Unmanned aerial vehicle water surface flow velocity measurement method and device based on double Doppler radars

Abstract

The invention provides an unmanned aerial vehicle water surface flow velocity measurement method and device based on double Doppler radars, and solves the technical problems of low flow velocity measurement precision and poor anti-interference performance caused by radar signal distortion caused by a complex water surface flow field when an unmanned aerial vehicle is used for measuring flow in the prior art. The method comprises the following steps: acquiring a reflected signal of a radar, performing preprocessing, denoising, reconstruction and space-time alignment on the reflected signal by using a multi-scale topological wavelet decomposition technology to obtain a water surface flow velocity and a water level height, and performing flow velocity vector synthesis on the water surface flow velocity to obtain a water surface flow velocity; performing prediction and time averaging processing to obtain a time average flow rate; correcting the water surface flow velocity, converting the corrected water surface flow velocity into a vertical average flow velocity, and carrying out space grid division and local flow velocity calculation to obtain a local flow velocity of a space grid; and weighting the time average flow velocity and the local flow velocity of the space grid to obtain a final flow velocity, and performing flow integration to obtain section flow. The method can be widely applied to the technical field of hydrological monitoring.

Description

Technical Field

[0001] This application belongs to the technical field of hydrological monitoring. More specifically, it relates to a method and device for measuring the water surface velocity of an unmanned aerial vehicle (UAV) based on dual Doppler radar. Background Art

[0002] Traditional water surface velocity measurement techniques mainly include contact mechanical instruments, acoustic Doppler current profilers, and non-contact radar speed measurement. Contact current meters rely on the direct interaction between mechanical rotors and water flow, and there are problems such as easy wear and being easily stuck by floating objects, and they cannot adapt to high flow velocities or complex water flow environments; although Doppler current profilers can obtain vertical velocity profiles, they have defects such as large volume, high cost, and severe acoustic wave attenuation in turbid water bodies.

[0003] Currently, non-contact radar is applied to the speed measurement field. For example, in the Chinese invention patent with the application number CN202411049268.4, a multi-element UAV flow measurement method and system are proposed. By taking advantage of the light weight of the UAV and the absence of mechanical contact, the accuracy and reliability of water body velocity measurement are improved by integrating multiple data sources. However, when measuring speed in the case of turbulent water surfaces, the clutter interference caused by water surface wind waves significantly reduces the signal-to-noise ratio, and the strong rainfall attenuation effect and the existence of gas-liquid two-phase flow further exacerbate the data distortion. Therefore, there is an urgent need to provide a UAV water surface velocity measurement method and device to solve the above problems. Summary of the Invention

[0004] The purpose of the embodiments of this application is to provide a method and device for measuring the water surface velocity of an unmanned aerial vehicle based on dual Doppler radar, so as to solve the technical problems of low flow velocity measurement accuracy and poor anti-interference ability caused by radar signal distortion due to complex water surface flow fields when using UAVs to measure flow in the prior art.

[0005] To achieve the above purpose, the first aspect of the embodiments of this application provides a method for measuring the water surface velocity of an unmanned aerial vehicle based on dual Doppler radar, including Obtaining the reflected signal of the radar, and preprocessing the reflected signal by using multi-scale topological wavelet decomposition technology to obtain the corrected detail coefficients and multi-scale topological features; Denosing and reconstructing the multi-scale topological features and the corrected detail coefficients to obtain a reconstructed signal, performing spatio-temporal alignment to obtain the flow velocity and water level height of the water surface, and synthesizing the flow velocity vectors of the water surface to obtain the water surface velocity; Predicting and performing time averaging on the water surface velocity and water level height to obtain the time-averaged flow velocity; Converting the water surface velocity after correction into the vertical average flow velocity, performing spatial grid division and local flow velocity calculation to obtain the local flow velocity of the spatial grid; The final flow velocity is obtained by weighting the time-averaged flow velocity and the local flow velocity of the spatial grid, and the cross-sectional flow rate is obtained by integrating the flow rate. The preprocessing process includes: decomposing according to the number of layers of the reflection signal to obtain detail coefficients, extracting topological features to obtain multi-scale topological features, and performing denoising and then fusing to obtain the corrected detail coefficients.

[0006] Preferably, the Kalman state equation introducing the gravitational acceleration and the water surface slope is used to predict the water surface flow velocity and the water level height, and the formula is as follows: ; In the formula, is the water surface flow velocity at the next moment (k + 1), is the water level height at the next moment (k + 1), is the time step, g is the gravitational acceleration, S is the water surface slope parameter, is the water surface flow velocity at the current moment k, is the water level height at the current moment k, is the process noise.

[0007] Preferably, before performing spatio-temporal alignment, calculate the mutual information between the reconstructed signal and the multi-scale topological features, and use the mutual information to verify the topological consistency of the reconstructed signal to obtain the verified reconstructed signal, including the verified water level signal and the verified water surface reflection signal; Perform attitude compensation on the verified water level signal to obtain the attitude-compensated water level signal; Perform spatio-temporal alignment on the verified water surface reflection signal and the attitude-compensated water level signal to obtain the flow velocity and water level height on the water surface.

[0008] Preferably, the process of topological feature extraction includes: constructing a complex according to the detail coefficients, and calculating the time series of betti numbers for the complex at each scale to obtain multi-scale topological features.

[0009] Preferably, the process of denoising and then fusing includes: incorporating the multi-scale topological features into the noise variance and calculating the corrected noise standard deviation; Using the corrected noise standard deviation, fuse the multi-scale topological features with the detail coefficients to obtain the corrected detail coefficients.

[0010] Preferably, use the signal-to-noise ratio factor as the turbulence correction factor to correct the water surface flow velocity to obtain the corrected water surface flow velocity, and use the vertical flow velocity distribution coefficient to convert the corrected water surface flow velocity to obtain the vertical average flow velocity.

[0011] Preferably, the formula for spatio-temporal alignment is as follows: ; In the formula, The reconstructed signals of the dual Doppler radar and the water level radar, which are the fixed delays of the reconstructed forward water surface reflection signal, the reconstructed backward water surface reflection signal, and the water level signal respectively, The flow velocity of the water surface corresponding to the forward Doppler radar, The flow velocity of the water surface corresponding to the backward Doppler radar, The water level height in the vertical direction of the water level radar, The reconstructed signal of the forward Doppler radar after removing the delay, The reconstructed signal of the backward Doppler radar after removing the delay, The reconstructed signal of the water level radar after removing the delay and attitude compensation.

[0012] Preferably, the process of spatial grid division and local flow velocity calculation includes: dividing the vertical average flow velocity according to equal width and equal depth and then assigning values to obtain the local flow velocity of the spatial grid.

[0013] Preferably, the process of denoising the multi-scale topological features and the modified detail coefficients includes: dynamically denoising the multi-scale topological features to generate a hybrid threshold function, and performing joint threshold processing on the modified detail coefficients according to the hybrid threshold function to obtain the detail coefficients after joint threshold processing.

[0014] The second aspect of the embodiments of the present application provides an unmanned aerial vehicle water surface flow velocity measurement device based on a dual Doppler radar, which uses the above-mentioned unmanned aerial vehicle water surface flow velocity measurement method based on a dual Doppler radar, including: an unmanned aerial vehicle, with a control terminal inside, a positioning device on the top, an adjusting device and a transmitting and receiving device on the bottom, and both the adjusting device and the transmitting and receiving device are electrically connected to the control terminal; The control terminal controls the adjusting device and the transmitting and receiving device to obtain the radar reflection signal; The control terminal preprocesses the reflection signal by using the multi-scale topological wavelet decomposition technology to obtain the modified detail coefficients and multi-scale topological features; Denoise and reconstruct the multi-scale topological features and the modified detail coefficients to obtain a reconstructed signal, perform spatio-temporal alignment to obtain the flow velocity and water level height of the water surface, and perform flow velocity vector synthesis on the flow velocity of the water surface to obtain the water surface flow velocity; Perform prediction and time averaging processing on the water surface flow velocity and the water level height to obtain the time-averaged flow velocity; Convert the water surface flow velocity after correction into the vertical average flow velocity, perform spatial grid division and local flow velocity calculation to obtain the local flow velocity of the spatial grid; Weight the time-averaged flow velocity and the local flow velocity of the spatial grid to obtain the final flow velocity, and perform flow integration to obtain the cross-sectional flow rate.

[0015] The beneficial effects of the present application are as follows: The present application proposes a method and device for measuring the water surface velocity of an unmanned aerial vehicle (UAV) based on dual Doppler radars. The dual Doppler radars are used to simultaneously obtain the reflection signals from different azimuths of the water surface. At the same time, a multi-scale topological wavelet decomposition technology that combines the four-layer decomposition method of Daubechies6 wavelet basis and multi-scale topological technology is used to preprocess the obtained reflection signals to obtain the corrected detail coefficients and multi-scale topological features. Through topological feature extraction, multi-scale topological features are obtained. Then, the multi-scale topological features are introduced into the joint noise estimation of the noise variance, and denoising and reconstruction are performed using dynamic topological threshold calculation and joint threshold processing to achieve topological-wavelet joint denoising and signal reconstruction. The data accuracy is ensured through topological consistency verification, and finally, the data of each sensor is output after spatio-temporal alignment. It effectively cancels the dynamic disturbance of the UAV and environmental interference, significantly improves the measurement stability and anti-noise ability under complex water surface flow fields, and improves the measurement efficiency and measurement accuracy.

[0016] Moreover, the present application also sets a vector synthesis algorithm for the dual Doppler radars to fuse the measurement results of multiple Doppler radars to achieve high-precision estimation and output of the water surface velocity and cross-sectional flow rate. In addition, the adjustment device adjusts the radar installation angle and position in real time, effectively cancels the measurement errors caused by unstable UAV hovering, environmental disturbance, and water surface turbulence, significantly improves the stability and accuracy of the flow velocity measurement data, overcomes the defects of low signal-to-noise ratio compensation, large blind area, and untimely dynamic response of traditional single radars, realizes robust signal extraction and accurate flow rate determination in complex water surface flow field environments, greatly shortens the device debugging and parameter calibration time, reduces the hardware cost, and improves the applicability, reliability, and anti-interference ability of the UAV water surface velocity measurement device in various actual application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a method for measuring the water surface velocity of an unmanned aerial vehicle based on dual Doppler radars provided by an embodiment of the present application; Figure 2 It is a schematic diagram of the water depth and vector synthesis angle of the water surface velocity measurement of an unmanned aerial vehicle provided by an embodiment of the present application; Figure 3 It is a schematic diagram of the attitude compensation of the water level signal for the water surface velocity measurement of an unmanned aerial vehicle provided by an embodiment of the present application; Figure 4Structural diagram of an unmanned aerial vehicle (UAV) water surface flow velocity measurement device based on dual Doppler radar provided by an embodiment of the present application; Figure 5 Exploded view of an unmanned aerial vehicle (UAV) water surface flow velocity measurement device based on dual Doppler radar provided by an embodiment of the present application; Figure 6 Enlarged view of a flowmeter in an unmanned aerial vehicle (UAV) water surface flow velocity measurement device based on dual Doppler radar provided by an embodiment of the present application; Figure 7 Perspective view of the upper part of a flight platform in an unmanned aerial vehicle (UAV) water surface flow velocity measurement device based on dual Doppler radar provided by an embodiment of the present application; Figure 8 Perspective view of the lower part of a flight platform in an unmanned aerial vehicle (UAV) water surface flow velocity measurement device based on dual Doppler radar provided by an embodiment of the present application.

[0019] In the figure: 1. Positioning device; 10. Central processing unit; 11. Data receiver; 12. FPGA; 13. Soft shell protective cover; 14. Bearing; 15. Turntable; 16. Upper connecting pin; 17. Telescopic rod; 18. Rotating shaft; 19. Lower connecting pin; 2. Protective cover; 20. Upper clamping plate; 21. Flight rod; 22. Lower clamping plate; 23. Rotor; 24. Motor; 25. Support plate; 26. Upper clamping block; 27. Lower clamping block; 3. Flowmeter; 4. Flight platform; 5. Dual Doppler radar; 6. Water level radar; 9. IMU sensor. Detailed implementation manners

[0020] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0021] Please refer to Figure 1 , which is a flowchart of a method for measuring the water surface flow velocity of an unmanned aerial vehicle (UAV) based on dual Doppler radar provided by the first aspect of an embodiment of the present application, including: S1: Obtain the reflection signals of the dual Doppler radar and the water level radar located at the monitoring section, including the water surface reflection signal and the water level signal. Among them, the water surface reflection signal includes the forward water surface reflection signal and the backward water surface reflection signal.

[0022] Specifically, the UAV is controlled to hover above the monitoring section, and the UAV's planar coordinates are obtained through the positioning device. Meanwhile, the data receiver collects the real-time pitch angle and roll angle data of the flowmeter provided by the IMU sensor and sends the pitch angle and roll angle data to the FPGA. After the FPGA correlates the pitch angle and roll angle data with the UAV's planar coordinates, it controls the turntable and telescopic rod of the adjusting device according to the pitch angle and roll angle data to adjust the radar installation angle in real time, so that the dual Doppler radar forms a fixed angle with the water flow direction, and the water level radar is perpendicular to the water surface.

[0023] Among them, the self-stabilizing gimbal adjusts the angles and positions of the water level radar and the dual Doppler radar in real time through the PID algorithm (Proportional-Integral-Derivative algorithm) to ensure that the water level radar is always perpendicular to the water surface. At the same time, the dual Doppler radar is positioned along the water flow direction and maintains a fixed angle between its transmitted wave and the water surface. In this application, the PID algorithm of the self-stabilizing gimbal adopts incremental PID control, and its control quantity output formula is as follows: ; In the formula, is the angle deviation at the current moment (the difference between the target angle and the angle measured by the IMU sensor), is the proportional gain, and its value range is (0.5, 0.8), is the integral gain, and its value range is (0.05, 0.1), is the derivative gain, and its value range is (0.1, 0.2). p represents proportion, i represents integral, d represents derivative, and k is the discrete time step, representing the current moment. According to the control quantity the position of the flowmeter is adjusted in real time to make the water level radar always perpendicular to the water surface, and the dual Doppler radar maintains a preset angle with the water flow direction.

[0024] The FPGA synchronously triggers the dual Doppler radar to emit electromagnetic waves at 24 GHz and the water level radar to emit pulsed waves at 120 GHz to obtain radar reflection signals, including the forward water surface reflection signal, the backward water surface reflection signal, and the water level signal. Here, the frequency of the radar emission band is not limited and can be set according to the actual situation. Within the synchronous time window, the reflection signals obtained by the radar include the following three types:

[0025] In the formula, are the water surface reflection signal obtained by the forward Doppler radar in the dual Doppler radar, the water surface reflection signal obtained by the backward Doppler radar in the dual Doppler radar, and the water level signal of the vertical distance obtained by the water level radar in sequence. f represents the forward Doppler radar, r represents the backward Doppler radar, h is the vertical ranging of the water level radar, is the synchronous time window, t0 is the initial time, is the sampling time interval, where the time window length is synchronized with the radar sampling frequency. is the radial velocity component of the forward Doppler radar in the dual-Doppler radar. is the radial velocity component of the backward Doppler radar in the dual-Doppler radar. is the true water level. is the additive noise.

[0026] S2: Preprocess the acquired reflection signal using the multi-scale topological wavelet decomposition technique to obtain the corrected detail coefficients and multi-scale topological features.

[0027] The process of preprocessing using the multi-scale topological wavelet decomposition technique includes: determining the scale function, wavelet function, and the number of decomposition layers of the reflection signal, decomposing according to the number of layers to obtain the approximation coefficients and detail coefficients, constructing a complex based on the detail coefficients and calculating the time series of the Betti numbers of each scale complex to obtain the multi-scale topological features, incorporating the noise variance to calculate the corrected noise standard deviation, and fusing each scale topological feature with the detail coefficients using the noise standard deviation to obtain the corrected detail coefficients.

[0028] Specifically, use the multi-scale topological wavelet decomposition technique based on the four-layer decomposition method of the Daubechies6 wavelet basis to decompose the acquired reflection signal simultaneously to obtain the approximation coefficients and detail coefficients of the th layer after decomposition. First, the scale function of the reflection signal before decomposition and the wavelet function need to satisfy the following conditions:

[0029] where and are the coefficients of the filter, defined by the Daubechies6 wavelet standard, , is the support length, that is, the non-zero interval of the scale function in the time domain, and t is the synchronized time window.

[0030] Then, determine the number of wavelet decomposition layers:

[0031] where J is the number of wavelet decomposition layers (decomposition scale number), and N is the number of sampling points of the reflection signal (the length of the radar reflection signal), and it needs to satisfy (the wavelet decomposition feasibility condition).

[0032] Next, decompose the jth layer of any reflection signal to obtain the approximation coefficients and detail coefficients of the jth layer, and the decomposition formula is as follows:

[0033] Among them, is the approximation coefficient of the j-th layer, is the detail coefficient of the j-th layer, n is the discrete-time index of the output coefficient, and the actual range is , retain the coefficients of . is the noise estimation based on wavelet coefficients, is the approximation coefficient of the (j - 1)-th layer, w is the low-pass filter for extracting the low-frequency components of the signal, g is the high-pass filter for extracting the high-frequency components of the signal, and k is the discrete-time index of the input signal.

[0034] Finally, the FPGA is used to extract the topological features of the detail coefficients to obtain multi-scale topological features, and after denoising, they are fused to obtain the corrected detail coefficients. The specific steps are as follows: First, an alpha complex is constructed based on the filtered detail coefficients at wavelet decomposition scales j = 1,..., J to identify the signal mutation features caused by surface turbulence. The formula for constructing the alpha complex is as follows:

[0035] In the formula, is the average amplitude of the radius adaptation and the detail coefficients, which controls the generation of simplices, is the detail coefficient of the j-th layer, is the radius of the closed ball, is the number of coefficients.

[0036] Next, for each scale complex , calculate the time series of its 0-dimensional Betti numbers to obtain multi-scale topological features, and distinguish the noise in the reflected signal from the true flow velocity signal. Among them, the formula for obtaining the multi-scale topological features is as follows:

[0037] In the formula, is the multi-scale topological feature, K is the number of topological features detected in the j-th layer, calculated by the Alpha complex, I is the indicator function, which is 1 when the condition is satisfied and 0 otherwise, is the survival time of the k-th topological feature, t is the time series, and is the adaptive persistence threshold for analyzing the persistence of topological features. Making and have the same formula means that the connectivity of the ALpha complex at radius is consistent with the result when the threshold in persistent homology is , ensuring the logical self-consistency of the topological analysis and simplifying the parameter process.

[0038] Furthermore, the multi-scale topological features are incorporated into the noise variance calculation to obtain the corrected noise standard deviation. Here, the topological structure information contained in the multi-scale topological features is used to optimize the noise modeling, improving the robustness and estimation accuracy of the system. The formula is as follows:

[0039] Among them, is the corrected noise standard deviation, is the standard deviation of the input noise, is the local time window length, N is the number of sampling points of the reflected signal, taking 10% of the sampling points of the reflected signal, is the multi-scale topological feature corresponding to the nth discrete time point.

[0040] Finally, using the corrected noise standard deviation, each scale topological feature is fused with the detail coefficient of the jth layer of wavelet decomposition to obtain the corrected detail coefficient. The formula is as follows:

[0041] Among them, is the corrected detail coefficient, is the hyperbolic tangent function, which normalizes the topological feature to [-1, 1], is the detail coefficient of the jth layer, is the multi-scale topological feature corresponding to the nth discrete time point.

[0042] S3: Denoise the multi-scale topological features according to the corrected detail coefficient and reconstruct to obtain the reconstructed signal. After verification, perform spatio-temporal alignment to obtain the flow velocity and the radar vertical water level height corresponding to the water surface of the dual Doppler radar, and perform flow velocity vector synthesis to obtain the water surface flow velocity.

[0043] Specifically, perform dynamic denoising on the multi-scale topological features according to the corrected detail coefficient to obtain the hybrid threshold function, and perform joint threshold processing to obtain the detail coefficient after joint threshold processing; perform wavelet inverse transform reconstruction on the detail coefficient after joint threshold processing according to the approximation coefficient to obtain the reconstructed signal. After verification, perform spatio-temporal alignment to obtain the flow velocity and the radar vertical water level height corresponding to the water surface of the dual Doppler radar, and perform flow velocity vector synthesis to obtain the water surface flow velocity. Among them, the reconstructed signal includes the reconstructed water level signal and the water surface reflection signal.

[0044] Input the multi-scale topological features into the FPGA according to the corrected detail coefficient for dynamic denoising, generating a hybrid threshold function related to the radar signal-to-noise ratio for suppressing the vibration of the unmanned aerial vehicle and environmental noise interference. Among them, the formula for generating the hybrid threshold function is as follows:

[0045] In the formula, is the basic threshold function, is the hybrid threshold function corresponding to the nth discrete time point, is the number of sampling points in the jth layer of the reflected signal, is the multi-scale topological feature corresponding to the nth discrete time point.

[0046] Next, based on the corrected detail coefficients, the hybrid threshold function generated by dynamic denoising is subjected to joint threshold processing to generate a scene-adaptive detection threshold to avoid the limitations of the fixed threshold. Among them, the formula for the joint threshold is as follows: ; In the formula, is the corrected detail coefficient, is the detail coefficient after joint threshold processing.

[0047] Then, the inverse wavelet transform is performed on the approximation coefficients and the detail coefficients after joint threshold processing to obtain a reconstructed signal, including the reconstructed water level signal and the water surface reflection signal. The formula for obtaining the reconstructed signal is as follows:

[0048] In the formula, is the reconstructed signal, is the approximation coefficient, is the scaling function, J is the number of layers of wavelet decomposition, is the detail coefficient after joint threshold processing in the reconstruction, is the wavelet function, t is the time window, and n is the number of sampling points of the transmitted signal.

[0049] Again, the topological consistency is verified by calculating the mutual information (MI) between the reconstructed signal and the multi-scale topological feature to obtain the verified reconstructed signal, including the verified water level signal and the water surface reflection signal. The verification formula is as follows:

[0050] In the formula, is the mutual information calculation, is the reconstructed signal, is the topological feature, and t is the time window. When the signal re-acquisition mechanism is triggered, , where J is the number of layers of wavelet decomposition, N is the number of sampling points of the reflected signal, v is the topological consistency verification index, is the reconstructed signal, For multi-scale topological features, I is , referring to mutual information calculation, where p(x, y) is the joint probability distribution of X and Y, p(x) is the marginal probability distribution of X, and p(y) is the marginal probability distribution of Y.

[0051] Furthermore, please refer to Figure 2 to perform attitude compensation on the verified water level signal , and restore the tilted beam to the true height perpendicular to the water surface through geometric correction to obtain the water level signal after attitude compensation , where the formula for attitude compensation is as follows:

[0052] In the formula, are respectively the pitch angle and roll angle of the UAV measured by the IMU sensor in sequence, is the reconstructed water level signal, is the water level signal after attitude compensation.

[0053] Perform spatio-temporal alignment on the verified water surface reflection signal and the water level signal after attitude compensation, including timestamp alignment and delay compensation alignment, to obtain the flow velocity and the vertical water level height of the water radar corresponding to the water surface of the dual Doppler radar. Among them, the formula for spatio-temporal alignment is as follows:

[0054] In the formula, are the reconstructed signals of the dual Doppler radar and the water level radar, respectively corresponding to the fixed delays of the reconstructed forward water surface reflection signal, forward water surface reflection signal and water level signal, is the flow velocity of the forward Doppler radar corresponding to the water surface, is the flow velocity of the backward Doppler radar corresponding to the water surface, is the vertical water level height of the water radar, is the reconstructed signal of the forward Doppler radar after removing the delay, is the reconstructed signal of the backward Doppler radar after removing the delay, is the reconstructed signal of the water radar after removing the delay and attitude compensation.

[0055] S4: Synthesize the flow velocity vectors according to the flow velocity of the water surface of the dual Doppler radar to obtain the water surface flow velocity; construct a Kalman state equation to predict the water surface flow velocity and water level height at the current moment to obtain the water surface flow velocity and water level height at the next moment, and perform time averaging on the water surface flow velocity at the next moment to obtain the time-averaged flow velocity.

[0056] Please refer to Figure 2 , is the angle between the forward Doppler radar beam and the water flow direction, is the included angle between the backward Doppler radar beam and the water flow direction, is the forward Doppler radar radial velocity component calculated from the water surface velocity of the dual-Doppler radar, is the backward Doppler radar radial velocity component calculated from the water surface velocity of the dual-Doppler radar; according to and , the water surface velocities corresponding to the dual-Doppler radar are synthesized by velocity vectors to obtain the water surface velocity. The specific calculation formula is as follows:

[0057] In the formula, is the water surface velocity, is the forward Doppler radar radial velocity component, is the backward Doppler radar radial velocity component, is the included angle between the forward Doppler radar beam and the water flow direction, is the included angle between the backward Doppler radar beam and the water flow direction, are the signal-to-noise ratios of the forward Doppler radar and the backward Doppler radar, respectively.

[0058] Construct a Kalman state equation that dynamically adjusts based on process noise and observation noise (where the gravitational acceleration and water surface slope parameters are introduced) to predict the water surface velocity and water level height at the current moment and obtain the water surface velocity and water level height at the next moment. The specific formula is as follows:

[0059] In the formula, is the water surface velocity at the next moment (k + 1), is the water level height at the next moment (k + 1), is the time step, g is the gravitational acceleration, S is the water surface slope parameter, is the water surface velocity at the current moment (k), is the water level height at the current moment (k), is the process noise.

[0060] Among them, the observation equation is used to associate the predicted state of the system (water surface velocity and water level) with the actual sensor measurement data (radar signal), and the state estimation is optimized through data assimilation. The observation equation is as follows:

[0061] In the formula, is the process noise, is the observation noise, g is the gravitational acceleration, S is the water surface slope parameter, is the included angle between the beam direction of the forward radar and the water flow direction, is the included angle between the beam direction of the backward radar and the water flow direction, is the pitch angle of the UAV. is the water surface velocity of the forward Doppler radar after spatio-temporal alignment, is the water surface velocity of the backward Doppler radar after spatio-temporal alignment, is the vertical water level height of the water level radar after spatio-temporal alignment.

[0062] Furthermore, time averaging is performed on the water surface velocity at the next moment obtained from the Kalman state equation to obtain the time-averaged velocity.

[0063] Time averaging refers to performing time-weighted averaging on the water surface velocity at the next moment, and the formula is as follows:

[0064] In the formula, is the time-averaged velocity, K is the average time window, is the index quantity, is the water surface velocity at the next moment, is the pitch angle of the UAV.

[0065] S5: Using the turbulence correction factor and the vertical velocity distribution coefficient, the water surface velocity is corrected and converted into the vertical average velocity, spatial grid division and local velocity calculation are performed to obtain the local velocity of the spatial grid; the time-averaged velocity and the local velocity of the spatial grid are weighted to obtain the final velocity, and flow integration is performed to obtain the sectional flow rate.

[0066] First, the signal-to-noise ratio factor is introduced into the water surface velocity as the turbulence correction factor, the water surface velocity is corrected to obtain the corrected water surface velocity, and according to the vertical velocity distribution coefficient the corrected water surface velocity is converted to obtain the vertical average velocity, and the specific formula is as follows:

[0067] In the formula, is the vertical average velocity, is the water surface velocity, is the vertical velocity distribution coefficient, which is dynamically calculated according to the bed roughness and the hydraulic radius and obtained, is the signal-to-noise ratio factor, which is used as the turbulence correction factor, is the signal-to-noise ratio of the forward Doppler radar, is the signal-to-noise ratio of the backward Doppler radar.

[0068] Then, the vertical average velocity Divide it into a grid of N×M, where the horizontal direction is divided with equal width and the vertical direction is divided with equal depth to obtain the vertically averaged flow velocity after division .

[0069] Among them, assign the averaged flow velocity after division to obtain the local flow velocity of the spatial grid , and the formula is as follows:

[0070] In the formula, is the local flow velocity of the spatial grid, is the averaged flow velocity after division, h is the vertical ranging of the water level radar, is the flow velocity exponent (default =7), is the water depth at the center point of the grid, i is the horizontal grid unit, and j is the vertical grid unit.

[0071] Finally, take the time-averaged flow velocity as the baseline value, and perform weighting with the local flow velocity of the spatial grid to obtain the final flow velocity of the grid cell , and perform flow integration to obtain the sectional flow. The specific formula is as follows:

[0072] In the formula, is the final flow velocity of the grid cell, is the local flow velocity of the spatial grid, is the time-averaged flow velocity.

[0073] Among them, the calculation formula for flow integration is as follows:

[0074] In the formula, Q is the sectional flow, is the number of horizontal grids, is the number of vertical grids, i is the horizontal grid unit, j is the vertical grid unit, is the final flow velocity of the spatial grid cell, is the area of the grid cell, , is the horizontal grid size, is the vertical grid size; is the included angle between the water flow direction and the normal direction of the monitoring section, which is obtained through the PID angle calibration result.

[0075] Please refer to Figure 4, which is a structural diagram of an unmanned aerial vehicle (UAV) water surface velocity measurement system provided in the second aspect of an embodiment of the present application, including: a UAV, with a control terminal inside the UAV, a positioning device 1 at the top, an adjusting device and a transmitting and receiving device at the bottom. The transmitting and receiving device, the positioning device 1 and the adjusting device are all electrically connected to the control terminal.

[0076] The control terminal controls the transmitting and receiving device, the positioning device 1 and the adjusting device to obtain the reflection signals of the dual Doppler radar 5 and the water level radar 6 located at the monitoring section.

[0077] The control terminal preprocesses the reflection signals by using the multi-scale topological wavelet decomposition technology to obtain the corrected detail coefficients and multi-scale topological features; The multi-scale topological features and the corrected detail coefficients are denoised and then reconstructed by inverse wavelet transform to obtain a reconstructed signal. Space-time alignment is performed to obtain the flow velocity of the water surface corresponding to the dual Doppler radar 5 and the vertical water level height of the water level radar 6. Flow velocity vector synthesis is performed to obtain the water surface flow velocity. A Kalman state equation is constructed to predict the water surface flow velocity and water level height at the next moment, and time-averaging processing is performed to obtain the time-averaged flow velocity; The water surface flow velocity is corrected by using the turbulence correction factor and the vertical flow velocity distribution coefficient and then converted into the vertical average flow velocity. Space grid division and local flow velocity calculation are performed to obtain the local flow velocity of the space grid; The time-averaged flow velocity and the local flow velocity of the space grid are weighted to obtain the final flow velocity, and flow integration is performed to obtain the sectional flow rate.

[0078] Specifically, please refer to Figure 5 , the UAV includes a flight platform 4, and a number of flight rods 21 extend outward from the flight platform 4. A flight mechanism is provided at one end of the flight rod 21 away from the flight platform 4. The flight mechanism is successively provided with a support plate 25, a motor 24 and a rotor 23 from bottom to top. Among them, a control terminal is provided inside the UAV, a positioning device 1 is provided at the top, and a transmitting and receiving device and an adjusting device are provided at the bottom. The positioning device 1, the transmitting and receiving device and the adjusting device are all electrically connected to the control terminal.

[0079] Further, a protective cover 2 is installed on the top of the flight platform 4, and the positioning device 1 is arranged on the top of the protective cover 2. RTK (Real-Time Kinematic) is selected as the positioning device 1 in this application to receive the plane coordinates of the UAV in real time. There is no limitation on the positioning device here, and it can be selected according to actual situations.

[0080] The transmitting and receiving device includes an IMU sensor 9 and a flowmeter 3. The IMU sensor 9 is used to receive the pitch angle and roll angle data of the radar obtained by the flowmeter 3 in real time, and at the same time communicate with the control terminal through the CAN bus.

[0081] Please refer toFigure 6 , the flowmeter 3 includes a dual Doppler radar 5 and a water level radar 6. Among them, the dual Doppler radar 5 includes a forward Doppler radar and a backward Doppler radar, which are symmetrically installed on the front and back sides of the flowmeter 3 respectively, and the included angle between the axis of its transmitting beam and the water flow direction is adjustable, with a range of 30° to 60°. The water level radar 6 is installed at the bottom of the flowmeter 3 for measuring the vertical distance of the water surface.

[0082] Please refer to Figure 5 , the control terminal is set inside the unmanned aerial vehicle and is located above the center of the flight platform 4. Among them, the upper surface and the lower surface of the flight platform 4 are respectively covered with an upper clamping plate 20 and a lower clamping plate 22, and a number of upper clamping blocks 26 and lower clamping blocks 27 for clamping the upper clamping plate 20 and the lower clamping plate 22 are correspondingly arranged. Among them, through holes are symmetrically arranged at the centers of the upper clamping plate 20 and the lower clamping plate 22 for positioning the control terminal at the center of the unmanned aerial vehicle.

[0083] Please refer to Figure 7 , the control terminal includes a central processor 10, a data receiver 11 and an FPGA 12. Among them, the data receiver 11 is connected to the FPGA 12 through an SPI interface, used to receive the planar coordinates of the unmanned aerial vehicle transmitted by the positioning device 1 and the original signals of the dual Doppler radar 5 and the water level radar 6, and communicates with the FPGA 12 through a CAN bus to real-time feedback the received data of the flowmeter 3.

[0084] Please refer to Figure 8 , the adjusting device is set at the bottom of the unmanned aerial vehicle and is located between the lower clamping plate 22 and the transmitting and receiving device. The adjusting device includes a bearing 14, a turntable 15 and a telescopic rod 17. The turntable 15 is connected to the telescopic rod 17 through an upper connecting pin 16, and the telescopic rod 17 is rigidly connected to the rotating shaft 18 and the transmitting device through a lower connecting pin 19. A soft shell protective cover 13 is sleeved outside the adjusting device to protect the adjusting device, prevent external interference and make the adjustment more precise.

[0085] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0086] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included within the protection scope of the present application.

Claims

1. A method for measuring water surface velocity of an unmanned aerial vehicle based on dual Doppler radar, characterized in that: include: Acquire the reflected signal of the radar, and pre-process the reflected signal using a multi-scale topological wavelet decomposition technique to obtain a corrected detail coefficient and a multi-scale topological feature; Denoising and reconstructing the multi-scale topological features and the corrected detail coefficients to obtain a reconstructed signal, performing spatiotemporal alignment to obtain the flow velocity and water level height on the water body surface, and performing flow velocity vector synthesis on the flow velocity on the water body surface to obtain the water surface flow velocity; Predicting and time-averaging the water surface velocity and the water level to obtain a time-averaged velocity; The water surface velocity is corrected and converted into a vertical average velocity, and spatial grid division and local velocity calculation are performed to obtain a local velocity of the spatial grid; The time average flow velocity and the local flow velocity of the spatial grid are weighted to obtain a final flow velocity, and flow integration is performed to obtain a cross-sectional flow rate; The preprocessing process includes: decomposing the reflected signal according to the number of layers to obtain detail coefficients, extracting topological features to obtain multi-scale topological features, and performing denoising and fusion to obtain corrected detail coefficients.

2. A method for measuring water surface velocity of an unmanned aerial vehicle based on a dual Doppler radar as claimed in claim 1, characterized in that: The water surface velocity and the water level are predicted by using the Kalman state equation that introduces gravity acceleration and water surface gradient. The formula is as follows: ； In the formula, is the water surface velocity at the next moment (k+1), is the water level at the next moment (k+1), is the time step, g is the gravitational acceleration, S is the water surface gradient parameter, is the water surface velocity at the current time k, is the water level at the current time k, is the process noise.

3. The method for measuring water surface velocity of an unmanned aerial vehicle based on a dual Doppler radar as claimed in claim 1, characterized in that: Before performing the spatiotemporal alignment, the mutual information between the reconstructed signal and the multi-scale topological feature is calculated, and the topological consistency of the reconstructed signal is verified using the mutual information to obtain a verified reconstructed signal, including a verified water level signal and a verified water surface reflection signal; Performing attitude compensation on the verified water level signal to obtain an attitude-compensated water level signal; The verified water surface reflection signal and the water level signal after attitude compensation are aligned in time and space to obtain the flow velocity and water level height of the water body surface.

4. The method for measuring water surface velocity of an unmanned aerial vehicle based on a dual Doppler radar as claimed in claim 1, characterized in that: The process of extracting the topological features includes: constructing a complex according to the detail coefficients, calculating the time series of Betti numbers for the complex at each scale, and obtaining the multi-scale topological features.

5. The method for measuring water surface velocity of an unmanned aerial vehicle based on a dual Doppler radar as claimed in claim 1, characterized in that: The post-denoising fusion process includes: incorporating the multi-scale topological features into the noise variance, and calculating a corrected noise standard deviation; The multi-scale topological feature is fused with the detail coefficient by using the corrected noise standard deviation to obtain the corrected detail coefficient.

6. The method for measuring water surface velocity of an unmanned aerial vehicle based on dual Doppler radar as claimed in claim 1, characterized in that: The water surface velocity is corrected by using a signal-to-noise ratio factor as a turbulence correction factor to obtain a corrected water surface velocity, and the vertical velocity distribution coefficient is used to convert the corrected water surface velocity to obtain the vertical average velocity.

7. The method for measuring water surface velocity of an unmanned aerial vehicle based on dual Doppler radar as claimed in claim 1, characterized in that: The formula for the spatiotemporal alignment is as follows: ； In the formula, are the reconstructed signals of the dual Doppler radar and the water level radar, corresponding to the fixed delays of the reconstructed forward water surface reflection signal, backward water surface reflection signal and water level signal, respectively. is the velocity of the water surface corresponding to the forward Doppler radar, is the velocity of the water surface corresponding to the backward Doppler radar, is the vertical water level height of the water level radar, De-delayed reconstructed signal for forward Doppler radar, Reconstruct the signal by removing the delay for backward Doppler radar, Reconstructed signal after delay removal and attitude compensation for water level radar.

8. The method for measuring water surface velocity of an unmanned aerial vehicle based on dual Doppler radar as claimed in claim 1, characterized in that: The process of spatial grid division and local flow velocity calculation includes: dividing the vertical average flow velocity according to equal width and equal depth and then assigning values ​​to obtain the local flow velocity of the spatial grid.

9. The method for measuring water surface velocity of an unmanned aerial vehicle based on dual Doppler radar as claimed in claim 1, characterized in that: The process of denoising the multi-scale topological features and the corrected detail coefficients includes: dynamically denoising the multi-scale topological features to generate a mixed threshold function, and performing joint threshold processing on the corrected detail coefficients according to the mixed threshold function to obtain the detail coefficients after joint threshold processing.

10. A UAV water surface flow velocity measurement device based on dual Doppler radar, using the UAV water surface flow velocity measurement method based on dual Doppler radar described in claim 1, characterized in that: include: A drone, wherein a control terminal is provided inside the drone, a positioning device is provided on the top, and an adjustment device and a transmitting and receiving device are provided on the bottom, and the adjustment device and the transmitting and receiving device are both electrically connected to the control terminal; The control terminal controls the adjustment device and the transmitting and receiving device to obtain the reflected signal of the radar; The control terminal pre-processes the reflected signal using a multi-scale topological wavelet decomposition technique to obtain a corrected detail coefficient and a multi-scale topological feature; Denoising and reconstructing the multi-scale topological features and the corrected detail coefficients to obtain a reconstructed signal, performing spatiotemporal alignment to obtain the flow velocity and water level height on the water body surface, and performing flow velocity vector synthesis on the flow velocity on the water body surface to obtain the water surface flow velocity; Predicting the water surface velocity and the water level height and performing time-average processing to obtain a time-average velocity; The water surface velocity is corrected and converted into a vertical average velocity, and spatial grid division and local velocity calculation are performed to obtain a local velocity of the spatial grid; The time average flow velocity and the local flow velocity of the space grid are weighted to obtain the final flow velocity, and the cross-sectional flow is obtained by flow integration.

Images