Generative artificial intelligence ammunition damage assessment method, system and program product

Abstract

A generative artificial intelligence ammunition damage assessment method, system and program product belong to the field of artificial intelligence technology. The method includes: step 1: establishing a Gaussian mixture model, which includes N Gaussian distributions; step 2: obtaining P physical parameters of ammunition; step 3: attributing the pth physical parameter to the data of the yth physical index according to the data classification model; step 4: generating multiple data respectively according to the N Gaussian distributions to supplement the data of each physical index, so that the data under the N ammunition indexes include M data, forming an M-row and N-column data matrix, and inputting each row of the data matrix into the ammunition damage assessment model. The ammunition damage assessment model estimates the M performance index sets of the ammunition according to each row of the input data matrix, and then calculates the R performance indexes of the ammunition. The present invention supplements the physical parameters of the ammunition through the Gaussian mixture model, and then estimates the damage performance index of the ammunition, which reduces the cost and improves the efficiency.

Description

Technical Field

[0001] The present invention relates to a generative artificial intelligence ammunition damage assessment method, system and computer program product, belonging to the technical field of data processing. Background Art

[0002] In modern military tactics and weapon system analysis, it is crucial to accurately evaluate the damage capability of ammunition to targets. For example, the Chinese invention patent application with publication number CN118673264A discloses a fragment cumulative damage assessment method and system, which includes step S1: setting a fragment warhead for experiment, presetting the parameters of the fragment warhead, and constructing a fragment warhead damage capability matrix function; step S2: setting a simulated target area, defining that the performance of each part in the target area being damaged by the impact is uniform, and calculating the core damage area in combination with the material, shape and anti-destructiveness of the target material of the fragment; step S3: constructing a fragment cumulative damage model, combining the damage capability of the fragment, the regional characteristics of the target and the error factor of the shooting, and calculating and predicting the cumulative damage probability caused by the warhead fragments to the target in a single shot and multiple shots.

[0003] This invention patent application does not disclose a technical solution on how to classify the physical parameters of the warhead using a data classification model, nor does it disclose a technical solution on how to supplement the physical parameters when the parameters are unbalanced. Summary of the invention

[0004] To achieve the above-mentioned purpose of the invention, the present invention provides a generative artificial intelligence ammunition damage assessment method, system and program product, which constructs a data classification model, classifies physical parameters through the data classification model, and generates data through Gaussian distribution to supplement the physical parameters, thereby estimating the damage performance indicators of the ammunition, greatly reducing the manpower and material costs required to obtain multiple data for each physical indicator, and improving efficiency.

[0005] To achieve the above-mentioned purpose, the present invention provides a generative artificial intelligence ammunition damage assessment method, which comprises the following steps:

[0006] Step 1: Establish a Gaussian mixture model, wherein the Gaussian mixture model includes N Gaussian distributions, where N is a positive integer greater than or equal to 6;

[0007] Step 2: Get the P physical parameters of the ammunition and get the parameter set , where represents the pth physical parameter, where P is a positive integer greater than or equal to N;

[0008] Step 3: According to the following data classification model, the pth physical parameter Data belonging to the yth physical indicator:

[0009] ,

[0010] In the formula, For the pth data The probability that the component belongs to the nth Gaussian and the category belongs to the yth physical indicator, C is the set of physical indicators, ; , and are the parameters of the data classification model; is a Gaussian distribution set, ;

[0011] Step 4: Generate multiple data according to N Gaussian distributions to supplement the data of each physical indicator, so that the data of N physical indicators include M data, thereby forming an M-row and N-column data matrix, and input each row of the data matrix into the ammunition damage assessment model. The ammunition damage assessment model estimates the M performance indicator set of the ammunition according to each row of the input data matrix. , m=1,...,M.

[0012] To achieve the above-mentioned purpose of the invention, the present invention also provides a system, which includes a storage medium and one or more processors, wherein the storage medium stores a computer program, and the computer program is called by one or more processors to implement the above-mentioned generative artificial intelligence ammunition damage assessment method.

[0013] To achieve the above-mentioned purpose of the invention, the present invention provides a computer program product, which uses computer language to compile the above-mentioned generative artificial intelligence ammunition damage assessment method into a computer program that is called and executed by one or more processors.

[0014] Compared with the prior art, the generative artificial intelligence ammunition damage assessment method, system and computer program product provided by the present invention construct a data classification model, classify physical parameters through the data classification model, and generate data through Gaussian distribution to supplement the physical parameters, thereby estimating the damage performance indicators of the ammunition, greatly reducing the manpower and material costs required to obtain multiple data for each physical indicator and improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the generative artificial intelligence ammunition damage assessment method provided by the present invention. DETAILED DESCRIPTION

[0016] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0017] In the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance.

[0018] Figure 1 : is a flow chart of the generative artificial intelligence ammunition damage assessment method provided by the present invention. Figure 1 As shown, the generative artificial intelligence ammunition damage assessment method provided by the present invention comprises the following steps:

[0019] Step 1: Establish a Gaussian mixture model, wherein the Gaussian mixture model includes N Gaussian distributions, and the Gaussian distributions correspond to the physical indicators of the ammunition one by one, and N is a positive integer greater than or equal to 6;

[0020] Step 2: Get the P physical parameters of the ammunition and get the parameter set , where represents the pth physical parameter, where P is a positive integer greater than or equal to N;

[0021] Step 3: According to the following data classification model, the pth physical parameter Data belonging to the yth physical indicator:

[0022] ,

[0023] In the formula, For the pth data The probability that the component belongs to the nth Gaussian and the category belongs to the yth physical indicator, C is the set of physical indicators, ; is the mixing coefficient parameter, and are the two parameters of the nth Gaussian distribution; is a Gaussian distribution set, ;

[0024] Step 4: Generate multiple data according to N Gaussian distributions to supplement the data of each physical index, so that the data under the N physical indexes include M data, thereby forming an M-row and N-column data matrix. Input each row of the data matrix into the ammunition damage assessment model. The ammunition damage assessment model estimates the M performance index sets of the ammunition according to each row of the input data matrix: , m=1,...,M.

[0025] The generative artificial intelligence ammunition damage assessment method provided by the present invention also includes the following steps:

[0026] Step 5: Calculate the performance index set of the ammunition according to the following formula:

[0027] ,

[0028] In the formula, .

[0029] In the present invention, the data classification model The parameters of are determined by the following steps:

[0030] S1-1: Divide the training sample set into a labeled training sample set and unlabeled training sample set , where Indicates that there are labeled training samples, Represents training samples Physical indicators data; represents unlabeled training samples;

[0031] S1-2: Classification model based on current data Parameter calculation for unlabeled training samples The probability of belonging to the nth Gaussian distribution component:

[0032] ,

[0033] In the formula, is the mixing coefficient; Unlabeled training samples The probability of a component belonging to the nth Gaussian distribution; is the parameter of the nth Gaussian distribution; N is the number of Gaussian distributions;

[0034] S1-3: Based on Update the current data classification model Parameters:

[0035] ,

[0036] ,

[0037] ,

[0038] In the formula, Indicates the number of labeled training samples belonging to the nth Gaussian distribution;

[0039] S1-4: Repeat steps S1-2 and S1-3 until convergence to obtain a data classification model Parameters , and .

[0040] The present invention constructs the above-mentioned data classification model, classifies the physical parameters through the data classification model, generates data through Gaussian distribution to supplement the physical parameters, and then estimates the damage performance index of the ammunition, which greatly reduces the manpower and material costs required to obtain multiple data for each physical index and improves efficiency.

[0041] In the present invention, the ammunition performance evaluation model is trained by a BP neural network, which includes the following steps:

[0042] S2-1: Take out multiple sets of data sets from the database, each set of data sets includes the current k pre-training data sets X k and the current k validation dataset , the pre-training dataset X k Input to the input layer of the BP neural network, and output the data set from the output layer of the BP neural network :

[0043] ,

[0044] In the formula, , is the current parameter set of the BP neural network;

[0045] S2-2: Assume the loss function L is:

[0046]

[0047] The gradient is calculated according to the following formula:

[0048] ,

[0049] S2-3: Update the current parameter set of the BP neural network according to the following formula:

[0050] ;

[0051] In the formula, is the learning rate, Express The gradient of

[0052] S2-4: Determine whether the loss function L is the minimum. If so, end the training. If not, replace the current parameter set with a new parameter set and return to S2-1.

[0053] The present invention trains the BP neural network into an ammunition performance evaluation model through the above method, and estimates the damage performance index of the ammunition through the ammunition performance evaluation model, which is more scientific than manually setting parameters.

[0054] In the present invention, physical parameters include the diameter of the bullet, the length of the cartridge case, the weight of the bullet and the cartridge case, the composition of the filling powder, the smoothness of the bullet and the cartridge case, and the muzzle kinetic energy of the gun using the ammunition. In addition, physical parameters also include the shape data of the bullet.

[0055] In the present invention, the performance indicators of ammunition include: range, armor-piercing ability, killing range, accuracy, stability, etc. Range refers to the maximum distance that the ammunition can reach; accuracy refers to the accuracy of the ammunition hitting the target, including the deviation between the impact point and the target; stability refers to the stability of the explosive during storage and transportation; armor-piercing ability refers to the ability of the ammunition to penetrate armor.

[0056] In the present invention, the range of the ammunition is obtained by the following method:

[0057] S3-1: The rotor drone is made to fly and rotate to a set position, so that a first camera and a second camera carried by the drone and separated by a set distance maintain a fixed posture, and then first image information and second image information including an image of a bullet landing point are respectively obtained;

[0058] S3-2: Using an image processor to process the first image of the first camera through a segmentation unit, a pixel window including K×K pixels is used to segment a frame of image acquired by the first camera and the second camera into Regions, , , where K is the number of pixels in a row and a column of the pixel window, C is the number of pixels in a row of a frame image, and D is the number of pixels in a column of a frame image. represents the floor function;

[0059] S3-3: Calculate the first value of the pixel window at the ath row and the bth column according to the following formula:

[0060] ,

[0061] In the formula, , The center pixel of the pixel window at row a and column b The value of Pixel The value of

[0062] S3-4: Calculate the second value of the pixel window at the ath row and the bth column according to the following formula:

[0063] ;

[0064] S3-5: Comparison and ,when When , the pixel value in the pixel window of the ath row and the bth column is set to 1, otherwise it is set to 0, so that the one-frame image acquired by the first target detection device and the second target detection device is binarized, and the position of the bullet landing point image in the first image is further acquired from the binarized image. Repeat steps 3-2 to 3-5, and acquire the position of the bullet landing point image in the second image in a similar way;

[0065] S3-6: Determine the distance from the rotorcraft to the drop point by binocular intersection measurement according to the position of the rotorcraft, the attitude of the first camera, the attitude of the second camera, the position of the drop point image in the first image, the position of the drop point image in the second image, and the distance between the first camera and the second camera;

[0066] S3-7: Determine the range based on the distance from the rotorcraft to the bullet drop point and the distance from the gun muzzle to the rotorcraft.

[0067] In the present invention, for ammunition with a shorter range, the rotor drone can be made to hover over the muzzle, and for ammunition with a longer range, the rotor drone can be made to hover over a position between the muzzle and the bullet landing point.

[0068] The present invention measures the range of ammunition by the above method without being affected by the surface of the bullet landing point.

[0069] The present invention also provides a system comprising a storage medium and one or more processors, wherein the storage medium stores a computer program, and the computer program is called by one or more processors to implement the above-mentioned generative artificial intelligence ammunition damage assessment method.

[0070] The present invention also provides a computer program product, which uses computer language to compile the above-mentioned generative artificial intelligence ammunition damage assessment method into a computer program that is called and executed by one or more processors.

[0071] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A generative artificial intelligence ammunition damage assessment method, characterized in that: The steps include: Step 1: Establish a Gaussian mixture model, wherein the Gaussian mixture model includes N Gaussian distributions, where N is a positive integer greater than or equal to 6; Step 2: Get J physical parameters of the ammunition and get the parameter set , where represents the jth physical parameter, where J is a positive integer greater than or equal to N; Step 3: According to the following data classification model, the jth physical parameter Data belonging to the yth physical indicator: ， In the formula, is the jth data The probability that the component belongs to the nth Gaussian and the category belongs to the yth physical indicator, C is the set of physical indicators, ; , and are the parameters of the data classification model; is a Gaussian distribution set, ; Step 4: Generate multiple data according to N Gaussian distributions to supplement the data of each physical indicator, so that the data of N physical indicators include M data, thereby forming a data matrix with M rows and N columns. Input each row of the data matrix into the ammunition damage assessment model. The ammunition damage assessment model estimates the M performance indicator sets of the ammunition according to each row of the input data matrix. , m=1,...,M; Step 5: Calculate the performance index set of the ammunition according to the following formula: , In the formula, .

2. The generative artificial intelligence ammunition damage assessment method according to claim 1, characterized in that: The ammunition performance evaluation model is trained by BP neural network, which includes the following steps: S2-1: Take out multiple sets of data sets from the database, each set of data sets includes the current k pre-training data sets X k and the current k validation dataset , the pre-training dataset X k Input to the input layer of the BP neural network, and output the data set from the output layer of the BP neural network : , In the formula, , is the current parameter set of the BP neural network; S2-2: Assume the loss function L is: ; The gradient is calculated according to the following formula: , S2-3: Update the current parameter set of the BP neural network according to the following formula: ； , In the formula, is the learning rate, Express The gradient of S2-4: Determine whether the loss function L is the minimum. If so, end the training. If not, replace the current parameter set with a new parameter set and return to S2-1.

3. The generative artificial intelligence ammunition damage assessment method according to any one of claims 1-2, characterized in that: Physical parameters include the diameter of the bullet, the length of the cartridge case, the weight of the bullet and cartridge case, the composition of the filler, the finish of the bullet and cartridge case, and the muzzle energy of the gun using the ammunition.

4. The generative artificial intelligence ammunition damage assessment method according to claim 3 is characterized in that: Performance indicators include: accuracy, stability, range, armor penetration and killing range.

5. A generative artificial intelligence ammunition damage assessment system, characterized in that: It includes a storage medium and one or more processors, wherein the storage medium stores a computer program, and the computer program is called by one or more processors to implement the generative artificial intelligence ammunition damage assessment method according to any one of claims 1 to 4.

6. A computer program product, characterized in that The generative artificial intelligence ammunition damage assessment method described in any one of claims 1 to 4 is compiled into a computer program called and executed by one or more processors using a computer language.

Images