Implementation method of optic nerve link
Through the implementation method of visual neural links, the multi-level data processing and standardized interaction mechanism are used to extract and transmit necessary visual information, solving the problem of wireless communication bandwidth bottleneck, improving the efficiency and reliability of information transmission, and supporting battlefield decision-making.
Patent Information
- Application Number
- CN202411101399.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-29
AI Technical Summary
Wireless communication bandwidth is subject to the coordination problem of perceived data growth. Especially under high confrontation conditions, the bandwidth pressure and low efficiency of data link transmission video images are high, forming a communication bottleneck, making it difficult to meet the needs of information transmission under the existing communication bandwidth limitations.
Designing the implementation method of visual neural links, including visual information interaction architecture design, visual link interaction standards, visual information extraction and application verification environment, extract and transmit necessary visual information through multi-level data processing and standardized interaction mechanisms, and reduce communication Bandwidth pressure, deep convolutional neural network and other technologies are used to improve information processing accuracy and efficiency.
It effectively alleviates the bottleneck of communication bandwidth, improves the reliability and efficiency of information transmission, ensures timely transmission and high accuracy of battlefield situation information, supports tactical decision-making, reduces manual errors and data volume, and improves the real-time and reliability of the system. .
Smart Images

Figure CN119052418A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of visual neural technology, and in particular to a method for realizing a visual neural link. Background Art
[0002] The value of data is to provide it to the object in need, and data needs to be transmitted from the generating end to the application end. The basic function of communication is to move data from one end to another, and the size of the bandwidth determines the data throughput rate. Radio communication faces a problem, which is the coordination problem between the limited communication bandwidth and the unrestricted growth of perception data.
[0003] In principle, the wireless communication bandwidth has the boundary described by Shannon's theorem, and it is affected by the environment. The improvement of sensor performance can easily lead to a doubling of data, but it is difficult to increase the communication bandwidth after reaching the extreme value. Therefore, communication seems to have become a "barrier" from the data generation end to the application end. For sensors on unmanned platforms, other platforms hope to obtain all its data and the information it contains in a timely manner, and the communication link is the bottleneck that restricts this desire. Especially under high confrontation conditions, the bandwidth pressure of data link transmission of video images is greater and the efficiency is lower. In order to overcome this bottleneck problem, we provide a method for implementing a visual neural link. Summary of the invention
[0004] In view of the above-mentioned shortcomings of the prior art, the first purpose of the present invention is to provide a method for implementing a visual neural link to solve the problems in the above-mentioned background technology.
[0005] To achieve the above object, the present invention provides the following technical solutions: A method for implementing a visual neural link comprises the following steps: S1, visual information interaction architecture design; S2, visual link interaction standard; S3, visual information extraction; S4, visual information application verification environment; Among them: Step S1, visual information interaction architecture design involves designing and planning the overall architecture of the visual information system, including determining the functions, interfaces and interaction methods of each component in the system; Step S2, visual link interaction standards focus on formulating standards and protocols for visual information interaction, including defining the format of data transmission, communication protocols, and data security requirements to ensure visual information interaction between various systems or devices; Step S3, visual information extraction is to extract useful visual information from raw data or input, involving image processing, feature extraction, target recognition and other technologies; Step S4, visual information application verification environment is to verify the designed visual information in actual application scenarios, including developing a test environment, executing experiments and evaluating system performance.
[0006] The present invention is further configured as follows: in step S1: S1.1, Perception layer data collection and storage: Collect raw data through a variety of sensors (such as infrared, SAR, photoelectric) to ensure that diverse unstructured data is obtained; S1.2, Visual perception data processing: According to the data types of different sensors, use the corresponding data processing mechanism to perform preliminary processing on the raw data to generate initial visual information; S1.3, Information complementary layer processing: Multi-level processing of the initially generated visual information, including information fusion, association verification and screening, to ensure the accuracy and credibility of the information, and encapsulate the information into an unstructured text format that is easy to process; S1.4, message generator processing: The unstructured text is processed by the text fusion parser to parse out key information (such as target information or event information), and combined with data from other platforms to perform information fusion and optimization to generate more accurate visual information; S1.5, Visual information encoding and filling: Match the generated visual information with the message standard platform encoding library, and fill the encoding into the corresponding fields of the relevant message to complete the conversion from unstructured data to distributable structured visual messages.
[0007] The present invention is further configured as follows: in step S2: S2.1. Visual information representation format design: Design a set of unified encoding format rules and data element dictionaries so that visual sensors and control systems can interact efficiently under the same set of standards; S2.2, Visual interaction coding format rules: Design the visual message interaction format between sensors and control systems according to different task requirements; S2.3, Visual message dictionary design: Create and improve the visual message dictionary and determine the specific value range of each field attribute; S2.4, Visual Link Interaction Protocol Development: Design the interaction protocol between the visual sensor and the controller, and clarify the protocol sub-items in different task processes; S2.5, Protocol Interaction Flow Diagram: Provide intuitive reference and guidance by drawing and explaining the commonly used visual link interaction protocol flow chart.
[0008] The present invention is further configured as follows: in step S3: S3.1. Direct extraction architecture design: Analyze the visual information required for the task and build the corresponding visual information extraction framework; S3.2, Direct extraction process design: Design a logically rigorous extraction process, using advanced methods such as deep convolutional neural networks to ensure the generation and iterative fine-tuning of candidate bounding boxes; S3.3, indirect extraction reasoning process: using the extracted visual information, through the combination of sensor parameters, platform parameters and image parameters, other required visual information is inferred to expand the scope of information acquisition; S3.4, Estimation error analysis: Analyze in detail the possible sources of errors in the estimation process and clarify the impact of the errors on the final visual information; S3.5. Schematic diagram: The calculation process and results are presented through schematic diagrams, which intuitively explain the calculation process and accuracy-dependent factors.
[0009] The present invention is further configured as follows: in step S4: S4.1. Design of application effect feedback mechanism: Design an evaluation and feedback mechanism for the application effect of visual information, and adjust the visual information extraction method based on the feedback; S4.2. Laboratory environment simulation verification: simulate small target detection tasks in a laboratory environment, use deep learning technology to discover and identify targets, and verify the application effect of visual information; S4.3, data processing forward realization: by realizing the forward shift of image data processing, the timeliness of data processing is enhanced, the efficiency of information flow is improved, and the timely processing and transmission of visual information is ensured; S4.4, simulation platform verification: the processed visual information is transmitted to the simulation platform, and the task is adjusted through multiple feedbacks to optimize the visual information; S4.5. Situation map generation and verification: Generate high-precision situation maps, evaluate and verify the effectiveness of the visual information application effect feedback mechanism.
[0010] The present invention is further configured as follows: the visual information application effect feedback is an evaluation of the effect of visual information use, which is fed back to the visual sensor. The visual sensor adjusts the transformation information extraction method based on the feedback information to provide visual information that meets the needs.
[0011] Beneficial Effects Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects: 1. The present invention is realized by forwarding data processing and simplifying information. First, the visual neural link forwards data processing and distributes concise information on the link, avoiding the need for lossless transmission of original data and reducing the pressure on communication bandwidth. For example, in high-resolution video images captured by drones, the visual neural link only extracts and transmits necessary visual information rather than the entire video data. This method not only reduces the amount of transmitted data and improves communication efficiency, but also meets user needs under the existing communication bandwidth limitations. Secondly, by performing preliminary processing at the data generation end, extracting key information and performing necessary compression and encoding, and then transmitting it to the application end, the transmission of information is faster and more efficient, ensuring the timeliness of information. Under high confrontation conditions, this method avoids the problem of excessive bandwidth pressure and improves the reliability and efficiency of information transmission. By forwarding data processing and simplifying information, the present invention effectively alleviates the bottleneck problem of communication bandwidth, ensures the timely transmission of battlefield situation information, and provides strong support for tactical decision-making.
[0012] 2. The present invention uses advanced computer vision and machine learning technologies, and the system can automatically analyze and process sensor data, extract key information and generate high-value visual messages. This automatic processing not only improves the accuracy and efficiency of data processing, but also avoids misjudgments and omissions caused by human participation. Secondly, intelligent processing technology analyzes and screens sensor data at the sensor data generation stage, and only transmits valuable information, reducing the burden of wireless communication, thereby greatly reducing the amount of data transmitted, reducing the load on the communication link, and also improving the speed and reliability of information transmission. In addition, autonomous data processing technology liberates a large amount of human resources and reduces dependence on manual interpretation. The efficiency of machine data processing is much higher than that of manual processing, and the stability of processing accuracy can be maintained. This not only improves the real-time and effectiveness of data processing, but also enables unmanned platforms to operate efficiently in more complex and dynamic battlefield environments. Finally, through automated and intelligent data processing, the present invention ensures high accuracy and reliability of battlefield information, reduces information distortion caused by human errors, and ensures that commanders can make timely tactical decisions based on accurate information. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 A schematic diagram of a flow chart of a method for implementing a visual neural link of the present invention; Figure 2 A schematic diagram of the overall architecture of visual information interaction of a method for implementing a visual neural link of the present invention; Figure 3 A description of an air platform-fighter message dictionary and a filling schematic diagram of a method for implementing a visual neural link of the present invention; Figure 4A schematic diagram of a visual neural link interaction protocol of a method for implementing a visual neural link of the present invention; Figure 5 A schematic diagram of a large-viewing-field small-object detection architecture of a method for implementing a visual neural link of the present invention; Figure 6 A schematic diagram of a visual information extraction process of a method for realizing a visual neural link of the present invention; Figure 7 A schematic diagram of a target movement dynamic calculation schematic diagram of a method for realizing a visual neural link of the present invention; Figure 8 A schematic diagram of the design and implementation of a visual information application effect feedback mechanism verification environment for a method for implementing a visual neural link of the present invention. DETAILED DESCRIPTION
[0014] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments, and the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.
[0015] It should be further explained that the drawings and implementation modes of the present invention mainly describe the concept of the present invention. On the basis of this concept, the specific forms and settings of some connection relationships, positional relationships, power mechanisms, power supply systems, hydraulic systems and control systems may not be fully described. However, on the premise that those skilled in the art understand the concept of the present invention, those skilled in the art can implement the above-mentioned specific forms and settings in a well-known manner.
[0016] When an element is referred to as being "fixed to" or "disposed on" another element, it may be directly on the other element or indirectly on the other element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or indirectly connected to the other element.
[0017] The directional words "inside and outside" refer to the inside and outside relative to the outline of each component itself. The directions or positional relationships indicated by the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" are based on the directions or positional relationships shown in the drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.
[0018] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used here to describe the spatial positional relationship between a device or feature and other devices or features as shown in the figure. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figure. For example, if the device in the accompanying drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" may include both "above" and "below". The device may also be positioned in other different ways, and the spatially relative descriptions used here are interpreted accordingly.
[0019] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "multiple" means two or more, and "several" means one or more, unless otherwise clearly and specifically defined.
[0020] The present invention will be further described below in conjunction with the embodiments.
[0021] Example 1 A method for implementing a visual neural link comprises the following steps: S1, visual information interaction architecture design; S2, visual link interaction standard; S3, visual information extraction; S4, visual information application verification environment; Among them: the step S1, visual information interaction architecture design involves designing and planning the overall architecture of the visual information system, including determining the functions, interfaces and interaction methods of each component in the system; the step S2, visual link interaction standards focus on formulating standards and protocols for visual information interaction, including defining the format of data transmission, communication protocols, and data security requirements to ensure visual information interaction between various systems or devices; the step S3, visual information extraction is to extract useful visual information from raw data or input, involving image processing, feature extraction, target recognition and other technologies; the step S4, visual information application verification environment is to verify the designed visual information in actual application scenarios, including developing a test environment, executing experiments and evaluating system performance; In step S1: S1.1, Perception layer data collection and storage: Raw data is collected through a variety of sensors (such as infrared, SAR, photoelectric) to ensure that diverse unstructured data is obtained; S1.2, Visual perception data processing: According to the data types of different sensors, use the corresponding data processing mechanism to perform preliminary processing on the raw data to generate initial visual information; S1.3, Information complementary layer processing: Multi-level processing of the initially generated visual information, including information fusion, association verification and screening, to ensure the accuracy and credibility of the information, and encapsulate the information into an unstructured text format that is easy to process; S1.4, message generator processing: The unstructured text is processed by the text fusion parser to parse out key information (such as target information or event information), and combined with data from other platforms to perform information fusion and optimization to generate more accurate visual information; S1.5, Visual information encoding and filling: Match the generated visual information with the message standard platform encoding library, and fill the encoding into the corresponding fields of the relevant message to complete the conversion from unstructured data to distributable structured visual messages; In step S2: S2.1. Visual information representation format design: Design a set of unified encoding format rules and data element dictionaries so that visual sensors and control systems can interact efficiently under the same set of standards; S2.2, Visual interaction coding format rules: Design the visual message interaction format between sensors and control systems according to different task requirements; S2.3, Visual message dictionary design: Create and improve the visual message dictionary and determine the specific value range of each field attribute; S2.4, Visual Link Interaction Protocol Development: Design the interaction protocol between the visual sensor and the controller, and clarify the protocol sub-items in different task processes; S2.5, Protocol Interaction Flow Diagram: By drawing and explaining the commonly used visual link interaction protocol flow chart (such as Figure 5 ), providing intuitive reference and guidance; In step S3: S3.1. Direct extraction architecture design: Analyze the visual information required for the task and build the corresponding visual information extraction framework; S3.2. Direct extraction process design: Design a logically rigorous extraction process, adopt advanced methods such as deep convolutional neural networks to ensure the generation and iterative fine-tuning of candidate bounding boxes, and use deep convolutional neural networks to achieve candidate bounding box generation and fine-tuning (such as Figure 7 ); S3.3, indirect extraction reasoning process: using the extracted visual information, through the combination of sensor parameters, platform parameters and image parameters, other required visual information is inferred to expand the scope of information acquisition; S3.4, Estimation error analysis: Analyze in detail the possible sources of errors in the estimation process and clarify the impact of the errors on the final visual information; S3.5, Schematic diagram: The calculation process and results are presented through schematic diagrams, and the calculation process and accuracy dependence factors are explained intuitively; In step S4: S4.1. Design of application effect feedback mechanism: Design an evaluation and feedback mechanism for the application effect of visual information, and adjust the visual information extraction method based on the feedback; S4.2. Laboratory environment simulation verification: simulate small target detection tasks in a laboratory environment, use deep learning technology to discover and identify targets, and verify the application effect of visual information; S4.3, data processing forward realization: by realizing the forward shift of image data processing, the timeliness of data processing is enhanced, the efficiency of information flow is improved, and the timely processing and transmission of visual information is ensured; S4.4, simulation platform verification: the processed visual information is transmitted to the simulation platform, and the task is adjusted through multiple feedbacks to optimize the visual information; S4.5. Situation map generation and verification: Generate high-precision situation maps, evaluate and verify the effectiveness of the visual information application effect feedback mechanism.
[0022] In this embodiment: In the modern battlefield environment, timely and accurate visual information is crucial for tactical decision-making. Through the advanced visual information interaction architecture, various visual sensors on the battlefield (such as drones, reconnaissance satellites, and ground monitoring equipment) can efficiently interact with the control system to improve combat effectiveness. The architecture involves multiple links such as the perception layer, visual perception data processing layer, information complementation layer, and message generator. Through a multi-level information processing mechanism, it ensures that the information collected from the sensor can be accurately transmitted to the command system.
[0023] First, in the perception layer, various sensors (such as infrared, SAR, and optoelectronics) are responsible for collecting raw unstructured data on the battlefield, which are mainly stored in the form of images and videos. These sensors can capture key information such as the movement of enemy forces and equipment deployment, forming first-hand intelligence on the battlefield. Next, the visual perception data processing layer uses specialized data processing mechanisms (such as infrared imaging processing, SAR image synthesis, and optoelectronic image enhancement) to process the raw data according to the data type of the sensor, generate preliminary visual information, and lay the foundation for subsequent information processing.
[0024] At the information complementation layer, the initially generated visual information will undergo multi-level processing, including information fusion, association verification and screening, to remove noise and redundant information, form more accurate battlefield situation information, and encapsulate it in an unstructured text format. These processes ensure the credibility and accuracy of the information and provide commanders with a reliable battlefield situation map. In the message generator, the text fusion parser further processes these unstructured text information, parses and fuses information provided by other platforms (such as ground forces and satellite reconnaissance), and generates accurate target and event information. By matching with the message standard platform encoding library, this information is converted into structured visual messages and filled into the corresponding message fields, thus completing the whole process from unstructured data to distributable structured visual messages.
[0025] In order to ensure efficient interaction between visual sensors and control systems, standardization of visual link interaction is essential. First, it is necessary to design the coding format rules and data element dictionary of visual information to ensure that sensors and control systems can exchange information under the same set of standards. According to different battlefield tasks (such as searching for enemy command centers and monitoring enemy movements), design the interaction format of visual messages and define the structure and content of messages. Create a visual message dictionary to standardize the specific value range of each field attribute to ensure the consistency and readability of information. Design interaction protocols in different task processes, such as drones guiding ground forces, satellite monitoring and target locking, etc., to clarify the close connection of each link. By drawing and explaining the commonly used visual link interaction protocol flow chart, operators can quickly understand and execute tasks, improving the accuracy and efficiency of task execution.
[0026] In terms of visual information extraction, we first analyze the visual information required for battlefield tasks and build corresponding extraction frameworks, such as the small object detection architecture in target search. We design a logical extraction process and use advanced methods such as deep convolutional neural networks to achieve the generation of candidate bounding boxes and multiple iterations of fine-tuning to improve the accuracy and efficiency of target detection. In some cases, we use the extracted visual information, combined with sensor and platform parameters, to infer information such as the direction and speed of movement of enemy targets, thereby expanding the scope of information acquisition. We analyze the sources of errors in the inference process in detail, such as initial information errors and parameter accuracy, to ensure the accuracy and reliability of the inferred information. We use schematic diagrams to display complex processes, such as dynamic inference of target movements, and intuitively explain the inference process and accuracy-dependent factors for easy understanding and application.
[0027] In order to ensure that the visual information extraction method can continuously adapt to battlefield needs and improve the accuracy and practicality of information, it is necessary to design an evaluation and feedback mechanism for the application effect of visual information. In a laboratory environment, the battlefield small target detection task is simulated, and deep learning technology is used to discover and identify targets, verify and optimize the detection algorithm. By realizing the forward shift of image data processing, the timeliness of data processing and the efficiency of information flow are enhanced, the delay of data transmission is reduced, and the battlefield information can be quickly processed and transmitted. The processed visual information is transmitted to the battlefield simulation platform, and the tasks are adjusted through multiple feedbacks to optimize the visual information to ensure its high accuracy and reliability. Generate a high-precision battlefield situation map, evaluate and verify the effectiveness of the feedback mechanism for the application effect of visual information, provide an accurate battlefield situation map, provide decision support for commanders, and improve the efficiency and effectiveness of battlefield command.
[0028] Through these steps, the visual information interaction architecture can efficiently realize the entire process from sensor data acquisition, processing to information encoding and application, which not only improves the accuracy and timeliness of information acquisition, but also ensures the efficient transmission and application of information through standardized interaction protocols and feedback mechanisms, providing strong support for battlefield decision-making.
[0029] Example 2 A method for realizing a visual neural link, wherein the visual information application effect feedback is an evaluation of the effect of visual information use, which is fed back to a visual sensor. The visual sensor adjusts and transforms the information extraction method based on the feedback information to provide visual information that meets the needs.
[0030] In this embodiment: the small target detection task in the search, the feedback of the application effect of visual information needs to be simulated and verified in a laboratory environment. Small target detection uses deep learning technology to solve the problem that target discovery and identification rely on rear processing, avoids the original image of the sensor being sent down, realizes the forward processing of image data, and achieves the application effect of increasing the timeliness of data processing and the efficiency of information flow; then the processed visual information is transmitted to the simulation display, and the task is adjusted through multiple feedbacks to improve the visual information and obtain a high-precision situation map.
[0031] Working principle: like Figure 1-8 As shown in the figure, the implementation method of the visual neural link extracts, processes and distributes structured visual information from the raw data collected by the visual sensor through multi-level information processing and standardized interaction mechanisms to improve the accuracy and timeliness of information transmission, which is particularly suitable for environments such as battlefields that require real-time decision-making. First, the system collects raw unstructured data on the battlefield through various sensors such as infrared, SAR, and optoelectronics at the perception layer. These data are mainly stored in the form of images and videos. Then, according to the data type of the sensor, a special data processing mechanism is used to process them at the visual perception data processing layer to generate preliminary visual information. Next, the information complementation layer performs multi-level processing on the preliminary generated visual information, including information fusion, association verification and screening, removes noise and redundant information, forms credible battlefield situation information, and encapsulates it into an unstructured text format. In the message generator, the text fusion parser further processes these unstructured text information, parses and fuses information from other platforms, generates accurate target and event information, and converts this information into structured visual messages by matching with the message standard platform encoding library. In order to ensure efficient interaction between visual sensors and control systems, standardization is essential. It is necessary to design unified encoding format rules and data element dictionaries, and design the interactive format and interactive protocol of visual messages according to different battlefield tasks. In terms of visual information extraction, by analyzing the visual information required for battlefield tasks, constructing an extraction framework, designing a logically rigorous extraction process, and using methods such as deep convolutional neural networks, the generation and fine-tuning of candidate bounding boxes are realized, the scope of information acquisition is expanded, and the errors in the calculation process are analyzed in detail to ensure the accuracy of information. Finally, in order to ensure that the visual information extraction method can continuously adapt to battlefield needs and improve the accuracy and practicality of information, it is necessary to design an evaluation and feedback mechanism for the application effect of visual information, simulate battlefield small target detection tasks in a laboratory environment, verify and optimize the detection algorithm, and enhance the timeliness of data processing by realizing the forward shift of image data processing. The processed visual information is transmitted to the battlefield simulation platform, and the task is adjusted through multiple feedback to generate a high-precision battlefield situation map, provide accurate battlefield situation, provide decision support for commanders, and improve battlefield command efficiency.
[0032] In the whole process, the data flow of coarse-grained data only needs a general channel, while the data of fine-grained data needs a dedicated channel. The granularity in the data is the degree to which the data can be directly applied. Some data can be directly applied and executed, such as instructions, spatiotemporal information, etc., but most data can only be applied after being processed, such as voice, text, images, videos, etc. The coarse-grained data flow only needs to be compressed and unpacked, and the communication system does not need to know its content. Both ends of the communication need to be equipped with a processing system to apply it. The fine-grained data flow needs to be encoded and decoded, and can be applied without too much processing. In future wars, the transmission of coarse-grained data between unmanned platforms should be minimized as much as possible, because the battlefield requires data to be timely and concise. The information service concept of thinking for the other party and providing it with the information it needs will be a new highlight in the development of data links. Think for the other party and process the data, and only fine-grained information enters the link, so that the data link becomes an information link, and there is no nonsense in the link, which is simple and efficient.
[0033] Information in the information chain is the soul. Everything revolves around the efficiency, simplicity, and value maximization of information, including the design of links. Different tasks and data determine different information interaction mechanisms, and the way they are presented in the information chain is also different, resulting in dedicated links. The information chain is like a "nervous system", which hinges all platforms and systems together to form a "body". Controlling the information means controlling the "body", which is in line with the concept of Data Link 2.0 that emphasizes the right to control information. Through these steps, the visual information interaction architecture can efficiently realize the entire process from sensor data acquisition, processing to information encoding and application. It not only improves the accuracy and timeliness of information acquisition, but also ensures the efficient transmission and application of information through standardized interaction protocols and feedback mechanisms, providing strong support for battlefield decision-making.
[0034] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
[0035] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0036] Unless otherwise specifically stated, the relative arrangement, numerical expressions and numerical values of the parts and steps set forth in these embodiments do not limit the scope of the present application. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, method and equipment known to those of ordinary skill in the relevant field may not be discussed in detail, but in appropriate cases, the technology, method and equipment should be considered as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so that once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.
Claims
1. A method for implementing a visual neural link, characterized in that: The following steps are involved: S1, visual information interaction architecture design; S2, visual link interaction standard; S3, visual information extraction; S4, visual information application verification environment; Among them: Step S1, visual information interaction architecture design involves designing and planning the overall architecture of the visual information system, including determining the functions, interfaces and interaction methods of each component in the system; Step S2, visual link interaction standards focus on formulating standards and protocols for visual information interaction, including defining the format of data transmission, communication protocols, and data security requirements to ensure visual information interaction between various systems or devices; Step S3, visual information extraction is to extract useful visual information from raw data or input, involving image processing, feature extraction, target recognition and other technologies; Step S4, visual information application verification environment is to verify the designed visual information in actual application scenarios, including developing a test environment, executing experiments and evaluating system performance.
2. The method for implementing a visual neural link according to claim 1, characterized in that: In step S1: S1.1, Perception layer data collection and storage: Collect raw data through a variety of sensors (such as infrared, SAR, photoelectric) to ensure that diverse unstructured data is obtained; S1.2, Visual perception data processing: According to the data types of different sensors, use the corresponding data processing mechanism to perform preliminary processing on the raw data to generate initial visual information; S1.3, Information complementary layer processing: Multi-level processing of the initially generated visual information, including information fusion, association verification and screening, to ensure the accuracy and credibility of the information, and encapsulate the information into an unstructured text format that is easy to process; S1.4, message generator processing: The unstructured text is processed by the text fusion parser to parse out key information (such as target information or event information), and combined with data from other platforms to perform information fusion and optimization to generate more accurate visual information; S1.5, Visual information encoding and filling: Match the generated visual information with the message standard platform encoding library, and fill the encoding into the corresponding fields of the relevant message to complete the conversion from unstructured data to distributable structured visual messages.
3. The method for implementing a visual neural link according to claim 1, characterized in that: In step S2: S2.
1. Visual information representation format design: Design a set of unified encoding format rules and data element dictionaries so that visual sensors and control systems can interact efficiently under the same set of standards; S2.2, Visual interaction coding format rules: Design the visual message interaction format between sensors and control systems according to different task requirements; S2.3, Visual message dictionary design: Create and improve the visual message dictionary and determine the specific value range of each field attribute; S2.4, Visual Link Interaction Protocol Development: Design the interaction protocol between the visual sensor and the controller, and clarify the protocol sub-items in different task processes; S2.5, Protocol Interaction Flow Diagram: Provide intuitive reference and guidance by drawing and explaining the commonly used visual link interaction protocol flow chart.
4. The method for implementing a visual neural link according to claim 1, characterized in that: In step S3: S3.
1. Direct extraction architecture design: Analyze the visual information required for the task and build the corresponding visual information extraction framework; S3.2, Direct extraction process design: Design a logically rigorous extraction process, using advanced methods such as deep convolutional neural networks to ensure the generation and iterative fine-tuning of candidate bounding boxes; S3.3, indirect extraction reasoning process: using the extracted visual information, through the combination of sensor parameters, platform parameters and image parameters, other required visual information is inferred to expand the scope of information acquisition; S3.4, Estimation error analysis: Analyze in detail the possible sources of errors in the estimation process and clarify the impact of the errors on the final visual information; S3.
5. Schematic diagram: The estimation process and results are presented through schematic diagrams, the estimation process and accuracy-dependent factors are explained intuitively, and a schematic diagram of the dynamic estimation of target movement is drawn (as shown in Figure 8).
5. The method for implementing a visual neural link according to claim 1, characterized in that: In step S4: S4.
1. Design of application effect feedback mechanism: Design an evaluation and feedback mechanism for the application effect of visual information, and adjust the visual information extraction method based on the feedback; S4.
2. Laboratory environment simulation verification: simulate small target detection tasks in a laboratory environment, use deep learning technology to discover and identify targets, and verify the application effect of visual information; S4.3, data processing forward realization: by realizing the forward shift of image data processing, the timeliness of data processing is enhanced, the efficiency of information flow is improved, and the timely processing and transmission of visual information is ensured; S4.4, simulation platform verification: the processed visual information is transmitted to the simulation platform, and the task is adjusted through multiple feedbacks to optimize the visual information; S4.
5. Situation map generation and verification: Generate high-precision situation maps, evaluate and verify the effectiveness of the visual information application effect feedback mechanism.
6. The method for implementing a visual neural link according to claim 5, characterized in that: The visual information application effect feedback is an evaluation of the effect of visual information use, which is fed back to the visual sensor. The visual sensor adjusts the information extraction method based on the feedback information to provide visual information that meets the needs.
