Vehicle off-line detection method, detection device, detection equipment and storage medium

Abstract

The invention discloses a vehicle offline detection method, a detection device, detection equipment and a storage medium. The method comprises the following steps: acquiring a current vehicle image of a current vehicle; the current vehicle image is input to a target vehicle part installation position detection model, so that the target vehicle part installation position detection model outputs a corresponding detection result, and the target vehicle part installation position detection model is obtained based on a single-step multi-frame target detection model and a multi-head attention model; the detection result comprises a first detection result used for representing that the installation position of the current vehicle part of the current vehicle is not deviated and a second detection result used for representing that the installation position of the current vehicle part is deviated; and when a first detection result output by the target vehicle part installation position detection model is received, generating a corresponding offline detection qualified result. According to the method, vehicle off-line detection is carried out on the current vehicle based on the target vehicle component installation position detection model, and the detection efficiency and the detection accuracy of vehicle off-line detection are improved.

Description

Technical Field

[0001] The present application belongs to the field of vehicle technology, and in particular, relates to a vehicle offline detection method, detection device, detection equipment and storage medium. Background Art

[0002] Vehicle off-line inspection refers to the process in which the vehicle needs to go through the corresponding inspection processes in sequence during vehicle production to ensure the vehicle production quality.

[0003] At present, vehicle off-line inspection mainly relies on inspection personnel to manually measure the installation positions of vehicle components, and determine whether the vehicle has passed the off-line inspection based on the measurement results.

[0004] However, when inspectors are in a fatigued state and manually measure the installation positions of vehicle parts, the measurement efficiency is low and the measurement results are prone to errors, which reduces the efficiency and accuracy of vehicle off-line inspection. Summary of the invention

[0005] In view of this, the embodiments of the present application provide a vehicle offline detection method, a detection device, a detection equipment and a storage medium to overcome or at least partially solve the above problems of the prior art.

[0006] In a first aspect, an embodiment of the present application provides a vehicle offline detection method, comprising: obtaining a current vehicle image of a current vehicle; inputting the current vehicle image into a target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result based on the current vehicle image, and the target vehicle component installation position detection model is obtained based on a single-step multi-frame target detection model and a multi-head attention model, and the detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased; when the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline detection qualified result is generated.

[0007] Among them, in some optional embodiments, before inputting the current vehicle image into the target vehicle component installation position detection model, the vehicle offline detection method also includes: constructing a target vehicle component installation position detection model based on a single-step multi-frame target detection model and a multi-head attention model; determining whether the target vehicle component installation position detection model converges; inputting the current vehicle image into the target vehicle component installation position detection model, including: when it is determined that the target vehicle component installation position detection model converges, inputting the current vehicle image into the target vehicle component installation position detection model.

[0008] Among them, in some optional embodiments, a target vehicle part installation position detection model is constructed according to a single-step multi-box target detection model and a multi-head attention model, including: fusing the single-step multi-box target detection model and the multi-head attention model to obtain an initial vehicle part installation position detection model; obtaining a corresponding sample set according to historical vehicle images, the sample set at least including a training set; inputting the training set to the initial vehicle part installation position detection model for training to obtain a corresponding target vehicle part installation position detection model.

[0009] Among them, in some optional embodiments, before inputting the training set to the initial vehicle component installation position detection model for training and obtaining the corresponding target vehicle component installation position detection model, the vehicle offline detection method also includes: data enhancement processing of the training set to obtain the corresponding enhanced training set; inputting the training set to the initial vehicle component installation position detection model for training and obtaining the corresponding target vehicle component installation position detection model, including: inputting the enhanced training set to the initial vehicle component installation position detection model for training and obtaining the corresponding target vehicle component installation position detection model.

[0010] Among them, in some optional embodiments, the sample set also includes a test set, and determining whether the target vehicle component installation position detection model converges, including: inputting the test set to the target vehicle component installation position detection model for testing to obtain the corresponding test result; determining the loss value between the target result corresponding to the test set and the test result; when the loss value is less than or equal to a preset threshold, determining that the target vehicle component installation position detection model converges; when the loss value is greater than the preset threshold, determining that the target vehicle component installation position detection model has not converged.

[0011] Among them, in some optional embodiments, before generating the corresponding off-line inspection qualified result, the vehicle off-line inspection method also includes: determining whether the current vehicle sensor of the current vehicle has passed the performance test; generating the corresponding off-line inspection qualified result, including: when it is determined that the current vehicle sensor of the current vehicle has passed the performance test, generating the corresponding off-line inspection qualified result.

[0012] Among them, in some optional embodiments, before determining whether the current vehicle sensor of the current vehicle has passed the performance test, the vehicle offline detection method also includes: obtaining the operating frequency of the current vehicle sensor of the current vehicle; determining whether the current vehicle sensor is in a normal online state according to the operating frequency; determining whether the current vehicle sensor of the current vehicle has passed the performance test, including: when it is determined that the current vehicle sensor is in a normal online state according to the operating frequency, determining whether the current vehicle sensor has passed the performance test.

[0013] In the second aspect, the embodiment of the present application provides a vehicle offline detection device, including an image acquisition module, an image input module and a generation module. The image acquisition module is used to acquire the current vehicle image of the current vehicle; the image input module is used to input the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image, and the target vehicle component installation position detection model is obtained based on the single-step multi-frame target detection model and the multi-head attention model, and the detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased; the generation module is used to generate a corresponding offline detection qualified result when receiving the first detection result output by the target vehicle component installation position detection model.

[0014] In a third aspect, an embodiment of the present application provides a detection device, comprising a memory; one or more processors coupled to the memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the vehicle offline detection method provided in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides a vehicle offline detection system, including a vehicle and a detection device as provided in the third aspect above, the vehicle is connected to the detection device, and the detection device is used to perform vehicle offline detection on the vehicle.

[0016] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a program code is stored. The program code can be called by a processor to execute the vehicle offline detection method provided in the first aspect above.

[0017] In a sixth aspect, an embodiment of the present application provides a computer program product. When the computer program product is run on a computer device, the computer device executes the vehicle offline detection method provided in the first aspect above.

[0018] The solution provided by the present application obtains a current vehicle image of the current vehicle and inputs the current vehicle image into a target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model is obtained based on a single-step multi-frame target detection model and a multi-head attention model. The detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased. When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline inspection qualified result is generated, thereby realizing a target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model, and using the target vehicle component installation position detection model to perform vehicle offline inspection on the current vehicle, thereby improving the detection efficiency and detection accuracy of vehicle offline inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A schematic diagram of a scenario of a vehicle offline detection system provided in an embodiment of the present application is shown.

[0021] Figure 2 A schematic flow chart of a vehicle offline detection method provided in an embodiment of the present application is shown.

[0022] Figure 3 Another flow chart of the vehicle offline detection method provided in an embodiment of the present application is shown.

[0023] Figure 4 A structural block diagram of a vehicle offline detection device provided in an embodiment of the present application is shown.

[0024] Figure 5 A functional block diagram of a detection device provided in an embodiment of the present application is shown.

[0025] Figure 6 A computer-readable storage medium provided in an embodiment of the present application for storing or carrying a program code for implementing a vehicle offline detection method provided in an embodiment of the present application is shown.

[0026] Figure 7A computer program product provided in an embodiment of the present application and used for storing or carrying program codes for implementing the vehicle offline detection method provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0027] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] It should be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0029] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms unless the context clearly indicates other working conditions.

[0030] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0031] In addition, in the description of the present application, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0032] Vehicle off-line inspection refers to the process in which the vehicle needs to go through the corresponding inspection processes in sequence during vehicle production to ensure the vehicle production quality.

[0033] At present, vehicle off-line inspection mainly relies on inspection personnel to manually measure the installation positions of vehicle components, and determine whether the vehicle has passed the off-line inspection based on the measurement results.

[0034] However, when inspectors are in a fatigued state and manually measure the installation positions of vehicle parts, the measurement efficiency is low and the measurement results are prone to errors, which reduces the efficiency and accuracy of vehicle off-line inspection.

[0035] In response to the above problems, the vehicle offline detection method, detection device, detection equipment and storage medium provided in the embodiments of the present application obtain the current vehicle image of the current vehicle and input the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model is obtained based on the single-step multi-frame target detection model and the multi-head attention model. The detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased. When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline detection qualified result is generated, and the target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model is implemented, and the target vehicle component installation position detection model is used to perform vehicle offline detection on the current vehicle, thereby improving the detection efficiency and detection accuracy of the vehicle offline detection.

[0036] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0037] See also Figure 1 , which shows a schematic diagram of an application scenario of the vehicle offline detection system provided in an embodiment of the present application, which may include a vehicle 100 and a detection device 200. The vehicle 100 is communicatively connected to the detection device 200 and exchanges data with the detection device 200. The detection device 200 can be used to perform vehicle offline detection on the vehicle 100.

[0038] Among them, the vehicle 100 can be any one of an electric vehicle (for example, an electric car, a battery car, etc.), a hybrid vehicle (for example, a hybrid electric vehicle (HEV)), a fuel vehicle or a gas vehicle, etc. The type of vehicle 100 is not limited here and can be specifically set according to actual needs.

[0039] The detection device 200 may be a server or a terminal device, etc., which is not limited here and may be specifically configured according to actual needs.

[0040] The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), big data, and artificial intelligence platforms.

[0041] The terminal device may be a mobile terminal device (for example, a vehicle-mounted terminal, a PDA (Personal Digital Assistant), a tablet computer (Tablet Personal Computer, Tablet PC), a laptop computer, etc.), or a fixed terminal device (desktop computer, smart panel, etc.), etc.

[0042] In some embodiments, the vehicle 100 may include a vehicle body, a vehicle sensor, and an on-board display screen, wherein the vehicle sensor and the on-board display screen are mounted on the vehicle body, and the vehicle body provides mounting support for the vehicle sensor and the on-board display screen.

[0043] The vehicle sensors and the vehicle-mounted display screen are communicatively connected to the detection device 200 and perform data exchange with the detection device 200 .

[0044] For example, the vehicle sensor and the vehicle display screen can be connected to the detection device 200 through the vehicle communication protocol, and exchange data with the detection device 200 through the vehicle communication protocol. The vehicle communication protocol can be any one of the controller area network (CAN) protocol, the controller area network with flexible data rate (CAN FD) protocol, the Ethernet protocol, the vehicle area network (VAN) protocol or the multimedia transmission system (MOST protocol).

[0045] Vehicle sensors may include at least any one of a vehicle speed sensor, a temperature sensor, a rotation speed sensor, a pressure sensor, a rotation angle sensor, a torque sensor, a hydraulic sensor, a vehicle camera sensor, a radar sensor, and a combined inertial navigation sensor. The type of vehicle sensor is not limited here and can be set according to actual needs.

[0046] The vehicle display screen can be any one of LED display screen, liquid crystal display screen, plasma display screen or laser display screen, etc. The type of vehicle display screen is not limited here and can be set according to actual needs.

[0047] In some embodiments, the detection device 200 may include a data transmission module (Data module), a human-machine communication module (Hmi module) and a monitoring module (Guard module). The Data module and the Hmi module are communicatively connected to the Guard module and perform data exchange with the Guard module.

[0048] For example, the Data module and the Hmi module are connected to the Guard module through the Protocol Buffer (PB) protocol, and perform data exchange with the Guard module through the PB protocol.

[0049] The Data module and the Hmi module are also communicatively connected to the vehicle 100, and the Guard module is used to control the Data module and the Hmi module to perform data interaction with the vehicle 100. For example, the Data module is connected to the vehicle sensor via the vehicle communication protocol, and the Guard module is used to control the Data module to perform data interaction with the vehicle sensor via the vehicle communication protocol. The Hmi module is connected to the vehicle display screen via the Ethernet protocol, and the Guard module is used to control the Hmi module to perform data interaction with the vehicle display screen via the Ethernet protocol.

[0050] See also Figure 2 , which shows a flow chart of a vehicle offline detection method provided by an embodiment of the present application. In a specific embodiment, the vehicle offline detection method can be applied to Figure 1 The detection device 200 in the vehicle offline detection system shown in FIG. 2 is taken as an example to describe the detection device 200 in the following. Figure 2 The process shown is described in detail, and the vehicle offline detection method may include the following steps S110 to S130.

[0051] Step S110: Acquire a current vehicle image of the current vehicle.

[0052] In the embodiment of the present application, when the user needs to perform an offline vehicle inspection on the current vehicle, a detection instruction can be sent to the detection device, and the detection device receives and responds to the detection instruction to obtain the current vehicle image of the current vehicle. The current vehicle image may include an internal structure image and an external structure image of the current vehicle.

[0053] In some embodiments, the vehicle offline detection system may further include a camera, which is mounted on the vehicle body, and the vehicle body may provide mounting support for the camera, and the camera may be used to collect images of the current vehicle. The camera is connected to the detection device via a network, and exchanges data with the detection device via the network.

[0054] When the user needs to perform an offline inspection on the current vehicle, a detection instruction can be sent to the detection device. The detection device receives and responds to the detection instruction, sends an image acquisition instruction to the camera via the network, and the camera receives and responds to the image acquisition instruction, acquires the current vehicle image of the current vehicle, and sends the current vehicle image to the detection device via the network. The detection device receives the current vehicle image returned by the camera.

[0055] Among them, the camera can be any one of a white light camera, an infrared camera, a laser camera, etc. The type of camera is not limited here and can be set according to actual needs.

[0056] The network can be any one of a ZigBee network, a Bluetooth (BT) network, a Wireless Fidelity (Wi-Fi) network, a Thread network, a Long Range Radio (LoRa) network, a Low-Power Wide-Area Network (LPWAN), an infrared network, a Narrow Band Internet of Things (NB-IoT), a Controller Area Network (CAN), a Digital Living Network Alliance (DLNA) network, a Wide Area Network (WAN), a Local Area Network (LAN), a Metropolitan Area Network (MAN) or a Wireless Personal Area Network (WPAN), etc. The type of network is not limited here and can be set according to actual needs.

[0057] In some embodiments, the detection device may include an input panel. When a user needs to perform an off-line vehicle detection on the current vehicle, the user may input a detection instruction on the input panel of the detection device, and the detection device receives the detection instruction through the input panel.

[0058] In some embodiments, the detection device may be provided with a voice recognition module. When a user needs to perform a vehicle offline detection on the current vehicle, the user can send a voice message within the voice collection range of the voice recognition module. The voice recognition module collects the voice information sent later, and performs voice recognition on the collected voice information. Based on the recognition result of the voice recognition, it is determined that the recognition result contains keywords for indicating that a vehicle offline detection should be performed on the current vehicle, for example, "vehicle offline detection", and for example, "vehicle" and "offline detection", etc., then it is determined that a detection instruction for performing an offline detection on the current vehicle has been received.

[0059] As an example, the voice message sent by the user is: perform an offline test on the current vehicle, and the recognition result of the voice recognition includes the keyword "vehicle offline test", then it is determined that a test instruction for performing an offline test on the current vehicle is received.

[0060] In some embodiments, the vehicle offline detection system may further include a client, which is connected to the detection device via a network and exchanges data with the detection device via the network.

[0061] When the user needs to perform an offline inspection on the current vehicle, a detection instruction can be sent to the client. The client receives and responds to the detection instruction, and forwards the detection instruction to the detection device through the network. The detection device receives the detection instruction forwarded by the client.

[0062] Among them, the client can be any one of a mobile client (for example, a mobile phone client, a personal digital assistant (PDA) client, a tablet computer (Tablet Personal Computer, Tablet PC) client, a laptop computer client, a smart watch client, a smart bracelet client or a wearable client, etc.) or a fixed client (for example, a desktop computer client, a smart panel client, etc.). The type of client is not limited here and can be set according to actual needs.

[0063] In some embodiments, after acquiring the current vehicle image of the current vehicle, the detection device can construct a target vehicle component installation position detection model based on a single shot multi-box target detection (Single Shot MultiBox Detector, SSD) model and a multi-head attention model, and determine whether the target vehicle component installation position detection model converges.

[0064] Among them, the SSD model uses different feature layers to detect different scales, so the SSD model can accommodate many object contours of different shapes and sizes, thereby realizing the detection of multiple targets. The SSD model can include a backbone network module, a top-down module, a bottom-up module, and a network output layer module connected in sequence.

[0065] The multi-head attention module can be connected between the backbone module and the top-down module of the SSD model. The multi-head attention module can be used to mine the feature relationship of the backbone module and output the feature relationship to the top-down module. By adding the multi-head attention module, the expression ability of the target vehicle component installation position detection model can be enhanced, and the detection accuracy of vehicle off-line detection can be improved.

[0066] As an implementation method, the detection device can integrate the SSD model and the multi-head attention model to obtain an initial vehicle part installation position detection model, and obtain a corresponding sample set based on historical vehicle images. The sample set at least includes a training set, and the training set is input into the initial vehicle part installation position detection model. The initial vehicle part installation position detection model receives and responds to the training set, is trained according to the training set, and obtains the corresponding target vehicle part installation position detection model.

[0067] As an implementation method, the detection device can integrate the SSD model and the multi-head attention model to obtain an initial vehicle part installation position detection model, and obtain a corresponding sample set based on historical vehicle images, and perform data enhancement on the training set to obtain a corresponding enhanced training set, and input the enhanced training set into the initial vehicle part installation position detection model. The initial vehicle part installation position detection model receives and responds to the enhanced training set, and is trained based on the enhanced training set to obtain a corresponding target vehicle part installation position detection model, thereby realizing the training of the initial vehicle part installation position detection model based on the enhanced training set, avoiding the low robustness of the target vehicle part installation position detection model obtained by training the initial vehicle part installation position detection model with a smaller training set due to fewer historical vehicle images, and increasing the robustness of the target vehicle part installation position detection model.

[0068] Among them, data enhancement processing may include at least any one of brightness enhancement processing, grayscale enhancement processing, contrast enhancement processing, and transparency enhancement processing. The type of data enhancement processing is not limited here and can be set according to actual needs.

[0069] As an implementation method, the sample set may also include a test set. The detection device may input the test set into the target vehicle component installation position detection model. The target vehicle component installation position detection model receives and responds to the test set, performs testing according to the test set, obtains corresponding test results, and determines the loss value between the target result corresponding to the test set and the test result, and determines whether the target vehicle component installation position detection model converges based on the loss value.

[0070] When the loss value is less than or equal to a preset threshold, it is determined that the target vehicle component installation position detection model has converged; when the loss value is greater than the preset threshold, it is determined that the target vehicle component installation position detection model has not converged.

[0071] Among them, the preset threshold can be used to characterize the maximum loss value when the target vehicle component installation position detection model converges. The preset threshold can be a loss value pre-set by the user, or it can be a loss value automatically generated by the detection equipment according to multiple vehicle offline detection processes, etc. The setting method of the preset threshold is not limited here, and it can be set according to actual needs.

[0072] Step S120: inputting the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result according to the current vehicle image.

[0073] In an embodiment of the present application, the detection device can input the current vehicle image into the target vehicle component installation position detection model. The target vehicle component installation position detection model receives and responds to the current vehicle image, performs a vehicle off-line inspection on the current vehicle based on the current vehicle image, obtains the corresponding detection result, and outputs the detection result to the detection device. The detection device receives the detection result output by the target vehicle component installation position detection model.

[0074] The detection result may include a first detection result for characterizing that the installation position of the current vehicle component of the current vehicle has no deviation, and a second detection result for characterizing that the installation position of the current vehicle component of the current vehicle has a deviation.

[0075] In some embodiments, after determining whether the target vehicle component installation position detection model has converged, the detection device can input the current vehicle image into the target vehicle component installation position detection model when it is determined that the target vehicle component installation position detection model has converged. The target vehicle component installation position detection model receives and responds to the current vehicle image, performs a vehicle offline inspection on the current vehicle according to the current vehicle image, obtains a corresponding inspection result, and outputs the inspection result to the detection device. The detection device receives the inspection result output by the target vehicle component installation position detection model, and realizes that when it is determined that the target vehicle component installation position detection model has converged, a vehicle offline inspection is performed on the current vehicle based on the target vehicle component installation position detection model, thereby ensuring the stability of the target vehicle component installation position detection model and improving the detection credibility of the vehicle offline inspection of the current vehicle.

[0076] Step S130: When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline detection qualified result is generated.

[0077] In an embodiment of the present application, after the detection device inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result based on the current vehicle image, when the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline inspection qualified result can be generated, thereby realizing the target vehicle component installation position detection model constructed according to the SSD model and the multi-head attention model, and using the target vehicle component installation position detection model to perform vehicle offline inspection on the current vehicle, thereby improving the detection efficiency and detection accuracy of the vehicle offline inspection.

[0078] In some embodiments, when the detection device receives the first detection result output by the target vehicle component installation position detection model, after generating the corresponding offline inspection qualified result, the vehicle identification of the current vehicle can be obtained, and a qualified inspection mark corresponding to the vehicle identification can be generated and stored, which is conducive to tracing the vehicle offline inspection of the current vehicle according to the qualified inspection mark, avoiding repeated vehicle offline inspection of the current vehicle, and improving the inspection efficiency of the vehicle offline inspection.

[0079] As an implementation mode, when the Guard module receives the first detection result output by the target vehicle component installation position detection model, after generating the corresponding offline detection qualified result, it can send an identification acquisition instruction to the Hmi module, the Hmi module receives and responds to the identification acquisition instruction, forwards the identification acquisition instruction to the vehicle display screen, the vehicle display screen receives and responds to the identification acquisition instruction, generates and displays the corresponding first identification input prompt information, so that the user inputs the corresponding vehicle identification to the vehicle display screen according to the first identification input prompt information, the vehicle display screen receives the vehicle identification input by the user, sends the vehicle identification to the Hmi module, the Hmi module receives and forwards the vehicle identification to the Guard module, and the Guard module receives the vehicle identification returned by the Hmi module.

[0080] As an implementation method, the vehicle offline detection system may further include a client associated with the user. When the Guard module receives the first detection result output by the target vehicle component installation position detection model, after generating the corresponding offline detection qualified result, it may send a second identification input prompt information to the client via the network. The client receives and responds to the second identification input prompt information, prompting the user to input the corresponding vehicle identification to the client according to the second identification input prompt information. The client receives the vehicle identification input by the user and sends the vehicle identification to the Guard module. The Guard module receives the vehicle identification returned by the client.

[0081] As an implementation method, the vehicle offline detection system may further include a camera, which is connected to the Guard module via a network and exchanges data with the Guard module via the network. The camera is used to collect vehicle images of a preset area of ​​the current vehicle to obtain a corresponding preset vehicle area image. The preset vehicle area image includes a vehicle identification image of the current vehicle.

[0082] When the Guard module receives the first detection result output by the target vehicle component installation position detection model, after generating the corresponding offline detection qualified result, it can send an image acquisition instruction to the camera through the network. The camera receives and responds to the image acquisition instruction, collects vehicle images of the preset area of ​​the current vehicle, obtains the corresponding preset vehicle area image, and sends the preset vehicle area image to the Guard module through the network. The Guard module receives the preset vehicle area image returned by the camera, inputs the preset vehicle area image into the pre-trained deep learning network model, the deep learning network model receives and responds to the preset vehicle area image, detects the vehicle identification of the preset vehicle area image, obtains the corresponding vehicle identification, and outputs the vehicle identification to the Guard module. The Guard module receives the vehicle identification output by the deep learning network model.

[0083] Among them, the deep learning network model can be a Convolutional Neural Networks (CNN) model, a Deep Belief Networks (DBN) model, a Stacked Auto Encoder Networks (SAE) model, a Recurrent Neural Networks (RNN) model, a Deep Neural Networks (DNN) model, a Long Short-Term Memory (LSTM) network model or a Gated Recurring Units (GRU) model, etc. The type of deep learning network model is not limited here and can be set according to actual needs.

[0084] As an implementation method, the detection device may further include a local memory, which is communicatively connected to the Guard module and performs data exchange with the Guard module. After the Guard module obtains the vehicle identification of the current vehicle, a corresponding detection pass identification may be generated according to the vehicle identification, and the detection pass identification may be sent to the local memory, which receives the detection pass identification and stores the detection pass identification.

[0085] In some embodiments, the vehicle offline detection system may further include a server, which is connected to the Guard module via a network and exchanges data with the Guard module via the network. After the Guard module obtains the vehicle identification of the current vehicle, it can generate a corresponding qualified detection identification according to the vehicle identification, and send the qualified detection identification to the server via the network. The server receives and responds to the qualified detection identification and stores the qualified detection identification.

[0086] In some embodiments, after the detection device inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result based on the current vehicle image, when receiving the second detection result output by the target vehicle component installation position detection model, a corresponding first abnormal prompt information can be generated. The first abnormal prompt information can be used to prompt the user to adjust the current vehicle component installation position, thereby ensuring that the current vehicle that has passed the vehicle offline inspection is a qualified vehicle, and ensuring that the qualified rate of vehicles that have passed the vehicle offline inspection is high.

[0087] The first abnormal prompt information may be at least any one of the first abnormal sound prompt information, the first abnormal light prompt information, or the first abnormal text prompt information, and is not limited here.

[0088] The solution provided in this embodiment obtains the current vehicle image of the current vehicle and inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model is obtained based on the single-step multi-frame target detection model and the multi-head attention model. The detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased. When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline inspection qualified result is generated, thereby realizing the target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model, and using the target vehicle component installation position detection model to perform vehicle offline inspection on the current vehicle, thereby improving the detection efficiency and detection accuracy of vehicle offline inspection.

[0089] See also Figure 3 , which shows a flow chart of a vehicle offline detection method provided by another embodiment of the present application. In a specific embodiment, the vehicle offline detection method can be applied to Figure 1 The detection device 200 in the vehicle offline detection system shown in FIG. 2 is taken as an example to describe the detection device 200 in the following. Figure 3 The process shown is described in detail, and the vehicle offline detection method may include the following steps S210 to S240.

[0090] Step S210: Acquire a current vehicle image of the current vehicle.

[0091] Step S220: inputting the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result according to the current vehicle image.

[0092] In this embodiment, step S210 and step S220 may refer to the contents of the corresponding steps in the aforementioned embodiment, which will not be described in detail here.

[0093] Step S230: When receiving the first detection result output by the target vehicle component installation position detection model, determining whether the current vehicle sensor of the current vehicle passes the performance test.

[0094] In this embodiment, when the detection device receives the first detection result output by the target vehicle component installation position detection model, it can determine whether the current vehicle sensor of the current vehicle passes the performance test.

[0095] Among them, the current vehicle sensor may include at least any one of a vehicle speed sensor, a temperature sensor, a rotation speed sensor, a pressure sensor, a rotation angle sensor, a torque sensor, a hydraulic sensor, a vehicle-mounted camera sensor, a radar sensor, and a combined inertial navigation sensor, and the acquisition accuracy of the current vehicle sensor can be used to characterize the detection performance of the current vehicle sensor.

[0096] Specifically, when the Guard module receives the first detection result output by the target vehicle component installation position detection model, it can send a data acquisition instruction to the Data module. The Data module receives and responds to the data acquisition instruction, obtains the collected data of the current vehicle sensor, and sends the collected data to the Guard module. The Guard module receives the collected data returned by the Data module and determines whether the current vehicle sensor passes the performance test based on the collected data.

[0097] In some embodiments, the vehicle offline detection system may further include a client associated with the user, the client is connected to the Guard module via a network, and performs data interaction with the Guard module via the network. When the Guard module receives the first detection result output by the target vehicle component installation position detection model, it may send a data acquisition instruction to the Data module, the Data module receives and responds to the data acquisition instruction, acquires the collected data of the current vehicle sensor, and sends the collected data to the Guard module, the Guard module receives the collected data returned by the Data module, and sends the collected data to the client via the network, the client receives and responds to the collected data, prompts the user to determine whether the collection accuracy of the current vehicle sensor meets the requirements based on the collected data, and sends the corresponding determination result to the client, the client receives the determination result returned by the user, and forwards the determination result to the Guard module, and the Guard module determines whether the current vehicle sensor passes the performance test based on the determination result returned by the client.

[0098] The determination result may include a first determination result used to indicate that the acquisition accuracy of the current vehicle sensor meets the requirement, and a second determination result used to indicate that the acquisition accuracy of the current vehicle sensor does not meet the requirement.

[0099] When the Guard module receives the first determination result returned by the client, it determines that the current vehicle sensor has passed the performance test; when the Guard module receives the second determination result returned by the client, it determines that the current vehicle sensor has not passed the performance test.

[0100] In some embodiments, when the Guard module receives the first detection result output by the target vehicle component installation position detection model, it can send a data acquisition instruction to the Data module. The Data module receives and responds to the data acquisition instruction, acquires the collected data of the current vehicle sensor, and sends the collected data to the Guard module. The Guard module receives the collected data returned by the Data module, sends the collected data to the Hmi module, the Hmi module receives and responds to the collected data sent by the Guard module, sends the collected data to the vehicle-mounted display screen, the vehicle-mounted display screen receives and responds to the collected data, displays the collected data, and generates corresponding judgment prompt information, so that the user can determine whether the collection accuracy of the current vehicle sensor meets the requirements according to the judgment prompt information, and input the corresponding judgment result to the vehicle-mounted display screen. The vehicle-mounted display screen receives and responds to the judgment result input by the user, sends the judgment result to the Hmi module, the Hmi module forwards the judgment result to the Guard module, and the Guard module determines whether the current vehicle sensor passes the performance test according to the judgment result returned by the Hmi module.

[0101] When the Guard module receives the first determination result returned by the Hmi module, it determines that the current vehicle sensor passes the performance test; when the Guard module receives the second determination result returned by the Hmi module, it determines that the current vehicle sensor fails the performance test.

[0102] The determination prompt information may be at least any one of sound determination prompt information, light determination prompt information or text determination prompt information, etc. The type of the determination prompt information is not limited here and may be set according to actual needs.

[0103] In some embodiments, when the detection device receives the first detection result output by the target vehicle component installation position detection model, it can obtain the operating frequency of the current vehicle sensor of the current vehicle, and determine whether the current vehicle sensor is in a normal online state based on the operating frequency. When it is determined that the current vehicle sensor is in a normal online state based on the operating frequency, it is determined whether the current vehicle sensor has passed the performance test. This achieves performance testing of the current vehicle sensor when it is detected that there is no deviation in the installation position of the current vehicle component of the current vehicle and the current vehicle sensor is in a normal online state, thereby ensuring a high pass rate for vehicles that have passed the vehicle offline inspection.

[0104] When the operating frequency is within the preset frequency range, it is determined that the current vehicle sensor is in a normal online state; when the operating frequency is not within the preset frequency range, it is determined that the current vehicle sensor is in an abnormal online state.

[0105] Among them, the preset frequency range can be used to characterize the communication frequency range when the current vehicle sensor is in a normal communication state. The preset frequency range can be a communication frequency range pre-set by the user, or it can be a communication frequency range automatically generated by the detection equipment according to multiple vehicle offline detection processes, etc., which is not limited here.

[0106] Specifically, the Guard module can send a frequency acquisition instruction to the Data module, the Data module receives and responds to the frequency acquisition instruction, acquires the operating frequency of the current vehicle sensor, and sends the operating frequency to the Guard module, the Guard module receives the operating frequency returned by the Data module, and determines whether the current vehicle sensor is in a normal online state based on the operating frequency, and when it is determined that the current vehicle sensor is in a normal online state based on the operating frequency, sends a data acquisition instruction to the Data module, the Data module receives and responds to the data acquisition instruction, acquires the collected data of the current vehicle sensor, and sends the collected data to the Guard module, the Guard module receives the collected data returned by the Data module, and determines whether the current vehicle sensor passes the performance test based on the collected data.

[0107] Step S240: When it is determined that the current vehicle sensor of the current vehicle passes the performance test, a corresponding offline test qualified result is generated.

[0108] In this embodiment, when the detection equipment determines that the current vehicle sensor of the current vehicle has passed the performance test, a corresponding offline test qualified result can be generated, so that when it is detected that the installation position of the current vehicle components of the current vehicle is without deviation and the current vehicle sensor has passed the performance test, an offline test qualified result is generated, thereby ensuring that the qualified rate of vehicles that have passed the vehicle offline test is high.

[0109] In some embodiments, when the detection device determines that the current vehicle sensor is in an abnormal online state based on the working frequency, it indicates that the communication state of the current vehicle sensor is abnormal, and a corresponding second abnormal prompt information can be generated. The second abnormal prompt information can be used to prompt the user to adjust the online state of the current vehicle sensor, and when the user's first confirmation instruction is received, the step of obtaining the working frequency of the current vehicle sensor is executed, thereby realizing continuous offline detection of the current vehicle, ensuring that the offline vehicles that have passed the vehicle offline detection are qualified vehicles, and further ensuring that the qualification rate of vehicles that have passed the vehicle offline detection is high.

[0110] Among them, the second abnormal prompt information can be at least any one of the second abnormal sound prompt information, the second abnormal light prompt information or the second abnormal text prompt information, and the first confirmation instruction can be used to indicate that the user has completed the adjustment of the online status of the current vehicle sensor.

[0111] In some embodiments, when the detection device determines that the current vehicle sensor has not passed the performance test, it means that the acquisition accuracy of the current vehicle sensor does not meet the requirements, and a corresponding third abnormal prompt information can be generated. The third abnormal prompt information can be used to prompt the user to calibrate the acquisition accuracy of the current vehicle sensor, and when the user's second confirmation instruction is received, the step of determining whether the current vehicle sensor of the current vehicle has passed the performance test is executed, thereby realizing continuous offline testing of the current vehicle, ensuring that the offline vehicles that have passed the vehicle offline test are qualified vehicles, and further ensuring that the pass rate of vehicles that have passed the vehicle offline test is high.

[0112] Among them, the third abnormal prompt information can be at least any one of the third abnormal sound prompt information, the third abnormal light prompt information or the third abnormal text prompt information, and the second confirmation instruction can be used to indicate that the user has completed the calibration of the acquisition accuracy of the current vehicle sensor.

[0113] In some embodiments, when the detection device receives the second detection result output by the target vehicle component installation position detection model, it can also obtain the operating frequency of the current vehicle sensor of the current vehicle, and determine whether the current vehicle sensor is in a normal online state based on the operating frequency. Regardless of whether the current vehicle sensor is in a normal online state based on the operating frequency, it is further determined whether the current vehicle sensor passes the performance test, thereby realizing the sequential determination of the current vehicle component installation position, whether the current vehicle sensor is in a normal online state, and whether the current vehicle sensor passes the performance test. In this way, all reasons causing the current vehicle to fail the offline inspection can be determined through a single vehicle offline inspection process, making it convenient for staff to uniformly adjust all reasons causing the current vehicle to fail the offline inspection, which not only improves the comprehensiveness and accuracy of locating the reasons for the failure of the offline inspection, but also avoids the waste of labor costs caused by repeated adjustments to the reasons for the failure of the offline inspection.

[0114] The solution provided in this embodiment obtains the current vehicle image of the current vehicle and inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image, and when receiving the first detection result output by the target vehicle component installation position detection model, determines whether the current vehicle sensor of the current vehicle passes the performance test, and when it is determined that the current vehicle sensor of the current vehicle passes the performance test, generates the corresponding offline detection qualified result, realizes the target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model, and uses the target vehicle component installation position detection model to perform vehicle offline detection on the current vehicle, thereby improving the detection efficiency and detection accuracy of the vehicle offline detection.

[0115] Furthermore, when it is detected that there is no deviation in the installation position of the current vehicle components of the current vehicle and the current vehicle sensors pass the performance test, an offline test qualified result is generated, ensuring a high pass rate for vehicles that have passed the offline test.

[0116] See also Figure 4 , which shows a vehicle offline detection device 300 provided by an embodiment of the present application. The vehicle offline detection device 300 can be applied to Figure 1 The detection device 200 in the vehicle offline detection system shown in FIG. 2 is taken as an example to describe the detection device 200 in the following. Figure 4 The vehicle offline detection device 300 shown in the figure is described in detail. The vehicle offline detection device 300 may include an image acquisition module 310 , an image input module 320 and a generation module 330 .

[0117] The image acquisition module 310 can be used to acquire the current vehicle image of the current vehicle; the image input module 320 can be used to input the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model can be obtained based on the single-step multi-frame target detection model and the multi-head attention model. The detection result may include a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased; the generation module 330 can be used to generate a corresponding offline detection qualified result when receiving the first detection result output by the target vehicle component installation position detection model.

[0118] In some implementations, the vehicle offline detection device 300 may further include a construction module and a convergence determination module.

[0119] The construction module can be used to construct the target vehicle component installation position detection model according to the single-step multi-frame target detection model and the multi-head attention model before the image input module 320 inputs the current vehicle image to the target vehicle component installation position detection model; the convergence determination module can be used to determine whether the target vehicle component installation position detection model converges.

[0120] In some embodiments, the image input module 320 may include an image input unit.

[0121] The image input unit may be used to input the current vehicle image into the target vehicle component installation position detection model when it is determined that the target vehicle component installation position detection model has converged.

[0122] In some embodiments, the construction module may include a fusion unit, an acquisition unit, and a training set input unit.

[0123] The fusion unit can be used to fuse the single-step multi-box target detection model and the multi-head attention model to obtain an initial vehicle part installation position detection model; the acquisition unit can be used to obtain a corresponding sample set based on historical vehicle images, and the sample set at least includes a training set; the training set input unit can be used to input the training set to the initial vehicle part installation position detection model for training to obtain the corresponding target vehicle part installation position detection model.

[0124] In some implementations, the vehicle offline detection device 300 may further include an enhanced processing module.

[0125] The enhancement processing module can be used for the training set input unit to input the training set to the initial vehicle component installation position detection model for training. Before obtaining the corresponding target vehicle component installation position detection model, the training set is data enhanced to obtain the corresponding enhanced training set.

[0126] In some implementations, the training set input unit may include an input subunit.

[0127] The input subunit can be used to input the enhanced training set into the initial vehicle component installation position detection model for training to obtain the corresponding target vehicle component installation position detection model.

[0128] In some implementations, the sample set may further include a test set, and the convergence determination module may include a test set input unit, a first determination unit, a second determination unit, and a third determination unit.

[0129] The test set input unit can be used to input the test set into the target vehicle component installation position detection model for testing to obtain the corresponding test results; the first determination unit can be used to determine the loss value between the target result corresponding to the test set and the test result; the second determination unit can be used to determine that the target vehicle component installation position detection model has converged when the loss value is less than or equal to a preset threshold; the third determination unit can be used to determine that the target vehicle component installation position detection model has not converged when the loss value is greater than a preset threshold.

[0130] In some implementations, the vehicle offline detection device 300 may further include a performance determination module.

[0131] The performance determination module may be used to determine whether a current vehicle sensor of a current vehicle passes a performance test before the generation module 330 generates a corresponding offline test qualified result.

[0132] In some implementations, the generation module 330 may include a generation unit.

[0133] The generating unit may be configured to generate a corresponding off-line test qualified result when it is determined that the current vehicle sensor of the current vehicle passes the performance test.

[0134] In some implementations, the vehicle offline detection device 300 may further include a frequency acquisition module and a state determination module.

[0135] The frequency acquisition module can be used to obtain the operating frequency of the current vehicle sensor of the current vehicle before the performance determination module determines whether the current vehicle sensor of the current vehicle passes the performance test; the state determination module can be used to determine whether the current vehicle sensor is in a normal online state based on the operating frequency.

[0136] In some implementations, the performance determination module may include a fourth determination unit.

[0137] The fourth determination unit may be configured to determine whether the current vehicle sensor passes the performance test when it is determined according to the working frequency that the current vehicle sensor is in a normal online state.

[0138] The solution provided in this embodiment obtains the current vehicle image of the current vehicle and inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model is obtained based on the single-step multi-frame target detection model and the multi-head attention model. The detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased. When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline inspection qualified result is generated, thereby realizing the target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model, and using the target vehicle component installation position detection model to perform vehicle offline inspection on the current vehicle, thereby improving the detection efficiency and detection accuracy of vehicle offline inspection.

[0139] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. Any processing method described in the method embodiment can be implemented by the corresponding processing module in the device embodiment, and will not be repeated one by one in the device embodiment.

[0140] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0141] See also Figure 5 , which shows a functional block diagram of a detection device 200 provided by an embodiment of the present application, and the detection device 200 may include one or more of the following components: a memory 210, a processor 220, and one or more applications, wherein the one or more applications may be stored in the memory 210 and configured to be executed by one or more processors 220, and the one or more applications are configured to execute the method described in the aforementioned method embodiment.

[0142] The memory 210 may include a random access memory (RAM) or a read-only memory (ROM). The memory 210 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 210 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as obtaining a current vehicle image, inputting a current vehicle image, outputting a detection result, receiving a first detection result, generating a qualified offline detection result, constructing a target vehicle component installation position detection model, determining whether the target vehicle component installation position detection model converges, determining that the target vehicle component installation position detection model converges, fusing a single-step multi-frame target detection model and a multi-head attention model, obtaining an initial vehicle component installation position detection model, obtaining a sample set, inputting a sample set, training an initial vehicle component installation position detection model, obtaining a target vehicle component installation position detection model, data enhancement processing a training set, obtaining an enhanced training set, inputting an enhanced training set, inputting a test set, testing the target vehicle component installation position detection model, obtaining a test result, determining a loss value, determining that the target vehicle component installation position detection model has not converged, determining whether the current vehicle sensor passes the performance test, determining that the current vehicle sensor passes the performance test, obtaining an operating frequency, determining whether the vehicle sensor is in a normal online state, and determining that the vehicle sensor is in a normal online state, etc.), instructions for implementing the following various method embodiments, etc. The storage data area can also store data created by the detection device 200 during use (such as the current vehicle, the current vehicle image, the target vehicle component installation position detection model, the detection results, the single-step multi-box target detection model, the multi-head attention model, the current vehicle component installation position, the first detection result, the second detection result, the offline detection qualified result, the initial vehicle component installation position detection model, the historical vehicle image, the sample set, the training set, the enhanced training set, the test set, the test results, the target results, the loss value, the preset threshold value, the current vehicle sensor and the operating frequency), etc.

[0143] The processor 220 may include one or more processing cores. The processor 220 uses various interfaces and lines to connect the various parts of the entire detection device 200, and executes various functions and processes data of the detection device 200 by running or executing instructions, programs, code sets or instruction sets stored in the memory 210, and calling data stored in the memory 210. Optionally, the processor 220 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 220 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It is understandable that the above-mentioned modem may not be integrated into the processor 220, but may be implemented separately through a communication chip.

[0144] Please refer to Figure 6 , which shows a structural block diagram of a computer-readable storage medium provided in an embodiment of the present application. The computer-readable storage medium 500 stores a program code 510, which can be called by a processor to execute the method described in the above method embodiment.

[0145] The computer-readable storage medium 500 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. Optionally, the computer-readable storage medium 500 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 500 has storage space for program code 510 that performs any method steps in the above method. These program codes can be read from or written to one or more computer program products. The program code 510 can be compressed, for example, in an appropriate form.

[0146] Please refer to Figure 7, which shows a structural block diagram of a computer program product 600 provided in an embodiment of the present application. The computer program product 600 includes a computer program / instruction 610, and the computer program / instruction 610 is stored in a computer-readable storage medium of a computer device. When the computer program product 600 is run on a computer device, a processor of the computer device reads the computer program / instruction 610 from the computer-readable storage medium, and the processor executes the computer program / instruction 610, so that the computer device performs the method described in the above method embodiment.

[0147] The solution provided in this embodiment obtains the current vehicle image of the current vehicle and inputs the current vehicle image into the target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs the corresponding detection result according to the current vehicle image. The target vehicle component installation position detection model is obtained based on the single-step multi-frame target detection model and the multi-head attention model. The detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased. When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline inspection qualified result is generated, thereby realizing the target vehicle component installation position detection model constructed according to the single-step multi-frame target detection model and the multi-head attention model, and using the target vehicle component installation position detection model to perform vehicle offline inspection on the current vehicle, thereby improving the detection efficiency and detection accuracy of vehicle offline inspection.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle offline detection method, characterized in that: include: Get the current vehicle image of the current vehicle; Input the current vehicle image to a target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result according to the current vehicle image, the target vehicle component installation position detection model is obtained based on a single-step multi-frame target detection model and a multi-head attention model, and the detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased; When the first detection result output by the target vehicle component installation position detection model is received, a corresponding offline detection qualified result is generated.

2. The vehicle offline detection method according to claim 1, characterized in that: Before inputting the current vehicle image into the target vehicle component installation position detection model, the vehicle offline detection method further includes: Based on the single-step multi-box target detection model and the multi-head attention model, a target vehicle component installation position detection model is constructed; Determining whether the target vehicle component installation position detection model converges; The step of inputting the current vehicle image into a target vehicle component installation position detection model comprises: When it is determined that the target vehicle component installation position detection model has converged, the current vehicle image is input into the target vehicle component installation position detection model.

3. The vehicle offline detection method according to claim 2, characterized in that: The target vehicle component installation position detection model is constructed according to the single-step multi-frame target detection model and the multi-head attention model, including: The single-step multi-box object detection model and the multi-head attention model are integrated to obtain the initial vehicle component installation position detection model; According to the historical vehicle images, a corresponding sample set is obtained, wherein the sample set at least includes a training set; The training set is input into the initial vehicle component installation position detection model for training to obtain a corresponding target vehicle component installation position detection model.

4. The vehicle offline detection method according to claim 3, characterized in that: Before inputting the training set to the initial vehicle component installation position detection model for training to obtain the corresponding target vehicle component installation position detection model, the vehicle offline detection method further includes: Data enhancement processing is performed on the training set to obtain a corresponding enhanced training set; The step of inputting the training set to the initial vehicle component installation position detection model for training to obtain a corresponding target vehicle component installation position detection model includes: The enhanced training set is input into the initial vehicle component installation position detection model for training to obtain a corresponding target vehicle component installation position detection model.

5. The vehicle offline detection method according to claim 3, characterized in that: The sample set also includes a test set, and determining whether the target vehicle component installation position detection model converges includes: Input the test set to the target vehicle component installation position detection model for testing to obtain corresponding test results; Determine the target result corresponding to the test set and the loss value of the test result; When the loss value is less than or equal to a preset threshold, it is determined that the target vehicle component installation position detection model converges; When the loss value is greater than a preset threshold, it is determined that the target vehicle component installation position detection model has not converged.

6. The vehicle offline detection method according to any one of claims 1 to 5, characterized in that: Before generating the corresponding off-line test qualified result, the vehicle off-line test method further includes: determining whether a current vehicle sensor of the current vehicle passes a performance test; The generating of the corresponding offline test qualified result includes: When it is determined that the current vehicle sensor of the current vehicle passes the performance test, a corresponding off-line test qualified result is generated.

7. The vehicle offline detection method according to claim 6, characterized in that: Before determining whether the current vehicle sensor of the current vehicle passes the performance test, the vehicle offline detection method further includes: Acquire an operating frequency of a current vehicle sensor of the current vehicle; Determining whether the current vehicle sensor is in a normal online state according to the operating frequency; The determining whether the current vehicle sensor of the current vehicle passes the performance test includes: When it is determined according to the working frequency that the current vehicle sensor is in a normal online state, it is determined whether the current vehicle sensor passes a performance test.

8. A vehicle offline detection device, characterized in that: include: An image acquisition module, used to acquire a current vehicle image of a current vehicle; An image input module, used for inputting the current vehicle image into a target vehicle component installation position detection model, so that the target vehicle component installation position detection model outputs a corresponding detection result according to the current vehicle image, wherein the target vehicle component installation position detection model is obtained based on a single-step multi-frame target detection model and a multi-head attention model, and the detection result includes a first detection result for characterizing that the current vehicle component installation position of the current vehicle is unbiased, and a second detection result for characterizing that the current vehicle component installation position is biased; A generation module is used to generate a corresponding offline inspection qualified result when receiving the first detection result output by the target vehicle component installation position detection model.

9. A detection device, characterized in that: include: Memory; One or more processors coupled to the memory; One or more applications, wherein the one or more applications are stored in the memory and are configured to be executed by one or more processors, and the one or more applications are configured to execute the vehicle offline detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, and the program codes can be called by a processor to execute the vehicle offline detection method according to any one of claims 1 to 7.

Images