Intelligent aiming and shooting method and system

Abstract

The present invention discloses an intelligent aiming and shooting method and system, which relates to the field of intelligent aiming and shooting technology. The method and system include a software update module that can regularly review and optimize software codes, repair known vulnerabilities and defects, continuously improve algorithms such as target identification, tracking, and trajectory calculation, and use a large amount of measured data and simulated data to train and verify the algorithms, thereby improving the accuracy and stability of the algorithms. A sensor fusion module is used to more accurately detect and identify targets in foggy weather, where optical sensor imaging is limited but radar and infrared thermal imagers may still be able to work effectively. A central processing module is used to automatically correct errors or take safety measures, such as temporarily suspending the automatic aiming function. An environmental perception module is used to collect and expand target data in different environments, including target images, motion trajectories, and other information under different weather, terrain, and lighting conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent aiming and shooting, and in particular to an intelligent aiming and shooting method and system. Background Art

[0002] The existing public document - CN118657933B, discloses an intelligent aiming and shooting method and system, including acquiring videos and images of shooting scenes through an image acquisition device, using a target recognition algorithm to identify target attributes and relative position information, the target attributes include target type recognition, posture recognition and action recognition, the relative position information includes the relative distance and direction to the shooting position, and calculating the actual position coordinates; specifically, it includes: performing target detection and classification; identifying target types through color features and texture features; extracting target key points, inferring target posture and performing action recognition through a long short-term memory network; predicting the next position and motion trajectory of the target through Kalman filtering, calculating the shooting timing and trajectory based on the real-time position of the target, shooting through a laser transmitter according to the calculated trajectory and timing, and providing feedback and correction for each shot to dynamically adjust the shooting parameters; the application can achieve precise strikes on targets and can be applied to military training, simulated confrontation, and actual combat;

[0003] However, the existing intelligent aiming and shooting methods and systems still have the following drawbacks in actual use:

[0004] The process is greatly affected by environmental factors. Under complex meteorological conditions, such as heavy rain, heavy snow, and dense fog, the detection accuracy of the sensor will decrease, and the image recognition algorithm may not be able to accurately identify the target; and in complex terrain environments, there are many background objects, which can easily lead to misidentification or target loss. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent aiming and shooting method and system to solve the problems existing in the prior art.

[0006] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: an intelligent aiming and shooting method and system, comprising:

[0007] S1: Establish a comprehensive software update mechanism, regularly review and optimize software codes, fix known vulnerabilities and defects, and obtain software update modules;

[0008] S2: Acquire data from multiple different types of sensors, including optical sensors, radars, and infrared thermal imagers, and integrate the data from the multiple different sensors through sensor fusion technology to obtain a sensor fusion module;

[0009] S3: Set up a powerful exception handling procedure. When encountering abnormal situations such as erroneous data input and electromagnetic interference, it can automatically correct the error or take safety measures, such as temporarily stopping the automatic aiming function and prompting the user to intervene manually.

[0010] S4: Establish an environmental perception module to collect surrounding environmental data, and transmit the environmental data to the environmental analysis engine. Based on the deep learning model, the engine can quickly determine the type and complexity of the environment and identify and aim at the target.

[0011] Furthermore, the establishment of a comprehensive software update mechanism, regular review and optimization of software codes, repair of known vulnerabilities and defects, and acquisition of software update modules specifically include:

[0012] Utilize a large amount of measured and simulated data to train and verify algorithms, continuously improve algorithms such as target recognition, tracking, and trajectory calculation, and push software updates in a timely manner through the network or external storage devices. For example, when a new software vulnerability is discovered or a better algorithm version becomes available, the system can automatically prompt the user to update to ensure that the software is always in the best condition.

[0013] Furthermore, the acquisition of multiple different types of sensor data, such as optical sensors, radars, infrared thermal imagers, etc., and integration of the multiple different sensor data through sensor fusion technology to obtain a sensor fusion module specifically includes:

[0014] Based on the sensor fusion module, when in dense fog, optical sensor imaging is limited, but radar and infrared thermal imagers can still work effectively to detect and identify targets.

[0015] Furthermore, a powerful exception handling program is provided. When encountering abnormal situations such as erroneous data input and electromagnetic interference, it can automatically correct the error or take safety measures, such as temporarily stopping the automatic aiming function and prompting the user to manually intervene. Specifically, it includes:

[0016] Based on the error data input detection and correction algorithm;

[0017] Assume that the input data is a series of data vectors X = [x1, x2…, x n ], first, for each data point x i Set a reasonable value range, [min i ,max i ], which can be determined through historical data statistics or equipment specifications;

[0018] When inputting data, determine whether each data point meets the min i ≤x i ≤maxi If it is not satisfied, the data point is judged to be erroneous data. For example, for the target distance data d, if the normal range is [0,1000] meters (assuming this is the effective range of the weapon), when the input d = -50 meters or d = 1500 meters, it is judged to be abnormal.

[0019] If erroneous data is detected, for a single abnormal data point, there are multiple related data points around it (such as other sensor data at the same moment), and the average or weighted average can be used to replace the abnormal data. Let the normal data point be x j1 ,x j2 ,…,x jk , the corrected x i (assuming an exception) can be (average method), or assign different weights w according to the reliability of the data l , (weighted average method);

[0020] Electromagnetic interference detection and response algorithms;

[0021] Assume that the sensor signal strength is S during normal operation, and its amplitude, frequency and other characteristics have a normal range. For example, for a radio frequency sensor, the normal operating frequency is f0, and the amplitude range is [A min ,A max ], the signal strength S(t), its frequency f(t) and amplitude A(t) can be collected regularly (every T time interval);

[0022] When |A(t)-A0|>ΔA (ΔA is the set amplitude deviation threshold) or |f(t)-f0|>Δf, (Δf is the set frequency deviation threshold), it is determined that there may be electromagnetic interference. At the same time, it can also be judged by the noise level of the signal. Let the normal noise power be N0. If the detected noise power N(t) satisfies N(t)>N0+ΔN (ΔN is the noise deviation threshold), it is also determined that there may be interference;

[0023] Once it is determined that there is electromagnetic interference, the safety measures are automatically triggered. If the intensity of the electromagnetic interference exceeds a certain severity threshold, th , the automatic aiming function can be directly suspended, and a warning signal can be sent to the user at the same time. Assume that the electromagnetic interference intensity I can be obtained by a comprehensive evaluation of the degree of signal abnormality (such as I = w1ΔA + w2Δf + w3ΔN, w1, w2, w3 are weight coefficients). When I>I th , executes the operation of pausing automatic aiming and prompting the user to intervene manually.

[0024] Furthermore, the environment perception module is established to collect surrounding environment data, and the environmental data is transmitted to the environment analysis engine. The engine is based on a deep learning model to quickly determine the type and complexity of the environment and identify and aim at the target, specifically including:

[0025] The target recognition algorithm formula is based on a deep learning convolutional neural network. The calculation process mainly involves the operations of convolution, pooling, and fully connected layers. Assume that the input image is I, and its size is H×W×C (height H, width W, number of channels C);

[0026] The convolution kernel is K and the size is k h ×k w ×C(height k h , width k w ), the step size is s, the padding is p, and the size calculation formula of the convolutional layer output feature map O is: Each element o of the output feature map O i,j The calculation of is obtained through convolution operation:

[0027] Taking the maximum pooling as an example, the pooling kernel size is p h ×p w , the step size is s, and the calculation formula for the size of the output feature map P after pooling is: Each element p of the output feature map P i,j It takes the maximum value within the pooling window:

[0028] After multiple convolution and pooling layers, the feature map is flattened into a vector x, and the output y of the fully connected layer is passed through the weight matrix W f and the bias vector b f Calculation: y = f (W f x+b f ), where f is the activation function. In target recognition, the output of the last fully connected layer is usually the category probability distribution, and the target category is determined by comparing the probability size;

[0029] Aiming algorithm formula, based on the basic principles of exterior ballistics, assumes that the distance to the target is d, the initial velocity of the bullet is v0, and the gravitational acceleration of the bullet is g (usually 9.8m / s 2 ), without considering the influence of other factors such as wind, the ballistic equations in the horizontal direction x and vertical direction y are:

[0030] Horizontal direction: x = v 0x t, where v ox is the horizontal component of the bullet's initial velocity, and t is the flight time;

[0031] Vertical direction: where v 0y is the vertical component of the bullet's initial velocity. When considering wind direction and wind speed (assuming wind speed is v w , the angle between the wind direction and the shooting direction is θ), the horizontal speed needs to be corrected: v′ 0x =v 0x -v w cosθ;

[0032] Finally, the intelligent aiming system also needs to combine the initial posture of the gun (pitch angle α and yaw angle β). After obtaining the gun posture data through the sensor, the aiming point is corrected accordingly. The correction values ​​Δx and Δy of the aiming point in the horizontal and vertical directions can be calculated based on the above ballistic equation and the gun posture, thereby determining the final aiming point position.

[0033] An intelligent aiming and shooting system includes a software update module, a sensor fusion module, a central processing module and an environmental perception module;

[0034] The software update module is connected to a sensor fusion module, the sensor fusion module is connected to a central processing module, and the central processing module is connected to an environment perception module.

[0035] Compared with the prior art, the intelligent aiming and shooting method and system provided by the present invention has the following beneficial effects:

[0036] This intelligent aiming and shooting method and system, through the software update module, can regularly review and optimize the software code, fix known vulnerabilities and defects, and continuously improve algorithms such as target recognition, tracking, and trajectory calculation. It uses a large amount of measured data and simulated data to train and verify the algorithms, thereby improving the accuracy and stability of the algorithms. Through the sensor fusion module, in dense fog, optical sensor imaging is limited, but radar and infrared thermal imagers may still be able to work effectively. After fusing the data of these sensors, targets can be detected and identified more accurately. Through the central processing module, errors can be corrected automatically or safety measures can be taken, such as temporarily suspending the automatic aiming function and prompting the user to intervene manually to avoid misidentification or target loss. Through the environmental perception module, target data in different environments are collected and expanded, including target images, motion trajectories, and other information under different weather, terrain, and lighting conditions, for training and optimizing the system's recognition and aiming capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0038] Figure 1 Schematic diagram of the method flow of the present invention;

[0039] Figure 2 Schematic diagram of the system structure of the present invention;

[0040] Figure 3 It is a schematic structural diagram of the sight of the present invention. DETAILED DESCRIPTION

[0041] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1-3 , an intelligent aiming and shooting method and system, comprising:

[0043] S1: Establish a comprehensive software update mechanism, regularly review and optimize software codes, fix known vulnerabilities and defects, and obtain software update modules;

[0044] S2: Acquires data from various types of sensors, including optical sensors, radars, and infrared thermal imagers. Sensor fusion technology is used to integrate these data to create a sensor fusion module.

[0045] S3: Set up a powerful exception handling procedure. When encountering abnormal situations such as erroneous data input and electromagnetic interference, it can automatically correct the error or take safety measures, such as temporarily stopping the automatic aiming function and prompting the user to intervene manually.

[0046] S4: Establish an environmental perception module to collect surrounding environmental data, and transmit the environmental data to the environmental analysis engine. Based on the deep learning model, the engine can quickly determine the type and complexity of the environment and identify and aim at the target.

[0047] The aforementioned establishment of a comprehensive software update mechanism involves regular review and optimization of software codes, repair of known vulnerabilities and defects, and acquisition of software update modules, specifically including:

[0048] Utilize a large amount of measured and simulated data to train and verify algorithms, continuously improve algorithms such as target recognition, tracking, and trajectory calculation, and push software updates in a timely manner through the network or external storage devices. For example, when a new software vulnerability is discovered or a better algorithm version becomes available, the system can automatically prompt the user to update to ensure that the software is always in the best condition.

[0049] The sensor fusion module is obtained by acquiring data from various types of sensors, such as optical sensors, radars, and infrared thermal imagers, and integrating the data through sensor fusion technology to obtain the sensor fusion module, which specifically includes:

[0050] Based on the sensor fusion module, when in dense fog, optical sensor imaging is limited, but radar and infrared thermal imagers can still work effectively to detect and identify targets.

[0051] The aforementioned powerful exception handling procedure can automatically correct errors or take safety measures when encountering abnormal situations such as erroneous data input and electromagnetic interference, such as temporarily suspending the automatic aiming function and prompting the user to manually intervene. Specifically, it includes:

[0052] Based on erroneous data input detection and correction algorithms;

[0053] Assume that the input data is a series of data vectors X=[x1,x2…,x n ], first, for each data point x i Set a reasonable value range, [min i ,max i ], which can be determined through historical data statistics or equipment specifications;

[0054] When inputting data, determine whether each data point meets the min i ≤x i ≤max i If it is not satisfied, the data point is judged to be erroneous data. For example, for the target distance data d, if the normal range is [0,1000] meters (assuming this is the effective range of the weapon), when the input d = -50 meters or d = 1500 meters, it is judged to be abnormal.

[0055] If erroneous data is detected, for a single abnormal data point, there are multiple related data points around it (such as other sensor data at the same moment), and the average or weighted average can be used to replace the abnormal data. Let the normal data point be x j1 ,x j2 ,…,x jk , the corrected x i (assuming an exception) can be (average method), or assign different weights w according to the reliability of the data l , (weighted average method);

[0056] Electromagnetic interference detection and response algorithms;

[0057] Assume that the sensor signal strength is S during normal operation, and its amplitude, frequency and other characteristics have a normal range. For example, for a radio frequency sensor, the normal operating frequency is f0, and the amplitude range is [A min ,A max ], the signal strength S(t), its frequency f(t) and amplitude A(t) can be collected regularly (every T time interval);

[0058] When |A(t)-A0|>ΔA (ΔA is the set amplitude deviation threshold) or |f(t)-f0|>Δf, (Δf is the set frequency deviation threshold), it is determined that there may be electromagnetic interference. At the same time, it can also be judged by the noise level of the signal. Let the normal noise power be N0. If the detected noise power N(t) satisfies N(t)>N0+ΔN (ΔN is the noise deviation threshold), it is also determined that there may be interference;

[0059] Once it is determined that there is electromagnetic interference, the safety measures are automatically triggered. If the intensity of the electromagnetic interference exceeds a certain severity threshold, th , the automatic aiming function can be directly suspended, and a warning signal can be sent to the user at the same time. Assume that the electromagnetic interference intensity I can be obtained by a comprehensive evaluation of the degree of signal abnormality (such as I = w1ΔA + w2Δf + w3ΔN, w1, w2, w3 are weight coefficients). When I>I th , executes the operation of pausing automatic aiming and prompting the user to intervene manually.

[0060] The environment perception module is established to collect surrounding environment data, and the environmental data is transmitted to the environment analysis engine. The engine is based on a deep learning model to quickly determine the type and complexity of the environment and identify and aim at the target. Specifically, it includes:

[0061] The target recognition algorithm formula is based on a deep learning convolutional neural network. The calculation process mainly involves the operations of convolution, pooling, and fully connected layers. Assume that the input image is I, and its size is H×W×C (height H, width W, number of channels C);

[0062] The convolution kernel is K and the size is k h ×k w ×C(height k h , width k w ), the step size is s, the padding is p, and the size calculation formula of the convolutional layer output feature map O is: Each element o of the output feature map O i,j The calculation of is obtained through convolution operation:

[0063] Taking the maximum pooling as an example, the pooling kernel size is p h ×p w , the step size is s, and the calculation formula for the size of the output feature map P after pooling is: Each element p of the output feature map P i,j It takes the maximum value within the pooling window:

[0064] After multiple convolution and pooling layers, the feature map is flattened into a vector x, and the output y of the fully connected layer is passed through the weight matrix W f and the bias vector b f Calculation: y = f (W f x+b f ), where f is the activation function. In target recognition, the output of the last fully connected layer is usually the category probability distribution, and the target category is determined by comparing the probability size;

[0065] Aiming algorithm formula, based on the basic principles of exterior ballistics, assumes that the distance to the target is d, the initial velocity of the bullet is v0, and the gravitational acceleration of the bullet is g (usually 9.8m / s 2 ), without considering the influence of other factors such as wind, the ballistic equations in the horizontal direction x and vertical direction y are:

[0066] Horizontal direction: x = v 0x t, where v ox is the horizontal component of the bullet's initial velocity, and t is the flight time;

[0067] Vertical direction: where v 0y is the vertical component of the bullet's initial velocity. When considering wind direction and wind speed (assuming wind speed is v w , the angle between the wind direction and the shooting direction is θ), the horizontal speed needs to be corrected: v′ 0x =v 0x -v w cosθ;

[0068] Finally, the intelligent aiming system also needs to combine the initial posture of the gun (pitch angle α and yaw angle β). After obtaining the gun posture data through the sensor, the aiming point is corrected accordingly. The correction values ​​Δx and Δy of the aiming point in the horizontal and vertical directions can be calculated based on the above ballistic equation and the gun posture, thereby determining the final aiming point position.

[0069] An intelligent aiming and shooting system includes a software update module, a sensor fusion module, a central processing module and an environmental perception module;

[0070] The software update module is connected to a sensor fusion module, the sensor fusion module is connected to a central processing module, and the central processing module is connected to an environment perception module.

[0071] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. An intelligent aiming and shooting method, characterized in that: include: S1: Establish a software update mechanism to regularly review and optimize software code, fix known vulnerabilities and defects, obtain software update modules, utilize a large amount of measured and simulated data for algorithm training and verification, continuously improve target recognition, tracking, and trajectory calculation, and push software updates in a timely manner via the network or external storage devices. When new software vulnerabilities are discovered or a better algorithm version becomes available, the system automatically prompts users to update to ensure that the software is always in optimal condition. S2: Acquire data from multiple different types of sensors, including optical sensors, radars, and infrared thermal imagers. These data are integrated using sensor fusion technology to create a sensor fusion module. Based on the sensor fusion module, even in dense fog where optical sensor imaging is limited, radars and infrared thermal imagers can still function to detect and identify targets. S3: Establish an exception handling procedure to automatically correct errors or take safety measures when encountering erroneous data input or electromagnetic interference. Safety measures include temporarily suspending the automatic aiming function or prompting the user for manual intervention. The input data is the target position, speed, and environmental parameters obtained by the sensor. S4: Establish an environmental perception module to collect surrounding environmental data, and transmit the environmental data to the environmental analysis engine. Based on the deep learning model, the engine determines the type and complexity of the environment and identifies and targets the target.

2. The intelligent aiming and shooting method according to claim 1, wherein: The exception handling program is established to automatically correct errors or take safety measures when encountering erroneous data input or electromagnetic interference anomalies. The safety measures include temporarily stopping the automatic aiming function or prompting the user to manually intervene. The input data is the target position, speed, and environmental parameters obtained by the sensor; specifically, they include: Based on the error data input detection and correction algorithm; Input data vector , first, for each data point Set a reasonable value range. ,determined through historical data statistics or equipment specifications; When inputting data, determine whether each data point satisfies , when it is not satisfied, the data point is judged to be erroneous data; for target distance data , when the normal range is meters, input rice or When the number of meters is less than 1, it is considered abnormal; When erroneous data is detected, for a single abnormal data point, there are multiple data points related to it. The average value or weighted average value is used to replace the abnormal data. When the normal data point is , after correction It is obtained by the average method according to the following formula: ； Or assign different weights based on the reliability of the data , obtained by the weighted average method of the following formula: ； Electromagnetic interference detection and response algorithms; When working normally, the sensor signal strength is , its amplitude and frequency characteristics have a normal range. For a radio frequency sensor, the normal operating frequency is , the amplitude range is , regularly collect signal strength When the frequency and amplitude ; when , The amplitude deviation threshold is set or , When the frequency deviation threshold is set, it is judged to be electromagnetic interference. At the same time, when judging by the noise level of the signal, the normal noise power is , when the noise power is detected satisfy , When it reaches the noise deviation threshold, it is also determined to be interfered; Once it is determined that there is electromagnetic interference, safety measures are automatically triggered. When the intensity of the electromagnetic interference exceeds a certain severity threshold When the automatic aiming function is directly suspended, or a warning signal is sent to the user, the electromagnetic interference intensity Through comprehensive evaluation of the degree of signal abnormality, , is the weight coefficient, , executes the operation of pausing automatic aiming and prompting the user to intervene manually.

3. The intelligent aiming and shooting method according to claim 1, wherein: The environment perception module is established to collect surrounding environment data, and the environmental data is transmitted to the environment analysis engine. The engine uses a deep learning model to determine the type and complexity of the environment and identify and target the target. Specifically, it includes: The target recognition algorithm formula is based on the convolutional neural network of deep learning. The calculation process involves the operations of convolution, pooling and full connection layers. When the input image is , whose size is , Indicates height, Indicates width, Indicates the number of channels; The convolution kernel is , the size is , the height is , width is , the step size is , filled with , the convolutional layer outputs feature maps The size calculation formula is: ； ； Output feature map Each element of The calculation of is obtained through convolution operation: ； The pooling kernel size of the maximum pooling is , the step size is , output feature map after pooling The size calculation formula is: ； ； Output feature map Each element of It takes the maximum value within the pooling window: ； After multiple convolution and pooling layers, the resulting feature map is flattened into a vector , the output of the fully connected layer Through the weight matrix and the bias vector Calculation yields: ,in For target recognition, the output of the last fully connected layer is usually the category probability distribution, and the target category is determined by comparing the probability size. Aiming algorithm formula, based on the basic principles of exterior ballistics, the distance to the known target is , the initial velocity of the bullet is , the gravitational acceleration of the bullet is , Pick , without considering the influence of other factors, in the horizontal direction and vertical direction The ballistic equation on is: Horizontal direction: ,in is the horizontal component of the bullet's initial velocity, is the flight time; Vertical direction: ,in is the vertical component of the bullet's initial velocity. When the wind direction and speed are considered, the wind speed is , the angle between wind direction and shooting direction is , the horizontal velocity needs to be corrected: ; Finally, the intelligent aiming system combines the initial posture of the gun, the pitch angle and yaw angle After obtaining the gun posture data through the sensor, the aiming point is corrected accordingly. The correction amount of the aiming point in the horizontal and vertical directions is and The final aiming point position is determined based on the above ballistic equation and gun posture.

4. An intelligent aiming and shooting system for implementing the method of claims 1-3, characterized in that: Includes software update module, sensor fusion module, central processing module and environmental perception module; The software update module is connected to a sensor fusion module, the sensor fusion module is connected to a central processing module, and the central processing module is connected to an environment perception module.

Images