Embedded information system satellite positioning mode control method

Abstract

The invention provides a satellite positioning mode control method for an embedded information system, relates to the technical field of image processing, and is used for monitoring and management of position determination in the flight process of an unmanned aerial vehicle. The embedded information system satellite positioning mode control method specifically comprises the following steps: acquiring position information of to-be-positioned unmanned aerial vehicle equipment; shooting by using a camera of the to-be-positioned unmanned aerial vehicle equipment to obtain an overlook flight panoramic image; acquiring a high-resolution image corresponding to the position information by using a satellite system; and sending the high-resolution image to an embedded information system in unmanned aerial vehicle equipment to be positioned, wherein the embedded information system determines a positioning mode by using a high-resolution image recognition result. Through the steps, the identification accuracy of the positioning system can be evaluated and corrected, and the flight safety of the unmanned aerial vehicle in an unknown area is ensured.

Description

Technical Field

[0001] The present invention relates to the field of image recognition, and in particular to a satellite positioning mode control method for an embedded information system, which is used for monitoring and managing the position determination during the flight of an unmanned aerial vehicle. Background Art

[0002] As a versatile aerial platform, drones have been widely used in many fields of production and life. With the support of modern technology, drones can perform a variety of tasks in different environments, such as damage assessment after natural disasters, environmental monitoring, wildlife protection, etc. Its ability to be deployed quickly makes drones an ideal choice for preliminary situation assessment in emergency situations. In the agricultural field, drones collect data on soil, crop growth conditions, and pests and diseases through high-definition cameras to help farmers make more accurate decisions. Drones also play an important role in urban planning and construction. Drones are used to monitor construction progress, conduct land surveying and urban space planning. By shooting from high altitudes, drones can provide a faster, safer and lower-cost data collection method than traditional ground measurements. In addition, in the power and oil industries, drones are used to inspect transmission lines and pipelines, which not only improves the efficiency of inspections, but also greatly reduces the risk of personnel working in dangerous environments.

[0003] Despite the ever-expanding application scope and technological advancement of drones, the positioning accuracy of drones during missions remains a key issue. Traditional methods for positioning drones rely mainly on their own global positioning systems (GPS). This system determines the geographic location of drones by receiving signals transmitted from satellites in geosynchronous orbit, but in complex or unknown environments, such as urban areas with tall buildings or mountain jungles, GPS signals may be shielded or interfered with, resulting in inaccurate positioning. In addition, the GPS system provides the absolute geographic location of the drone, and is powerless for relative position information in the environment, such as the location of nearby buildings, ground obstacles or other aircraft. This limits the application of drones in unknown or complex environments because it cannot effectively identify and avoid obstacles that may be encountered, which is crucial to ensuring the flight safety of drones and the successful execution of missions.

[0004] In view of the shortcomings of the prior art, the present invention proposes a satellite positioning mode control method for an embedded information system. The method integrates satellite positioning technology and image recognition technology to provide more accurate location information, and can also identify and process potential risks near the location of the drone in real time, thereby more effectively avoiding risks and improving flight safety. Summary of the invention

[0005] The present invention provides a satellite positioning mode control method for an embedded information system, the method specifically comprising the following steps:

[0006] S1: Obtain the location information of the drone device to be located;

[0007] S2: Using the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view of the flight panoramic image;

[0008] S3: Using a satellite system to capture a high-resolution image corresponding to the location information obtained in step S1;

[0009] S4: Sending the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines a positioning mode using the high-resolution image recognition result.

[0010] The present invention provides an embodiment of the specification and also proposes a satellite positioning mode control device for an embedded information system, the device comprising:

[0011] Position positioning module: The position positioning module obtains the position information of the unmanned aerial vehicle device to be positioned;

[0012] Panoramic image generation module: the panoramic image generation module uses the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view panoramic image;

[0013] Satellite image acquisition module: The satellite image acquisition module uses high-resolution images corresponding to location information acquired during shooting by a satellite system;

[0014] Mode control module: The mode control module sends the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines the positioning mode using the high-resolution image recognition result.

[0015] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned satellite positioning mode control method of an embedded information system is implemented.

[0016] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program implements the above-mentioned embedded information system satellite positioning mode control method.

[0017] Compared with the prior art, the present invention proposes a satellite positioning mode control method for an embedded information system. During the process of UAV positioning, the present invention first uses the positioning system configured by the UAV device itself to perform position positioning. At the same time, based on the position information, a high-resolution image taken by a satellite with the position as the geometric center is obtained, that is, the position information is identified as the geometric center point of the high-resolution image taken by the satellite. By judging the displacement deviation between the UAV position in the actual high-resolution image taken by the satellite and the center point, it is judged whether the current positioning system is accurate, and other positioning systems are replaced if it is inaccurate.

[0018] In the process of high-resolution image recognition, the present invention uses the overhead panoramic view taken by the drone itself for guidance. The panoramic view taken by the drone equipment should be included in the high-resolution image taken by the satellite. The present invention guides by finding similar areas in the panoramic image and the high-resolution image to improve the accuracy of drone recognition. At the same time, the high-resolution image is obtained by taking pictures based on the position information sent by the drone. When the position information is not biased, the drone should be located at the center of the high-resolution image, that is, the probability of the drone appearing near the center of the high-resolution image is higher than the probability of appearing at the edge of the image. Therefore, the present invention establishes a Gaussian distribution function with the geometric center position of the high-resolution image as the midpoint to guide the recognition network to perform key target detection in the central area. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments 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 creative work.

[0020] Figure 1 This is a flow chart of the satellite positioning mode control of the present invention; DETAILED DESCRIPTION

[0021] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0022] The following describes the implementation methods of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, in the absence of conflict, the following embodiments and the features in the embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work belong to the scope of protection of the present application.

[0023] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present application, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspect described herein can be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this device and / or practice this method.

[0024] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the examples can be practiced without these specific details.

[0025] The embodiment of this specification proposes a satellite positioning mode control method for an embedded information system, and the method specifically includes the following steps:

[0026] S1: Obtain the location information of the drone device to be located;

[0027] S2: Using the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view of the flight panoramic image;

[0028] S3: Using a satellite system to capture a high-resolution image corresponding to the location information obtained in step S1;

[0029] S4: Sending the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines a positioning mode using the high-resolution image recognition result.

[0030] Specifically, in one embodiment, the unmanned aerial vehicle device to be positioned should be equipped with a variety of position positioning systems, such as the GPS of the United States, the GLONASS of Russia, the BeiDou of China, and the Galileo system of the European Union. Preferably, the unmanned aerial vehicle device to be positioned is equipped with a global navigation satellite system (GNSS), which can receive signals from multiple satellite systems around the world. The multi-system receiving capability ensures that positioning information can be obtained at any location in the world, especially when a system signal is weak or unavailable, other systems can supplement the positioning information and improve the overall positioning reliability.

[0031] In addition, the drone to be positioned should be equipped with an inertial measurement unit (IMU), which is a combination of accelerometers, gyroscopes, and sometimes magnetometers, used to provide the drone's instant motion status, including speed, direction, and attitude. The IMU uses micro-electromechanical system (MEMS) technology. When GNSS signals are unavailable or unstable, IMU data is crucial and can assist or independently provide position and navigation information.

[0032] The drone device to be located should be equipped with a communication module, which is responsible for exchanging data with the ground control station or Internet cloud service. The data exchange includes flight data, mission control instructions, real-time video stream and location information data. The communication module includes one or more technologies of Wi-Fi, Bluetooth, LTE and 5G. For missions that require long distances or high reliability, LTE and 5G provide stable connections across wide areas and support high data rate transmission, which are very suitable for transmitting high-definition video or large amounts of sensor data. At the same time, these modules should be configured with security protocols to ensure the security and privacy of data transmission.

[0033] The method of obtaining a bird's-eye view of the flight panoramic image by using the camera of the unmanned aerial vehicle device to be positioned specifically includes:

[0034] S2-1: When the location information of the drone device to be located is obtained, the camera of the drone device is started and video recording is started, and the start recording time is recorded;

[0035] S2-2: Record a video of preset length, extract key frames and pre-process the video;

[0036] S2-3: stitching the preprocessed key frames to obtain a bird's-eye view panoramic image.

[0037] The key frame extraction of the video comprises:

[0038] S2-2-1: Initialize the key frame list L and set the first frame F of the video 1 is the first key frame K 1 , and add it to the key frame list L;

[0039] S2-2-2: For the i-th frame F i , calculate F i With key frame K j The moving component MV, the local description difference LD and the global description difference FD and the similarity S between the key frames K j is the key frame most recently added to the key frame list L, 2≤i≤N, 1≤j<i, N is the total number of video frames;

[0040] S2-2-3: Calculate the comprehensive difference D, if the comprehensive difference D is greater than the preset difference threshold δ D And the similarity S is greater than the preset overlap threshold δ S When the i-th frame F i As a potential key frame, update i=i+1, and repeat steps S2-2-2 and S2-2-3 until the number of iterations reaches the preset time window τ or i>N;

[0041] S2-2-4: Within the preset time window τ frames, select the frame with the largest comprehensive difference D among the potential key frames as the key frame K j+1 Add to L, update j=j+1, and repeat steps S2-2-2, S2-2-3 and S2-2-4 until all video frames are traversed and the key frame list L is output.

[0042] The motion component MV is defined as:

[0043]

[0044] Among them, N f is the number of feature points, and The i-th frame F i and the jth key frame K j The coordinates of the nth matching feature point in ;

[0045] The local description difference LD is defined as:

[0046]

[0047] Among them, N p is the number of key points, Dim represents the local description dimension, desc i,n [d] and desc j，n [d] represents the i-th frame F i and the jth key frame K j The d-th dimension of the n-th matching key point in to;

[0048] In one embodiment, the local description may use a SIFT algorithm, where a SIFT descriptor is a 128-dimensional vector that describes image gradient information around a key point;

[0049] The global description difference FD is defined as:

[0050]

[0051] Among them, N B is the number of histogram buckets, hist i [b] and hist j [b] are the i-th frame F i and the jth key frame K j The b-th bucket of the b-th histogram in .

[0052] When acquiring a panoramic image, the present invention first extracts the key frames of the video shot by the drone device. During the video image shot by the present invention, there is no need to control the flight speed, flight direction, etc. of the drone device. In the process of selecting the key frames, the present invention respectively calculates the inter-frame difference using the moving component, local description difference, and global description difference to ensure the computational efficiency of extracting the key frames. At the same time, the inter-frame similarity is calculated to ensure the smoothness of the subsequent panoramic image construction.

[0053] The stitching of the pre-processed key frames to obtain the overhead flight panoramic image specifically includes:

[0054] S2-3-1: For the previous and next key frames K j and K j+1 Perform feature extraction to obtain feature map f j and f j+1 ;

[0055] S2-3-2: For feature map f j and f j+1 Perform feature matching to obtain a related feature description graph;

[0056] S2-3-3: Outputting an affine transformation parameter matrix according to the associated feature description graph using an analytical network;

[0057] S2-3-4: performing image registration according to the affine transformation parameter matrix, and fusing to obtain the i-th panoramic image;

[0058] S2-3-5: Repeat steps S2-3-1 to S2-3-4 until all key frames are traversed and stitched to output a bird's-eye view flight panoramic image.

[0059] The pair of previous and next key frames K j and K j+1 Perform feature extraction to obtain feature map f j and fj+1 , the feature extraction network is composed of 4 residual modules, each of which is designed as:

[0060] O t =ReLU(BN(W 3*3 *ReLU(BN(W 1*1 *I t +b 1*1 ))+b 3*3 )+I t );

[0061] Among them, I t and O t represents the input and output of the tth residual module, W 1*1 and W 3*3 are the convolution kernel weights of 1*1 and 3*3 respectively, b 1*1 and b 3*3 is the corresponding bias term, BN and ReLU are batch normalization and activation functions respectively, the step size and padding in the convolution calculation are both 1, and the input of the first residual module is any key frame K in the key frame list L j , the output of the fourth residual module is the feature map f j ;

[0062] The calculation method of the associated feature description graph is:

[0063]

[0064] Among them, f j and f j+1 The feature map size is w*h*N c , the size of the associated feature description graph Crl is w*h*(w*h), w and h represent the length and width respectively, (w*h) is the number of channels of Crl, N c Represents the feature map f j and f j+1 The number of channels, x, y, z represent the length, width, and channel index of the feature map respectively;

[0065] The resolution network is set as:

[0066] f mid =ReLU(W 2 *(MaxPool(ReLU(W 1 *Crl+b 1 )))+b 2 );

[0067] f affine =FC 2 (Flatten(f mid ));

[0068] Among them, W 1 and W 2 The first and second layer convolution kernel parameters, b 1 and b 2 are the first and second layer bias terms, respectively, and f mid represents the hidden features in the middle of the parsing network, f affine represents the affine transformation parameter matrix, Flatten represents the flattening operation, FC 2 Represents two fully connected mapping operations.

[0069] The present invention first extracts the feature points of the key frames to be spliced ​​during the panoramic image stitching process. Different from the traditional feature point extraction method, the present application constructs a residual module for extracting the feature points of the key frames. The panoramic image of the present invention is spliced ​​from the key frame images, and there is a large difference between the key frame images. Therefore, the feature point extraction method based on the neural network can extract the deep semantic features between the key frames, and obtain the affine transformation parameter matrix after the key point extraction is completed to realize the key frame image stitching.

[0070] In one embodiment, the high-resolution image corresponding to the location information obtained in step S1 is captured by the satellite system. The location information collected in step S1 needs to be transmitted to the ground control center through the communication equipment configured by the unmanned access. The location information includes the longitude, latitude and altitude information of the drone. After receiving the location information, the control center verifies the accuracy and completeness of the location information and performs format conversion adapted to the satellite control system. The control center calculates the shooting adjustment amount of the satellite camera by analyzing the drone location information and satellite status information. The shooting adjustment amount includes the pitch angle and yaw angle of the satellite lens to ensure that the satellite camera can be aligned with the location of the drone. When the ground control center completes the processing, the control information is transmitted to the satellite control system through the communication equipment to adjust the lens angle. Once the satellite adjusts the lens angle and resolution normally, it enters the shooting program. At this stage, the satellite operating system controls the camera to capture high-resolution images according to the control parameters, and transmits the high-resolution image data to the ground control center. After receiving the image data, the ground control center will perform preliminary image processing and image enhancement.

[0071] The method of adjusting the attitude of remote sensing satellites through ground position information is an existing technology and will not be described in detail here. In one embodiment, the process of photographing drone equipment using a satellite system also includes adjusting the resolution of the satellite system camera and fine-tuning the satellite orbit. The above-mentioned resolution adjustment and orbit satellite are both existing technologies and will not be described in detail here. References: "On-orbit autonomous focusing method for agile remote sensing satellites based on lunar imaging", Yin Yanhe, Proceedings of the 8th Annual Conference on High-Resolution Earth Observation, "Single-orbit scheduling method for agile imaging satellites for regional targets", Song Jinyun, Digital Technology and Applications.

[0072] In one embodiment, the embedded information system uses the high-resolution image recognition result to determine the positioning mode, and uses an image fusion network to identify the position of the drone;

[0073] The image fusion network includes a high-resolution image feature extraction module, a multi-scale fusion module, an information guidance module and a positioning module;

[0074] The high-resolution image feature extraction module includes four enhanced residual jump structures, wherein the first enhanced residual jump structure is defined as:

[0075]

[0076] Among them, I H represents a high-resolution image, Represents the output feature map of the first enhanced residual skip structure;

[0077] The second to fourth enhanced residual skip structures are defined as:

[0078]

[0079] in, represents the output feature map of the mth enhanced residual skip structure, 2≤m≤4;

[0080] The multi-scale fusion module includes three parallel branches for feature enhancement, and the input of the three parallel branches is the output of the fourth enhanced residual jump structure. The first and second layers of the three parallel branches are defined as:

[0081]

[0082]

[0083] Among them, 1≤p≤3, and Represent the first and second layer output feature maps of the pth parallel branch, Conv p,1 and Convp,2 They represent the convolution calculations of the first and second layers of the pth parallel branch, respectively, where the convolution kernel sizes of the first to third parallel branches are 3*3, 5*5, and 7*7, respectively;

[0084] The outputs of the three parallel branches are added pixel by pixel to obtain the output of the multi-scale fusion module. M ;

[0085] The input of the information guidance module is the output of the multi-scale fusion module. M , the information guidance module is defined as:

[0086]

[0087]

[0088]

[0089]

[0090] Among them, O inf The output feature map F represents the information guidance module s 、F loc and F img represent location satellite image guidance, location information guidance, and panoramic information guidance, σ, add, and Represents activation function, pixel-by-pixel addition and pixel-by-pixel multiplication respectively Represents the 1*1 convolution operations for query, key, value, and output channel adjustment, respectively. dim represents the number of channels after dimensionality reduction, expand and reshape represent channel expansion and size adjustment respectively, x c and c Indicates O M The center point coordinates σ x and σ y Respectively represent the standard deviation of Gaussian distribution along the horizontal and vertical axes, I par For a bird's-eye view of the flight panorama, max axis＝1 Indicates taking the maximum value by row;

[0091] The output of the information guide module O inf The input is sent to the positioning module composed of a fully connected network to obtain the results of the drone equipment to be located in the satellite image;

[0092] When the distance between the unmanned aerial vehicle device to be located in the satellite image and the center position of the image is less than a preset threshold, the current positioning system is used to continue positioning the unmanned aerial vehicle device; otherwise, other positioning systems are used to continue positioning the unmanned aerial vehicle device.

[0093] During the process of UAV positioning, the present invention first uses the positioning system configured by the UAV equipment itself to perform position positioning. At the same time, based on the position information, a high-resolution image taken by a satellite with the position as the geometric center is obtained, that is, the position information is identified as the geometric center point of the high-resolution image taken by the satellite. By judging the displacement deviation between the UAV position in the actual high-resolution image taken by the satellite and the center point, it is judged whether the current positioning system is accurate, and other positioning systems are replaced if it is inaccurate.

[0094] In the process of high-resolution image recognition, the present invention uses the overhead panoramic view taken by the drone itself for guidance. The panoramic view taken by the drone equipment should be included in the high-resolution image taken by the satellite. The present invention guides by finding similar areas in the panoramic image and the high-resolution image to improve the accuracy of drone recognition. At the same time, the high-resolution image is obtained by taking pictures based on the position information sent by the drone. When the position information is not biased, the drone should be located at the center of the high-resolution image, that is, the probability of the drone appearing near the center of the high-resolution image is higher than the probability of appearing at the edge of the image. Therefore, the present invention establishes a Gaussian distribution function with the geometric center position of the high-resolution image as the midpoint to guide the recognition network to perform key target detection in the central area.

[0095] The embodiment of this specification also proposes a satellite positioning mode control device for an embedded information system, the device comprising:

[0096] Position positioning module: The position positioning module obtains the position information of the unmanned aerial vehicle device to be positioned;

[0097] Panoramic image generation module: the panoramic image generation module uses the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view panoramic image;

[0098] Satellite image acquisition module: The satellite image acquisition module uses high-resolution images corresponding to location information acquired during shooting by a satellite system;

[0099] Mode control module: The mode control module sends the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines the positioning mode using the high-resolution image recognition result.

[0100] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned satellite positioning mode control method of an embedded information system is implemented.

[0101] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program implements the above-mentioned embedded information system satellite positioning mode control method.

[0102] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0103] In this specification, the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments described later, the description is relatively simple, and the relevant parts can be referred to the partial description of the previous embodiments.

[0104] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A satellite positioning mode control method for an embedded information system, characterized in that: The method comprises the following steps: S1: Obtain the location information of the drone device to be located; S2: Using the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view of the flight panoramic image; S3: Using a satellite system to capture a high-resolution image corresponding to the location information obtained in step S1; S4: Sending the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines a positioning mode using the high-resolution image recognition result.

2. The embedded information system satellite positioning mode control method according to claim 1, characterized in that: The method of obtaining a bird's-eye view of the flight panoramic image by using the camera of the unmanned aerial vehicle device to be positioned specifically includes: S2-1: When the location information of the drone device to be located is obtained, the camera of the drone device is started and video recording is started, and the start recording time is recorded; S2-2: Record a video of preset length, and extract and pre-process the video; S2-3: stitching the preprocessed key frames to obtain a bird's-eye view panoramic image.

3. The embedded information system satellite positioning mode control method according to claim 2, characterized in that: The key frame extraction of the video comprises: S2-2-1: Initialize the key frame list L, set the first frame F1 of the video as the first key frame K1, and add it to the key frame list L; S2-2-2: For the i-th frame F i , calculate the motion vector MV, local description difference LD, global description difference FD, and similarity S between F i and the key frame K j . The key frame K j is the key frame newly added to the key frame list L, 2 ≤ i ≤ N, 1 ≤ j < i, and N is the total number of video frames; S2-2-3: Calculate the comprehensive difference D, if the comprehensive difference D is greater than the preset difference threshold δ D And the similarity S is greater than the preset overlap threshold δ S When the i-th frame F i As a potential key frame, update i=i+1, and repeat steps S2-2-2 and S2-2-3 until the number of iterations reaches the preset time window τ or i>N; S2-2-4: In the preset time window τ frames, select the frame with the largest comprehensive difference D among the potential key frames as the key frame K j+1 Add to L, update j=j+1, and repeat steps S2-2-2, S2-2-3 and S2-2-4 until all video frames are traversed and the key frame list L is output.

4. The satellite positioning mode control method of an embedded information system according to claim 3, characterized in that: The stitching of the pre-processed key frames to obtain the overhead flight panoramic image specifically includes: S2-3-1: For the previous and next key frames K j and K j+1 Perform feature extraction to obtain feature map f j and f j+1 ; S2-3-2: For feature map f j and f j+1 Perform feature matching to obtain a related feature description graph; S2-3-3: Outputting an affine transformation parameter matrix according to the associated feature description graph using an analytical network; S2-3-4: performing image registration according to the affine transformation parameter matrix, and fusing to obtain the j-th panoramic image; S2-3-5: Repeat steps S2-3-1 to S2-3-4 until all key frames are traversed and stitched to output a bird's-eye view panoramic image.

5. The satellite positioning mode control method of an embedded information system according to claim 4, characterized in that: The embedded information system uses the high-resolution image recognition result to determine the positioning mode, and adopts an image fusion network to identify the position of the drone; the image fusion network includes a high-resolution image feature extraction module, a multi-scale fusion module, an information guidance module and a positioning module; the high-resolution image feature extraction module includes 4 enhanced residual jump structures, wherein the first enhanced residual jump structure is defined as: Among them, I H represents a high-resolution image, Represents the output feature map of the first enhanced residual skip structure; The second to fourth enhanced residual skip structures are defined as: in, represents the output feature map of the mth enhanced residual skip structure, 2≤m≤4.

6. The satellite positioning mode control method of an embedded information system according to claim 5, characterized in that: The multi-scale fusion module includes three parallel branches for feature enhancement, and the input of the three parallel branches is the output of the fourth enhanced residual jump structure. The first and second layers of the three parallel branches are defined as: Among them, 1≤p≤3, and Represent the first and second layer output feature maps of the pth parallel branch, Conv p,1 and Conv p,2 They represent the convolution calculations of the first and second layers of the pth parallel branch, respectively, where the convolution kernel sizes of the first to third parallel branches are 3*3, 5*5, and 7*7, respectively; The outputs of the three parallel branches are added pixel by pixel to obtain the output of the multi-scale fusion module. M .

7. The satellite positioning mode control method of an embedded information system according to claim 6, characterized in that: The input of the information guidance module is the output of the multi-scale fusion module. M , the information guidance module is defined as: Among them, O inf The output feature map F represents the information guidance module s 、F loc and F img represent location satellite image guidance, location information guidance, and panoramic information guidance, σ, add, and Represents activation function, pixel-by-pixel addition and pixel-by-pixel multiplication respectively Represents the 1*1 convolution operations for query, key, value, and output channel adjustment, respectively. dim represents the number of channels after dimensionality reduction, expand and reshape represent channel expansion and size adjustment respectively, x c and c Indicates O M The center point coordinates σ x and σ y Respectively represent the standard deviation of Gaussian distribution along the horizontal and vertical axes, I par For a bird's-eye view of the flight panorama, max axis＝1 Indicates taking the maximum value by row; The output of the information guide module O inf The input is sent to the positioning module composed of a fully connected network to obtain the results of the drone equipment to be located in the satellite image.

8. An embedded information system satellite positioning mode control device, characterized in that: The device is used to execute the satellite positioning mode control method of the embedded information system according to any one of claims 1 to 7, and the system comprises: Position positioning module: The position positioning module obtains the position information of the unmanned aerial vehicle device to be positioned; Panoramic image generation module: the panoramic image generation module uses the camera of the unmanned aerial vehicle device to be positioned to obtain a bird's-eye view panoramic image; Satellite image acquisition module: The satellite image acquisition module uses high-resolution images corresponding to location information acquired during shooting by a satellite system; Mode control module: The mode control module sends the high-resolution image to an embedded information system in the unmanned aerial vehicle device to be positioned, and the embedded information system determines the positioning mode using the high-resolution image recognition result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a satellite positioning mode control method for an embedded information system as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements a satellite positioning mode control method for an embedded information system according to any one of claims 1 to 7.

Images