted face image to improve the image clarity and feature recognizability. In this way, through cropping and preprocessing, a clearer and more accurate input is provided to the face recognition model, so that the identity of people approaching the vehicle can be more accurately verified.

[0044] In some embodiments, the face recognition model mentioned in this step uses a pre-trained FaceNet model to extract the feature vector of the target face image, and then calculates the similarity between the extracted feature vector and each feature vector in the car owner database, and authenticates the person corresponding to the target face image based on the similarity; each feature vector in the car owner database represents the facial features of each car owner. That is to say, when confirming whether the person in the image is the owner of the car, the pre-trained FaceNet model is first used to extract the feature vector of the face image. The FaceNet model is a deep learning model for face recognition and clustering. Its main idea is to map the face image to a multidimensional space and represent the similarity of the face through spatial distance. After extracting the feature vector of the face image, it is compared with the feature vector in the car owner database. The feature vector in the car owner database can be collected and stored during the car owner registration process, representing the unique facial features of each car owner. In implementation, the system can use cosine similarity to compare the extracted feature vector with the feature vector in the car owner database to evaluate the similarity between them. For example, when the similarity between the extracted feature vector and the feature vector of a registered car owner is greater than or equal to the target threshold, it can be determined that the person currently approaching the vehicle is the registered car owner. In this way, the identity verification of the person approaching the vehicle is effectively achieved.

[0045] Step 103: If the judgment result is yes, a door opening signal is sent to the central control system of the vehicle so that the central control system releases the door locking state.

[0046] In this embodiment, a visual method is used for identity authentication. When the person approaching the vehicle is confirmed to be the owner, the system generates a door opening signal and sends it to the central control system of the vehicle. After receiving the door opening signal, the central control system controls the door lock and unlock actuator to release the door lock state, so that the door opens automatically. In this way, the owner does not need to carry additional items, and can open the door automatically just by approaching the car, which improves the convenience of use. Moreover, compared with the automatic door opening method relying on mobile terminals, the visual method is more difficult to be hacked or simulated by illegal users, thereby effectively improving the safety of the vehicle.

[0047] In some embodiments, it also includes: detecting whether there is a person approaching the vehicle and detecting the distance between the person and the vehicle through ultrasonic sensors installed around the vehicle; the step of identifying the detected face through the face recognition model is performed when the ultrasonic sensor detects that a person is approaching the vehicle and the distance reaches a preset distance threshold. In other words, the system can monitor the distance of a person approaching the vehicle through ultrasonic sensors installed around the vehicle. When the ultrasonic sensor detects that a person is approaching the vehicle and the distance reaches a preset distance threshold, the system triggers the face recognition model to perform face recognition on the person approaching the vehicle through the face recognition model. In this way, the uninterrupted operation of the face recognition model can be avoided, thereby effectively saving computing resources.

[0048] In addition, in some embodiments, it also includes: if the similarity between the extracted feature vector and each feature vector in the car owner database is lower than the preset similarity threshold, the abnormal warning mechanism is activated; the abnormal warning mechanism includes at least one of: sending a reminder message to the car owner, recording abnormal events, and sending an alarm notification. In other words, a user abnormality detection mechanism is set, that is, during the feature comparison process, if the similarity between the extracted feature vector and the feature vector in the car owner database is low, it indicates that the detected face image is abnormal. In this case, the system activates the abnormal warning mechanism and takes corresponding measures to deal with it, including reminding the car owner, recording abnormal events, sending alarm notifications, etc. In this way, by timely identifying and responding to abnormal situations, the safety of the vehicle is effectively improved, and the safety of the vehicle and its internal items is guaranteed.

[0049] In the embodiment of the present application, an environmental image around the vehicle is acquired, and after face detection is performed on the environmental image, the detected face is recognized through a face recognition model to confirm whether it is the owner of the vehicle. When it is confirmed that the person approaching the vehicle is the owner, a door opening signal is sent to the central control system of the vehicle to unlock the door. In this way, by using visual methods for identity authentication, the owner only needs to approach the car to open the door automatically, which improves the convenience of use and user experience. Moreover, compared with the method of using a mobile terminal as a smart key for the car, the visual method is more difficult to be invaded or simulated by illegal users, and there is no risk of theft or loss, so it can effectively improve the safety of the vehicle.

[0050] In order to explain the solution of this application in more detail, a specific embodiment is introduced below:

[0051] This embodiment provides a new energy vehicle automatic door opening control solution based on a visual method. This embodiment is implemented as a vehicle automatic door opening control system, hereinafter referred to as the system. The technical process of this embodiment includes:

[0052] S21, image data collection;

[0053] Specifically, this step may include the following:

[0054] S211. Install high-definition cameras around the vehicle to ensure that images of people approaching the vehicle can be captured in all directions, wherein the front camera is installed on the top of the front windshield of the vehicle to capture images of people in front of the vehicle, especially when the owner approaches the vehicle from the front; the rear camera is installed on the trunk lid of the vehicle to capture images of people behind the vehicle, which is suitable for the case where the owner approaches the vehicle from the rear; the left camera and the right camera are installed on the left and right rearview mirrors of the vehicle respectively to capture images of people approaching the vehicle from both sides;

[0055] S212, setting the frame rate and resolution of the camera to ensure a balance between image quality and processing speed. Optionally, the frame rate is set to 30 frames per second and the resolution is set to 1080p to ensure that the captured image is clear and has sufficient real-time performance;

[0056] S213, using the OpenCV library to obtain a video stream from the camera in real time. When implementing, the camera can be initialized and video stream capture can be started through the VideoCapture class of OpenCV to ensure that the camera works normally and can continuously provide stable video data. After that, the images are sorted and cleaned by frame-by-frame analysis, including deleting low-quality or duplicate images;

[0057] S214, using the image labeling tool LabelImg to label the position of the face in the image;

[0058] S22. Design a face localization model based on deep learning;

[0059] Specifically, in this embodiment, YOLOv8n is selected as the basic model of the face positioning model, which consists of the Backbone part, the Neck part and the Head part. Generally, the size of the basic model input image is 640×640×3 pixels, and the sizes of the three feature maps output by the Head part are 80×80×128 pixels, 40×40×256 pixels and 20×20×512 pixels respectively. In the face positioning task, when the user is far away from the car, the target to be detected can be regarded as a small target, and 20×20×512 pixels will cause a large loss of feature information, thereby affecting the detection accuracy;

[0060] Based on this, the present embodiment removes the convolution part in Stage Layer 4 of the Backbone part and only retains the SPPF structure. The structure of the Backbone part in the improved YOLOv8n is as follows: Figure 2 As shown in the figure. Through this improvement, the size of the feature map can be enlarged to 160×160×64 pixels, 80×80×128 pixels, and 40×40×256 pixels;

[0061] When building the face location model, the data set obtained by S214 is divided into training set, validation set and test set in a ratio of 7:2:1, and then the samples in the data set are randomized to ensure the effectiveness of the training and the generalization ability of the model. During training, the input size is 640×640×3 pixels, the batch size can be set to 32, the training lasts for 200 epochs, the Adam optimizer is used, the initial learning rate is 0.001, and the cosine annealing scheduler is used to update the learning rate, annealing every 10 epochs;

[0062] S23. Design a face recognition model based on deep learning;

[0063] Specifically, the system uses the face positioning model constructed by S22 to determine the position of the face annotation box in the original image. The annotation box is a rectangular box, and its position information can be represented by the positions of four vertices relative to the origin of the original image, namely (X1, Y1), (X2, Y2), (X3, Y3) and (X4, Y4). Based on this information, the system can accurately calculate the position and size of the face part in the original image and perform corresponding cropping operations;

[0064] After that, the system preprocesses the cropped face image. In order to maintain consistency, the system adjusts the cropped face image to a fixed size, such as 160×160 pixels. At the same time, a high-pass filter is used to improve the image clarity and feature recognizability.

[0065] The face recognition model constructed in the embodiment scheme uses the pre-trained FaceNet model to extract the feature vector of the face image. The FaceNet model includes a batch input layer and a deep CNN (Convolutional Neural Networks, convolutional neural network), wherein the batch input layer is used to receive the pre-processed face image data and pass it to the deep CNN for feature extraction; the deep CNN is composed of multiple convolutional layers and pooling layers, which are used to gradually extract the features of the face image; in the final stage of feature extraction, the model performs L2 norm normalization to normalize the length of the feature vector to make it have unit length; the feature vector processed by L2 norm normalization is regarded as the embedding of the face, indicating the position of the face in the feature space; the FaceNet model uses a triple loss function to train the model;

[0066] After extracting the feature vector from the face image, the system accesses the car owner database to extract the face feature vector of the registered car owner, and then uses cosine similarity to compare the feature vector extracted from the face image with the face feature vector in the car owner database to evaluate the similarity between them;

[0067] S24. Design a user anomaly detection mechanism;

[0068] Specifically, the system monitors the distance of a person approaching the vehicle through ultrasonic sensors installed around the vehicle. When the ultrasonic sensor detects that a person is approaching the vehicle and the distance reaches a preset distance threshold of 0.5m, the system triggers the face recognition model constructed in S23. The face recognition model will perform face recognition on the person approaching the vehicle and compare it with the face feature vector pre-stored in the vehicle owner database.

[0069] During the comparison process, if the similarity between the feature vector extracted from the face image and the majority of face feature vectors in the car owner database is low, it means that the detected face image is abnormal. At this time, the system will activate the abnormality warning mechanism and take corresponding measures to deal with it, including reminding the car owner, recording abnormal events, sending alarm notifications, etc.

[0070] S25, identity authentication and door opening control;

[0071] Specifically, when identity authentication and information verification are passed, the system generates a door opening signal and sends the door opening signal to the central control system of the door through the CAN bus. After receiving the door opening signal, the central control system will issue corresponding instructions to the door control unit through the vehicle network, so that the door control unit controls the door lock and unlock actuator to release the door lock state; in this process, if the new energy vehicle is equipped with a vehicle anti-theft system, the central control system will first check the anti-theft status of the door before releasing the door lock state to ensure that the unlocking operation will not trigger the alarm or locking function of the anti-theft system;

[0072] In addition, the system uses an efficient database management system, such as MySQL, to store recognition records, including video data, recognition time, results and vehicle owner information. The stored recognition records are desensitized to protect user privacy. The video data records the details of the recognition process, which are compressed and encoded with H.264 and indexed according to timestamps. At the same time, strict access rights and audit mechanisms are set to ensure that only authorized personnel can access and modify data, and all access and operations are recorded to maintain data security and compliance.

[0073] This embodiment has at least the following advantages: First, visual methods are used for identity authentication. Compared with the method of using mobile terminals such as mobile phones as smart keys, visual methods are more difficult to be invaded or simulated by illegal users, thereby improving the security of the system; second, car owners do not need to carry additional items, and can automatically open and lock doors by just approaching the car, which effectively improves the convenience of use and user experience; third, the face positioning model uses YOLOv8n as the basic model, which improves the recognition speed and efficiency; a user anomaly detection mechanism is introduced to improve the safety of the vehicle and ensure the safety of items inside the vehicle; fourth, the system has a staged trigger mechanism, that is, the face recognition model is triggered only when the distance between the person and the vehicle reaches a preset threshold, avoiding unnecessary continuous operation, improving the intelligence of the system and resource utilization efficiency, thereby improving the performance and reliability of the system.

[0074] Corresponding to the embodiments of the aforementioned method, the present application also provides embodiments of an automobile control device and a terminal using the same:

[0075] like Figure 3 As shown, Figure 3 : is a block diagram of an automobile control device provided in an embodiment of the present application, the device comprising:

[0076] The detection module 31 is used to obtain an environmental image around the vehicle and perform face detection on the environmental image;

[0077] The recognition module 32 is used to recognize the detected face through a face recognition model to determine whether the person corresponding to the face is the vehicle owner; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on the feature vector extracted from the input face image;

[0078] The control module 33 is used to send a door opening signal to the central control system of the vehicle if the judgment result is yes, so that the central control system releases the door locking state.

[0079] The implementation process of the functions and effects of each module in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, which will not be repeated here.

[0080] This application also provides an electronic device, see Figure 4 , Figure 4 A block diagram of an electronic device provided in an embodiment of the present application. The electronic device may include a processor 410, a communication interface 420, a memory 430, and at least one communication bus 440. The communication bus 440 is used to realize direct connection and communication between these components. The communication interface 420 of the electronic device in the embodiment of the present application is used to communicate signaling or data with other node devices. The processor 410 may be an integrated circuit chip with signal processing capabilities.

[0081] The processor 410 mentioned above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor 410 can also be any conventional processor, etc.

[0082] The memory 430 may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electric erasable programmable read-only memory (EEPROM), etc. The memory 430 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 410, the electronic device can execute the above-mentioned Figure 1 The method embodiment involves various steps.

[0083] Optionally, the electronic device may further include a storage controller and an input / output unit.

[0084] The memory 430, storage controller, processor 410, peripheral interface, input and output unit components are directly or indirectly electrically connected to each other to achieve data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses 440. The processor 410 is used to execute executable modules stored in the memory 430, such as software function modules or computer programs included in the electronic device.

[0085] The input and output unit is used to provide users with the task creation and to create a start optional time period or preset execution time for the task to realize the interaction between the user and the server. The input and output unit can be, but is not limited to, a mouse and a keyboard.

[0086] Understandably, Figure 4 The structure shown is for illustration only, and the electronic device may also include Figure 4 More or fewer components as shown, or with Figure 4 Different configurations are shown. Figure 4 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0087] An embodiment of the present application further provides a storage medium having instructions stored thereon. When the instructions are run on a computer, the computer program is executed by a processor to implement the method described in the method embodiment. To avoid repetition, the method will not be described here.

[0088] The present application also provides a computer program product, which, when executed on a computer, enables the computer to execute the method described in the method embodiment.

[0089] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0090] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0091] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0092] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0094] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

Claims

1. A vehicle control method, characterized in that: include: Acquire an environmental image around the vehicle, and perform face detection on the environmental image; The detected face is identified by a face recognition model to determine whether the person corresponding to the face is the vehicle owner; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on the feature vector extracted from the input face image; If the judgment result is yes, a door opening signal is sent to the central control system of the vehicle so that the central control system releases the door locking state.

2. The method according to claim 1, characterized in that The performing face detection on the environment image includes: The face detection is performed on the environment image by using a face positioning model; during the training process, the face positioning model learns to detect the position information of the face in the input image according to the features of the input image.

3. The method according to claim 2, characterized in that The face positioning model includes a YOLOv8n model; the backbone network of the YOLOv8n model includes four Stage Layers, wherein the first, second and third Stage Layers each include a convolution module and a C2f module, and the fourth Stage Layer includes an SPPF module.

4. The method according to claim 2, characterized in that: Before the detected face is recognized by the face recognition model, the method includes: Cropping the environment image according to the position information detected by the face positioning model; Resize the cropped face image to the target size; Processing the adjusted face image using a high-pass filter to generate a target face image; The step of identifying the detected face by using a face recognition model includes: The target facial image is output to a facial recognition model so that the facial recognition model performs facial recognition on the target facial image.

5. The method according to claim 4, characterized in that The face recognition model uses a pre-trained FaceNet model to extract the feature vector of the target face image, then calculates the similarity between the extracted feature vector and each feature vector in the car owner database, and authenticates the person corresponding to the target face image based on the similarity; each feature vector in the car owner database represents the facial features of each car owner.

6. The method according to claim 5, characterized in that Also includes: Detecting whether a person is close to the vehicle and detecting the distance between the person and the vehicle by means of ultrasonic sensors installed around the vehicle; The step of identifying the detected face by using the face recognition model is performed when the ultrasonic sensor detects that a person is approaching the vehicle and the distance reaches a preset distance threshold.

7. The method according to claim 6, characterized in that Also includes: If the similarity between the extracted feature vector and each feature vector in the vehicle owner database is lower than a preset similarity threshold, an abnormal warning mechanism is activated; the abnormal warning mechanism includes at least one of: sending a reminder message to the vehicle owner, recording abnormal events, and sending an alarm notification.

8. An automobile control device, characterized in that: include: A detection module, used to acquire an environmental image around the vehicle and perform face detection on the environmental image; The recognition module is used to recognize the detected face through the face recognition model to determine whether the person corresponding to the face is the vehicle owner; during the training process, the face recognition model learns to authenticate the person corresponding to the face image based on the feature vector extracted from the input face image; The control module is used to send a door opening signal to the central control system of the vehicle if the judgment result is yes, so that the central control system releases the door locking state.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images