Directional sound wave dispersing method and system

Abstract

The invention discloses a directional sound wave dispersing method and system, and relates to the technical field of security and protection, and the method comprises the steps: recognizing a target in a real-time image; determining a target angle of the target relative to the carrying platform; controlling the sound production array to generate a plurality of beams of directional sound waves, wherein each beam of directional sound wave points to one target; analyzing the dispersing effect; if the dispersing effect is not good, the threat level of the target is evaluated, and a preferential dispersing target is determined by considering the rotation time of rotating the loudspeaker and the threat level of each target; and rotating the loudspeaker and the sound production array to respectively emit dispersion sound waves and directional sound waves to the preferential dispersion target. Directional sound waves are generated by adopting a phased array technology, dispersion is performed without rotating a loudspeaker, the dispersion response speed is improved, the dispersion process is divided into primary dispersion and advanced dispersion, the rotating loudspeaker and the sound production array are used for joint dispersion at the same time during advanced dispersion, and the dispersion effect of a high-threat target is ensured.

Description

Technical Field

[0001] This application relates to the field of security technology, and in particular to a method for repelling crowds or birds using directional sound waves. Background Technology

[0002] In security operations, certain targets require specific methods to disperse or remove them to prevent these potentially dangerous targets from causing further serious consequences. For example, near airports, it is necessary to scare away birds in the air to prevent bird strikes. Similarly, during holidays or gatherings, unexpected events can easily lead to crowds gathering and potentially causing stampedes, requiring the dispersal of the gathered crowds.

[0003] Among the many dispersal methods, acoustic dispersal is a feasible approach. By installing speakers on platforms such as helicopters and drones, directional sound wave transmission can be achieved by adjusting the orientation of the speakers, thereby achieving a directional dispersal effect. This directional dispersal technology only acts on the target that needs to be dispersed, and has little or no effect on other objects, greatly improving the efficiency of energy utilization.

[0004] However, current methods for directional dispersal using rotatable loudspeakers are limited by the loudspeaker's rotation speed and the number of loudspeakers, making it impossible to achieve rapid and accurate dispersal of multiple targets. In particular, it is impossible to quickly adjust the loudspeaker's direction to achieve real-time dispersal after the platform's attitude changes. Summary of the Invention

[0005] This application provides a directional acoustic wave dispersal method and system to solve the problems in speed and quantity that exist in the prior art when dispersing using rotatable loudspeakers.

[0006] On one hand, embodiments of this application provide a directional acoustic wave dispersal method, including: Acquire real-time images and identify targets within them; Determine the three-dimensional spatial position of each target relative to the platform, and determine the target angle based on the three-dimensional control position; The sound array is controlled to generate multiple directional sound waves, each pointing to a target. The sound array is fixedly installed at the bottom of the mounting platform and includes multiple fixed speakers. By delaying the sound waves of each fixed speaker, the generated sound waves are superimposed to form directional sound waves. Continuously acquire dispersal images after directional sound wave emission, and analyze to determine the dispersal effect; If the dispersal effect is not good, the threat level of each target in the dispersal image is assessed, taking into account the rotation time of the rotating speaker and the threat level of each target, to determine the priority targets to be dispersed; the rotating speaker is mounted at the bottom of the platform, and after rotating the speaker in the direction of rotation and aiming it at the target, it can emit dispersal sound waves towards the target. The rotating loudspeaker and the sound array emit dispersive sound waves and directional sound waves to the priority target to be dispersed, respectively. For other targets, a sound array emits directional sound waves to each target.

[0007] On the other hand, embodiments of this application also provide a directional acoustic wave dispersal system, including: A camera is used to capture real-time images; The data processing module is used to identify targets in real-time images, determine the three-dimensional spatial position of each target relative to the mounting platform, and determine the target angle based on the three-dimensional control position. The sound array is fixedly installed at the bottom of the mounting platform. The sound array includes multiple fixed speakers. After delaying the sound waves of each fixed speaker, the resulting sound waves are superimposed to form directional sound waves. The speaker is rotated and mounted on the bottom of the platform. After rotating the speaker and pointing it at the target, it can emit dispersive sound waves towards the target. The acoustic control unit is used to control the acoustic array to generate multiple directional acoustic waves, each of which is directed at a target. The camera continuously captures dispersal images after the directional sound waves are emitted. The data processing module analyzes and determines the dispersal effect. If the dispersal effect is not good, the data processing module assesses the threat level of each target in the dispersal image, taking into account the rotation time of the rotating speaker and the threat level of each target, and determines the priority targets to be dispersed. The sound wave control unit controls the rotating speaker and the sound array to emit dispersal sound waves and directional sound waves to the priority targets to be dispersed, respectively. For other targets, the sound wave control unit controls the sound array to emit directional sound waves to each target.

[0008] The directional acoustic wave dispersal method and system disclosed in this application have the following advantages: Using phased array technology to generate directional sound waves eliminates the need for rotating loudspeakers to disperse targets, thus improving the dispersal response speed. Furthermore, the dispersal process is divided into initial dispersal and advanced dispersal. During advanced dispersal, rotating loudspeakers and sound arrays are used simultaneously to jointly disperse priority targets, ensuring the dispersal effect on high-threat targets. For targets other than priority targets, the sound array is used to disperse them, ensuring that all targets are dispersed in a timely manner. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of a directional acoustic wave dispersal method provided in an embodiment of this application. Detailed Implementation

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0012] Figure 1 A flowchart illustrating a directional acoustic wave dispersal method provided in this application embodiment. This application embodiment provides a directional acoustic wave dispersal method, including: S100: Acquire real-time images and identify targets in the real-time images.

[0013] For example, after acquiring a real-time image, the real-time image is sequentially processed by noise reduction, grayscale conversion, and edge enhancement, and then the target is identified using a target recognition model.

[0014] Specifically, noise reduction can be achieved using Gaussian filtering to remove environmental noise, edge enhancement can be achieved using the Sobel operator to highlight the target contour, and grayscale conversion is used to simplify calculations.

[0015] When identifying a target, a target recognition model (such as YOLOv8) needs to be invoked to classify the target and annotate the bounding box on the preprocessed real-time image, outputting the target type C and the target's pixel coordinates in the image (u). i ,v i For example, C=1 represents birds, C=2 represents crowds gathered on the ground, and C=3 represents other objects that need to be dispersed. Pixel coordinates (u i ,v i In the diagram, i represents the target number (i = 1, 2, ..., n, where n is the total number of targets identified), and u... i and v i These are the horizontal and vertical pixel coordinates (in pixels: px) of the i-th target, respectively.

[0016] In the embodiments of this application, the platform needs to be initialized before it is put into formal use to ensure that each device is in normal working condition.

[0017] Specifically, the initialization operation includes the following: (1) Platform startup and self-test Start the helicopter / drone's power system and flight control system, complete attitude calibration (such as hovering stability and flight path control accuracy), and ensure that the platform can operate stably in the target area (such as the area around the airport or the assembly site).

[0018] The self-test platform's power supply system was verified to provide a continuous and stable power supply (e.g., voltage fluctuation ≤ ±5%) to the acoustic wave emitting device (rotating speaker, sound array), camera, and data processing module.

[0019] (2) Core equipment functions and parameter presets Camera: Activate the high-definition industrial camera (resolution ≥1920×1080, frame rate ≥25fps). Preset parameters include focal length (adjust according to the monitoring distance, such as f=12mm for 10-50m monitoring), exposure time (avoid overexposure in strong light or underexposure in weak light), and white balance to ensure clear image acquisition.

[0020] Sound Array: Initialize the hardware status of the sound array (such as whether the unit speakers are normal and whether the beamforming module is activated), preset the initial power P0 (set according to the target type, such as bird dispersal P0=50dB, crowd dispersal P0=70dB, where dB is the sound pressure level), and beam pointing accuracy (≤±1°, to ensure directionality).

[0021] Rotate the speaker: Self-test the speaker's rotation mechanism (such as a stepper motor or servo motor), confirm the rotation angular velocity ω (inherent parameter, such as ω=15° / s, i.e., 15 degrees per second), and the maximum rotation range (azimuth 0-360°, pitch -30° to +60°). Calibrate the speaker's initial pointing angle to be directly below the mounting platform.

[0022] Data processing module: Start the embedded computing unit (such as an edge AI chip), load the target recognition model (such as YOLOv8), relative position calculation algorithm, genetic algorithm and power optimization program to ensure that the real-time data processing latency is ≤100ms.

[0023] After the platform is initialized, the camera will continuously capture real-time images of the ground / air below or around the platform at a preset frame rate of 25fps. The real-time images will be transmitted to the data processing module in real time (transmission bandwidth ≥100Mbps to avoid lag).

[0024] S110, determine the three-dimensional spatial position of each target relative to the mounting platform, and determine the target angle based on the three-dimensional control position.

[0025] For example, after the target is identified, an image coordinate system, a body coordinate system, and a target coordinate system are established. The target coordinate system is parallel to the body coordinate system. The coordinates of the target in the image coordinate system are converted into three-dimensional coordinates in the body coordinate system using a pinhole camera model. Then, the azimuth and pitch angles of the target relative to the mounting platform are calculated based on the three-dimensional coordinates. The azimuth and pitch angles together form the target angle.

[0026] In the embodiments of this application, the image coordinate system has its origin at the upper left corner of the camera imaging plane, with the u-axis pointing horizontally to the right and the v-axis pointing vertically downward (unit: px); the body coordinate system has its origin O at the center of gravity of the mounting platform, with the X-axis along the platform's forward direction, the Y-axis pointing horizontally to the right, and the Z-axis pointing vertically downward (unit: m); and the target coordinate system has its origin O at the geometric center of each target i. i This coordinate system is parallel to the body coordinate system (unit: m).

[0027] The pixel coordinates of the target are converted into three-dimensional coordinates in the body coordinate system using a pinhole camera model (x). i , y i ,z i The formula is as follows:

[0028] Where u0 and v0 are the coordinates of the camera's principal point, i.e., the center pixel coordinates of the imaging plane, obtained from camera calibration, in pixels (px). i ,y i ,z i Let x be the three-dimensional coordinates of target i in the body coordinate system, where x i Let y be the distance along the direction of travel of the platform. i z represents the horizontal distance. i The vertical downward distance is expressed in meters (m). K is the camera intrinsic parameter matrix, in the form of:

[0029] Among them, f x =f / dx、f y =f / dy, where f is the camera focal length in meters (m); dx and dy are the physical dimensions of a single pixel in meters (m / px).

[0030] Based on three-dimensional coordinates (x) i , y i , z i ), calculate the azimuth angle θ of target i relative to the mounting platform. i (Angle between the horizontal plane and the X-axis) and pitch angle (The angle between the vertical plane and the X-axis) serves as the angular basis for the directionality of sound waves:

[0031]

[0032] θ i , Converting to angles (1 rad = (180 / π)°), we obtain the composite pointing angle α of target i. i The composite pointing angle is the total angle combining the azimuth and elevation angles, in degrees.

[0033] S120 controls the sound-generating array to generate multiple directional sound waves, each directional sound wave pointing to a target; the sound-generating array is fixedly installed at the bottom of the mounting platform, and the sound-generating array includes multiple fixed speakers. After delaying the sound waves of each fixed speaker, the generated sound waves are superimposed to form directional sound waves.

[0034] For example, the above operation is the initial dispersion. Through the initial dispersion, the beamforming advantage of the sound array can be used to disperse multiple targets simultaneously, preventing these targets from moving further or gathering near the critical area.

[0035] Specifically, based on the angle θ of each target i i , By using array signal processing algorithms such as delay-sum beamforming, the sound-generating array generates multiple directional sound waves, with the main lobe of each sound wave pointing towards the corresponding target i at (θ). i , Furthermore, the sidelobe suppression ratio of the directional acoustic wave is ≥20dB to avoid interference with non-target areas.

[0036] Furthermore, when the acoustic array initially disperses the target, it emits directional sound waves using an initial power P0, which is set according to the target's category.

[0037] If C=1 (birds): P0=45-60dB, sound frequency 2-5kHz, this is the sensitive frequency range for birds. If C=2 (humans): P0=65-80dB, sound frequency 500-2000Hz, a frequency range that is perceptible to humans and harmless. Simultaneously, the power application time, i.e., the continuous working time of the sound array during the initial dispersal, is 30-60s, which is sufficient time to observe the effect of the initial dispersal.

[0038] S130 continuously acquires dispersion images after emitting directional sound waves and analyzes them to determine the dispersion effect.

[0039] For example, the method for analyzing the dispersing effect includes: Set performance evaluation indicators, including the rate of change in the target quantity and the rate of change in the target distance; The effectiveness evaluation indicators are compared with the pre-set effectiveness thresholds, and the dispersing effect is determined to be good or bad based on the comparison results.

[0040] Specifically, the rate of change of the target quantity ΔN i The formula used to assess whether the target has been reduced is as follows:

[0041] Where, N i,t0 Let N be the number of target i at time t0 (before the initial dispersal). i,t1 Let ΔN be the number of target i at time t1 (after the initial dispersal); i <0 indicates a reduction in the target, ΔN i ≤-30% indicates a significant decrease in quantity.

[0042] Target distance change rate ΔD i The formula used to assess whether a target is far from the platform (or critical area) is as follows:

[0043] Among them, D i,t0 Let D be the straight-line distance (in meters) between target i and the platform (or the center of the critical area) at time t0. i,t1 The distance at time t1; ΔD i >0 indicates that the target is moving away, ΔD i ≥20% indicates a significant distance.

[0044] For all objectives i, if ΔN i ≤-30% and ΔD i If the success rate is ≥20%, the effect is considered good, and continued monitoring is sufficient without further action. If at least one target i does not satisfy ΔN... i ≤-30% or ΔD i If the percentage is ≥20%, the effect is considered unsatisfactory, and the process proceeds to the advanced dispersal stage, i.e., steps S140-S160.

[0045] S140, if the dispersal effect is not good, the threat level of each target in the dispersal image is evaluated, and the rotation time of the rotating speaker and the threat level of each target are taken into account to determine the priority targets to be dispersed; the rotating speaker is mounted at the bottom of the mounting platform, and after the rotating speaker is rotated and aligned with the target, it can emit dispersal sound waves towards the target.

[0046] Examples of methods for assessing threat levels include: Set threat assessment metrics, including target size, target distance factor, and target activity. The threat level is obtained by weighting and summing the threat assessment indicators with their respective weight coefficients.

[0047] Specifically, for the set of targets S that are not effective, S={i|target i is not effective}, |S|≥1, they are sorted according to the dimension of threat level to provide a basis for priority dispersal.

[0048] Three threat assessment indicators were selected and normalized to the [0,1] interval. The three threat assessment indicators are as follows.

[0049] Target size S i : Represents the proportion of target i, as shown in the following formula:

[0050] Where, N j,t1 Let t1 be the number of targets j in the target set S.

[0051] Target distance factor D' i : Represents the reciprocal of the distance between target i and the critical region, as shown in the following formula:

[0052] Among them, D min This refers to the warning radius of a critical area, such as the warning radius D of an airport runway. min =50m; if D i,t1 ≤D min Then D' i =1 ； If D i,t1 ≥2D min ,D' i =0.

[0053] Target Activity A i : Represents the movement trend of target i, as shown in the following formula:

[0054] Among them, v i,toward v is the speed at which target i approaches the platform or critical area, in m / s. max The maximum possible speed at which a target can approach, such as the v of a bird. max =10m / s, the v of the crowd max =2m / s; A i =0 indicates distance, A i =1 indicates a rapid approach.

[0055] The threat level T of target i is calculated using a weighted summation model. i T i ∈[0,1], the larger the value, the higher the threat:

[0056] Among them, w1, w2, and w3 are weight coefficients that satisfy w1+w2+w3=1. They are obtained by training from expert experience or historical data. For example, in the airport scenario: bird threat distance is prioritized, w2=0.5, w1=0.3, w3=0.2; in the crowd scenario: size is prioritized, w1=0.5, w2=0.3, w3=0.2.

[0057] Furthermore, methods for determining priority targets for dispersal include: Determine the time required for the rotating loudspeaker to rotate toward each target within its rotational speed limit; Using the maximum difference between threat level and rotation time as the objective function, a genetic algorithm is used to solve the objective function, and the target that maximizes the objective function is selected as the priority target to be dispersed.

[0058] Specifically, rotating the speaker requires turning it from the current pointing angle α0 (initially 0°) to the resultant pointing angle α of target i. i Calculate the rotation time t i , as a time cost indicator for priority dispersal.

[0059] The rotational angular velocity ω of the rotating speaker is an inherent parameter, and the rotation time t... i The ratio of angular difference to angular velocity:

[0060] in, Indicates taking the absolute value, t i Rotation time, in seconds.

[0061] Overall Threat Level T i (Maximize) and rotation time t i (Minimize), and use a genetic algorithm to select the priority target i* to ensure efficient use of resources.

[0062] In genetic algorithms, the population consists of one target ID (i∈S) for each individual, and the population size is M=20 (adjusted according to |S|, M≥|S|). The fitness function is used to measure the priority of an individual (i.e., target i), and the formula is as follows:

[0063] Where F(i) is the fitness of target i, and a and b are weight coefficients, where a>0 and b>0, such as a=10 and b=1, used to balance the impact of threat and time; the larger F(i) is, the higher the priority of target i. Iteration parameters include the maximum number of iterations G=50 and the crossover probability P. c =0.8, mutation probability P m =0.1.

[0064] The steps of a genetic algorithm are as follows: (1) Initialize the population: randomly generate M individuals to form the initial population; (2) Calculate fitness: For each individual i, substitute into the formula F(i) to calculate fitness; (3) Selection operation: Two individuals are randomly selected using the tournament selection method, and the individuals with higher fitness are retained. This process is repeated M times to form a new population. (4) Crossover operation: For P in the new population c For each individual pair in the specified ratio, exchange their target IDs to generate offspring individuals; (5) Mutation operation: For P in the population m The proportion of individuals is randomly replaced with the target numbers that were not selected in S; Iteration termination: Repeat steps (2)-(5) until the number of iterations reaches G, and output the individual with the highest fitness, which is the priority target i* to be dispelled.

[0065] S150 uses a rotating loudspeaker and a sound array to emit dispersing sound waves and directional sound waves to the priority dispersing target, respectively.

[0066] For example, when the rotating loudspeaker and the sound array emit dispersing sound waves and directional sound waves to the priority dispersing target respectively, an optimization function is established with the minimum total power as the optimization objective. Under the constraints, the particle swarm optimization algorithm is used to solve the optimization function to obtain the optimal power when the rotating loudspeaker and the sound array emit sound waves. The rotating loudspeaker and the sound array emit dispersing sound waves and directional sound waves to the priority dispersing target with their respective optimal power. The constraints include dispersing effect constraints, power upper limit constraints, and power lower limit constraints.

[0067] Specifically, the optimization objective is to minimize the total power P. total ,Right now:

[0068] Among them, P s To increase the power of the loudspeaker, P a This represents the power of the sound-generating array.

[0069] The dispersal effect constraint is that for the priority dispersal target i*, ΔD' i* ≥ΔD min That is, it needs to be significantly far away, where ΔD' i* ΔD represents the rate of change of distance after the combined dispersion of the rotating loudspeaker and the sound array. min =30%. Power upper limit constraint is P s ≤P s_max P s_max To rotate the speaker to its maximum power, such as P s_max=100dB, P a ≤P a_max P a_max The maximum power of a single beam of the sound-emitting array, such as P a_max =90dB. The lower power limit constraint is P. s ≥P s_min P s_min The minimum effective power required to rotate the loudspeaker, such as P s_min =40dB, P a ≥P a_min P a_min The minimum effective power of the sound-generating array, such as P a_min =30dB.

[0070] The process of Particle Swarm Optimization (PSO) is as follows: (1) Particle coding: Each particle corresponds to a set of (P s , Pa), with a particle dimension of 2; (2) Fitness function: If the constraints are satisfied, fitness = 1 / P total (That is, the smaller the total power, the higher the fitness); if this condition is not met, the fitness is set to 0. (3) Iterative solution: Set the number of particles = 30, the maximum number of iterations = 20, iterate until the fitness converges, and output the optimal power P. s* , P a* .

[0071] After determining the optimal power for the rotating loudspeaker and the sound array, the rotating loudspeaker is rotated to the resultant directivity angle α of target i. i* (Time consumed t) i* ), with optimal power P s* The sound wave is emitted in a directional manner, and the sound array adjusts the beam direction to α. i* With optimal power P a* The dispersing sound wave is emitted, and the combined dispersing time is t2=60s.

[0072] S160, for other targets, the sound array emits directional sound waves to each target.

[0073] For example, when the sound-emitting array emits directional sound waves to each other target, the remaining power is allocated to each target according to the threat level of each target, and the sound-emitting array emits directional sound waves to each target with the corresponding allocated power.

[0074] Specifically, for targets j (j∈S, j≠i*) other than i* in the target set S, multiple directional sound waves are generated using the remaining power of the sound-emitting array to achieve synchronous dispersal.

[0075] Multi-beam power allocation: based on the threat level T of target j jDistribute the power P of the sound-emitting array aj :

[0076] Among them, P a,residue =P a_max -P a* , representing the remaining usable power of the sound-generating array, in dB; P aj ≥P a_min To ensure effective dispersal.

[0077] Multi-beam directional execution: The acoustic array generates a directional sound wave corresponding to target j, and the resulting pointing angle α points towards the target. j With power P aj A directional sound wave is emitted for a duration consistent with the combined dispersal time, i.e., 60 seconds.

[0078] After completing the advanced dispel, repeat step S140 to determine the dispel effect of all targets in S. If there are still targets with poor effects, update the target set S and repeat steps S150-S160 until the dispel effect of all targets meets the standard, or the platform reaches the maximum battery life / mission time limit, and end the dispel mission.

[0079] This application embodiment also provides a directional acoustic wave dispersal system, the system comprising: A camera is used to capture real-time images; The data processing module is used to identify targets in real-time images, determine the three-dimensional spatial position of each target relative to the mounting platform, and determine the target angle based on the three-dimensional control position. The sound array is fixedly installed at the bottom of the mounting platform. The sound array includes multiple fixed speakers. After delaying the sound waves of each fixed speaker, the resulting sound waves are superimposed to form directional sound waves. The speaker is rotated and mounted on the bottom of the platform. After rotating the speaker and pointing it at the target, it can emit dispersive sound waves towards the target. The acoustic control unit is used to control the acoustic array to generate multiple directional acoustic waves, each of which is directed at a target. The camera continuously captures dispersal images after the directional sound waves are emitted. The data processing module analyzes and determines the dispersal effect. If the dispersal effect is not good, the data processing module assesses the threat level of each target in the dispersal image, taking into account the rotation time of the rotating speaker and the threat level of each target, and determines the priority targets to be dispersed. The sound wave control unit controls the rotating speaker and the sound array to emit dispersal sound waves and directional sound waves to the priority targets to be dispersed, respectively. For other targets, the sound wave control unit controls the sound array to emit directional sound waves to each target.

[0080] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0081] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for directional acoustic wave dispersion, characterized in that, include: Acquire real-time images and identify targets in the real-time images; Determine the three-dimensional spatial position of each target relative to the mounting platform, and determine the target angle based on the three-dimensional control position; The sound-emitting array is controlled to generate multiple directional sound waves, each of which points to a target. The sound-emitting array is fixedly installed at the bottom of the mounting platform and includes multiple fixed speakers. The directional sound waves are superimposed to form the sound waves after the sound waves of each fixed speaker are delayed. Continuously acquire dispersal images after the directional sound waves are emitted, and analyze and determine the dispersal effect; If the dispersal effect is not good, the threat level of each target in the dispersal image is evaluated, taking into account the rotation time of the rotating speaker and the threat level of each target, to determine the priority targets to be dispersed; the rotating speaker is rotatably mounted at the bottom of the mounting platform, and after the rotating speaker is rotated and aligned with the target, it can emit dispersal sound waves towards the target; The rotating loudspeaker and the sound array respectively emit the dispersing sound wave and the directional sound wave toward the priority dispersing target; For the other targets, the sound-emitting array emits directional sound waves to each target.

2. The directional acoustic wave dispersal method according to claim 1, characterized in that, After acquiring the real-time image, the real-time image is sequentially processed by noise reduction, grayscale conversion and edge enhancement, and then the target is identified using a target recognition model.

3. The directional acoustic wave dispersion method according to claim 1, characterized in that, After the target is identified, an image coordinate system, a body coordinate system, and a target coordinate system are established. The target coordinate system is parallel to the body coordinate system. The coordinates of the target in the image coordinate system are converted into three-dimensional coordinates in the body coordinate system using a pinhole camera model. Then, the azimuth and pitch angles of the target relative to the mounting platform are calculated based on the three-dimensional coordinates. The azimuth and pitch angles together form the target angle.

4. The directional acoustic wave dispersion method according to claim 1, characterized in that, When the sound array initially disperses the target, it emits the directional sound wave using an initial power, which is set according to the type of the target.

5. The directional acoustic wave dispersion method according to claim 1, characterized in that, The analysis method for the dispersing effect includes: Set performance evaluation indicators, including the rate of change of target quantity and the rate of change of target distance; The effect evaluation index is compared with a pre-set effect threshold, and the dispersing effect is determined to be good or bad based on the comparison result.

6. The directional acoustic wave dispersion method according to claim 1, characterized in that, Methods for assessing the threat level include: Define threat assessment metrics, including target size, target distance factor, and target activity. The threat level is obtained by weighting and summing the threat assessment indicators with their respective weighting coefficients.

7. The directional acoustic wave dispersion method according to claim 1, characterized in that, The method for determining the priority targets to be dispersed includes: Determine the rotation time required for the rotating loudspeaker to rotate toward each of the targets, within the limit of its rotation speed. Using the maximum difference between the threat level and the rotation time as the objective function, a genetic algorithm is used to solve the objective function to obtain the target that maximizes the objective function, which is then used as the priority target to disperse.

8. The directional acoustic wave dispersion method according to claim 1, characterized in that, When the rotating loudspeaker and the sound-emitting array emit the dispersing sound wave and the directional sound wave respectively towards the priority dispersing target, an optimization function is established with the minimum total power as the optimization objective. Under the constraints, the particle swarm optimization algorithm is used to solve the optimization function to obtain the optimal power when the rotating loudspeaker and the sound-emitting array emit sound waves. The rotating loudspeaker and the sound-emitting array emit the dispersing sound wave and the directional sound wave towards the priority dispersing target with their respective optimal power. The constraints include dispersing effect constraints, power upper limit constraints, and power lower limit constraints.

9. A directional acoustic wave dispersion method according to claim 1, characterized in that, When the sound-emitting array emits the directional sound wave at each of the other targets, it allocates the remaining power to each target according to the threat level of each target, and the sound-emitting array emits the directional sound wave to each target with the corresponding allocated power.

10. A system applying the directional acoustic wave dispersal method according to any one of claims 1-9, characterized in that, include: A camera is used to capture real-time images; The data processing module is used to identify targets in the real-time image, determine the three-dimensional spatial position of each target relative to the mounting platform, and determine the target angle based on the three-dimensional control position; A sound-emitting array is fixedly installed at the bottom of the mounting platform. The sound-emitting array includes multiple fixed speakers. After delaying the sound waves of each fixed speaker, the generated sound waves are superimposed to form a directional sound wave. A rotating speaker is mounted on the bottom of the mounting platform. After the rotating speaker is rotated and aligned with the target, it can emit dispersive sound waves toward the target. A sound wave control unit is used to control the sound-generating array to generate multiple beams of the directional sound waves, each beam of the directional sound waves being directed at a target; The camera continuously captures dispersal images after the directional sound waves are emitted. The data processing module analyzes and determines the dispersal effect. If the dispersal effect is not good, the data processing module evaluates the threat level of each target in the dispersal image, and takes into account the rotation time of the rotating speaker and the threat level of each target to determine the priority targets to be dispersed. The acoustic control unit controls the rotating loudspeaker and the sound-emitting array to emit the dispersing sound wave and the directional sound wave respectively towards the priority dispersing target; for the other targets, the acoustic control unit controls the sound-emitting array to emit the directional sound wave towards each target.

Images