Large model-based large-scale damage effect evaluation and management system

Abstract

The present application relates to the technical field of damage management, and particularly relates to a joint operation large-scale damage effect evaluation and management system based on a large model. The system comprises a combat target data processing module, a target damage feature merging module, a target damage effect evaluation module and a multi-dimensional damage management analysis module, and can acquire a combat target battlefield situation image sequence and combat target unit text data; the combat target battlefield situation image sequence and the combat target unit text data are subjected to data cleaning and standardization, and feature analysis and damage feature vector merging are simultaneously performed to generate a combat target damage feature vector; combat geographic coordinate data are acquired and coupled with a damage large model to construct, and damage effect evaluation analysis and multi-dimensional damage management decision analysis are simultaneously performed to generate a combat target multi-dimensional damage effect management decision, so as to execute a combat target damage effect management strategy work. The present application can provide accurate damage evaluation for combat command.

Description

Technical Field

[0001] The present invention relates to the field of damage management technology, and in particular to a large-scale damage effect assessment and management system for joint operations based on a large model. Background Art

[0002] In joint operations, the scale and complexity of combat operations are constantly increasing. The use of various weapons and equipment can cause large-scale damage to battlefield targets and the surrounding environment. Accurately assessing these damage effects is crucial for combat command, resource allocation, and the planning of subsequent combat operations. With the increasing informatization of modern joint operations, battlefield data has become massive, multi-source, and heterogeneous, including satellite reconnaissance data, drone monitoring data, ground sensor network data, and intelligence reports from various combat units. However, traditional damage effect assessment methods mainly rely on manual analysis and simple model calculations. In the military field, combatants must manually analyze large amounts of battlefield reconnaissance images and videos, as well as limited sensor data, and combine empirical formulas to estimate the extent and scope of damage. This method is extremely inefficient and cannot provide accurate damage management information to combat commanders in a timely manner in the rapidly changing joint combat environment. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide a large-scale damage effect assessment and management system based on a large model to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a large-scale damage effect assessment and management system based on a large model is proposed, which includes the following modules:

[0005] The combat target data processing module is used to obtain a combat target battlefield situation image sequence corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance, and collect and integrate combat intelligence reports, combat instructions, and weapon and equipment performance parameters corresponding to each combat unit in the joint combat target area to obtain combat target unit text data; the combat target battlefield situation image sequence and combat target unit text data are cleaned and standardized to generate a combat target situation standard image sequence and combat target standard text data;

[0006] The target damage feature merging module is used to perform target damage feature recognition analysis on the standard image sequence of the combat target situation to obtain the combat target damage image features; perform combat semantic damage feature analysis on the standard text data of the combat target to obtain the combat target text semantic damage features; and merge the combat target damage image features and combat target text semantic damage features into damage feature vectors according to the time sequence and spatial position to generate the corresponding combat target damage feature vector in the same time and space dimensions;

[0007] The target damage effect assessment module is used to obtain the combat geographic coordinate data corresponding to the joint combat target area, and to couple the combat geographic coordinate data with the damage model based on the corresponding combat target damage feature vector in the same space-time dimension to generate a combat target damage space coupling large model; the damage effect assessment and analysis of the combat target damage space coupling large model is performed to obtain the combat target damage effect;

[0008] The multi-dimensional damage management analysis module is used to conduct multi-dimensional damage management decision analysis on each combat target unit within the combat target damage space coupling model based on the combat target damage effect, generate multi-dimensional damage effect management decisions for combat targets, and execute corresponding combat target damage effect management strategies.

[0009] Furthermore, the combat target data processing module includes the following functions:

[0010] Acquire combat target battlefield situation image sequences corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance;

[0011] By collecting and integrating the combat intelligence reports, combat instructions and weapon and equipment performance parameters of each combat unit in the joint combat target area, the combat target unit text data can be obtained;

[0012] Performing target image denoising on each frame of the combat target battlefield situation image sequence to obtain a combat target situation denoised image sequence;

[0013] Performing an optical imaging characteristic analysis on each frame of the combat target situation denoised image sequence to analyze the corresponding propagation, scattering and reflection characteristics of light in the combat target battlefield environment, and calculating the light intensity and color distribution corresponding to the occluded portion of each frame of the image based on the corresponding propagation, scattering and reflection characteristics in combination with a ray tracing algorithm. Simultaneously, performing image loss repair on each frame of the combat target situation denoised image sequence based on the light intensity and color distribution corresponding to the occluded portion of each frame of the image to generate a combat target situation repaired image sequence; performing pixel normalization processing on each frame of the combat target situation repaired image sequence to generate a combat target situation standard image sequence;

[0014] The combat target unit text data is cleaned and standardized to generate combat target standard text data.

[0015] Furthermore, the target image denoising for each frame of the combat target battlefield situation image sequence includes:

[0016] Divide each frame of the combat target battlefield situation image sequence into each combat target situation image local area;

[0017] Performing neighborhood pixel grayscale analysis between local areas of each combat target situation image within each frame image to obtain the neighborhood pixel grayscale value distribution corresponding to each local area within each frame image;

[0018] Based on the grayscale value distribution of the neighboring pixels between each local area in each frame of the image, each frame of the combat target battlefield situation image sequence is repaired with salt and pepper noise. If the grayscale value distribution of the neighboring pixels between each local area is greater than the grayscale value distribution of the corresponding pixel in the local area, it is considered that there is a salt and pepper noise point in the local area, and the corresponding salt and pepper noise point in the local area is repaired according to the grayscale value distribution of the pixel in the local area to generate a combat target situation salt and pepper denoised image sequence;

[0019] Each frame of the combat target situation salt and pepper denoised image sequence is denoised based on wavelet transform, so as to analyze the corresponding edge and texture features in each frame image area and dynamically adjust the wavelet transform threshold. The corresponding Gaussian noise in each frame image is removed based on the wavelet transform threshold, so as to retain the corresponding details of each frame image to the greatest extent while removing the noise, and obtain the combat target situation denoised image sequence.

[0020] Furthermore, the text cleaning and standardization of the combat target unit text data includes:

[0021] Performing text segmentation on the combat target unit text data to obtain a combat target unit text segmentation set;

[0022] Perform semantic attribute analysis on each word in the combat target unit text word set to obtain the semantic attribute corresponding to each combat target unit text word;

[0023] Based on the semantic attributes corresponding to each combat target unit text word and combined with the text encoder corresponding to the CLIP model, each word in the combat target unit text word set is embedded with a part-of-speech to generate a combat target text embedding word set;

[0024] Perform semantic understanding and grammatical correction on each word in the word set embedded in the combat target text, and use the common data standards and coding system in the military field to convert the grammar of words corresponding to different sources and formats into a consistent vocabulary expression format and standardize it to generate standard text data of combat targets.

[0025] Furthermore, the target damage feature merging module includes the following functions:

[0026] Based on the large image model corresponding to the convolutional neural network, target damage feature recognition and analysis are performed on each frame of the standard image sequence of the combat target situation. The target shape, structure, position change and damage trace characteristics corresponding to the combat target at different time periods and different spatial positions are analyzed to obtain the combat target damage image characteristics;

[0027] Based on the large text model corresponding to the Transformer architecture, the combat semantic damage feature analysis of the standard text data of combat targets is carried out to extract the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects. The semantic feature analysis of the standard text data of combat targets is carried out based on the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects to extract the corresponding target type, combat mission, damage description and weapon and equipment usage status, so as to obtain the semantic damage features of the combat target text;

[0028] The damage feature vectors of combat target damage images and combat target text semantic damage features are merged according to time sequence and spatial position to generate the corresponding combat target damage feature vectors in the same time and space dimension.

[0029] Furthermore, the target damage effect assessment module includes the following functions:

[0030] Obtain the combat geographic coordinate data corresponding to the joint combat target area;

[0031] The operational geographic coordinate data corresponding to the joint operational target area is uniformly converted into the corresponding military coordinate system to obtain the corresponding joint operational target coordinate data in the military coordinate system;

[0032] Obtain the combat terrain and elevation model corresponding to the joint combat target area, and construct a combat geospatial model of the corresponding joint combat target coordinate data in the military coordinate system based on the combat terrain and elevation model to generate a large geospatial model of the combat target;

[0033] Based on the corresponding damage feature vectors of the combat target in the same space-time dimension, the damage large model is coupled with the combat target geographic space large model to generate the combat target damage space coupled large model;

[0034] The damage effect evaluation and analysis of the large-scale coupled damage space model of the combat target is carried out to obtain the damage effect of the combat target.

[0035] Furthermore, the damage effect evaluation and analysis of the combat target damage space coupling large model includes:

[0036] The target combat capability coefficient, target damage impact on the battle situation, target battlefield damage distribution area, and target combat defense level corresponding to each target unit are obtained through the large-scale coupled model of combat target damage space.

[0037] Based on the target combat capability coefficient corresponding to each combat target unit, the degree of impact of target damage on the battle situation, the target battlefield damage distribution area, and the target combat defense degree, the target damage degree calculation formula is used to calculate the target damage degree of each combat target unit in the combat target damage space coupling large model to obtain the combat target damage degree;

[0038] Based on the damage degree of the combat target, the damage effect evaluation and analysis of each combat target unit in the combat target damage space coupling large model is carried out to obtain the combat target damage effect.

[0039] Furthermore, the target damage degree calculation formula is specifically as follows:

[0040]

[0041] Where D is the damage degree of the combat target, M is the target battlefield damage distribution area corresponding to the combat target unit, ε is the target combat capability coefficient corresponding to the combat target unit, f is the target combat defense degree corresponding to the combat target unit, S is the impact degree of the target damage on the battle situation corresponding to the combat target unit, and e is the base of the natural logarithm.

[0042] Furthermore, the multi-dimensional damage management analysis module includes the following functions:

[0043] Based on the damage effects of combat targets in time and space, a damage diffusion analysis is conducted on the large-scale coupled damage model of combat targets to obtain the damage propagation and diffusion space range corresponding to the damage effects in time and space dimensions;

[0044] Based on the damage propagation and diffusion space range corresponding to the damage effect in the time and space dimensions, a large-scale damage reasoning and prediction analysis is conducted on each combat target unit within the large-scale coupled damage space model of the combat target to predict the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat missions.

[0045] Based on the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat tasks, a multi-dimensional damage management decision analysis is conducted on the corresponding combat target unit to generate a multi-dimensional damage effect management decision for the combat target to implement the corresponding combat target damage effect management strategy.

[0046] Furthermore, the multi-dimensional damage effect management decision-making of the combat objectives is specifically to analyze the impact of the damage effect on combat command and control, communication and liaison, and combat deployment in terms of combat system recovery, and formulate targeted recovery plans, including giving priority to repairing key communication nodes and redeploying troops to fill the defense loopholes caused by the damage effect; in terms of logistics support, to evaluate the impact of the damage effect on material reserves and transportation routes to formulate emergency material allocation plans and transportation route optimization plans; in terms of personnel safety, to formulate personnel evacuation, rescue and treatment plans based on the corresponding danger level of the damaged area and the distribution of personnel.

[0047] Beneficial effects of the present invention:

[0048] The large-scale damage effect evaluation and management system based on a large model comprises a combat target data processing module, a target damage feature merging module, a target damage effect evaluation module and a multi-dimensional damage management analysis module.Compared with the prior art, the beneficial effects of the present application are that the high-resolution satellite reconnaissance is used to obtain the battlefield situation image sequence of the combat target area, and the combat intelligence reports, combat orders and weapon equipment performance parameters of each combat unit in the area are integrated to finally form the standard text data of the combat target unit.The high-resolution satellite reconnaissance can provide accurate and real-time battlefield situation images, which are crucial for understanding the actual situation in the combat area.The image data can accurately reflect the changes of the enemy positions, equipment deployment, combat resources and geographical environment, thereby providing reliable information support for subsequent decision-making, and the integrated combat intelligence reports and orders can describe the combat intentions and actual performance of each combat unit in detail to form a comprehensive combat situation awareness, which can provide high-quality decision-making basis for commanders and improve combat effectiveness and coordination capability.The data cleaning and standardization processing can ensure the uniformity and consistency of various data to provide a stable data foundation for subsequent analysis and model construction.Secondly, the target damage feature recognition analysis is performed on the combat target situation standard image sequence to obtain the damage image features of the combat target, which can automatically identify the damage degree, position and form of the combat target in the image through image processing and pattern recognition technology to ensure that the damage state of each target on the battlefield can be clearly understood.The combat text semantic damage feature analysis can extract damage-related intelligence elements through the analysis of intelligence reports and orders, especially in complex combat conditions, which helps to judge the actual damage and threat level of the enemy target.The fusion of image features and text features can more comprehensively evaluate the target damage situation and provide multi-dimensional damage information.The combat target damage feature vector generated by combining time and space factors can quantitatively present the combat target damage situation, which helps to more accurately evaluate the damage.Finally, by conducting a multi-dimensional damage management decision analysis on the large model of combat target damage space coupling based on the damage effect of combat targets, a management strategy is generated. Through multi-dimensional damage management decision analysis, the damage status of combat targets can be comprehensively evaluated, and the use of resources and tactical execution can be optimized taking into account the damage degree, strategic value and combat requirements of different targets. This analysis can make detailed judgments in multiple dimensions, thereby balancing between various combat units to avoid waste of resources and missing the best opportunities. In this way, the optimal damage management strategy can be effectively identified and formulated, and flexible adjustments and decisions can be made based on this strategy to ensure that the execution of combat plans is more accurate and efficient. Through reasonable management decisions, not only can we respond quickly to dynamic changes in the battlefield, but we can also ensure the strike effect of various combat targets and the optimal use of resources, thereby providing accurate damage management information to combat command in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0050] Figure 1 This is a module diagram of the large-scale damage effect assessment and management system based on a large model of the present invention;

[0051] Figure 2 for Figure 1 Functional flow diagram of the combat target data processing module;

[0052] Figure 3 for Figure 1 Schematic diagram of the functional flow of the mid-target damage feature merging module. DETAILED DESCRIPTION

[0053] The following is a clear and complete description of the technical system of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0054] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0055] It should be understood that, although the terms "first", "second" and the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element without departing from the scope of the example embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0056] To achieve the above object, please refer to Figures 1 to 3 The application provides a large model-based large-scale damage effect evaluation and management system, which comprises the following modules:

[0057] A combat target data processing module is configured to acquire a combat target battlefield situation image sequence corresponding to a joint combat target region in real time through high-resolution satellite reconnaissance, and collect and integrate combat intelligence reports, combat orders and weapon equipment performance parameters corresponding to each combat unit in the joint combat target region to obtain combat target unit text data; and perform data cleaning and standardization on the combat target battlefield situation image sequence and the combat target unit text data to generate a combat target situation standard image sequence and combat target standard text data.

[0058] A target damage feature merging module is configured to perform target damage feature recognition analysis on the combat target situation standard image sequence to obtain combat target damage image features; perform combat semantic damage feature analysis on the combat target standard text data to obtain combat target text semantic damage features; and merge the combat target damage image features and the combat target text semantic damage features in time sequence and spatial position to generate corresponding combat target damage feature vectors in the same space-time dimension.

[0059] A target damage effect evaluation module is configured to acquire combat geographic coordinate data corresponding to the joint combat target region, and perform damage large model coupling construction on the combat geographic coordinate data based on the corresponding combat target damage feature vectors in the same space-time dimension to generate a combat target damage space coupling large model; and perform damage effect evaluation analysis on the combat target damage space coupling large model to obtain a combat target damage effect.

[0060] A multi-dimensional damage management analysis module is configured to perform multi-dimensional damage management decision analysis on each combat target unit in the combat target damage space coupling large model based on the combat target damage effect, generate a combat target multi-dimensional damage effect management decision, and execute a corresponding combat target damage effect management strategy.

[0061] In the embodiments of the application, please refer to Figure 1 FIG. 1 is a schematic diagram of the modules of the large-scale damage effect assessment and control system based on a large model of the present invention. In this example, the large-scale damage effect assessment and control system based on a large model includes the following modules:

[0062] S1: Combat target data processing module, used to obtain combat target battlefield situation image sequences corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance, and collect and integrate combat intelligence reports, combat instructions, and weapon and equipment performance parameters corresponding to each combat unit in the joint combat target area to obtain combat target unit text data; data cleaning and standardization of the combat target battlefield situation image sequences and combat target unit text data to generate combat target situation standard image sequences and combat target standard text data;

[0063] In an embodiment of the present invention, a reconnaissance system is formed by using high-resolution optical satellites and radar satellites. The optical satellite shoots the target area at a frequency of once every 3 minutes under good daylight conditions to obtain high-definition visible light images. The radar satellite is not restricted by weather and light and continuously scans the target area to obtain microwave images. The image data is transmitted to the ground station in real time via a satellite communication link. The ground station removes transmission interference and arranges the image data into a time sequence. The intelligent data acquisition terminal of each combat unit records combat intelligence reports, combat instructions and weapon equipment performance parameters in real time, and transmits them to the data center via the military communication network. The center unifies the data format and classifies and integrates them to obtain text data. Median filtering and wavelet transform are used to denoise the image sequence, and optical characteristics are analyzed to repair missing and standardize pixel values ​​to generate a standard image sequence. Special characters and stop words are removed from the text data, and stem extraction and terminology standardization are performed to generate standard text data.

[0064] S2: Target damage feature merging module, used to perform target damage feature recognition analysis on the standard image sequence of combat target situation to obtain combat target damage image features; perform combat semantic damage feature analysis on the standard text data of combat target to obtain combat target text semantic damage features; and merge the combat target damage image features and combat target text semantic damage features according to time sequence and spatial position to generate the corresponding combat target damage feature vector in the same time and space dimensions;

[0065] In an embodiment of the present invention, a pre-trained convolutional neural network is used to analyze a standard image sequence of combat target situations. The network's convolutional layer extracts local image features, the pooling layer reduces the amount of data, and the fully connected layer performs feature classification and regression to identify the target's shape, structure, position changes, and damage traces to obtain damage image features. A language model based on the Transformer architecture is used to process standard text data, and a multi-head self-attention mechanism is used to capture text dependencies, extract keywords, semantic relationships, and combat mission logical chains, parse the target type, combat mission, damage description, and weapon usage status, and obtain text semantic damage features. Timestamps and spatial coordinates are annotated for the two types of features. The image damage situation and text damage information are integrated according to the time sequence and spatial correspondence, and finally a damage feature vector of the same time and space dimension is generated.

[0066] S3: Target damage effect assessment module, used to obtain the combat geographic coordinate data corresponding to the joint combat target area, and to couple the combat geographic coordinate data with a damage model based on the corresponding combat target damage feature vector in the same space-time dimension to generate a combat target damage space coupling large model; and to perform damage effect assessment and analysis on the combat target damage space coupling large model to obtain the combat target damage effect;

[0067] In an embodiment of the present invention, combat geographic coordinate data is obtained with the help of a high-precision satellite positioning system, geographic information collection equipment, and military surveying and mapping. The positioning system provides approximate longitude and latitude, and the collection equipment obtains precise coordinates through triangulation. Surveying and mapping personnel correct the data in combination with topographic maps and field surveys, and use the geographic information system to associate the damage feature vectors of combat targets in the same time and space dimensions with geographic coordinates. A data fusion algorithm is used to integrate the damage feature values ​​with the coordinate data, and a large-scale spatial coupling model of combat target damage is constructed. The model is evaluated using a multi-index evaluation system and simulation analysis method, and evaluation indicators such as the capability coefficient of the combat target, the degree of impact on the battle situation, the damage distribution area, and the degree of combat defense are set. Different damage scenario parameters are input to simulate target state changes, the simulation results are statistically analyzed, and the damage effect of the combat target is obtained based on the evaluation criteria.

[0068] S4: Multi-dimensional damage management analysis module is used to perform multi-dimensional damage management decision analysis on each combat target unit within the combat target damage space coupling large model based on the combat target damage effect, generate combat target multi-dimensional damage effect management decision, and execute the corresponding combat target damage effect management strategy.

[0069] In an embodiment of the present invention, a multi-dimensional damage management decision analysis is conducted on the corresponding combat target unit by combining the results obtained from previous predictions. The multi-dimensional damage management decision analysis is conducted from three dimensions: combat system recovery, logistics support, and personnel safety. In terms of combat system recovery, the impact of damage on command and control, communication and combat deployment is analyzed, and troops are redeployed to fill defense gaps and determine the communication nodes that are repaired first. In terms of logistics support, the damage to the material storage depot is checked, and a deployment plan is formulated based on the remaining materials and combat needs. A new transportation route is planned using a geographic information system. In terms of personnel safety, an evacuation, rescue, and treatment plan is formulated based on the danger level of the damaged area and the distribution of personnel. The danger level areas are divided, evacuation routes are planned, and rescue teams are arranged to treat and transfer the wounded. The results of the analysis in each dimension are integrated to generate a management decision and implement the corresponding management strategy.

[0070] Furthermore, the combat target data processing module includes the following functions:

[0071] Acquire combat target battlefield situation image sequences corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance;

[0072] By collecting and integrating the combat intelligence reports, combat instructions and weapon and equipment performance parameters of each combat unit in the joint combat target area, the combat target unit text data can be obtained;

[0073] Performing target image denoising on each frame of the combat target battlefield situation image sequence to obtain a combat target situation denoised image sequence;

[0074] Performing an optical imaging characteristic analysis on each frame of the combat target situation denoised image sequence to analyze the corresponding propagation, scattering and reflection characteristics of light in the combat target battlefield environment, and calculating the light intensity and color distribution corresponding to the occluded portion of each frame of the image based on the corresponding propagation, scattering and reflection characteristics in combination with a ray tracing algorithm. Simultaneously, performing image loss repair on each frame of the combat target situation denoised image sequence based on the light intensity and color distribution corresponding to the occluded portion of each frame of the image to generate a combat target situation repaired image sequence; performing pixel normalization processing on each frame of the combat target situation repaired image sequence to generate a combat target situation standard image sequence;

[0075] The combat target unit text data is cleaned and standardized to generate combat target standard text data.

[0076] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Schematic diagram of the functional flow of the combat target data processing module in this embodiment. The combat target data processing module includes the following functions:

[0077] S11: Real-time acquisition of battlefield situation image sequences of combat targets corresponding to the joint combat target area through high-resolution satellite reconnaissance;

[0078] In an embodiment of the present invention, a satellite reconnaissance network is constructed using high-resolution optical and radar satellites. During daytime hours with good lighting conditions, the optical satellites capture visible light images of the joint combat target area at specific intervals, such as every five minutes. Radar satellites, unrestricted by lighting and weather conditions, continuously scan the target area, acquiring microwave images. The image data captured by the optical and radar satellites is transmitted in real time to a ground receiving station via a satellite communication link. The ground receiving station performs preliminary processing on the received image data to remove interference generated during transmission, and then arranges the data in chronological order to form a sequence of images of the combat target battlefield situation.

[0079] S12: Collect and integrate combat intelligence reports, combat instructions, and weapon and equipment performance parameters corresponding to each combat unit in the joint combat target area to obtain combat target unit text data;

[0080] In an embodiment of the present invention, each combat unit in the joint combat target area is equipped with a special data acquisition terminal for recording combat intelligence reports, combat instructions and weapon and equipment performance parameters in real time. These data acquisition terminals transmit data to a data center through a military communication network. The data center unifies the format of the received data and converts the text data provided by different combat units into the same encoding format. Then, the data is classified and sorted according to the type and source of the data, and the combat intelligence reports, combat instructions and weapon and equipment performance parameters are classified separately. These classified data are integrated together to finally form the combat target unit text data.

[0081] S13: performing target image denoising on each frame of the combat target battlefield situation image sequence to obtain a combat target situation denoised image sequence;

[0082] In an embodiment of the present invention, a denoising method combining median filtering and wavelet transform is used for each frame of an image in a combat target battlefield situation image sequence. First, the image is subjected to median filtering to remove salt and pepper noise in the image. For each pixel in the image, the median of the pixel values ​​in its neighborhood is taken as the new value of the pixel. Then, the image after median filtering is subjected to wavelet transform to decompose the image into sub-bands of different scales and directions. By setting an appropriate threshold, the noise coefficients in the sub-bands are removed. Finally, the processed wavelet coefficients are subjected to inverse wavelet transform to reconstruct the image to obtain a denoised image. This processing is performed on each frame of the sequence to ultimately form a combat target situation denoised image sequence.

[0083] S14: performing optical imaging characteristic analysis on each frame of the combat target situation denoised image sequence to analyze the corresponding propagation, scattering and reflection characteristics of light in the combat target battlefield environment, and calculating the light intensity and color distribution corresponding to the blocked portion in each frame of the image based on the corresponding propagation, scattering and reflection characteristics in combination with a ray tracing algorithm; and performing image missing repair on each frame of the combat target situation denoised image sequence based on the light intensity and color distribution corresponding to the blocked portion in each frame of the image to generate a combat target situation repaired image sequence; performing pixel normalization processing on each frame of the combat target situation repaired image sequence to generate a combat target situation standard image sequence;

[0084] In an embodiment of the present invention, for each frame of a combat target situation denoised image sequence, an optical model is used to analyze the propagation, scattering and reflection characteristics of light in the combat target battlefield environment. Based on factors such as the terrain, topography, and meteorological conditions of the battlefield environment, the propagation path and attenuation coefficient of the light are determined. A ray tracing algorithm is used to simulate the propagation process of light in the scene starting from the light source, and the light intensity and color distribution of the obscured portion are calculated. For missing portions of the image, the image is filled and repaired based on the calculated light intensity and color distribution. For example, for missing portions due to building obstruction, the color and brightness are matched based on the surrounding lighting conditions and the reflective characteristics of the object. After the repair is completed, pixel normalization is performed on each frame of the combat target situation repaired image sequence, and the pixel values ​​of the image are uniformly mapped to the range of 0-255, ultimately generating a combat target situation standard image sequence.

[0085] S15: Perform text cleaning and standardization on the combat target unit text data to generate combat target standard text data.

[0086] In an embodiment of the present invention, text cleaning is performed on the combat target unit text data. First, special characters, punctuation marks and redundant spaces in the text are removed, and these special characters and spaces are matched by using regular expressions and replaced with empty characters. Then, stop word filtering is performed to remove meaningless words in the text, such as "的", "了", "是", etc. Next, stem extraction is performed on the text to restore the words to their basic form, such as restoring "running" to "跑". Finally, according to the terminology standards and coding rules in the military field, the words in the text are standardized, and synonyms and antonyms are unified into standard terms. For example, "fighter" and "fighter jet" are unified into "fighter". After these processes, standard text data of combat targets are finally generated.

[0087] Furthermore, the target image denoising for each frame of the combat target battlefield situation image sequence includes:

[0088] Divide each frame of the combat target battlefield situation image sequence into each combat target situation image local area;

[0089] In an embodiment of the present invention, a fixed grid division method is used to process each frame of a combat target battlefield situation image sequence. The size of the grid is determined based on the image size and the approximate distribution of the combat targets. For example, the image is divided into 10×10 grids, and each grid is a local area of ​​the combat target situation image. By traversing the pixel coordinates of the image, the image is divided into multiple local areas according to the boundaries of the grid. The purpose of this is to provide a basis for subsequent local feature analysis of the image, because different local areas may contain different combat target information, such as military facilities, troop assembly points, etc.

[0090] Preferably, neighborhood pixel grayscale analysis is performed between each local area of ​​the combat target situation image in each frame image to obtain the neighborhood pixel grayscale value distribution corresponding to each local area in each frame image;

[0091] In an embodiment of the present invention, for each local area of ​​the combat target situation image divided in each frame image, with each local area as the center, the surrounding adjacent local areas are selected as neighborhoods, and the difference in pixel grayscale values ​​between each local area and its neighboring local areas is calculated. The distribution of the neighborhood pixel grayscale values ​​is obtained through statistical analysis methods. For example, for a certain local area, the difference in pixel grayscale values ​​between it and the four neighboring local areas above, below, left, and right is calculated, and then the frequency of occurrence of these differences is counted to form a neighborhood pixel grayscale value distribution histogram. In this way, the changing pattern of pixel grayscale values ​​between each local area can be clearly understood.

[0092] Preferably, salt and pepper noise restoration is performed on each frame of the combat target battlefield situation image sequence based on the grayscale value distribution of the neighboring pixels corresponding to each local area in each frame of the image. If it is determined that the grayscale value distribution of the neighboring pixels corresponding to each local area is greater than the grayscale value distribution of the corresponding pixel in the local area, it is considered that there is a salt and pepper noise point in the local area, and the corresponding salt and pepper noise point in the local area is restored according to the grayscale value distribution of the pixel corresponding to the local area, so as to generate a salt and pepper denoised image sequence of the combat target situation;

[0093] In an embodiment of the present invention, for each local area of ​​each frame image, the grayscale value distribution of its neighborhood pixels is compared with the grayscale value distribution of pixels in the local area. If the grayscale value distribution of the neighborhood pixels is larger than the grayscale value distribution of the pixels in the local area as a whole, it indicates that there are salt and pepper noise points in the local area. At this time, the median filtering method is used for repair. For each pixel in the local area, the median of the grayscale values ​​of its neighborhood pixels is taken as the new grayscale value of the pixel. For example, for a 3×3 neighborhood, the grayscale values ​​of these 9 pixels are sorted from small to large, and the middle value is taken as the new grayscale value of the center pixel. By performing such processing on all local areas of each frame image, a salt and pepper denoised image sequence of the combat target situation is finally generated.

[0094] Preferably, each frame image in the combat target situation salt and pepper denoised image sequence is denoised based on wavelet transform, so as to analyze the corresponding edge and texture features in each frame image area to dynamically adjust the wavelet transform threshold, and remove the corresponding Gaussian noise in each frame image based on the wavelet transform threshold, so as to retain the corresponding details of each frame image to the greatest extent while removing the noise, so as to obtain a combat target situation denoised image sequence.

[0095] In an embodiment of the present invention, a two-dimensional wavelet transform is performed on each frame of a combat target situation salt and pepper denoised image sequence to decompose the image into sub-bands of different scales and directions. By analyzing the coefficients of each sub-band, edge and texture features in the image are identified. For regions with rich edges and textures, the wavelet transform threshold is appropriately increased to reduce damage to details; for flat regions, the wavelet transform threshold is reduced to effectively remove Gaussian noise. In specific operations, an initial threshold is set, and then the threshold is dynamically adjusted according to the statistical characteristics of the sub-band coefficients. For wavelet coefficients smaller than the threshold, they are set to zero; for coefficients larger than the threshold, appropriate shrinkage processing is performed. Finally, the processed wavelet coefficients are subjected to an inverse wavelet transform to reconstruct the image, thereby obtaining a combat target situation denoised image sequence. The image retains important details of the image while removing Gaussian noise.

[0096] Furthermore, the text cleaning and standardization of the combat target unit text data includes:

[0097] Performing text segmentation on the combat target unit text data to obtain a combat target unit text segmentation set;

[0098] In an embodiment of the present invention, combat target unit text data is processed by using a word segmentation algorithm based on a combination of rules and statistics. First, a professional dictionary in the military field is constructed, which contains a large number of military terms, weapon and equipment names, combat operation names, and other vocabulary. When processing text data, a forward maximum match is first performed based on the dictionary. Starting from the beginning of the text, the longest word in the dictionary is matched as much as possible. For example, when encountering "air defense missile system", it is directly segmented as a complete word. For words that are not matched in the dictionary, a statistical model, such as a hidden Markov model (HMM), is used to segment the words based on the probability of word occurrence in the corpus and context information. After this processing, the combat target unit text data is segmented into independent words, and finally a combat target unit text word set is formed.

[0099] Preferably, a semantic attribute analysis is performed on each word in the combat target unit text word set to obtain a semantic attribute corresponding to each combat target unit text word;

[0100] In an embodiment of the present invention, a semantic attribute analysis is performed on each word in the combat target unit text word set by utilizing a semantic knowledge base and a machine learning model. The semantic knowledge base stores a large number of concepts in the military field and their semantic relationships, such as "tank" belongs to the "weapons and equipment" category, and "attack" belongs to the "combat action" category. For each word in the word set, a search and match is first performed in the semantic knowledge base to determine the basic semantic category to which it belongs. At the same time, a pre-trained machine learning model, such as the BERT model, is used to analyze the context of the word to further refine its semantic attributes. For example, for the word "troops", the context can be used to determine whether it is an "army troop", "naval troop" or "air force troop". In this way, the corresponding semantic attributes are determined for each combat target unit text word.

[0101] Preferably, based on the semantic attributes corresponding to each combat target unit text word and in combination with the text encoder corresponding to the CLIP model, each word in the combat target unit text word set is subjected to part-of-speech embedding to generate a combat target text embedding word set;

[0102] In an embodiment of the present invention, the text encoder of the CLIP model is applied to the combat target unit text word set. For each word in the word set, an initial feature vector is assigned to it according to its semantic attributes. The text encoder of the CLIP model further processes these words and embeds them into a high-dimensional semantic space. During the embedding process, the encoder considers the semantic relationship and contextual information between words, so that words with similar semantics are closer in the embedding space. For example, "fighter" and "bomber" both belong to "aviation weapons and equipment" semantically, and their vectors are relatively close in the embedding space. Through this part-of-speech embedding operation, the combat target unit text word set is converted into the combat target text embedded word set, and each word is represented in the form of a vector.

[0103] Preferably, semantic understanding and grammatical correction are performed on each word in the word segmentation set embedded in the combat target text, so as to convert the grammar of words corresponding to different sources and formats into a consistent vocabulary expression format and standardize them using the data standards and coding system commonly used in the military field, so as to generate standard text data of the combat target.

[0104] In an embodiment of the present invention, natural language processing technology is used to perform semantic understanding and grammatical correction on words in a word set embedded in a combat target text. First, semantic similarity calculation is used to determine whether the semantics between words are accurately expressed. For example, it is checked whether "combat aircraft" and "fighter plane" express the same concept. If so, they are uniformly expressed as "fighter plane". For grammatical errors, a grammatical rule library is used to check and correct them, such as correcting word collocation errors, part of speech errors, etc. Then, based on the data standards and coding systems commonly used in the military field, the words are formatted and standardized. For example, "air defense missiles" are uniformly numbered according to the coding system, and the names of weapons and equipment in different formats are converted into standard expressions. After these processes, the combat target text embedded in the word set is converted into combat target standard text data, so that it has a unified format and standardized expression.

[0105] Furthermore, the target damage feature merging module includes the following functions:

[0106] Based on the large image model corresponding to the convolutional neural network, target damage feature recognition and analysis are performed on each frame of the standard image sequence of the combat target situation. The target shape, structure, position change and damage trace characteristics corresponding to the combat target at different time periods and different spatial positions are analyzed to obtain the combat target damage image characteristics;

[0107] Based on the large text model corresponding to the Transformer architecture, the combat semantic damage feature analysis of the standard text data of combat targets is carried out to extract the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects. The semantic feature analysis of the standard text data of combat targets is carried out based on the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects to extract the corresponding target type, combat mission, damage description and weapon and equipment usage status, so as to obtain the semantic damage features of the combat target text;

[0108] The damage feature vectors of combat target damage images and combat target text semantic damage features are merged according to time sequence and spatial position to generate the corresponding combat target damage feature vectors in the same time and space dimension.

[0109] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Schematic diagram of the functional flow of the target damage feature merging module. In this embodiment, the target damage feature merging module includes the following functions:

[0110] S21: Based on the large image model corresponding to the convolutional neural network, target damage feature recognition and analysis are performed on each frame of the standard image sequence of the combat target situation. The target shape, structure, position changes and damage trace characteristics corresponding to the combat target at different time periods and different spatial positions are analyzed to obtain the combat target damage image characteristics;

[0111] In an embodiment of the present invention, a pre-trained convolutional neural network image model is used to input a standard image sequence of a combat target situation into the model frame by frame. The convolution layer of the model extracts features from the image and captures local features in the image through convolution kernels of different scales. The pooling layer downsamples the feature map to reduce the amount of data and retain important features. During the analysis process, the model will compare images from different time periods to see whether the appearance of the combat target is deformed or the structure is damaged. For images at different spatial positions, it will determine whether the position of the target has moved. At the same time, it will carefully identify whether there are craters, cracks and other damage marks in the image. For example, when analyzing an image of a military building, the model will detect whether the walls of the building are damaged or the roof has collapsed. Through detailed analysis of each frame of the image, the image characteristics of the combat target damage are finally obtained.

[0112] S22: Based on the large text model corresponding to the Transformer architecture, the standard text data of combat targets is analyzed for combat semantic damage characteristics to extract keywords, semantic relationships, and combat mission logic chains related to large-scale damage effects. Based on these keywords, semantic relationships, and combat mission logic chains related to large-scale damage effects, the standard text data of combat targets is analyzed for semantic features to extract the corresponding target type, combat mission, damage description, and weapon and equipment usage status, thereby obtaining the semantic damage characteristics of the combat target text.

[0113] In an embodiment of the present invention, standard text data of combat targets is input into a large text model based on the Transformer architecture. The encoder part of the model encodes the input text and learns the semantic information in the text. Through the self-attention mechanism, the model can capture the dependencies between different parts of the text, thereby extracting keywords related to large-scale damage effects, such as "explosion", "destruction", "strategic target", etc. Then, the semantic relationship between these keywords is analyzed. For example, "a certain target was destroyed using a certain weapon" reflects the association between the weapon and the target. At the same time, the logical chain of the combat mission is sorted out, such as conducting reconnaissance first and then launching an attack. Based on the extracted information, the text data is subjected to semantic feature analysis to determine whether the target type is a military base, a bridge or other; clarify whether the combat mission is offensive, defensive or support; describe the damage situation in detail, such as partial damage, complete damage, etc.; and record the use of weapons and equipment, such as what kind of weapons were used and how many were used, and finally obtain the semantic damage features of the combat target text.

[0114] S23: Merge the damage feature vectors of the combat target damage image features and the combat target text semantic damage features according to the time sequence and spatial position to generate the corresponding combat target damage feature vector in the same time and space dimension.

[0115] In an embodiment of the present invention, the timestamps and spatial coordinates are respectively marked on the damage image features of the combat target and the semantic damage features of the combat target text. The timestamp is accurate to a specific moment, and the spatial coordinates can be expressed by military coordinates such as longitude and latitude. Then, the two types of features are sorted according to the time sequence to ensure that the features in the same time period can correspond to each other. For spatial positions, features with the same or similar spatial coordinates are regarded as features of the same spatial area. In the merging process, information such as the appearance and structural changes in the image features are integrated with information such as the target type and damage description in the text features. For example, for a combat target at a certain moment and location, the damage to the building shown in the image is combined with the information about the attack on the building described in the text to form a comprehensive feature vector. In this way, all damage features are merged to finally generate a corresponding combat target damage feature vector in the same time and space dimension.

[0116] Furthermore, the target damage effect assessment module includes the following functions:

[0117] Obtain the combat geographic coordinate data corresponding to the joint combat target area;

[0118] In an embodiment of the present invention, the operational geographic coordinate data of the joint combat target area is obtained by using a high-precision satellite positioning system, geographic information collection equipment and military surveying and mapping means. The satellite positioning system can provide the approximate latitude and longitude information of the target area in real time, and professional geographic information collection equipment will be deployed around the target area. By measuring the precise coordinates of multiple control points and combining triangulation and other methods, the detailed geographic coordinates of important locations in the target area are obtained. Military surveying and mapping personnel will supplement and correct the collected data based on topographic maps and field surveys, and finally integrate comprehensive and accurate operational geographic coordinate data of the joint combat target area.

[0119] Preferably, the operational geographic coordinate data corresponding to the joint operational target area is uniformly converted into the corresponding military coordinate system to obtain the corresponding joint operational target coordinate data in the military coordinate system;

[0120] In an embodiment of the present invention, by adopting a special coordinate conversion algorithm and tool, the previously acquired combat geographic coordinate data is converted according to the parameters and conversion rules of the military coordinate system. First, the parameters such as the origin, coordinate axis direction, and projection mode of the military coordinate system used are clarified. Then, the original geographic coordinate data is calculated point by point through the coordinate conversion formula. For example, for plane coordinate conversion, the affine transformation formula is used; for three-dimensional coordinate conversion involving elevation, correction is performed in combination with the geoid model. During the conversion process, each coordinate point is checked multiple times to ensure that the converted joint combat target coordinate data is accurate, and finally the corresponding joint combat target coordinate data in the military coordinate system is obtained.

[0121] Preferably, the combat terrain and elevation model corresponding to the joint combat target area are obtained, and based on the combat terrain and elevation model, a combat geospatial model is constructed for the corresponding joint combat target coordinate data in the military coordinate system to generate a large geospatial model of the combat target;

[0122] In an embodiment of the present invention, high-resolution topographic data of the joint combat target area is obtained by utilizing technologies such as aerial photogrammetry and lidar scanning, and an accurate elevation model is created using a digital elevation model (DEM) generation tool. This data is combined with the coordinate data of the joint combat target in a military coordinate system, and spatial analysis and modeling are performed using geographic information system (GIS) software. In the GIS software, the topographic data is loaded in the form of a raster or vector, and a corresponding elevation value is assigned to each coordinate point according to the elevation model. By setting different layers and attributes, the three-dimensional geographic space of the combat target area is simulated, including elements such as mountains, rivers, and buildings, thereby generating a large geographic space model of the combat target.

[0123] Preferably, a large damage model coupling construction is performed on the combat target geographic space large model based on the corresponding combat target damage feature vector in the same space-time dimension to generate a combat target damage space coupling large model;

[0124] In an embodiment of the present invention, damage characteristic vectors of combat targets are analyzed and organized. These vectors contain key information such as damage type, damage range, and damage time. Based on a large geospatial model of combat targets, these damage characteristic vectors are associated with combat targets in geospatial space. For example, for explosion damage, the affected area is marked in the geospatial model based on its damage range and center location; and the change parameters in the time dimension are set based on the damage time. Through data fusion and algorithm optimization, the damage characteristic vectors are deeply coupled with the geospatial model, enabling the model to accurately simulate the damage to combat targets in the same time and space dimensions, thereby generating a large spatially coupled model of combat target damage.

[0125] Preferably, a damage effect evaluation and analysis is performed on a large-scale coupled damage space model of a combat target to obtain the damage effect of the combat target.

[0126] In an embodiment of the present invention, a large-scale model of the coupled damage space of combat targets is evaluated by using a multi-index evaluation system and a simulation analysis method. Evaluation indicators are set from multiple aspects such as combat capability, battle situation impact, damage distribution, and combat defense. By simulating different damage scenarios, corresponding parameters are input into the model, and the state changes of the combat targets under different damage levels are observed. For example, after simulating an airstrike, the capability coefficient of the combat target, the degree of impact on the battle situation, the damage distribution area, and the degree of combat defense of the combat target are analyzed. The simulation results are statistically analyzed and the damage effect of the combat target is finally obtained according to the preset evaluation criteria, providing a scientific basis for subsequent combat decision-making and resource allocation.

[0127] Furthermore, the damage effect evaluation and analysis of the combat target damage space coupling large model includes:

[0128] The target combat capability coefficient, the target damage battle influence degree, the target battlefield damage distribution area and the target combat defense degree corresponding to each combat target unit are obtained by the combat target damage space coupling large model;

[0129] In the embodiment of the present application, the combat target damage space coupling large model is a complex model integrating various data and rules, so as to obtain the target combat capability coefficient for each combat target unit, such as military base, armored cluster, etc., by database query in the model, which is calculated by the performance of weapon equipment, personnel quality and other factors of the combat unit, the target damage battle influence degree is obtained according to the preset strategic rules and combat situation evaluation in the model, such as the influence of the destruction of a key military base on the whole combat direction, the target battlefield damage distribution area is accurately measured by combining the geographic information system (GIS) with the model, the range of the damage area is counted, and the target combat defense degree is obtained by analyzing the data of defense work strength and defense weapon configuration, etc., which are stored in the related database of the large model and obtained through a specific query interface, and finally the target combat capability coefficient, the target damage battle influence degree, the target battlefield damage distribution area and the target combat defense degree corresponding to each combat target unit are obtained.

[0130] Preferably, the target damage degree of each combat target unit in the combat target damage space coupling large model is calculated by using a target damage degree calculation formula based on the target combat capability coefficient, the target damage battle influence degree, the target battlefield damage distribution area and the target combat defense degree corresponding to each combat target unit, so as to obtain the combat target damage degree.

[0131] In the embodiment of the present application, a suitable target damage degree calculation formula is constituted by combining the target combat capability coefficient, the target damage battle influence degree, the target battlefield damage distribution area and the target combat defense degree corresponding to each combat target unit, and a quantitative calculation is performed to calculate the combat target damage degree of the combat target unit group, so as to finally obtain the combat target damage degree, in addition, the target damage degree calculation formula can also use any damage evaluation algorithm in the art to replace the process of target damage degree calculation, and is not limited to the target damage degree calculation formula, for example, the target combat capability coefficient is C, the target damage battle influence degree is I, the target battlefield damage distribution area is A, and the target combat defense degree is D, the formula is damage degree = αC + βI + γA + δD, wherein α, β, γ and δ are preset weight coefficients.

[0132] Preferably, the damage effect of each combat target unit in the combat target damage space coupling large model is evaluated and analyzed based on the combat target damage degree, so as to obtain the combat target damage effect.

[0133] In an embodiment of the present invention, combat target units are divided into different damage levels according to the degree of damage to the combat targets, such as light damage, moderate damage, and heavy damage. For combat target units with light damage, their remaining combat capabilities and remaining contributions to the overall combat situation are analyzed. For example, after a communication base station is lightly damaged, although some functions are impaired, some communication lines can still be maintained. For units with moderate damage, the time and resources required to restore their combat capabilities are evaluated. For example, after a military warehouse is moderately damaged, the time and amount of materials required to repair the warehouse and replenish supplies are calculated. For units with heavy damage, it is determined whether they have completely lost their combat capabilities and the chain effects on surrounding combat units and the overall combat situation. For example, after a command center is severely damaged, the impact of its interruption on troop command and control and subsequent response measures are analyzed. Based on the above analysis, the damage effect of the combat target is finally obtained.

[0134] Furthermore, the target damage degree calculation formula is specifically as follows:

[0135]

[0136] Where D is the damage degree of the combat target, M is the target battlefield damage distribution area corresponding to the combat target unit, ε is the target combat capability coefficient corresponding to the combat target unit, f is the target combat defense degree corresponding to the combat target unit, S is the impact degree of the target damage on the battle situation corresponding to the combat target unit, and e is the base of the natural logarithm.

[0137] The present invention obtains a target damage degree calculation formula by using a specific mathematical model and verifying it, which is used to calculate the target damage degree of each combat target unit in the combat target damage space coupling large model. The formula fully considers the combat target damage degree D, the target battlefield damage distribution area M corresponding to the combat target unit, the target combat capability coefficient ε corresponding to the combat target unit, the target combat defense degree f corresponding to the combat target unit, the target damage battle situation impact degree S corresponding to the combat target unit, and the base of the natural logarithm e. According to the mutual correlation between the combat target damage degree D and the above parameters, a functional relationship is formed. This formula can calculate the target damage degree for each unit within a large, spatially coupled model of the target's damage. Furthermore, by fully considering multiple key factors related to the target, including the target's battlefield damage distribution area, combat capability coefficient, defense level, and impact on the battle situation, these factors reflect the comprehensive characteristics of the target, encompassing not only the local damage effect but also the impact of the global battle situation, enabling a more accurate assessment of the overall target damage degree. The exponential function in this formula reflects the nonlinear influence of various factors on the target damage degree, particularly the complex relationship between the target's combat capability coefficient, combat defense level, and battle situation impact on target damage. This nonlinear calculation more realistically reflects the target damage effect, avoiding the distortion caused by oversimplified linear models. By calculating the damage degree of target units and then conducting damage effect assessment and analysis based on this, real-time, dynamic target assessment is achieved during combat. This helps commanders understand the actual damage effect of each target and provides more scientific data support for subsequent operational decisions. This target damage degree formula specifically emphasizes the "battlefield impact" of combat targets. This parameter reflects the importance and criticality of the target within the overall combat environment. Therefore, the formula not only assesses the damage degree of a single target but also takes into account the target's role and influence within the overall combat situation, making the damage assessment results more accurate to the actual battlefield environment. This formula allows for personalized assessment and calculation of each combat target unit, as each target has a different damage distribution, combat capability, combat defense, and battlefield impact. This differentiated assessment allows for more refined adjustments to combat strategies and resource allocation, improving combat effectiveness. Furthermore, through the support of an effective calculation formula model, it facilitates rapid and accurate damage effect analysis, thereby enhancing the accuracy and applicability of the target damage degree calculation formula.

[0138] Furthermore, the multi-dimensional damage management analysis module includes the following functions:

[0139] Based on the damage effects of combat targets in time and space, a damage diffusion analysis is conducted on the large-scale coupled damage model of combat targets to obtain the damage propagation and diffusion space range corresponding to the damage effects in time and space dimensions;

[0140] In an embodiment of the present invention, computer simulation technology is used in conjunction with the parameters and rules of a large-scale coupled model of combat target damage space to simulate the diffusion process of different types of damage effects (such as explosions and electromagnetic interference) over time. The initial damage point is set, and the propagation distance and direction of the damage effect at different times are calculated based on physical laws, terrain factors, and the characteristics of the combat environment. For example, the explosion damage effect is centered on the explosion point and diffuses in all directions over time according to the shock wave propagation formula. The blocking and reflection of shock waves by terrain such as mountains and urban buildings are taken into account to determine the diffusion range in space. Through multiple rounds of simulation and data analysis, the damage propagation and diffusion spatial range corresponding to the damage effect in the time and space dimensions are obtained, providing basic data for subsequent analysis.

[0141] Preferably, based on the damage propagation and diffusion spatial range corresponding to the damage effect in the time and space dimensions, a large-scale damage reasoning and prediction analysis is performed on each combat target unit within the large-scale coupled damage space model of the combat target, so as to predict the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat missions;

[0142] In an embodiment of the present invention, previously acquired damage propagation and diffusion spatial range data is input into a large-scale coupled damage model for combat target damage. For each combat target unit within the model, such as a military base or a cluster of weapons and equipment, the damage assessment algorithm within the model is used to predict the large-scale damage range based on the intersection of its spatial location and the damage diffusion range. For example, if a military base is within the explosion damage diffusion range, the number and area of ​​damaged buildings and equipment within the base are calculated based on parameters such as the base's facility layout and protection capabilities. The impact of damaged equipment on the combat process is analyzed to assess the degree of impact on combat effectiveness. The relationship between the damage range and surrounding combat target units is observed to predict damage to them, such as the potential impact of a destroyed communication base station on communication interruption for surrounding troops. Simultaneously, combined with subsequent combat mission planning, the potential impact of the damage effect on mission execution is analyzed, such as the possibility that an offensive route is blocked by the damage range. Ultimately, the large-scale damage range corresponding to the combat target unit, its impact on combat effectiveness, damage to surrounding combat target units, and the potential impact on subsequent combat missions are predicted.

[0143] Preferably, a multi-dimensional damage management decision analysis is conducted on the corresponding combat target unit based on the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat tasks, and a multi-dimensional damage effect management decision for the combat target is generated to execute the corresponding combat target damage effect management strategy.

[0144] In an embodiment of the present invention, a multi-dimensional damage management decision analysis is conducted on the corresponding combat target unit by combining the results obtained from previous predictions to analyze and formulate corresponding damage management decisions. For the recovery of the combat system, the damage to key nodes such as the combat command center and the communication hub is deeply analyzed. If a key communication node is destroyed, resulting in the interruption of communication between the command and control and the front-line troops, a repair team is arranged first to carry professional equipment to repair it. At the same time, the communication frequency is temporarily adjusted or the backup communication line is activated. According to the defense loopholes caused by the damage, troops are transferred from other defense areas to fill them. In terms of logistics support, the damage to the material storage depot is checked. If some materials are damaged, If resources are destroyed, an emergency allocation plan is formulated based on the remaining materials and combat needs, materials are allocated from the rear warehouse, the traffic conditions of the transportation routes due to damage effects (such as roads being blown off) are evaluated, and new transportation routes are planned using the Geographic Information System (GIS). In terms of personnel safety, detailed evacuation routes are formulated based on the danger level of the damaged area (such as the high danger level in the explosion area) and personnel distribution data (such as the dense personnel at the troop assembly point), and rescue teams are arranged to enter the dangerous area with first aid equipment to provide emergency treatment and transfer the injured. Ultimately, a multi-dimensional damage effect management decision for the combat target is generated to implement the corresponding combat target damage effect management strategy.

[0145] Furthermore, the multi-dimensional damage effect management decision-making of the combat objectives is specifically to analyze the impact of the damage effect on combat command and control, communication and liaison, and combat deployment in terms of combat system recovery, and formulate targeted recovery plans, including giving priority to repairing key communication nodes and redeploying troops to fill the defense loopholes caused by the damage effect; in terms of logistics support, to evaluate the impact of the damage effect on material reserves and transportation routes to formulate emergency material allocation plans and transportation route optimization plans; in terms of personnel safety, to formulate personnel evacuation, rescue and treatment plans based on the corresponding danger level of the damaged area and the distribution of personnel.

[0146] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced within the present invention.

[0147] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A large-scale damage effect assessment and management system based on a large model, characterized by: Includes the following modules: The combat target data processing module is used to obtain a combat target battlefield situation image sequence corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance, and collect and integrate combat intelligence reports, combat instructions, and weapon and equipment performance parameters corresponding to each combat unit in the joint combat target area to obtain combat target unit text data; the combat target battlefield situation image sequence and combat target unit text data are cleaned and standardized to generate a combat target situation standard image sequence and combat target standard text data; The target damage feature merging module is used to perform target damage feature recognition and analysis on the standard image sequence of combat target situation to obtain the combat target damage image features; Conduct combat semantic damage feature analysis on combat target standard text data to obtain combat target text semantic damage features; Merge the damage feature vectors of the combat target damage image features and the combat target text semantic damage features according to the time sequence and spatial position to generate the corresponding combat target damage feature vector in the same time and space dimension; The target damage effect assessment module is used to obtain the combat geographic coordinate data corresponding to the joint combat target area, and to couple the combat geographic coordinate data with the damage feature vector corresponding to the combat target in the same space-time dimension to construct a large damage model to generate a large spatial coupling model of combat target damage; The damage effect evaluation and analysis of the large-scale coupled damage model of the combat target is carried out to obtain the damage effect of the combat target. This includes the following functions: Obtain the combat geographic coordinate data corresponding to the joint combat target area; The operational geographic coordinate data corresponding to the joint operational target area is uniformly converted into the corresponding military coordinate system to obtain the corresponding joint operational target coordinate data in the military coordinate system; Obtain the combat terrain and elevation model corresponding to the joint combat target area, and construct a combat geospatial model of the corresponding joint combat target coordinate data in the military coordinate system based on the combat terrain and elevation model to generate a large geospatial model of the combat target; Based on the corresponding damage feature vectors of the combat target in the same space-time dimension, the damage large model is coupled with the combat target geographic space large model to generate the combat target damage space coupled large model; Conduct damage effect evaluation and analysis on the large-scale coupled damage space model of combat targets to obtain the damage effect of combat targets; this includes: The target combat capability coefficient, target damage impact on the battle situation, target battlefield damage distribution area, and target combat defense level corresponding to each target unit are obtained through the large-scale coupled model of combat target damage space. Based on the target combat capability coefficient corresponding to each combat target unit, the target damage impact on the battle situation, the target battlefield damage distribution area, and the target combat defense level, the target damage level calculation formula is used to calculate the target damage level of each combat target unit in the combat target damage space coupling large model to obtain the combat target damage level. The target damage level calculation formula is specifically as follows: ； Where, The degree of damage to the combat target, is the target battlefield damage distribution area corresponding to the combat target unit, is the target combat capability coefficient corresponding to the combat target unit, is the target combat defense degree corresponding to the combat target unit, The degree of damage to the target corresponding to the combat target unit affects the battle situation. is the base of natural logarithms; Based on the damage degree of the combat target, the damage effect of each combat target unit in the large-scale coupled model of the combat target damage space is evaluated and analyzed to obtain the damage effect of the combat target; The multi-dimensional damage management analysis module is used to perform multi-dimensional damage management decision analysis on each combat target unit within the large-scale coupled combat target damage space model based on the combat target damage effect, generate combat target multi-dimensional damage effect management decisions, and implement the corresponding combat target damage effect management strategy. It includes the following functions: Based on the damage effects of combat targets in the time and space dimensions, a large-scale damage diffusion analysis is conducted on the coupled damage model of combat targets. The diffusion process of different types of damage effects under time advancement is simulated. The initial damage point is set. Based on physical laws, terrain factors and combat environment characteristics, the propagation distance and direction of the damage effect at different times are calculated to obtain the damage propagation and diffusion spatial range corresponding to the damage effect in the time and space dimensions. Based on the damage propagation and diffusion space range corresponding to the damage effect in the time and space dimensions, a large-scale damage reasoning and prediction analysis is conducted on each combat target unit within the large-scale coupled damage space model of the combat target to predict the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat missions. Based on the large-scale damage range corresponding to the combat target unit, the impact on combat effectiveness, the damage to surrounding combat target units, and the potential impact on subsequent combat tasks, a multi-dimensional damage management decision analysis is conducted on the corresponding combat target unit to generate a multi-dimensional damage effect management decision for the combat target to implement the corresponding combat target damage effect management strategy.

2. The large-scale damage effect assessment and management system based on a large model according to claim 1 is characterized in that: The combat target data processing module includes the following functions: Acquire combat target battlefield situation image sequences corresponding to the joint combat target area in real time through high-resolution satellite reconnaissance; By collecting and integrating the combat intelligence reports, combat instructions and weapon and equipment performance parameters of each combat unit in the joint combat target area, the combat target unit text data can be obtained; Performing target image denoising on each frame of the combat target battlefield situation image sequence to obtain a combat target situation denoised image sequence; Performing an optical imaging characteristic analysis on each frame of the combat target situation denoised image sequence to analyze the corresponding propagation, scattering and reflection characteristics of light in the combat target battlefield environment, and calculating the light intensity and color distribution corresponding to the occluded portion of each frame of the image based on the corresponding propagation, scattering and reflection characteristics in combination with a ray tracing algorithm. Simultaneously, performing image loss repair on each frame of the combat target situation denoised image sequence based on the light intensity and color distribution corresponding to the occluded portion of each frame of the image to generate a combat target situation repaired image sequence; performing pixel normalization processing on each frame of the combat target situation repaired image sequence to generate a combat target situation standard image sequence; The combat target unit text data is cleaned and standardized to generate combat target standard text data.

3. The large-scale damage effect assessment and management system based on a large model according to claim 2 is characterized in that: The target image denoising of each frame of the combat target battlefield situation image sequence comprises: Divide each frame of the combat target battlefield situation image sequence into each combat target situation image local area; Performing neighborhood pixel grayscale analysis between local areas of each combat target situation image within each frame image to obtain the neighborhood pixel grayscale value distribution corresponding to each local area within each frame image; Based on the grayscale value distribution of the neighboring pixels between each local area in each frame of the image, each frame of the combat target battlefield situation image sequence is repaired with salt and pepper noise. If the grayscale value distribution of the neighboring pixels between each local area is greater than the grayscale value distribution of the corresponding pixel in the local area, it is considered that there is a salt and pepper noise point in the local area, and the corresponding salt and pepper noise point in the local area is repaired according to the grayscale value distribution of the pixel in the local area to generate a combat target situation salt and pepper denoised image sequence; Each frame of the combat target situation salt and pepper denoised image sequence is denoised based on wavelet transform, so as to analyze the corresponding edge and texture features in each frame image area and dynamically adjust the wavelet transform threshold. The corresponding Gaussian noise in each frame image is removed based on the wavelet transform threshold, so as to retain the corresponding details of each frame image to the greatest extent while removing the noise, and obtain the combat target situation denoised image sequence.

4. The large-scale damage effect assessment and management system based on a large model according to claim 2 is characterized in that: The text cleaning and standardization of combat target unit text data includes: Performing text segmentation on the combat target unit text data to obtain a combat target unit text segmentation set; Perform semantic attribute analysis on each word in the combat target unit text word set to obtain the semantic attribute corresponding to each combat target unit text word; Based on the semantic attributes corresponding to each combat target unit text word and combined with the text encoder corresponding to the CLIP model, each word in the combat target unit text word set is embedded with a part-of-speech to generate a combat target text embedding word set; Perform semantic understanding and grammatical correction on each word in the word set embedded in the combat target text, and use the common data standards and coding system in the military field to convert the grammar of words corresponding to different sources and formats into a consistent vocabulary expression format and standardize it to generate standard text data of combat targets.

5. The large-scale damage effect assessment and management system based on a large model according to claim 1 is characterized in that: The target damage feature merging module includes the following functions: Based on the large image model corresponding to the convolutional neural network, target damage feature recognition and analysis are performed on each frame of the standard image sequence of the combat target situation. The target shape, structure, position change and damage trace characteristics corresponding to the combat target at different time periods and different spatial positions are analyzed to obtain the combat target damage image characteristics; Based on the large text model corresponding to the Transformer architecture, the combat semantic damage feature analysis of the standard text data of combat targets is carried out to extract the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects. The semantic feature analysis of the standard text data of combat targets is carried out based on the keywords, semantic relationships and combat mission logic chains related to the large-scale damage effects to extract the corresponding target type, combat mission, damage description and weapon and equipment usage status, so as to obtain the semantic damage features of the combat target text; The damage feature vectors of combat target damage images and combat target text semantic damage features are merged according to time sequence and spatial position to generate the corresponding combat target damage feature vectors in the same time and space dimension.

6. The large-scale damage effect assessment and management system based on a large model according to claim 1 is characterized in that: Specifically, the decision-making process for managing the multi-dimensional damage effects of combat objectives involves analyzing the impact of damage effects on combat command and control, communications, and deployment, and developing targeted recovery plans. This includes prioritizing the repair of key communication nodes and redeploying forces to fill defense gaps caused by damage effects. Regarding logistics support, the impact of damage effects on material reserves and transportation routes is assessed to develop emergency material deployment plans and transportation route optimization plans. In terms of personnel safety, personnel evacuation, rescue and treatment plans are formulated based on the corresponding danger level of the damaged area and the distribution of personnel.

Images