Multi-excitation source water wave generation control method and system based on image recognition analysis

Abstract

A multi-excitation source water wave generation control method and system based on image recognition analysis, the method comprising: real-time acquisition of water surface images by an image acquisition device and preprocessing, and demarcation of a region of interest according to the positions of each wave source; analysis of time sequence image sequences in each region to extract the instantaneous phase of water waves at each wave source; calculation of the measured phase difference of the remaining wave sources relative to a reference benchmark with one wave source as the reference benchmark; calculation of the excitation timing compensation amount required for each wave source according to the deviation of the preset target phase relationship and the measured phase difference, combined with the current working frequency; generation of multi-channel control signals according to the timing compensation amount to drive the actuators of each wave source to execute the excitation action at the corresponding compensation time; and continuous repetition of the above steps to form a real-time closed-loop feedback control. The present application realizes flexible and accurate control of the phases of multiple wave sources by real-time perception of the actual phase state of each wave source through visual feedback and the application of accurate timing compensation.

Description

Technical Field

[0001] This invention relates to water wave generation control technology, and in particular to a multi-source water wave generation control method and system based on image recognition analysis. Background Technology

[0002] Water wave generation control technology has wide applications in scientific research experiments, teaching demonstrations, and engineering testing. In a multi-source water wave generation system, multiple wave sources operate in a preset time sequence, forming specific wave surface patterns through water wave interference. The core control requirement is that each wave source must maintain a precise and stable phase relationship.

[0003] In existing technologies, multi-point water wave excitation sources typically employ open-loop control, meaning that each excitation source is triggered at a preset time point, lacking real-time monitoring and feedback of the actual water wave output state. When environmental conditions (such as water temperature, water depth, and water surface boundary conditions) change, the actual wave response of each point source may drift, leading to inaccurate preset phase relationships and wave surface interference patterns deviating from expectations.

[0004] At the mechanical level, the wave source of a small water wave generator typically does not have a fixed initial standby position (such as a zero-phase reference point). After each excitation, the initial state of the wave source may differ slightly, and in a multi-source system, these differences can accumulate and destroy the desired wavefront interference pattern.

[0005] In addition, in some existing solutions, each wave source is equipped with an integrated built-in control system, which can independently adjust the amplitude and frequency of a single device. However, there is a lack of a unified phase reference and time synchronization mechanism between the devices, making it impossible to generate water waves with the same phase or a fixed phase difference through multiple devices working together.

[0006] Another existing solution uses mechanical linkage to achieve synchronization of multiple wave sources. Although it can ensure a fixed phase relationship, it sacrifices the flexibility and reconfigurability of the system—it cannot dynamically switch between synchronous / asynchronous modes during operation, nor can it flexibly adjust the relative phase between each wave source.

[0007] It should be noted that the information disclosed in the background section above is only for understanding the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0008] The main objective of this invention is to overcome the deficiencies in the aforementioned background technology and provide a multi-excitation source water wave generation control method and system based on image recognition analysis.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: A multi-source water wave generation control method based on image recognition analysis includes the following steps: S1. Real-time image acquisition of the water surface is performed using an image acquisition device; S2. Preprocess the acquired images and delineate the corresponding regions of interest based on the known locations of each wave source; S3. Analyze the time-series image sequences in each region of interest and extract the instantaneous phase of the water waves at each wave source location; S4. Using one of the wave sources as the phase reference, calculate the measured phase difference of the other wave sources relative to the reference. S5. Based on the deviation between the preset target phase relationship and the measured phase difference, and in conjunction with the current operating frequency, calculate the required excitation timing compensation for each wave source; S6. Generate multi-channel control signals according to the timing compensation amount, and drive each wave source actuator to perform excitation actions at the corresponding compensation time; S7. Repeat steps S1 to S6 continuously to form real-time closed-loop feedback control.

[0010] A multi-source water wave generation control system based on image recognition analysis includes: An image acquisition device is set above or to the side of the water surface to acquire real-time images of the water surface, including the corresponding areas of each wave source. A calibration plate, fixed below the water surface or at the bottom of the tank, provides a reference texture for the image acquisition device. The image processing and phase calculation unit is connected to the image acquisition device and is used to preprocess the acquired image, calculate the optical flow displacement field, reconstruct the water surface height field and extract the instantaneous phase, and calculate the deviation between the measured phase difference between each wave source and the target phase. The control unit, connected to the image processing and phase calculation unit, is used to calculate the excitation timing compensation amount of each wave source based on the deviation amount and the current operating frequency, and to generate multi-channel control signals. Multiple wave source actuators are connected to each output channel of the control unit, respectively, to receive the control signal and perform excitation actions at the corresponding compensation time; The image acquisition device, image processing and phase calculation unit, control unit, and wave source actuator constitute a real-time closed-loop feedback control loop.

[0011] The present invention has the following beneficial effects: This invention proposes a multi-source water wave generation control method based on image recognition analysis. It introduces computer vision technology into the multi-source water wave generation control task, using a non-contact water surface state acquisition device (such as a high-speed camera) to acquire water wave images of the corresponding areas of each wave source in real time. Image recognition algorithms (including optical flow, FFT spectral analysis, edge detection, etc.) are used to directly extract the instantaneous phase, amplitude, and frequency parameters of the water waves at each wave source location from the water surface image sequence, constructing a full-field perception capability for the multi-source water wave state. The control unit, based on the deviation between the measured phase and the target phase and combined with the operating frequency, converts the phase angle difference into an independent excitation delay time for each wave source. Differential and precise timing compensation is applied to each wave source through multi-channel control signals, enabling multiple excitation sources to generate synchronous, out-of-phase, or arbitrarily fixed phase difference water wave outputs without mechanical linkage. The above-mentioned “image acquisition → phase detection → deviation calculation → timing compensation → signal output → re-acquisition” constitutes a continuously operating closed-loop control loop, enabling the system to adapt to environmental changes (such as changes in water temperature leading to changes in wave velocity, changes in response delay due to aging of wave source actuators, changes in water surface boundary conditions, etc.), fundamentally solving the problem of phase drift that cannot be detected and corrected in open-loop control.

[0012] Compared to existing solutions, this invention has the following significant advantages. First, this invention requires no modification to the mechanical structure of the water wave generator. The image acquisition device does not have physical contact with the controlled wave source, and no additional inertial load or friction is introduced. It can serve as an "external" upgrade module for existing water wave generator systems, greatly reducing system modification costs. Simultaneously, this invention decomposes multi-source water wave control into two decoupled layers: the physical layer and the information layer. Each wave source is only responsible for executing the excitation action according to instructions, while the camera and processor are responsible for sensing the phase state and generating control commands—making the system adaptable to various existing water wave generators without requiring mechanical modifications to the wave source hardware. Second, this invention can achieve flexible switching of arbitrary phase relationships through pure software. By simply modifying the target phase parameters, the system can operate in multiple modes, including synchronous wave generation, anti-phase wave generation, wave generation with arbitrary fixed phase difference, and even dynamic phase difference changes according to a preset sequence, without downtime or hardware adjustments, overcoming the bottlenecks of mechanical linkage solutions in terms of flexibility and reconfigurability. Furthermore, visual feedback combined with a high-speed camera can achieve millisecond-level temporal resolution. For example, the inter-frame time interval of a standard 120fps high-speed camera is approximately 8.3ms, while that of a professional 1000fps high-speed camera is only 1ms. The corresponding phase resolution reaches 3.6° at a working frequency of 10Hz, meeting the accuracy requirements of most water wave applications. When the number of wave sources needs to be increased, only ROI analysis needs to be added to the image region corresponding to each new wave source; no additional sensing hardware is required. The increase in system complexity mainly lies in the computational load of the algorithm, which can be flexibly addressed by upgrading the processor, demonstrating good scalability.

[0013] In a preferred embodiment of the present invention, the total system control delay includes image exposure time, image sensor readout time, image data transmission time, image processing and phase calculation time, control signal generation and transmission time, and wave source actuator response time. Taking a frame rate of 200 frames per second as an example, the image frame interval is 5ms, the total transmission, processing, and calculation time of the image processing platform is approximately 10ms, the STM32 control parameter reception and timer loading time is less than 1ms, the actuator response time is approximately 5ms, and the total control delay is approximately 23ms. Compared to the 250ms water wave period at the target frequency of 4.00Hz, the delay accounts for approximately 9.2%, which meets the real-time requirements of closed-loop control. To compensate for this total delay, the control unit calculates the predicted phase before output. By increasing the camera frame rate, reducing the ROI area, or using GPU-accelerated phase detection, the total delay can be further controlled to around 10ms, thereby achieving effective control of water waves in the 1-20Hz range. Experiments have shown that without compensation, the phase difference between the two wave sources is about 60°. After applying trigger timing compensation and continuously updating the closed loop for 2-3 cycles, the phase difference between the two wave sources converges to within 5°, the overall phase error is reduced by about 90%, and the water surface forms a planar waveform that is closer to the ideal state.

[0014] Other beneficial effects of the embodiments of the present invention will be further described below. Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the water wave control system according to an embodiment of the present invention.

[0016] Figure 2 This is a comparison diagram of the water wave control experiment before and after wave source adjustment, as shown in this embodiment of the invention. Detailed Implementation

[0017] The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary and not intended to limit the scope and application of the present invention.

[0018] This invention aims to address the shortcomings of existing water wave generation systems in multi-source collaborative operation, such as open-loop control phase drift and lack of flexibility in mechanical linkage. It proposes a multi-source water wave generation control method and system based on image recognition analysis. This method uses a high-speed camera system to acquire water surface images in real time, employs visual algorithms to extract the instantaneous phase of the water waves at each wave source, and applies differentiated and precise timing compensation to each wave source via software. This enables multiple excitation sources to generate synchronous, out-of-phase, or arbitrarily fixed phase difference water wave outputs without mechanical linkage, achieving hardware and software decoupling for multi-source control. It offers advantages such as non-contact measurement, real-time closed-loop anti-interference, arbitrarily programmable phase relationships, and good scalability.

[0019] See Figure 1 This invention provides a multi-source water wave generation control method based on image recognition analysis, comprising the following steps: Step S1. Real-time image acquisition of the water surface is performed using an image acquisition device.

[0020] In some embodiments, the image acquisition device in step S1 employs a high-speed camera, preferably with a frame rate not less than twice the excitation frequency of the wave source, satisfying the Nyquist sampling theorem. The high-speed camera can be mounted above the water surface to capture images of the water surface from a top-down perspective; or it can be mounted on the side of the water surface to capture images of the water surface contour from a side-view perspective. For large-scale systems containing multiple wave sources, multiple cameras can be used to capture images in different areas, or the field of view of a single camera can be configured to cover the entire area where the wave sources are located.

[0021] Step S2. Preprocess the acquired images and delineate the corresponding regions of interest based on the known locations of each wave source.

[0022] In some embodiments, the preprocessing in step S2 may include noise reduction and contrast enhancement of the acquired raw image. The noise reduction may employ Gaussian filtering or median filtering to remove reflections from water surface impurities and sensor noise; the contrast enhancement is used to improve the visual distinction between wave peaks and troughs. The region of interest (ROI) is defined by delineating corresponding sub-regions in the image based on the known locations of each wave source, thereby reducing computational load and improving the signal-to-noise ratio. Since depth detection visual schemes near wave sources are severely affected by wave source structure during water wave propagation, the ROI is avoided being placed directly above the wave source, but rather positioned along the wave propagation direction at approximately half a wavelength from the wave source to obtain a stable and reliable wave height signal. In side-shot scenes, the water surface contour line can also be extracted during the preprocessing stage through edge detection.

[0023] Step S3. Analyze the time-series image sequences in each region of interest and extract the instantaneous phase of the water waves at each wave source location.

[0024] In some embodiments, the step S3 of extracting the instantaneous phase of water waves at each wave source location specifically includes: acquiring a reference image frame in a static water surface state; calculating the optical flow displacement field between each real-time acquired wave image frame and the reference image to obtain displacement components in two orthogonal directions; converting the two displacement components of the optical flow displacement field into water surface height gradient components in two orthogonal directions respectively; performing two-dimensional integral reconstruction on the height gradient components in the two directions to obtain the water surface height field of the current frame; extracting the average height value from the region of interest corresponding to each wave source as the representative height of the wave source to form a height time series of each wave source; and constructing an analytical signal and extracting the instantaneous phase of water waves at each wave source location after bandpass filtering the height time series.

[0025] In some embodiments, the step of converting the two directional displacement components of the optical flow displacement field into water surface height gradient components in two orthogonal directions specifically includes: fixing a calibration pattern below the water surface or at the bottom of the tank, and obtaining a reference image of the calibration pattern in a static water surface state; calculating the two directional displacement components of the apparent displacement field using the optical flow method based on the apparent displacement of the calibration pattern relative to the reference image in each frame of wave-bearing images; determining a scaling factor based on the air refractive index, water refractive index, and the distance from the calibration pattern plane to the static water surface, dividing the two directional displacement components by the scaling factor and inverting them to obtain the water surface height gradient components in two orthogonal directions.

[0026] In some embodiments, the step of performing two-dimensional integral reconstruction on the height gradient components in two directions to obtain the water surface height field of the current frame specifically includes: transforming the water surface height gradient components in two orthogonal directions to the frequency domain; performing a combination operation on the spectrum of the gradient components in each direction using wavenumber in the frequency domain to achieve integration; and then performing a two-dimensional inverse transformation to obtain the water surface height field of the current frame.

[0027] In some embodiments, the step of constructing an analytical signal and extracting the instantaneous phase after bandpass filtering the altitude time series specifically includes: setting the passband range of the bandpass filter according to the target operating frequency of the wave source, and performing bandpass filtering on the altitude time series of each wave source; constructing a complex analytical signal from the filtered altitude time series using Hilbert transform; and calculating the instantaneous phase of each wave source at each moment using the four-quadrant arctangent function based on the real and imaginary parts of the complex analytical signal.

[0028] Step S4. Using one of the wave sources as the phase reference, calculate the measured phase difference of the other wave sources relative to the reference.

[0029] In some embodiments, the reference reference in step S4 is selected as follows: the wave source that emits the second wave among all wave sources is used as the phase reference reference, that is, the reference wave source is determined after all ROIs have detected the highest amplitude value, and the delay time of the remaining wave sources is calculated relative to the reference in the form of a delayed signal; or a pre-set fixed wave source is used as the phase reference reference, preferably a wave source with proven robustness, and the other wave sources complete the delay with reference to this wave source. In addition, when the calculated delay time is negative (i.e., it needs to be triggered in advance), the triggering time of the wave source can be delayed to the next cycle, that is, one water wave cycle is added to the delay time, thereby avoiding the "early" problem.

[0030] Step S5. Based on the deviation between the preset target phase relationship and the measured phase difference, and in conjunction with the current operating frequency, calculate the required excitation timing compensation for each wave source.

[0031] In some embodiments, the calculation of the required excitation timing compensation for each wave source in step S5 specifically includes: performing a phase winding operation between the measured phase of each wave source and the reference phase to obtain the measured phase difference of each wave source; comparing the measured phase difference with a preset target phase difference to obtain the phase deviation; and dividing the phase deviation by the angular frequency to convert it into the excitation delay time in the time domain, wherein the angular frequency is 2π multiplied by the current operating frequency.

[0032] Step S6. Generate multi-channel control signals according to the timing compensation amount, and drive each wave source actuator to perform excitation actions at the corresponding compensation time.

[0033] In some embodiments, generating multi-channel control signals in step S6 specifically includes: obtaining the total system control delay, which includes the sum of image exposure time, image sensor readout time, image data transmission time, image processing and phase calculation time, control signal generation and transmission time, and wave source actuator response time; calculating the predicted phase after the total control delay based on the instantaneous phase of each wave source at the current moment and the current operating frequency; determining the timing compensation amount based on the deviation between the predicted phase and the target phase, and generating multiple trigger pulse signals through a hardware timer and sending them to each wave source actuator respectively.

[0034] Step S7. Repeat steps S1 to S6 continuously to form real-time closed-loop feedback control.

[0035] This invention also provides a multi-excitation source water wave generation control system based on image recognition analysis. The system includes an image acquisition device, a calibration board, an image processing and phase calculation unit, a control unit, and multiple wave source actuators.

[0036] The image acquisition device is positioned above or to the side of the water surface to acquire real-time images of the water surface, including the corresponding areas of each wave source. For large-scale multi-source systems, multiple cameras can be used to capture images in sections, or the field of view of a single camera can cover the entire wave source area.

[0037] The calibration plate is fixed below the water surface or at the bottom of the tank, preferably a checkerboard calibration plate, to provide a reference texture for the image acquisition device.

[0038] The image processing and phase calculation unit is connected to the image acquisition device and is used to preprocess the acquired images, calculate the optical flow displacement field, reconstruct the water surface height field, and extract the instantaneous phase. It also calculates the deviation between the measured phase difference and the target phase relationship between each wave source. Specifically, the image processing and phase calculation unit performs optical flow matching between each frame of the wave image and a reference image under static water surface conditions to obtain the two directional displacement components of the apparent displacement field of the checkerboard texture. Based on the difference between the air refractive index and the water refractive index, and the distance from the checkerboard to the water surface, a scaling factor is determined. The two directional displacement components of the apparent displacement field are then converted into water surface height gradient components in two orthogonal directions. The water surface height field is reconstructed through frequency domain integration, and the average height value is extracted from the region of interest corresponding to each wave source to form a height time series. The instantaneous phase of each wave source is extracted through bandpass filtering and Hilbert transform.

[0039] The control unit is connected to the image processing and phase calculation unit and is used to calculate the excitation timing compensation of each wave source based on the deviation and the current operating frequency, and generate multi-channel control signals.

[0040] The control unit uses one wave source as a reference, calculates the measured phase difference of each of the other wave sources relative to the reference, and compares the phase difference with the preset target phase relationship after phase winding operation to obtain the phase deviation. The phase deviation is divided by the angular frequency to convert it into the excitation delay time, and then drives each wave source actuator through multi-channel trigger pulses.

[0041] The control unit is further configured to: calculate the predicted phase after the delay ends by forward extrapolating the current instantaneous phase and the operating frequency based on the total system control delay, and output a timing compensation amount based on the predicted phase. The plurality of wave source actuators are respectively connected to each output channel of the control unit, and are used to receive the control signal and perform excitation actions at the corresponding compensation time. The wave source actuators can be any actuator suitable for water wave excitation, such as pneumatic, electromagnetic, or loudspeaker type.

[0042] The image acquisition device, image processing and phase calculation unit, control unit, and wave source actuator constitute a real-time closed-loop feedback control loop of "image acquisition → phase detection → deviation calculation → timing compensation → signal output → re-acquisition", which enables the system to adapt to environmental changes and fundamentally solves the problem of phase drift that cannot be detected and corrected in open-loop control.

[0043] The method and system of this invention introduce computer vision technology into the control of multi-source water wave generation. A high-speed camera acquires real-time images of the water surface, and image recognition algorithms such as optical flow and Hilbert transform are used to extract the instantaneous phase of the water waves at each source, achieving full-field perception of the multi-source water wave state in a non-contact manner. Based on this, according to the deviation between the measured phase and the target phase, the phase angle difference is converted into an independent excitation delay time for each wave source, and differentiated precise timing compensation is applied. This allows multiple excitation sources to generate synchronous, out-of-phase, or arbitrarily fixed phase difference water wave outputs without mechanical linkage. The aforementioned closed-loop control architecture of "perception → calculation → feedback" enables the system to adapt to environmental changes in real time and achieves hardware and software decoupling between wave source synchronization control and mechanical structure—achieving high-precision, programmable multi-source phase collaborative control without any hardware modification to existing water wave generators. Experimental verification shows (more details below) that after applying trigger timing compensation and continuous closed-loop updates, the phase difference between two wave sources can converge from 60° to within 5°, reducing the overall phase error by approximately 90%.

[0044] The following further describes the implementation methods, algorithm examples, and technical advantages of specific embodiments of the present invention.

[0045] The core innovative principle of this invention is as follows: A non-contact water surface state acquisition device acquires water wave images or equivalent height signals of the corresponding areas of each wave source in real time. A phase detection algorithm is used to obtain the measured phase of each wave source. The control unit compares the measured phase with the target phase and calculates the timing compensation amount for each wave source in the next trigger cycle. This allows multiple excitation sources to generate synchronous, out-of-phase, or arbitrarily fixed phase difference water wave outputs without mechanical linkage. Essential steps include image / equivalent signal acquisition, spatial calibration and ROI determination, phase detection, phase deviation calculation, timing compensation, and multi-channel trigger output. Steps such as image enhancement and anomaly removal can be selectively enabled depending on the scenario.

[0046] The workflow of this invention can be divided into the following seven steps, forming a complete closed-loop control circuit: Step 1: Image Acquisition A high-speed camera (with a frame rate no less than twice the excitation frequency of the wave source, satisfying the Nyquist sampling theorem) is used to photograph the water surface. The camera can be mounted above the water surface (overhead view), with a calibration board or calibration grid paper placed below the water surface to capture water wave information through the horizontal displacement of the grid; alternatively, it can be mounted on the side of the water surface (side view), obtaining water wave profile information through the undulations of the water surface contour. For large-scale multi-source systems, multiple cameras can be used to photograph in different areas, or the field of view of a single camera can cover the entire wave source area.

[0047] Step 2: Image Preprocessing

[0048] The acquired raw images undergo noise reduction processing (such as Gaussian filtering and median filtering to remove reflections from water surface impurities and sensor noise), contrast enhancement (to improve the visual distinction between wave peaks and troughs), and region of interest (ROI) extraction. ROI extraction delineates corresponding sub-regions for analysis in the image based on the known locations of each wave source, thereby reducing computational load and improving the signal-to-noise ratio.

[0049] Step 3: Phase Detection

[0050] The temporal image sequence within each ROI is analyzed to extract the temporal characteristic parameters of the water waves at that location. Based on the optical flow method—the optical flow field between adjacent frames is calculated, and the wave propagation direction and phase distribution are indirectly derived through the water surface motion vector.

[0051] Using a checkerboard pattern as a static reference background, the water level height change is inferred from the checkerboard's visual displacement caused by water surface disturbance. The basic relationship is as follows: water waves cause the local water surface normal to tilt; after light is refracted at the air-water interface, the checkerboard texture observed by the camera exhibits an apparent displacement relative to the waveless reference image; this apparent displacement field can be obtained using the optical flow method, further converted into a water surface height gradient, and integrated to obtain the water wave field height distribution, ultimately yielding phase information. The calibration process is as follows: 1. Fix the checkerboard calibration plate at the bottom of the tank or below the water surface, ensuring that the distance between the checkerboard plane and the still water surface is H.

[0052] 2. Acquire a reference image I0(x,y) under still water conditions as an undisturbed background.

[0053] 3. For each frame of wave image It(x,y), perform optical flow matching with the reference image I0(x,y) to obtain the apparent displacement field of the checkerboard texture.

[0054] 4. Based on the refractive index of air, n a Water refractive index n wBy establishing an approximate relationship between the displacement field and the water surface height gradient, and using the distance H from the water surface to the checkerboard pattern, the water wave height field of the current frame is obtained by performing a two-dimensional integral on the height gradient field.

[0055] 5. After completing the water wave field height distribution, obtain the representative height s of the region within the ROI. i Perform bandpass filtering and obtain the current complex amplitude A from the reference excitation frequency. i Thus, the phase information Φ is obtained.

[0056] This invention preferably employs a combined algorithm of "chessboard calibration + optical flow displacement field + water surface height field reconstruction + ROI time series phase extraction". For steady-state periodic waves, FFT spectral analysis can be further combined to improve the main frequency identification and noise resistance; for scenes requiring frame-by-frame closed-loop control, Hilbert transform is preferably used to extract the instantaneous phase. Edge detection is mainly used to extract the water surface contour height sequence in side-shot scenes, serving as a substitute or supplement to the optical flow method. Since most water wave generation requirements are based on sinusoidal waves, they can generally be assumed to be ideal sinusoidal waves.

[0057] The algorithm pseudocode is as follows: Input: Reference image I0, image sequence I(tk), frame rate f_cam, target frequency f0, ROI region Ωi 1. Set the system to time zero t0 and synchronize the camera and controller clocks. 2. For each frame of image Ik: 2.1 Record timestamp tk = t0 + k / f_cam 2.2 Calculate the optical flow field u = OpticalFlow(I0, Ik) 2.3 Calculate the height gradient based on the checkerboard refraction model: dh / dx = -ux / C dh / dy = -uy / C 2.4 Integrating the height gradient yields h(x,y,tk). 2.5 Extract the height signal for each wave source ROI: si(tk) = meanROI[h(x,y,tk)] 3. Perform bandpass filtering on si(t) 4. Construct the analytic signal zi(t) = si(t) + j·Hilbert[si(t)] 5. Calculate the phase Φi(t) = atan2(Im[zi(t)], Re[zi(t)]). 6. Calculate the phase difference ΔΦi = wrap[Φi - Φref] 7. Calculate the deviation ei based on the target phase. 8. Calculate the trigger compensation time Δti = ei / (2πf0) 9. Output Δti to the control unit to perform multi-channel trigger compensation. Step 4: Phase difference calculation After obtaining the instantaneous phase values ​​at the locations of each wave source, the phase difference between the different wave sources is calculated. For a system with N wave sources, one wave source can be selected as the phase reference, and the phase difference ΔΦ of the remaining N-1 wave sources relative to this reference can be calculated. i Phase difference is expressed in units of time. Wave source excitation is set to sinusoidal form: Currently common single-point water wave excitation methods (including but not limited to mechanical, pneumatic, and suction wave sources) mostly use sinusoidal excitation to generate waves.

[0058] Step 5: Calculate the delay time

[0059] Based on the difference between the desired target phase relationship and the current measured phase difference, and combined with the current operating frequency f, the required excitation delay time Δt for each wave source is calculated.

[0060] Step 6: Control signal output

[0061] A microcontroller control unit is used to generate multi-channel control CAN bus signals based on the calculated delay time Δt, which are then sent to each wave source actuator. The signals contain trigger timing information, allowing each wave source to perform the excitation action at a precisely specified time.

[0062] Step 7: Continuous Closed-Loop Feedback

[0063] Steps 1-6 above form a continuously operating closed-loop control system. During system operation, the high-speed camera continuously acquires images, and the control unit continuously updates phase difference information and delay time. Even if environmental conditions change (such as changes in water temperature leading to changes in wave velocity, or attenuation of wave source performance), the system can automatically adapt and maintain the preset phase relationship. The delay of the entire closed-loop control cycle depends on the image acquisition frame rate and algorithm processing speed; using modern hardware, millisecond-level real-time control can be achieved.

[0064] Figure 1 The overall process of the water wave control system according to an embodiment of the present invention is shown.

[0065] The closed-loop control delay of this system is not solely determined by the image acquisition frame rate, but also includes multiple stages such as image acquisition, data transmission, algorithm processing, control signal generation, and wave source actuator response. The total control delay T_d consists of the following components: T_exp (camera exposure time), T_read (image sensor readout time), T_trans (image data transmission time), T_alg (image processing and phase calculation time), T_ctrl (control signal generation and transmission time), and T_act (wave source actuator response time). The control unit uses a microcontroller and utilizes a hardware timer to output trigger pulses. The delays for control parameter reception, timer loading, and pulse output can be controlled within 1ms. The wave source actuator can be of speaker type, pneumatic type, or electromagnetic type.

[0066] In a typical embodiment, the 200f / s camera frame interval is 5ms, the exposure time is 2ms, the total transmission, processing, and calculation time of the image processing platform is approximately 10ms, the actuator response time is approximately 5ms, and the total control delay T_d is approximately 23ms. Under this delay, the system can effectively control the water wave phase within the 1-8Hz range, with a delay error within 20%. Since the camera frame interval, exposure time, and calculation time are relatively fixed, manual adjustment of the fixed delay can further optimize it to within 10%. By increasing the camera frame rate, reducing the ROI area, or using GPU-accelerated phase detection, the total delay can be controlled to around 10ms, thereby achieving effective control of water waves within the 1-20Hz range.

[0067] Experimental Example

[0068] Two pneumatic water wave excitation sources, designated as wave source S1 and wave source S2, are used. They are installed in the same water tank, spaced 300 mm apart horizontally. The water depth in the tank is 80 mm, and a checkerboard calibration plate is fixed 100 mm below the water surface. A high-speed camera is used for image acquisition, with a frame rate of 200 frames / s (5 ms between adjacent frames). The camera resolution is 1920×1080 pixels, and the exposure time is 1 / 500 s. The target operating frequency of the two wave sources is set to 4.00 Hz, corresponding to a water wave period T = 1 / f = 0.25 s. The control unit uses an industrial computer for image processing and phase calculation, and generates two TTL trigger pulses via a microcontroller's hardware timer to drive the first and second wave sources respectively.

[0069] The specific operating steps are as follows.

[0070] Step A: Acquire a checkerboard reference image I0 in a static water surface state and record this moment as the system time base t0; simultaneously, align the trigger clocks of the industrial computer, camera, and STM32 control unit so that each subsequent image frame has a timestamp t.k =t0+k / 200, where k is the frame number.

[0071] Step B: Set up ROI1 and ROI2 at a distance of about half the wavelength from the two wave sources along the direction of water wave propagation. Each ROI is a 40×40 pixel area used to extract the source phase of the corresponding wave source.

[0072] Step C, for the current frame image I k Calculate the optical flow displacement field u(x,y,t) using the reference image I0 k )=[ux(x,y,t k ),uy(x,y,t k [ ], and based on the distance H between the chessboard grid and the water surface, and the air refractive index n a =1.00, water refractive index n w =1.33, calculate the proportionality coefficient C=H×(1-n a / n w ).

[0073] Step D: Convert the optical flow displacement field into a water surface height gradient: h / x=-u x / C, h / y=-u y / C; then integrate in the frequency domain to obtain the water surface height field h(x,y,t) of the current frame. k Its frequency domain expression is H(k) x ,k y )=j×[k x ×FFT( h / x)+k y ×FFT( h / y)] / (k x ²+k y ²), and obtain h(x,y,t) through two-dimensional inverse Fourier transform. k =Re[IFFT(H)].

[0074] Step E: Calculate the average height within ROI1 and ROI2 respectively, to obtain the height time series s1(t) of the two wave sources. k =meanROI1[h(x,y,t) k )]、s2(t k =meanROI2[h(x,y,t) k )).

[0075] Step F, for s1(t) k ) and s2(tk Perform a bandpass filter centered at 4.00 Hz, with a passband of 3.5-4.5 Hz; then construct the analytic signal z. i (t)=s1(t)+j·Hilbert[s i (t)], and calculate the instantaneous phase Φ of each wave source. i (t)=atan2(Im[z i (t)],Re[z i (t)]). When the first wave source S1 is used as the reference source, the measured phase difference between the second wave source and the first wave source is ΔΦ. 21 (t) = wrap[Φ2(t) - Φ1(t)]. If the target emits waves in phase, then the target phase difference ΔΦ target =0. The phase deviation is e 21 (t)=wrap[ΔΦ 21 (t)-ΔΦ target ].

[0076] Step G: Calculate the trigger compensation time based on the phase deviation. The compensation formula is Δt = e 21 / (2πf). The trigger compensation time signal is converted into a bus signal using STM32 and sent to each wave source actuator.

[0077] An experimental verification was conducted based on a 4.00Hz water wave. Two identical wave sources, pneumatically propelled to generate waves on the water surface, were used. A plane wave-like distribution was achieved using a 1-to-6 aquarium air pump splitter. Without generation control, a phase difference existed between the two wave sources, resulting in a discrepancy between the combined effect and a complete plane wave. Figure 2 As shown (left: before adjustment). Phase difference analysis using the image system determined the phase difference between the two wave sources to be approximately 60°. Using the later wave source as a reference, the excitation time of the earlier wave source was delayed by 0.04 seconds. After re-performing the phase difference analysis, the phase difference between the two wave sources was within 5°, and the overall water wave morphology was consistent with the plane wave height, as shown below. Figure 2 As shown on the right (after adjustment), the overall phase error is reduced by approximately 90%.

[0078] Based on the above experimental examples, this system handles control delay and time synchronization as follows.

[0079] Step H, control delay and time synchronization are handled as follows. In a specific example, the image frame interval is 5 ms, the image transmission time is approximately 2 ms, the optical flow calculation, height field reconstruction, and phase calculation time is approximately 10 ms, the time for the STM32 to receive compensation parameters and load the hardware timer is less than 1 ms, and the response time of the wave source drive circuit and actuator is approximately 5 ms. Therefore, the total control delay Td from image acquisition to the compensation trigger output taking effect is approximately 23 ms. Since the target water wave period T = 250 ms, Td / T is approximately 9.2%, which meets the real-time requirements of closed-loop control. To compensate for this total delay, the control unit calculates the predicted phase Φi(tk+Td) = Φi(tk) + 2πfTd before output; at f = 4.00 Hz and Td = 0.023 s, the predicted compensation angle is approximately 2π × 4.00 × 0.023 = 0.578 rad, or 33.1°.

[0080] Alternative embodiments

[0081] When the application scenario does not allow for optical imaging (such as turbid water, dark rooms without light sources, or the need to penetrate foam layers), ultrasonic arrays or millimeter-wave radar can be used instead of high-speed cameras. Typical center frequencies for ultrasonic arrays are 200 kHz-2 MHz, with sampling rates no less than 1 MS / s; millimeter-wave radar can use the 60 GHz or 77 GHz bands, with a preferred range resolution of 1-5 mm; both can cover the common frequency adjustment range of water waves with high accuracy. By transmitting ultrasonic / electromagnetic waves to the water surface through a transmitting array and receiving the echoes, the phase / time-of-flight changes of the echoes are analyzed to indirectly measure the height changes of various points on the water surface, thus achieving a phase feedback function similar to vision-based solutions.

[0082] Alternatively, the high-speed photography function built into a smartphone (currently, mainstream smartphones support 120fps-960fps slow-motion shooting) can be used as an image acquisition device. All algorithmic processing can be performed using an image recognition app running on the phone, which communicates with each wave source control module via Bluetooth / Wi-Fi. This solution integrates the core sensing and computing functions of the system into a regular smartphone, making the system highly portable and consumer-grade accessible.

[0083] In summary, this invention proposes a multi-source water wave generation control method and system based on image recognition analysis. The core innovation of this invention lies in the first-time introduction of computer vision technology (high-speed camera image acquisition + image recognition analysis) into the field of multi-source water wave generation control, constructing a closed-loop feedback control system. Essentially, it does not force wave source synchronization through physical means, but rather obtains the actual phase state of each wave source in real time through visual feedback, and then uses software algorithms to calculate and apply precise timing compensation, thereby achieving flexible and accurate control of the phase relationship of multiple wave sources.

[0084] The outstanding innovative contributions and technical advantages of this invention include: 1. Real-time non-contact detection of water wave phase using visual sensing Unlike traditional mechanical linkage or open-loop timing schemes, this project is the first to use a high-speed camera to capture real-time images of the water surface. Through image recognition algorithms (FFT (Fast Fourier Transform) spectrum analysis, edge detection, optical flow, etc.), the instantaneous phase, amplitude, and frequency parameters of water waves at each wave source location are directly extracted from the water surface image sequence, thus constructing a full-field perception capability for the state of multi-source water waves.

[0085] 2. Source-by-source timing compensation control mechanism based on phase deviation

[0086] Based on the deviation between the measured phase and the target phase, and combined with the operating frequency, the phase angle difference is converted into an independent excitation delay time for each wave source. Differentiated and precise timing compensation is applied to each wave source through multi-channel control signals, thereby achieving flexible control of arbitrary phase relationships (synchronization, anti-phase, fixed phase difference, and dynamically changing phase difference) between multiple sources without modifying the mechanical structure of the wave sources.

[0087] 3. Real-time closed-loop control architecture of sensing → calculation → feedback

[0088] The above-mentioned "image acquisition → phase detection → deviation calculation → timing compensation → signal output → re-acquisition" constitutes a continuously operating closed-loop control loop, enabling the system to adapt to environmental changes (water temperature, wave source aging, boundary condition changes), fundamentally solving the problem of phase drift not being detected and corrected in open-loop control.

[0089] 4. Hardware and software decoupling design for wave source synchronization and wave source excitation

[0090] Multi-source water wave control is decomposed into two decoupled layers: a physical layer (each wave source is only responsible for executing the excitation action according to the command) and an information layer (the camera and processor are responsible for sensing the phase state and generating control commands), so that the system can be adapted to various existing water wave generating devices without the need for mechanical modification of the wave source hardware.

[0091] Compared with existing solutions, the advantages of this invention are reflected in the following aspects: After following the above steps, the phase difference between the two wave sources is approximately 60° without compensation. After applying a trigger timing compensation of approximately 0.04 s and continuously updating in a closed loop for 2-3 cycles, the phase difference between the two wave sources converges to within 5°, and the water surface forms a planar waveform that is closer to in-phase superposition. The camera frame rate, number of wave sources, operating frequency, ROI size, total control delay, and compensation time mentioned above are specific parameters of this embodiment; in practical applications, they can be adjusted proportionally according to the tank size, wave source type, and target waveform.

[0092] 1. Non-contact measurement, compatible with existing hardware. Compared with mechanical linkage solutions, this invention requires no modification to the mechanical structure of the water wave generator. The image acquisition device (such as a high-speed camera) does not have physical contact with the controlled wave source, and does not introduce additional inertial loads or friction. This feature allows this invention to be used as an "external" upgrade module for existing water wave generator systems, greatly reducing system modification costs.

[0093] 2. Real-time closed-loop feedback, resistant to environmental interference. Compared with open-loop control schemes using precision timing circuits, the visual feedback mechanism of this invention enables the system to perceive the actual output state of each wave source in real time. When changes in water temperature cause changes in wave velocity, or when aging of the wave source actuator causes changes in response delay, the closed-loop system can automatically detect and correct deviations. This adaptive capability is unattainable by any open-loop scheme.

[0094] 3. Flexible phase relationship programmability. Unlike mechanical linkage schemes that require changes to the mechanical structure to alter the phase relationship, this invention enables the switching of any phase relationship through pure software. By simply modifying the target phase parameter, the system can switch between multiple modes, such as synchronous wave generation, anti-phase wave generation, wave generation with arbitrary fixed phase difference, and even dynamic phase difference changes according to a preset sequence, without requiring system downtime or hardware adjustments.

[0095] 4. High precision. Visual feedback combined with a high-speed camera can achieve millisecond-level temporal resolution. Taking a typical high-speed camera at 120fps as an example, the inter-frame time interval is approximately 8.3ms; taking a professional high-speed camera at 1000fps as an example, the inter-frame time interval is only 1ms. The corresponding phase resolution can reach 3.6° at an operating frequency of 10Hz, meeting the accuracy requirements of most water wave applications.

[0096] 5. Excellent scalability. When the number of wave sources needs to be increased, only ROI analysis needs to be added to the image region corresponding to each new wave source. No additional sensing hardware is required (the number of sensors does not increase linearly with the number of wave sources). The increase in system complexity is mainly reflected in the computational load of the algorithm, which can be addressed by upgrading the processor.

[0097] This invention also provides a storage medium for storing a computer program, which, when executed, performs at least the methods described above.

[0098] This invention also provides a control device, including a processor and a storage medium for storing a computer program; wherein the processor executes the computer program by performing at least the method described above.

[0099] This invention also provides a processor that executes a computer program, at least performing the methods described above.

[0100] The storage medium can be implemented by any type of non-volatile storage device, or a combination thereof. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc or CD-ROM; magnetic surface memory can be disk storage or magnetic tape storage. The storage media described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0101] In the several embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0102] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0103] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0104] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0106] The methods disclosed in the several method embodiments provided by this invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0107] The features disclosed in the several product embodiments provided by this invention can be arbitrarily combined without conflict to obtain new product embodiments.

[0108] The features disclosed in the several method or device embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0109] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various equivalent substitutions or obvious modifications can be made without departing from the concept of the present invention, and all such modifications, achieving the same performance or application, should be considered within the scope of protection of the present invention.

Claims

1. A multi-source water wave generation control method based on image recognition analysis, characterized in that, Includes the following steps: S1. Real-time image acquisition of the water surface is performed using an image acquisition device; S2. Preprocess the acquired images and delineate the corresponding regions of interest based on the known locations of each wave source; S3. Analyze the time-series image sequences in each region of interest and extract the instantaneous phase of the water waves at each wave source location; S4. Using one of the wave sources as the phase reference, calculate the measured phase difference of the other wave sources relative to the reference. S5. Based on the deviation between the preset target phase relationship and the measured phase difference, and in conjunction with the current operating frequency, calculate the required excitation timing compensation for each wave source; S6. Generate multi-channel control signals according to the timing compensation amount, and drive each wave source actuator to perform excitation actions at the corresponding compensation time; S7. Repeat steps S1 to S6 continuously to form real-time closed-loop feedback control.

2. The method as described in claim 1, characterized in that, Step S3, which involves extracting the instantaneous phase of the water waves at each wave source location, specifically includes: Acquire a reference image while the water surface is still; For each frame of real-time acquired wave image, the optical flow displacement field between it and the reference image is calculated to obtain displacement components in two orthogonal directions. Based on the two directional displacement components of the optical flow displacement field, they are respectively converted into water surface height gradient components in two orthogonal directions; Two-dimensional integral reconstruction is performed on the height gradient components in the two directions to obtain the water surface height field of the current frame; The average height value is extracted from the region of interest corresponding to each wave source, which is used as the representative height of the wave source to form the height time series of each wave source; After bandpass filtering the height time series, an analytical signal is constructed and the instantaneous phase of the water wave at each wave source location is extracted.

3. The method as described in claim 2, characterized in that, The process of converting the displacement components of the optical flow displacement field in two directions into water surface height gradient components in two orthogonal directions specifically includes: Fix the calibration pattern below the water surface or at the bottom of the tank, and obtain a reference image of the calibration pattern in a static water state; Based on the apparent displacement of the calibration pattern in each frame of wave-bearing image relative to the reference image, the two directional displacement components of the apparent displacement field are calculated using the optical flow method. The scaling factor is determined based on the air refractive index, the water refractive index, and the distance from the calibration pattern plane to the still water surface. The displacement components in the two directions are divided by the scaling factor and inverted to obtain the water surface height gradient components in the two orthogonal directions.

4. The method as described in claim 2, characterized in that, The step of performing two-dimensional integral reconstruction of the height gradient components in two directions to obtain the water surface height field of the current frame specifically includes: Transform the water surface height gradient components in the two orthogonal directions to the frequency domain; In the frequency domain, the wavenumber is used to combine the spectra of gradient components in each direction to achieve integration, and then a two-dimensional inverse transform is performed to obtain the water surface height field of the current frame.

5. The method as described in claim 2, characterized in that, After bandpass filtering the altitude time series, an analytical signal is constructed and the instantaneous phase is extracted, specifically including: The passband range of the bandpass filter is set according to the target operating frequency of the wave source, and bandpass filtering is performed on the height time series of each wave source; The filtered height time series is constructed into a complex analytic signal using Hilbert transform; Based on the real and imaginary parts of the complex analytic signal, the instantaneous phase of each wave source at each moment is calculated using the four-quadrant arctangent function.

6. The method as described in claim 1, characterized in that, The method for selecting the reference datum in step S4 is as follows: Using the wave source with the latest emitted wave as the phase reference, the delay time of the other wave sources is calculated relative to this reference; or A pre-defined fixed wave source is used as the phase reference. When the calculated delay time is negative, the triggering time of the wave source is delayed to the next cycle, that is, one water wave cycle is added to the delay time.

7. The method as described in claim 1, characterized in that, Step S5, which involves calculating the required excitation timing compensation for each wave source, specifically includes: The measured phase of each wave source is compared with the reference phase by performing a phase winding operation to obtain the measured phase difference of each wave source. The measured phase difference is compared with the preset target phase difference to obtain the phase deviation. Divide the phase deviation by the angular frequency to convert it into the excitation delay time in the time domain.

8. The method as described in claim 1, characterized in that, The generation of multi-channel control signals in step S6 specifically includes: The total control delay of the system is obtained, which includes the sum of image exposure time, image sensor readout time, image data transmission time, image processing and phase calculation time, control signal generation and transmission time, and wave source actuator response time. Based on the instantaneous phase of each wave source and the current operating frequency, calculate the predicted phase after the total control delay; The timing compensation amount is determined based on the deviation between the predicted phase and the target phase, and multiple trigger pulse signals are generated by a hardware timer and sent to each wave source actuator.

9. A multi-source water wave generation control system based on image recognition analysis, characterized in that, include: An image acquisition device is set above or to the side of the water surface to acquire real-time images of the water surface, including the corresponding areas of each wave source. A calibration plate, fixed below the water surface or at the bottom of the tank, provides a reference texture for the image acquisition device. The image processing and phase calculation unit is connected to the image acquisition device and is used to preprocess the acquired image, calculate the optical flow displacement field, reconstruct the water surface height field and extract the instantaneous phase, and calculate the deviation between the measured phase difference between each wave source and the target phase. The control unit, connected to the image processing and phase calculation unit, is used to calculate the excitation timing compensation amount of each wave source based on the deviation amount and the current operating frequency, and to generate multi-channel control signals. Multiple wave source actuators are connected to each output channel of the control unit, respectively, to receive the control signal and perform excitation actions at the corresponding compensation time; The image acquisition device, image processing and phase calculation unit, control unit, and wave source actuator constitute a real-time closed-loop feedback control loop.

10. The system as described in claim 9, characterized in that: The image acquisition device is a high-speed camera, and the calibration board is a checkerboard calibration board; The image processing and phase calculation unit performs optical flow matching on each frame of wave image and the reference image under static water surface state to obtain the two-directional displacement components of the apparent displacement field of checkerboard texture. Based on the difference between air refractive index and water refractive index and the distance from checkerboard to water surface, the scaling factor is determined. The two-directional displacement components of the apparent displacement field are converted into water surface height gradient components in two orthogonal directions. The water surface height field is reconstructed by frequency domain integration. The average height value is extracted from the region of interest corresponding to each wave source as a representative height to form a height time series. The instantaneous phase of each wave source is extracted by bandpass filtering and Hilbert transform. The control unit uses one wave source as a reference, calculates the measured phase difference of each of the other wave sources relative to the reference, and compares the phase difference with the preset target phase relationship after phase winding operation to obtain the phase deviation. The phase deviation is divided by the angular frequency to convert it into the excitation delay time, and then drives each wave source actuator through multi-channel trigger pulses. The control unit is further configured to: calculate the predicted phase after the delay ends by using the current instantaneous phase and the operating frequency based on the total system control delay, and output a timing compensation amount based on the predicted phase.

Images