Aircraft identification method, aircraft identification device, electronic equipment and storage medium

Abstract

The invention discloses an aircraft identification method, an aircraft identification device, electronic equipment and a storage medium, and the aircraft identification method comprises the steps: obtaining a radar reflected wave which is formed based on the reflection of a radar transmitted wave by an obstacle; when it is determined that the aircraft exists in the preset airspace according to the radar reflected waves, obtaining a current aircraft image of the aircraft; and inputting the current aircraft image to an aircraft identification model to obtain an aircraft identification result, the aircraft identification model being obtained by training a deep learning network model based on a historical aircraft image, the historical aircraft image being labeled with a historical aircraft type, and the aircraft identification result comprising a current aircraft type. According to the method, aircraft identification is carried out on the current aircraft image according to the aircraft identification model, and the identification accuracy and the identification efficiency of aircraft identification are improved.

Description

Technical Field

[0001] The present application belongs to the field of image recognition technology, and in particular, relates to an aircraft recognition method, an aircraft recognition device, an electronic device and a storage medium. Background Art

[0002] Aircraft (for example, drones, airplanes, helicopters, etc.) play an increasingly important role in military, civil, and scientific research fields. With the increase in the number of aircraft and the expansion of their application scope, how to identify aircraft in real time and accurately has become a key issue in ensuring airspace safety and improving management efficiency.

[0003] At present, aircraft identification mainly relies on manual judgment of aircraft types based on aircraft images. The judgment results are easily affected by the subjective experience of the judges, resulting in low accuracy in aircraft identification. In addition, the manual judgment method results in low efficiency in aircraft identification. Summary of the invention

[0004] In view of this, embodiments of the present application provide an aircraft identification method, an aircraft identification device, an electronic device, and a storage medium to overcome the above problems of the prior art.

[0005] In a first aspect, an embodiment of the present application provides an aircraft identification method, including: obtaining a radar reflection wave, wherein the radar reflection wave is formed based on the reflection of a radar transmission wave by an obstacle; when it is determined that an aircraft exists in a preset airspace based on the radar reflection wave, obtaining a current aircraft image of the aircraft; inputting the current aircraft image into an aircraft identification model to obtain an aircraft identification result, wherein the aircraft identification model is obtained by training a deep learning network model based on historical aircraft images, the historical aircraft images are annotated with historical aircraft types, and the aircraft identification result includes a current aircraft category.

[0006] Among them, in some optional embodiments, before obtaining the radar reflection wave, the aircraft identification method also includes: obtaining the historical aircraft image; determining an initial training set based on the historical aircraft image; and training the deep learning network model based on the initial training set to obtain the aircraft identification model.

[0007] Among them, in some optional embodiments, the deep learning network model is trained based on the initial training set to obtain the aircraft recognition model, including: performing transfer learning on the deep learning network model based on the initial training set to obtain a transfer learning model; performing data augmentation processing on the initial training set to obtain an augmented training set; inputting the augmented training set into the transfer learning model for fine-tuning training to obtain the aircraft recognition model.

[0008] Among them, in some optional embodiments, the deep learning network model is transferred based on the initial training set to obtain the transfer learning model, including: inputting the initial training set into the deep learning network model for pre-training to obtain a pre-trained model; freezing the convolution layer of the pre-trained model to obtain a frozen training model; adding a fully connected layer to the top of the frozen training model to obtain the transfer learning model.

[0009] Among them, in some optional embodiments, before inputting the current aircraft image into the aircraft recognition model to obtain the aircraft recognition result, the aircraft recognition method also includes: preprocessing the current aircraft image to obtain a preprocessed aircraft image; inputting the current aircraft image into the aircraft recognition model to obtain the aircraft recognition result includes: inputting the preprocessed aircraft image into the aircraft recognition model to obtain the aircraft recognition result.

[0010] Among them, in some optional embodiments, the aircraft identification method further includes: if the current aircraft category is a preset aircraft category, generating a warning message.

[0011] Among them, in some optional embodiments, the aircraft identification result also includes a current probability distribution corresponding to the current aircraft category, and the aircraft identification method also includes: calculating the current confidence of the current aircraft category based on the current probability distribution; and generating an identification record including the current aircraft category and the current confidence.

[0012] In a second aspect, an embodiment of the present application provides an aircraft identification device, the aircraft identification device comprising a first acquisition module, a second acquisition module and a first input module. The first acquisition module is used to acquire radar reflection waves, the radar reflection waves are formed based on the reflection of radar transmission waves by obstacles; the second acquisition module is used to acquire the current aircraft image of the aircraft when it is determined that there is an aircraft in a preset airspace according to the radar reflection waves; the first input module is used to input the current aircraft image into an aircraft identification model to obtain an aircraft identification result, the aircraft identification model is obtained by training a deep learning network model based on historical aircraft images, the historical aircraft images are annotated with historical aircraft types, and the aircraft identification result includes the current aircraft category.

[0013] In a third aspect, an embodiment of the present application provides an electronic 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 aircraft identification method provided in the first aspect above.

[0014] In a fourth 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 aircraft identification method provided in the first aspect above.

[0015] In a fifth 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 aircraft identification method provided in the first aspect above.

[0016] The solution provided by the present application obtains radar reflection waves, which are formed based on the reflection of radar transmission waves by obstacles. When it is determined that an aircraft exists in a preset airspace based on the radar reflection waves, a current aircraft image of the aircraft is obtained, and the current aircraft image is input into an aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft recognition result includes the current aircraft category, thereby realizing aircraft recognition of the current aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] 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.

[0018] Figure 1 A schematic diagram of a scenario of an aircraft identification system provided in an embodiment of the present application is shown.

[0019] Figure 2 A schematic flow chart of an aircraft identification method provided in an embodiment of the present application is shown.

[0020] Figure 3 Another flow chart of the aircraft identification method provided in an embodiment of the present application is shown.

[0021] Figure 4Another flowchart of the aircraft identification method provided in the embodiment of the present application is shown.

[0022] Figure 5 A structural block diagram of an aircraft identification device provided in an embodiment of the present application is shown.

[0023] Figure 6 A functional block diagram of an electronic device provided in an embodiment of the present application is shown.

[0024] Figure 7 A computer-readable storage medium provided in an embodiment of the present application for storing or carrying a program code for implementing an aircraft identification method provided in an embodiment of the present application is shown.

[0025] Figure 8 A computer program product provided in an embodiment of the present application and used to store or carry program codes for implementing the aircraft identification method provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] 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.

[0027] 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.

[0028] 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, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0029] 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.

[0030] 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.

[0031] Aircraft (for example, drones, airplanes, helicopters, etc.) play an increasingly important role in military, civil, and scientific research fields. With the increase in the number of aircraft and the expansion of their application scope, how to identify aircraft in real time and accurately has become a key issue in ensuring airspace safety and improving management efficiency.

[0032] At present, aircraft identification mainly relies on manual judgment of aircraft types based on aircraft images. The judgment results are easily affected by the subjective experience of the judges, resulting in low accuracy in aircraft identification. In addition, the manual judgment method results in low efficiency in aircraft identification.

[0033] In response to the above problems, the aircraft identification method, aircraft identification device, electronic device and storage medium provided in the embodiments of the present application obtain radar reflection waves, which are formed based on the reflection of radar transmission waves by obstacles. When it is determined that there is an aircraft in a preset airspace based on the radar reflection waves, the current aircraft image of the aircraft is obtained, and the current aircraft image is input into an aircraft identification model to obtain an aircraft identification result. The aircraft identification model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft identification result includes the current aircraft category, thereby realizing aircraft identification of the current aircraft image according to the aircraft identification model, thereby improving the recognition accuracy and efficiency of aircraft identification.

[0034] 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.

[0035] See also Figure 1 , which shows a schematic diagram of an application scenario of the aircraft identification system provided in an embodiment of the present application, which may include an aircraft 100, a camera 200 and a processing device 300. The camera 200 is connected to the processing device 300 via a network and exchanges data with the processing device 300 via the network.

[0036] The aircraft 100 may be any one of a drone, an airplane, a helicopter, a flying saucer or a spaceship, etc. The type of the aircraft 100 is not limited here and may be specifically configured according to actual needs.

[0037] The camera 200 can be used to capture images of the aircraft 100 and send the captured aircraft images to the processing device 300. The camera 200 can be any one of a wide-angle camera, a macro camera, an ultra-wide-angle camera, or a panoramic camera, etc. The type of the camera 200 is not limited here and can be set according to actual needs.

[0038] The processing device 300 can be used to receive the aircraft image sent by the camera 200 and identify the aircraft 100 according to the aircraft image. The processing device 300 can be a terminal device or a server, etc. The type of the processing device 300 is not limited here and can be set according to actual needs.

[0039] The terminal device can be a mobile terminal device (for example, a mobile phone, a personal digital assistant (PDA), a tablet computer (Tablet Personal Computer, Tablet PC), a laptop computer, a smart watch, a smart bracelet or a wearable device, etc.), or a fixed terminal device (desktop computer, smart panel, etc.), etc.

[0040] The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or any one of the cloud servers that provide 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 or artificial intelligence platforms, etc.

[0041] 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.

[0042] In some implementations, the camera 200 may be fixedly installed at a certain location, or may be fixedly installed on a drone, etc. The setting method of the camera 200 is not limited here, and may be specifically set according to actual needs.

[0043] See also Figure 2 , which shows a flow chart of an aircraft identification method provided by an embodiment of the present application. In a specific embodiment, the aircraft identification method can be applied to Figure 1 The processing device 300 in the aircraft identification system shown in FIG. Figure 2 The process shown is described in detail, and the aircraft identification method may include the following steps S110 to S130.

[0044] Step S110: Acquire radar reflected waves.

[0045] In the embodiment of the present application, the aircraft identification system may further include a radar sensor, which may be used to transmit radar transmission waves to a preset airspace and detect radar reflection waves in the preset airspace, wherein the radar reflection waves may be formed based on the reflection of the radar transmission waves by obstacles. The radar sensor is connected to the processing device via a network and exchanges data with the processing device via the network.

[0046] The processing device can send monitoring instructions to the radar sensor through the network. The radar sensor receives and responds to the monitoring instructions, transmits radar transmission waves to the preset airspace, detects radar reflection waves in the preset airspace, and sends the detected radar reflection waves to the processing device through the network. The processing device receives the radar reflection waves returned by the radar sensor.

[0047] Among them, the radar sensor can be any one of millimeter wave radar, laser radar or microwave radar, etc. The type of radar sensor is not limited here and can be set according to actual needs.

[0048] The preset airspace can be a spatial range pre-set by the user, or a spatial range automatically generated by the processing equipment according to the identification process of multiple aircraft identifications, etc. The setting method of the preset airspace is not limited here, and it can be set according to actual needs.

[0049] Obstacles may include at least any one of aircraft, buildings and natural obstacles (eg, mountains, trees, etc.). The types of obstacles are not limited here and can be set according to actual needs.

[0050] In some implementations, after the processing device acquires the radar reflection wave, it can determine whether there is an aircraft in a preset airspace based on the radar reflection wave.

[0051] Specifically, after the processing device obtains the radar reflection wave, it can determine whether there is an aircraft in the preset airspace based on the radar arrival angle of the radar reflection wave.

[0052] When the radar arrival angle changes, it is determined that there is an aircraft in the preset airspace; when the radar arrival angle remains unchanged, it is determined that there is no aircraft in the preset airspace.

[0053] Step S120: When it is determined that an aircraft exists in the preset airspace according to the radar reflection wave, a current aircraft image of the aircraft is acquired.

[0054] In an embodiment of the present application, when the processing device determines that there is an aircraft in a preset airspace based on the radar reflection wave, an image acquisition instruction can be sent to the camera via the network. The camera receives and responds to the image acquisition instruction, acquires an image of the aircraft, obtains the current aircraft image of the aircraft, and sends the current aircraft image to the processing device via the network. The processing device receives the current aircraft image returned by the camera. When the existence of the aircraft is determined based on the radar reflection wave, the aircraft image is acquired, which can avoid acquiring images that do not contain the aircraft image, thereby improving the acquisition accuracy of the aircraft image.

[0055] Among them, the current aircraft image can be any one of a drone image, an airplane image, a helicopter image, a flying saucer image or a spaceship image, etc. The type of the current aircraft image is not limited here and can be set according to actual needs.

[0056] Step S130: input the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result.

[0057] In an embodiment of the present application, after the processing device obtains the current aircraft image, the current aircraft image can be input into the aircraft recognition model. The aircraft recognition model receives and responds to the current aircraft image, identifies the aircraft according to the current aircraft image, and obtains an aircraft recognition result. The aircraft recognition model can be obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft recognition result includes the current aircraft category, thereby realizing aircraft recognition of the current aircraft image according to the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0058] Among them, the current aircraft category can be at least any one of a drone, an airplane, a helicopter, a flying saucer or a spaceship, and the historical aircraft category can be at least any one of a drone, an airplane, a helicopter, a flying saucer or a spaceship, etc., and there is no limitation here.

[0059] In some embodiments, the processing device inputs the current aircraft image into the aircraft recognition model, and after obtaining the current aircraft category, if the current aircraft category is a preset aircraft category, an alarm message can be generated so that the user can intercept the aircraft according to the alarm message, which is beneficial to improving safety.

[0060] The preset aircraft category may be at least one of a drone, an airplane, a helicopter, a flying saucer or a spaceship, etc., which is not limited here.

[0061] The warning information may be at least any one of text warning information, sound warning information or light warning information, etc., which is not limited here.

[0062] In some implementations, the aircraft identification system may further include a user client, which is connected to the processing device via a network and performs data exchange with the processing device via the network.

[0063] The processing device inputs the current aircraft image into the aircraft recognition model. After obtaining the current aircraft category, if the current aircraft category is the preset aircraft category, an alarm message can be generated and sent to the user client through the network. The user client receives and displays the alarm message to prompt the user to intercept the aircraft according to the alarm information, which is beneficial to improve safety.

[0064] Among them, the user client can be a mobile client (for example, any one of a mobile phone client, a PDA client, a Tablet PC client, a laptop client, a smart watch client, a smart bracelet client or a wearable device client, etc.), or a fixed client (a desktop computer client, a smart panel client, etc.), etc.

[0065] In some embodiments, the aircraft recognition result may further include a current probability distribution corresponding to the current aircraft category. After the processing device inputs the current aircraft image into the aircraft recognition model and obtains the current aircraft category and the current probability distribution, the current confidence of the current aircraft category may be calculated according to the current probability distribution, and an recognition record including the current aircraft category and the current confidence may be generated, so that the user can trace the aircraft recognition process according to the recognition record, which is conducive to improving the user experience during the aircraft recognition process.

[0066] The processing device may determine the maximum probability according to the current probability distribution, and determine the maximum probability as the current confidence level.

[0067] The identification record may also include an image identifier and a timestamp of the current aircraft image. The image identifier may be an image code automatically generated according to a coding sequence, and the timestamp may be a capture time of the current aircraft image.

[0068] In some embodiments, the aircraft recognition result may also include a current probability distribution corresponding to the current aircraft category. After the processing device inputs the current aircraft image into the aircraft recognition model and obtains the current aircraft category and the current probability distribution, the current confidence of the current aircraft category may be calculated according to the current probability distribution, and an identification record including the current aircraft category and the current confidence may be generated, and the identification record may be stored in a preset database so that the user can trace the aircraft recognition process according to the identification record, which is conducive to improving the user experience during the aircraft recognition process.

[0069] The preset database may be a real-time database or a relational database, etc., which is not limited here.

[0070] In some embodiments, the processing device inputs the current aircraft image into the aircraft recognition model, and after obtaining the current aircraft category and the current probability distribution, the current confidence of the current aircraft category can be calculated based on the current probability distribution. If the current aircraft category is a preset aircraft category and the current confidence is greater than or equal to the confidence threshold, an alarm message is generated and sent to the user client through the network. The user client receives and displays the alarm message, thereby providing an early warning when the confidence that the aircraft is identified as a preset aircraft is high, thereby improving the accuracy of the early warning.

[0071] The solution provided by the present application obtains radar reflection waves, which are formed based on the reflection of radar transmission waves by obstacles. When it is determined that an aircraft exists in a preset airspace based on the radar reflection waves, a current aircraft image of the aircraft is obtained, and the current aircraft image is input into an aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft recognition result includes the current aircraft category, thereby realizing aircraft recognition of the current aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0072] See also Figure 3 , which shows a flow chart of an aircraft identification method provided by another embodiment of the present application. In a specific embodiment, the aircraft identification method can be applied to Figure 1 The processing device 300 in the aircraft identification system shown in FIG. Figure 3 The process shown is described in detail, and the aircraft identification method may include the following steps S210 to S260.

[0073] Step S210: Acquire historical aircraft images.

[0074] In this embodiment, the processing device may acquire a historical aircraft image, wherein the historical aircraft image may include a historical aircraft type.

[0075] In some implementations, the processing device has historical aircraft images pre-stored therein, and the processing device may read the pre-stored historical aircraft images.

[0076] In some embodiments, the camera pre-stores historical aircraft images, and the processing device can send an acquisition instruction to the camera via the network. The camera receives and responds to the acquisition instruction and sends the pre-stored historical aircraft images to the processing device via the network. The processing device receives the historical aircraft images returned by the camera.

[0077] In some implementations, the processing device may generate upload prompt information, and receive historical aircraft images uploaded by the user according to the upload prompt information.

[0078] Among them, the upload prompt information can be used to prompt the user to upload historical aircraft images to the processing device according to the upload prompt information. The upload prompt information can be at least any one of text prompt information, sound prompt information or light prompt information, etc., which is not limited here.

[0079] Step S220: Determine an initial training set based on historical aircraft images.

[0080] In this embodiment, after the processing device acquires the historical aircraft images, it may determine an initial training set according to the historical aircraft images.

[0081] Specifically, after the processing device acquires the historical aircraft images, it can annotate the historical aircraft types and historical probability distributions on the historical aircraft images to obtain an initial training set.

[0082] Step S230: Train the deep learning network model based on the initial training set to obtain an aircraft recognition model.

[0083] In this embodiment, after the processing device determines the initial training set based on the historical aircraft images, the deep learning network model can be trained based on the initial training set to obtain an aircraft recognition model.

[0084] Among them, the deep learning network model can be a convolutional neural network (CNN) model, a deep belief network (Deep Belief The deep learning network model can be any of the following: DBN model, Stacked Auto Encoder Networks (SAE) model, Recurrent Neural Networks (RNN) model, Deep Neural Networks (DNN) model, Long Short-Term Memory (LSTM) network model or 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.

[0085] In some embodiments, after the processing device determines the initial training set based on the historical aircraft images, it can perform transfer learning on the deep learning network model based on the initial training set to obtain a transfer learning model, perform data augmentation processing on the initial training set to obtain an augmented training set, and input the augmented training set into the transfer learning model for fine-tuning training to obtain an aircraft recognition model. By performing transfer learning and fine-tuning on the deep learning network model and utilizing the existing knowledge in the pre-trained model, the deep learning network model is fine-tuned, thereby improving the training efficiency and model recognition performance of the aircraft recognition model.

[0086] Among them, the processing device can input the initial training set into the deep learning network model, the deep learning network model receives and responds to the initial training set, performs pre-training according to the initial training set to obtain a pre-trained model, freezes the convolution layer of the pre-trained model to obtain a frozen training model, and adds a fully connected layer to the top of the frozen training model to obtain a transfer learning model.

[0087] The data augmentation processing may be at least any one of rotation processing, flipping processing, size scaling processing, brightness scaling processing, sharpness scaling processing, etc., which is not limited here.

[0088] Step S240: Acquire radar reflected waves.

[0089] Step S250: When it is determined that an aircraft exists in the preset airspace based on the radar reflection wave, a current aircraft image of the aircraft is acquired.

[0090] Step S260: input the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result.

[0091] In this embodiment, step 240, step S250 and step S260 may refer to the contents of the corresponding steps in the aforementioned embodiments and will not be described in detail here.

[0092] The solution provided in this embodiment obtains historical aircraft images, determines an initial training set based on the historical aircraft images, trains a deep learning network model based on the initial training set to obtain an aircraft recognition model, obtains radar reflection waves, and when it is determined that there is an aircraft in a preset airspace based on the radar reflection waves, obtains the current aircraft image of the aircraft, and inputs the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result. This achieves the training of the deep learning network model based on historical aircraft images to obtain an aircraft recognition model, and performs aircraft recognition on the current aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0093] See also Figure 4 , which shows a flow chart of an aircraft identification method provided by another embodiment of the present application. In a specific embodiment, the aircraft identification method can be applied to Figure 1 The processing device 300 in the aircraft identification system shown in FIG. Figure 4 The process shown is described in detail, and the aircraft identification method may include the following steps S310 to S340.

[0094] Step S310: Acquire radar reflected waves.

[0095] Step S320: When it is determined that an aircraft exists in the preset airspace based on the radar reflection wave, a current aircraft image of the aircraft is acquired.

[0096] In this embodiment, step S310 and step S320 may refer to the contents of the corresponding steps in the aforementioned embodiments, which will not be described in detail here.

[0097] Step S330: preprocessing the current aircraft image to obtain a preprocessed aircraft image.

[0098] In this embodiment, when the processing device determines that there is an aircraft in a preset airspace based on the radar reflection wave, after obtaining the current aircraft image of the aircraft, the current aircraft image can be preprocessed to obtain a preprocessed aircraft image. The image quality of the preprocessed aircraft image is higher than the image quality of the current aircraft image, which is beneficial to improving the recognition accuracy of the aircraft.

[0099] The image quality may include at least one of clarity, contrast or brightness, etc., which is not limited here. The preprocessing may include at least one of image denoising, image enhancement or image cropping, etc., which is not limited here.

[0100] Image denoising processing can be a process of processing the noise in the image through a filter (such as a Gaussian filter), image enhancement processing can be at least any one of contrast adjustment, brightness adjustment or sharpness adjustment, and image cropping processing can crop the current aircraft image according to the aircraft contour, or crop the current aircraft image according to a preset multiple of the image area of ​​the current aircraft image, etc. The preset multiple can be 1.5 times, or 2 times, etc., which is not limited here.

[0101] Step S340: input the pre-processed aircraft image to the aircraft recognition model to obtain an aircraft recognition result.

[0102] In this embodiment, the processing device preprocesses the current aircraft image. After obtaining the preprocessed aircraft image, the preprocessed aircraft image can be input into the aircraft recognition model. The aircraft recognition model receives the preprocessed aircraft image, identifies the aircraft based on the preprocessed aircraft image, and obtains an aircraft recognition result. This realizes aircraft recognition of the preprocessed aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0103] The solution provided in this embodiment obtains radar reflection waves, and when it is determined that there is an aircraft in a preset airspace based on the radar reflection waves, obtains the current aircraft image of the aircraft, preprocesses the current aircraft image to obtain a preprocessed aircraft image, and inputs the preprocessed aircraft image to an aircraft recognition model to obtain an aircraft recognition result, thereby realizing aircraft recognition of the preprocessed aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and recognition efficiency of the aircraft.

[0104] See also Figure 5 , which shows an aircraft identification device 400 provided by an embodiment of the present application. The aircraft identification device 400 can be applied to Figure 1 The processing device 300 in the aircraft identification system shown in FIG. Figure 5 The aircraft identification device 400 shown is described in detail. The aircraft identification device 400 may include a first acquisition module 410 , a second acquisition module 420 and a first input module 430 .

[0105] The first acquisition module 410 can be used to acquire radar reflection waves, which can be formed based on the reflection of radar transmission waves by obstacles; the second acquisition module 420 can be used to acquire the current aircraft image of the aircraft when it is determined that there is an aircraft in the preset airspace based on the radar reflection waves; the first input module 430 can be used to input the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model can be obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images can be annotated with historical aircraft types, and the aircraft recognition result can include the current aircraft category.

[0106] In some implementations, the aircraft identification device 400 may further include a third acquisition module, a first determination module, and a training module.

[0107] The third acquisition module can be used to acquire historical aircraft images before the first acquisition module 410 acquires the radar reflection wave; the first determination module can be used to determine the initial training set based on the historical aircraft images; the training module can be used to train the deep learning network model based on the initial training set to obtain an aircraft recognition model.

[0108] In some implementations, the training module may include a learning unit, a processing unit, and a first input unit.

[0109] The learning unit can be used to perform transfer learning on the deep learning network model based on the initial training set to obtain a transfer learning model; the processing unit can be used to perform data augmentation processing on the initial training set to obtain an augmented training set; the first input unit can be used to input the augmented training set into the transfer learning model for fine-tuning training to obtain an aircraft recognition model.

[0110] In some implementations, the learning unit may include an input subunit, a freeze subunit, and an add subunit.

[0111] The input subunit can be used to input the initial training set into the deep learning network model for pre-training to obtain a pre-trained model; the freeze subunit can be used to freeze the convolution layer of the pre-trained model to obtain a frozen training model; the add subunit can be used to add a fully connected layer to the top of the frozen training model to obtain a transfer learning model.

[0112] In some implementations, the aircraft identification device 400 may further include a processing module.

[0113] The processing module may be used to pre-process the current aircraft image to obtain a pre-processed aircraft image before the first input module 430 inputs the current aircraft image to the aircraft recognition model to obtain an aircraft recognition result.

[0114] In some embodiments, the first input module 430 may include a second input unit.

[0115] The second input unit can be used to input the pre-processed aircraft image into the aircraft recognition model to obtain an aircraft recognition result.

[0116] In some implementations, the aircraft identification device 400 may further include a first generating module.

[0117] The first generating module may be configured to generate warning information if the current aircraft category is a preset aircraft category.

[0118] In some implementations, the aircraft identification result may further include a current probability distribution corresponding to the current aircraft category, and the aircraft identification device 400 may further include a calculation module and a second generation module.

[0119] The calculation module can be used to calculate the current confidence of the current aircraft category according to the current probability distribution; the second generation module can be used to generate an identification record including the current aircraft category and the current confidence.

[0120] The solution provided in this embodiment obtains radar reflection waves, which are formed based on the reflection of radar transmission waves by obstacles. When it is determined that an aircraft exists in a preset airspace based on the radar reflection waves, a current aircraft image of the aircraft is obtained, and the current aircraft image is input into an aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft recognition result includes the current aircraft category, thereby realizing aircraft recognition of the current aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0121] 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.

[0122] 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.

[0123] See also Figure 6 , which shows a functional block diagram of an electronic device 500 provided by an embodiment of the present application, and the electronic device 500 may include one or more of the following components: a memory 510, a processor 520, and one or more applications, wherein the one or more applications may be stored in the memory 510 and configured to be executed by the one or more processors 520, and the one or more applications are configured to execute the method described in the aforementioned method embodiment.

[0124] The memory 510 may include a random access memory (RAM) or a read-only memory (ROM). The memory 510 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 510 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 radar reflection waves, determining the presence of an aircraft, obtaining a current aircraft image, inputting a current aircraft image, obtaining an aircraft recognition result, training a deep learning network model, annotating historical aircraft images, obtaining historical aircraft images, determining an initial training set, transfer learning, obtaining a transfer learning model, augmentation processing, obtaining an augmented training set, inputting an augmented training set, Fine-tuning training, inputting an initial training set, pre-training a deep learning network model, obtaining a pre-trained model, freezing a convolutional layer, obtaining a frozen training model, adding a fully connected layer, pre-processing the current aircraft image, obtaining a pre-processed aircraft image, inputting a pre-processed aircraft image, generating warning information, calculating a current confidence level, generating an identification record, image acquisition, and obtaining a current aircraft image, etc.), instructions for implementing the following various method embodiments, etc. The data storage area can also store data created by the electronic device 500 during use (such as current aircraft images, aircraft identification models, aircraft identification results, historical aircraft images, deep learning network models, historical aircraft types, current aircraft categories, initial training sets, transfer learning models, augmented training sets, pre-trained models, frozen training models, fully connected layers, pre-processed aircraft images, preset aircraft categories, warning information, current probability distribution, current confidence, identification records, radar reflection waves, radar transmission waves, obstacles and aircraft), etc.

[0125] The processor 520 may include one or more processing cores. The processor 520 uses various interfaces and lines to connect various parts of the entire electronic device 500, and executes various functions and processes data of the electronic device 500 by running or executing instructions, programs, code sets or instruction sets stored in the memory 510, and calling data stored in the memory 510. Optionally, the processor 520 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), programmable logic array (PLA), etc. The processor 520 can integrate one or a combination of a central processing unit (CPU), a 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 can be understood that the above-mentioned modem may not be integrated into the processor 520, but may be implemented separately through a communication chip.

[0126] Please refer to Figure 7 , 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 600 stores a program code 610, which can be called by a processor to execute the method described in the above method embodiment.

[0127] The computer readable storage medium 600 can 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 600 includes a non-transitory computer-readable storage medium. The computer readable storage medium 600 has storage space for program code 610 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 610 can be compressed, for example, in an appropriate form.

[0128] Please refer to Figure 8, which shows a structural block diagram of a computer program product 700 provided in an embodiment of the present application. The computer program product 700 includes a computer program / instruction 710, and the computer program / instruction 710 is stored in a computer-readable storage medium of a computer device. When the computer program product 700 is run on a computer device, a processor of the computer device reads the computer program / instruction 710 from the computer-readable storage medium, and the processor executes the computer program / instruction 710, so that the computer device performs the method described in the above method embodiment.

[0129] The solution provided in this embodiment obtains radar reflection waves, which are formed based on the reflection of radar transmission waves by obstacles. When it is determined that an aircraft exists in a preset airspace based on the radar reflection waves, a current aircraft image of the aircraft is obtained, and the current aircraft image is input into an aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are annotated with historical aircraft types. The aircraft recognition result includes the current aircraft category, thereby realizing aircraft recognition of the current aircraft image based on the aircraft recognition model, thereby improving the recognition accuracy and efficiency of aircraft recognition.

[0130] 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 method for aircraft identification, characterized in that: include: Acquiring a radar reflected wave, wherein the radar reflected wave is formed based on the radar transmitted wave being reflected by an obstacle; When it is determined according to the radar reflected wave that an aircraft exists in the preset airspace, acquiring a current aircraft image of the aircraft; The current aircraft image is input into an aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are marked with historical aircraft types. The aircraft recognition result includes the current aircraft category.

2. The aircraft identification method according to claim 1, characterized in that: Before obtaining the radar reflected wave, the aircraft identification method further includes: acquiring the historical aircraft image; determining an initial training set based on the historical aircraft images; The deep learning network model is trained based on the initial training set to obtain the aircraft recognition model.

3. The aircraft identification method according to claim 2, characterized in that: The step of training the deep learning network model based on the initial training set to obtain the aircraft recognition model includes: Performing transfer learning on the deep learning network model based on the initial training set to obtain a transfer learning model; Performing data augmentation processing on the initial training set to obtain an augmented training set; The augmented training set is input into the transfer learning model for fine-tuning training to obtain the aircraft recognition model.

4. The aircraft identification method according to claim 3, characterized in that: The step of performing transfer learning on the deep learning network model based on the initial training set to obtain a transfer learning model includes: Inputting the initial training set into the deep learning network model for pre-training to obtain a pre-trained model; Freezing the convolutional layer of the pre-trained model to obtain a frozen training model; Add a fully connected layer to the top of the frozen training model to obtain the transfer learning model.

5. The aircraft identification method according to claim 1, characterized in that: Before inputting the current aircraft image into the aircraft recognition model to obtain the aircraft recognition result, the aircraft recognition method further includes: Preprocessing the current aircraft image to obtain a preprocessed aircraft image; The step of inputting the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result includes: The pre-processed aircraft image is input into the aircraft recognition model to obtain the aircraft recognition result.

6. The aircraft identification method according to claim 1, characterized in that: Also includes: If the current aircraft category is a preset aircraft category, a warning message is generated.

7. The aircraft identification method according to any one of claims 1 to 6, characterized in that: The aircraft identification result also includes a current probability distribution corresponding to the current aircraft category, and the aircraft identification method further includes: Calculating a current confidence level of the current aircraft category according to the current probability distribution; An identification record including the current aircraft category and the current confidence level is generated.

8. An aircraft identification device, characterized in that: include: A first acquisition module is used to acquire a radar reflected wave, wherein the radar reflected wave is formed based on the radar transmission wave being reflected by an obstacle; A second acquisition module is used to acquire a current aircraft image of the aircraft when it is determined that there is an aircraft in the preset airspace according to the radar reflection wave; The first input module is used to input the current aircraft image into the aircraft recognition model to obtain an aircraft recognition result. The aircraft recognition model is obtained by training a deep learning network model based on historical aircraft images. The historical aircraft images are marked with historical aircraft types. The aircraft recognition result includes the current aircraft category.

9. An electronic device, characterized in that: include: Memory; One or more processors coupled to the memory; One or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by one or more processors, and the one or more application programs are configured to execute the aircraft identification 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, which can be called by a processor to execute the aircraft identification method according to any one of claims 1 to 7.

Images