Airtight narrow space unmanned reconnaissance system

Abstract

The invention belongs to the technical field of communication, and particularly discloses a closed narrow space unmanned reconnaissance system, which comprises a control end, a 5G base station and an unmanned carrier, and is characterized in that the unmanned carrier is provided with a communication module, an image acquisition module, an encryption module and a data packet construction module; the 5G base station serves as a data transfer processing station between the control end and the unmanned carrier, decrypts, identifies and stores a data packet structure body obtained through a data channel, and then sends the data packet structure body to the control end; and the control end sends an instruction to the unmanned carrier based on the identified data packet structure body to complete reconnaissance. According to the invention, different data packet structural bodies are constructed by establishing a special communication channel, so that higher precision is improved for data processing, and battle requirements in a specific environment are met.

Description

Technical Field

[0001] The invention belongs to the field of communication technology, and in particular relates to an unmanned reconnaissance system for a closed and narrow space. Background Art

[0002] During field operations, it is inevitable to encounter special enclosed and narrow environments such as tunnels, pits, and electronic shielding areas. When it is necessary to conduct reconnaissance in these areas, due to the lack of understanding of the internal terrain of the area and the possible inability to communicate with the outside world, rash entry may pose a safety hazard.

[0003] Therefore, providing an unmanned reconnaissance system for use in confined and narrow spaces is a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the prior art that the internal conditions of a special environment cannot be obtained and that the external environment cannot be communicated with, and to provide an unmanned reconnaissance system and a reconnaissance method for a confined and narrow space.

[0005] In a first aspect of the present invention, there is provided an unmanned reconnaissance system for a confined and narrow space, comprising a control terminal, a 5G base station and an unmanned vehicle, wherein the unmanned vehicle is provided with a communication module, an image acquisition module, an encryption module, a data packet construction module and a 3D model construction module; the communication module is connected to the 5G base station by a leaky cable wired communication connection and a 5G wireless communication connection to form a wired channel and a wireless channel; the image acquisition module is used to collect data of the area to be reconnaissanced; the data packet construction module is used to construct the collected data into a data packet structure; the encryption module is used to encrypt the data packet structure; the 3D model construction module constructs a three-dimensional model based on the collected data;

[0006] The wired channel and the wireless channel respectively establish MIMO channel models, the MIMO channel model includes at least three channels, namely a data channel, a control channel and an RTS signal channel, and the three channels of the wireless transmission all use 4.9 GHz and 700 MHz dual-frequency electromagnetic waves;

[0007] The 5G base station serves as a data transfer processing station between the control end and the unmanned vehicle, decrypts, identifies and stores the data packet structure obtained through the data channel and then sends it to the control end;

[0008] The control end sends instructions to the unmanned vehicle based on the identified data packet structure to complete the reconnaissance.

[0009] A further solution is that the data packet structure includes a data number, a data type, a collection time and a priority;

[0010] The collection time is the number of seconds from 0:0:0:0 on January 1, 1970 to the collection time.

[0011] A further solution is that the encrypted data packet structure is decrypted through the IP of the control end; the recognition of the data packet structure includes but is not limited to extracting features from the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category, and the storage of the data packet structure includes sorting the identified data packet structure based on the weighting of priority and acquisition time and storing it on the control end.

[0012] A further solution is that the category includes at least video data and audio data, and the unmanned vehicle further includes a data processing module, and the data module has a built-in computer vision algorithm, a speech recognition algorithm, and a natural language processing algorithm;

[0013] Processing video information through computer vision algorithms: Recognition and tracking of reconnaissance targets are achieved based on target detection algorithms and target tracking algorithms. Counting and statistical analysis are performed based on the recognition results to calculate a modeling subset for 3D modeling. The modeling subset is a valid image set for each frame in the video data.

[0014] Process the audio data through the speech recognition algorithm: convert audio to text based on the speech recognition algorithm and simultaneous interpretation function; recognize the voiceprint of the person based on the voiceprint recognition algorithm and combine it with the video face recognition algorithm to confirm the identity of the speaker;

[0015] The text information is deduplicated, normalized, and disambiguated through the natural language processing algorithm: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

[0016] A further solution is that the 5G base station sends instructions to the unmanned vehicle based on the stored data packet structure recognition results, and each time the unmanned vehicle executes an instruction, the control end deletes the data packet structure corresponding to the instruction.

[0017] A second aspect of the present invention provides an unmanned reconnaissance method for a confined and narrow space, which uses the above-mentioned reconnaissance system to enter the area to be reconnaissanced through an unmanned vehicle and establish communication with a control end; the method comprises the following steps:

[0018] Step 1: Establishment of communication link;

[0019] Connect the unmanned vehicle to the 5G base station using a combination of wired and wireless transmission;

[0020] When the unmanned vehicle enters the area to be reconnaissanced, the unmanned vehicle immediately interacts with the 5G base station through signaling, and establishes MIMO channel models for wired transmission and wireless transmission respectively for communication between the unmanned vehicle and the 5G base station. The MIMO channel model includes at least three channels, namely, a data channel, a control channel and an RTS signal channel. The three channels of the wireless transmission all use 4.9GHz and 700MHz dual-frequency electromagnetic waves.

[0021] The unmanned vehicle obtains data of the area to be surveyed, and transmits it to the 5G base station via wired or wireless transmission in the form of an encrypted data packet structure according to the RTS / CTS protocol;

[0022] Step 2: Obtain the data packet structure of the area to be detected and perform target identification in real time;

[0023] The 5G base station obtains the data packet structure through the data channel, decrypts, identifies and stores the data packet structure, and then sends it to the control end;

[0024] Step 3: The control end remotely controls the unmanned vehicle;

[0025] The control end controls the unmanned vehicle based on the identified data packet structure.

[0026] A further solution is that the data packet structure includes a data number, a data type, a collection time and a priority;

[0027] The collection time is the number of seconds from 0:0:0:0 on January 1, 1970 to the collection time.

[0028] A further solution is that in step 2, the encrypted data packet structure is decrypted through the IP of the control end; the recognition of the data packet structure includes but is not limited to extracting features from the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category; the storage of the data packet structure includes sorting the identified data packet structure based on the weighting of priority and acquisition time and storing it on the control end.

[0029] A further solution is that the category includes at least video data and audio data, and the identification of the data packet structure further includes:

[0030] Processing video information through computer vision algorithms: Recognition and tracking of reconnaissance targets are achieved based on target detection algorithms and target tracking algorithms. Counting and statistical analysis are performed based on the recognition results to calculate a modeling subset for 3D modeling. The modeling subset is a valid image set for each frame in the video data.

[0031] Process the audio data through the speech recognition algorithm: convert audio to text based on the speech recognition algorithm and simultaneous interpretation function; recognize the voiceprint of the person based on the voiceprint recognition algorithm and combine it with the video face recognition algorithm to confirm the identity of the speaker;

[0032] The text information is deduplicated, normalized, and disambiguated through natural language processing algorithms: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

[0033] A further solution is that the 5G base station sends instructions to the unmanned vehicle based on the stored data packet structure recognition results, and each time the unmanned vehicle executes an instruction, the control end deletes the data packet structure corresponding to the instruction.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention improves the accuracy of data processing by establishing dedicated communication channels and constructing different data packet structures, meets the combat requirements in specific environments, ensures the synchronous interaction of business, measurement and control, and command and control information, and meets the message throughput requirements and anti-interference performance requirements.

[0036] The present invention uses the "leaky cable + 5G" transmission technology to enable each node of the unmanned system data link to automatically generate, process, and exchange tactical information in accordance with the operation control and timing regulations of the communication protocol, establish communication links, and form a communication network with a certain topology structure to meet the real-time and reliable communication needs of combat missions and achieve combat missions; through the 3D acquisition equipment carried by the unmanned vehicle, the environment can be modeled to obtain the scene distribution; through AI intelligent data analysis, a large amount of important information can be obtained from the original data without manual intervention, saving a lot of information extraction time. The system will eventually display the extracted available information to the command hall so that experts can make decisions, which is beneficial to combat. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The following drawings are only used to illustrate and explain the present invention, and are not used to limit the scope of the present invention, wherein:

[0038] Figure 1 : Schematic diagram of the connection principle of the system of the present invention;

[0039] Figure 2 : Schematic diagram of the process of the present invention;

[0040] In the picture: 1. Unmanned vehicle; 2. 5G base station; 3. Control terminal. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution, design method and advantages of the present invention clearer, the present invention is further described in detail by specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0042] Example 1

[0043] like Figure 1 As shown, the present invention provides an unmanned reconnaissance system for a confined and narrow space, including a control terminal 3, a 5G base station 2 and an unmanned vehicle 1, wherein the unmanned vehicle 1 can be a robot, a robot dog, a drone, etc. Since the robot dog is small in size and flexible in action, the robot dog is taken as an example in this embodiment; the robot dog carries an optical fiber coil and enters special environments such as tunnels, pits, and electronic shielding areas to conduct reconnaissance, complete video collection, real-time map modeling, and transmit it to the 5G base station 2 and the control terminal 3. Through artificial intelligence target detection, the robot dog completes the discovery of the enemy and can complete tasks such as bomb throwing according to the instructions of the command terminal.

[0044] The present invention is provided with a communication module, an image acquisition module, an encryption module, a data packet construction module, and a 3D model construction module on the robot dog; the communication module is connected to the 5G base station 2 by a leakage cable wired communication connection and a 5G wireless communication connection to form a wired channel and a wireless channel; the image acquisition module is used to collect data of the area to be surveyed; the data packet construction module is used to construct the collected data into a data packet structure; the encryption module is used to encrypt the data packet structure; the 3D model construction module constructs a three-dimensional model based on the collected data.

[0045] MIMO channel models are established for the wired channel and the wireless channel respectively. The MIMO channel model includes at least three channels, namely, a data channel, a control channel and an RTS signal channel. The three channels of wireless transmission all use 4.9GHz and 700MHz dual-frequency electromagnetic waves; the data channel is a dedicated time slot for the unmanned vehicle end to transmit the data packet structure to the 5G base station 2, the control channel is a dedicated time slot for the 5G base station 2 to forward the control terminal 3 instructions to the unmanned vehicle end, and the RTS signal channel is the remaining time slots, which ensures that the unmanned vehicle end and the 5G base station 2 establish a communication connection in real time. In this embodiment In the process, wired transmission and wireless transmission communicate according to the RTS / CTS protocol, which can effectively avoid conflicts. For example, first, the robot dog sends an RTS signal to the 5G base station 2 through wireless transmission, indicating that the robot dog wants to send data to the 5G base station 2. After receiving the RTS, the 5G base station 2 sends a CTS signal to the robot dog through wireless transmission, indicating that it is ready and the robot dog can send. At the same time, the wired transmission path of the robot dog suspends sending data to the 5G base station 2. It should be noted that the premise of this process is that the robot dog maintains wireless communication with the 5G base station 2 and the communication is good. When the area to be investigated cannot communicate normally with the 5G base station 2, wired transmission is required as the first choice. At this time, the robot dog sends an RTS signal to the 5G base station 2 through wired transmission, indicating that the robot dog wants to send data to the 5G base station 2. After receiving the RTS, the 5G base station 2 sends a CTS signal to the robot dog through wired transmission, indicating that it is ready and the robot dog can send. At the same time, the wireless transmission path of the robot dog suspends sending data to the 5G base station 2, and completes the free switching of the "leaky cable + 5G" transmission technology through the communication protocol, ensuring uninterrupted communication and improving communication stability.

[0046] 5G base station 2 acts as a data transfer processing station between control terminal 3 and the robot dog, decrypts and identifies the data packet structure obtained through the data channel and sends it to control terminal 3;

[0047] The control terminal 3 sends instructions to the robot dog based on the identified data packet structure to complete the reconnaissance.

[0048] In the above, the data packet structure includes data number, data type, collection time and priority; as shown in the following table:

[0049] serial number Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ type video Audio picture Audio video time 1651911056 1651911084 1651911096 1651911108 1651911120 Priority 1 2 3 2 1

[0050] The collection time is the number of seconds from 0:0:0:0 on January 1, 1970 to the collection time.

[0051] The encryption module encrypts the data collected by the image acquisition module. In order to decrypt quickly and accurately, in this embodiment, the decryption key is the IP of the control terminal 3. When the encrypted data is transmitted to the 5G base station 2, the 5G base station 2 obtains the IP of the command terminal, and can quickly and accurately decrypt the encrypted file; the recognition of the data packet structure includes extracting features of the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category. The storage of the data packet structure includes sorting the identified data packet structure based on the weight of priority and acquisition time and storing it in the control terminal 3. The category includes at least video data and audio data. In view of the unreliable network transmission in a strong confrontation environment, the status quo of the multi-link and incomplete information situation of the unmanned reconnaissance and attack system, and the precise control under restricted conditions, the mapping of tactical and other messages to grammatical and semantic information is studied, so that the unmanned combat units can form a unified situation understanding, decision-making control and tactical coordination, thereby realizing "machine mutual recognition", forming a unified situation understanding and combat intent, and realizing consistent intelligent interaction between unmanned system units, platforms, and people. Accelerating the link from terminal perception to strike action is also the key to realizing manned / unmanned collaboration. Therefore, the robot dog must also include a data processing module, which is composed of a core algorithm server and an edge computing device. The core algorithm server provides powerful computing power, and the edge computing provides a convenient data acquisition function, which is then sent to the server for data integration. The operation of the entire system is completed through the data interaction between them. The data processing module has built-in computer vision algorithms, speech recognition algorithms, and natural language processing algorithms;

[0052] The video information is processed by computer vision algorithms: the reconnaissance targets (including people, equipment, etc.) are recognized and tracked based on the target detection algorithm (YOLOv5) and the target tracking algorithm (Deep Sort). The number of enemy troops and the enemy firepower can be calculated based on the recognition results. The modeling subset is obtained based on the data collected by the image acquisition module through the 3D model construction module for 3D modeling. The modeling subset is a valid image set for each frame in the video data.

[0053] Process the audio data through the speech recognition algorithm: convert audio to text based on the speech recognition algorithm and simultaneous interpretation function; recognize the voiceprint of the person based on the voiceprint recognition algorithm and combine it with the video face recognition algorithm to confirm the identity of the speaker;

[0054] The text information is deduplicated, normalized, and disambiguated through the natural language processing algorithm: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

[0055] Since the unmanned vehicle end will send data to the control end 3 in real time, in order to avoid storing too much data, the control end 3 will delete the data packet structure according to a certain strategy. For example, the control end 3 sends an instruction to the robot dog based on the stored data packet structure recognition result. Every time the robot dog executes an instruction, the control end 3 deletes the data packet structure corresponding to the instruction.

[0056] Example 2

[0057] The present application provides an unmanned reconnaissance method for a confined and narrow space, using the above-mentioned reconnaissance system, an unmanned vehicle 1 enters the area to be reconnaissanced, and establishes communication with a control terminal 3; Figure 2 As shown, the following steps are included:

[0058] Step 1: Establishment of communication link;

[0059] The unmanned vehicle 1 is connected to the 5G base station 2 by adopting a fusion mode of wired and wireless transmission;

[0060] When the unmanned vehicle 1 enters the area to be reconnaissanced, the unmanned vehicle 1 immediately performs signaling interaction with the 5G base station 2, and establishes MIMO channel models for wired transmission and wireless transmission respectively for communication between the unmanned vehicle 1 and the 5G base station 2. The MIMO channel model includes at least three channels, namely, a data channel, a control channel and an RTS signal channel. The three channels of the wireless transmission all use 4.9GHz and 700MHz dual-frequency electromagnetic waves;

[0061] The unmanned vehicle 1 obtains the data of the area to be surveyed, and transmits it to the 5G base station 2 via wired or wireless transmission in the form of an encrypted data packet structure according to the RTS / CTS protocol;

[0062] Step 2: Obtain the data packet structure of the area to be surveyed and perform target identification in real time; the data packet structure includes data number, data type, collection time and priority;

[0063] The 5G base station 2 obtains the data packet structure through the data channel, decrypts, identifies and stores the data packet structure, and then sends it to the control terminal 3; specifically, the encrypted data packet structure is decrypted through the IP of the control terminal 3; the identification of the data packet structure includes but is not limited to extracting features from the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category, and the storage of the data packet structure includes sorting the identified data packet structure based on the weight of priority and acquisition time and storing it in the control terminal 3;

[0064] The categories include at least video data and audio data, and the identification of the data packet structure also includes:

[0065] Processing video information through computer vision algorithms: Recognition and tracking of reconnaissance targets are achieved based on target detection algorithms and target tracking algorithms. Counting and statistical analysis are performed based on the recognition results to calculate a modeling subset for 3D modeling. The modeling subset is a valid image set for each frame in the video data.

[0066] Processing audio data through speech recognition algorithms: Converting audio to text based on speech recognition algorithms and simultaneous interpretation functions; Recognizing people’s voiceprints based on voiceprint recognition algorithms and combining them with video face recognition algorithms to confirm the speaker’s identity;

[0067] The text information is deduplicated, normalized, and disambiguated through natural language processing algorithms: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

[0068] Step 3, the control terminal 3 remotely controls the unmanned vehicle 1;

[0069] The control end 3 sends instructions to the unmanned vehicle 1 based on the identified data packet structure. Each time the unmanned vehicle 1 executes an instruction, the control end 3 deletes the data packet structure corresponding to the instruction.

[0070] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. An unmanned reconnaissance system for confined and narrow spaces, comprising a control terminal, a 5G base station and an unmanned vehicle, characterized in that: The unmanned vehicle is provided with a communication module, an image acquisition module, an encryption module, a data packet construction module, and a 3D model construction module; the communication module is connected to the 5G base station through a leaky cable wired communication connection and a 5G wireless communication connection to form a wired channel and a wireless channel; the image acquisition module is used to collect data of the area to be surveyed; the data packet construction module is used to construct the collected data into a data packet structure; the encryption module is used to encrypt the data packet structure; the 3D model construction module constructs a three-dimensional model based on the collected data; The wired channel and the wireless channel respectively establish MIMO channel models, the MIMO channel model includes at least three channels, namely a data channel, a control channel and an RTS signal channel, and the three channels of the wireless transmission all use 4.9 GHz and 700 MHz dual-frequency electromagnetic waves; The 5G base station serves as a data transfer processing station between the control end and the unmanned vehicle, decrypts, identifies and stores the data packet structure obtained through the data channel and then sends it to the control end; The control end sends instructions to the unmanned vehicle based on the identified data packet structure to complete the reconnaissance.

2. The unmanned reconnaissance system for confined and narrow spaces according to claim 1, characterized in that: The data packet structure includes data number, data type, collection time and priority; The collection time is the number of seconds from 0:0:0:0 on January 1, 1970 to the collection time.

3. The unmanned reconnaissance system for confined and narrow spaces according to claim 2 is characterized in that: The encrypted data packet structure is decrypted through the IP of the control end; the recognition of the data packet structure includes but is not limited to extracting features from the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category, and the storage of the data packet structure includes sorting the identified data packet structure based on the weighting of priority and acquisition time and storing it on the control end.

4. The unmanned reconnaissance system for confined and narrow spaces according to claim 3 is characterized in that: The categories include at least video data and audio data, and the unmanned vehicle further includes a data processing module, and the data module has built-in computer vision algorithms, speech recognition algorithms, and natural language processing algorithms; Processing video information through computer vision algorithms: Recognition and tracking of reconnaissance targets are achieved based on target detection algorithms and target tracking algorithms. Counting and statistical analysis are performed based on the recognition results to calculate a modeling subset for 3D modeling. The modeling subset is a valid image set for each frame in the video data. Processing audio data through speech recognition algorithms: Converting audio to text based on speech recognition algorithms and simultaneous interpretation functions; Based on the voiceprint recognition algorithm, the voiceprint of the person is recognized and combined with the video face recognition algorithm to confirm the identity of the speaker; The text information is deduplicated, normalized, and disambiguated through natural language processing algorithms: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

5. The unmanned reconnaissance system for confined and narrow spaces according to claim 4 is characterized in that: The 5G base station sends instructions to the unmanned vehicle based on the stored data packet structure recognition results. Every time the unmanned vehicle executes an instruction, the control end deletes the data packet structure corresponding to the instruction.

6. An unmanned reconnaissance method for a confined and narrow space, characterized in that: Using the reconnaissance system described in any one of claims 1 to 5, entering the area to be reconnaissance through an unmanned vehicle and establishing communication with the control end; comprising the following steps: Step 1: Establishment of communication link; Connect the unmanned vehicle to the 5G base station using a combination of wired and wireless transmission; When the unmanned vehicle enters the area to be reconnaissanced, the unmanned vehicle immediately interacts with the 5G base station through signaling, and establishes MIMO channel models for wired transmission and wireless transmission respectively for communication between the unmanned vehicle and the 5G base station. The MIMO channel model includes at least three channels, namely, a data channel, a control channel and an RTS signal channel. The three channels of the wireless transmission all use 4.9GHz and 700MHz dual-frequency electromagnetic waves. The unmanned vehicle obtains data of the area to be surveyed, and transmits it to the 5G base station via wired or wireless transmission in the form of an encrypted data packet structure according to the RTS / CTS protocol; Step 2: Obtain the data packet structure of the area to be detected and perform target identification in real time; The 5G base station obtains the data packet structure through the data channel, decrypts, identifies and stores the data packet structure, and then sends it to the control end; Step 3: The control end remotely controls the unmanned vehicle; The control end sends instructions to the unmanned vehicle based on the identified data packet structure to complete the reconnaissance.

7. The unmanned reconnaissance method for a confined and narrow space according to claim 6, characterized in that: The data packet structure includes data number, data type, collection time and priority; The collection time is the number of seconds from 0:0:0:0 on January 1, 1970 to the collection time.

8. The unmanned reconnaissance method for a confined and narrow space according to claim 7, characterized in that: In step 2, the encrypted data packet structure is decrypted through the IP of the control end; the identification of the data packet structure includes but is not limited to extracting features from the input data through a feature extraction model to obtain corresponding feature data, and then classifying the feature data through a classification model to obtain the corresponding category, and the storage of the data packet structure includes sorting the identified data packet structure based on the weighting of priority and acquisition time and storing it on the control end.

9. The unmanned reconnaissance method for a confined and narrow space according to claim 8, characterized in that: The categories include at least video data and audio data, and the identification of the data packet structure also includes: Processing video information through computer vision algorithms: Recognition and tracking of reconnaissance targets are achieved based on target detection algorithms and target tracking algorithms. Counting and statistical analysis are performed based on the recognition results to calculate a modeling subset for 3D modeling. The modeling subset is a valid image set for each frame in the video data. Process the audio data through the speech recognition algorithm: convert audio to text based on the speech recognition algorithm and simultaneous interpretation function; recognize the voiceprint of the person based on the voiceprint recognition algorithm and combine it with the video face recognition algorithm to confirm the identity of the speaker; The text information is deduplicated, normalized, and disambiguated through natural language processing algorithms: the text data information obtained based on speech recognition is translated into a recognizable language using machine translation, and the summary information of the text is extracted based on the summary extraction algorithm; the main keywords in the text are extracted based on the named entity recognition algorithm.

10. The unmanned reconnaissance method for a confined and narrow space according to claim 9, characterized in that: The 5G base station sends instructions to the unmanned vehicle based on the stored data packet structure recognition results. Every time the unmanned vehicle executes an instruction, the control end deletes the data packet structure corresponding to the instruction.

Images