Position data weight calibration system for SAR imaging of unmanned aerial vehicle
By constructing experimental scenarios and using guide rail and data fusion algorithms, the problem of inaccurate motion compensation in drone SAR imaging is solved, and high-precision and quasi-real-time SAR imaging effect is achieved.
Patent Information
- Application Number
- CN202410017267.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-05
- Publication Date
- 2025-07-08
AI Technical Summary
During SAR imaging, due to fluctuations in flight trajectory and speed, the prior art cannot accurately compensate motion, resulting in low imaging accuracy, especially in scenes without obvious reflector targets, which is poor in focus and difficult to deal with in real time.
By constructing an experimental scenario, a variety of position data are obtained using the first linear guide rail, the second linear guide rail, the lifting guide rail, the motion sensor and the synthetic aperture radar, and the weight of each data is calculated using the global optimization algorithm through the weight calculation module to perform data fusion and compensation.
It improves the accuracy of data fusion and motion compensation effect, realizes high-precision SAR imaging, has quasi-real-time imaging capabilities without increasing hardware burden, and is suitable for drones with smaller load capacity.
Smart Images
Figure CN120279104A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of radar imaging, and in particular to a position data weight calibration system for UAV SAR imaging. Background Art
[0002] Multi-rotor UAVs mostly fly in the troposphere, and are greatly affected by air flow disturbances during flight. In addition, due to factors such as UAV performance and operation level, the UAV cannot move forward in a uniform straight-line motion state. However, the SAR imaging technology is based on a uniform straight-line flight state. The fluctuations in the flight trajectory and flight speed of the UAV cause the SAR imaging system to be unable to directly operate. Therefore, motion compensation of the flight trajectory is required.
[0003] Conventional solutions usually use one type of data for motion compensation. For example, the position and speed data of the multi-rotor UAV measured by motion sensors (such as inertial navigation, GPS, integrated navigation, etc.) are used for motion parameter compensation, or radar echo data is used for motion compensation.
[0004] However, motion sensors usually have problems such as low positioning accuracy and cumulative errors, and cannot meet the actual requirements of high-precision motion compensation for UAV SAR imaging systems; when using radar echo data, if there are no obvious reflector targets in the imaging scene, it will lead to poor focusing effect and there are problems that cannot be processed in real time. Therefore, there is a lack of fusion processing of multiple types of data in the prior art, resulting in inaccurate motion compensation data and low SAR imaging accuracy. Summary of the Invention
[0005] In view of the above analysis, an embodiment of the present invention aims to provide a position data weight calibration system for UAV SAR imaging to solve the problem that the motion parameter compensation of the UAV is inaccurate due to the lack of fusion of data from multiple sources in the prior art.
[0006] An embodiment of the present invention provides a position data weight calibration system for UAV SAR imaging, including: a first linear guide rail, a second linear guide rail, a lifting guide rail, a motion sensor, a synthetic aperture radar, a data acquisition module, and a weight calculation module; the second linear guide rail is installed on the slider of the first linear guide rail and is perpendicular to the first linear guide rail on the horizontal plane; the lifting guide rail is installed on the slider of the second linear guide rail and is perpendicular to the second linear guide rail on the vertical plane; the motion sensor and the synthetic aperture radar are installed on the slider of the lifting guide rail;
[0007] The data acquisition module is configured to obtain the first position data collected by the motion sensor and the radar echo data collected by the synthetic aperture radar when the sliders of the three guide rails move simultaneously in each experiment, and send them to the weight calculation module;
[0008] The weight calculation module is configured to receive and convert the first position data in each experiment, obtain the second position data according to the received radar echo data, and obtain the third position data according to the accelerations of the sliders of the three guide rails; calculate three position weights by using a global optimization algorithm based on the errors between the three types of position data and the expected position data in multiple experiments.
[0009] For a further improvement based on the above system, the sliders of the three guide rails move along the accelerations in three directions respectively, simulating the motion mode of an unmanned aerial vehicle. The accelerations in the three directions are pre-measured by the inertial navigation equipment mounted when the unmanned aerial vehicle flies uniformly along the azimuth direction at a set speed.
[0010] For a further improvement based on the above system, the motion direction of the first linear guide rail is taken as the azimuth direction; at each experiment, the three types of position data and the expected position data at each radar sampling point are obtained according to the time of the radar sampling points; the position data all include the position data in two dimensions of the line of sight direction and the azimuth direction.
[0011] For a further improvement based on the above system, the weight calculation module calculates three position weights by using a global optimization algorithm based on the errors between the three types of position data and the expected position data in multiple experiments, including:
[0012] Obtain the variances of the three position errors at each radar sampling point according to the three position errors at each radar sampling point in multiple experiments;
[0013] Construct an objective function according to the three position weights, the three position errors and the variances of the position errors, calculate the optimal solutions of the three position weights in the multiple experiment data respectively by using a global optimization algorithm, and take the average to obtain the final three position weights.
[0014] For a further improvement based on the above system, construct an objective function according to the three position weights, the three position errors and the variances of the position errors through the following formula:
[0015]
[0016] where M represents the number of radar sampling points, Q1, Q2 and Q3 respectively represent the first position weight from the motion sensor, the second position weight from the synthetic aperture radar, and the third position weight from the simulated unmanned aerial vehicle flight, Δp ni , Δq ni and Δs ni respectively represent the position errors from the motion sensor, the synthetic aperture radar and the simulated unmanned aerial vehicle at the i-th radar sampling point in a set of experiment data, R ci , R di and R eirespectively represent the variances of the position errors from the motion sensor, synthetic aperture radar, and simulated UAV flight at the i-th radar sampling point.
[0017] Based on the further improvement of the above system, the desired azimuth position data in the desired position data is calculated according to the set speed of the UAV and the time of each radar sampling point, and the desired line-of-sight position data in the desired position data is 0.
[0018] Based on the further improvement of the above system, the weight calculation module converts the first position data, including: synchronizing the first position data with the radar sampling points according to the time stamp, and for the radar sampling points lacking the first position data among two adjacent first position data, filling in the data by linear interpolation, and then converting the first position data at each radar sampling point into the first line-of-sight position data and the first azimuth position data in the SAR image coordinate system according to the azimuth.
[0019] Based on the further improvement of the above system, the second position data obtained by the weight calculation module according to the received radar echo data includes the second line-of-sight position data and the second azimuth position data, and is obtained by calculating the displacement data of the received radar echo data in the line-of-sight and azimuth directions.
[0020] Based on the further improvement of the above system, calculating the displacement data of the received radar echo data in the line-of-sight and azimuth directions includes:
[0021] Segmenting the radar echo data in the azimuth direction, calculating the correlation peak of adjacent two azimuth spectra to obtain the frequency shift amount; using a high-pass filter and a low-pass filter to extract the frequency change values in the line-of-sight and azimuth directions from the frequency shift amount, and then multiplying by the corresponding line-of-sight and azimuth acceleration coefficients to obtain two accelerations, and respectively performing double integration on the two accelerations to obtain the displacement data of the radar echo data in the line-of-sight and azimuth directions.
[0022] Based on the further improvement of the above system, obtaining the third position data according to the accelerations of the sliders of the three guide rails includes: respectively performing double integration on the accelerations of the sliders of the first linear guide rail, the second linear guide rail, and the lifting guide rail to obtain the first distance, the second distance, and the third distance; the first distance is used as the third azimuth position data in the third position data; calculating the square root of the sum of the squares of the second distance and the third distance to obtain the third line-of-sight position data in the third position data.
[0023] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects:
[0024] 1. Construct an experimental scenario through guide rails to simulate the movement mode of the UAV, reducing the experimental cost; consider the influence of data from multiple sources on the UAV flight, and through the statistical analysis of multiple test data, obtain reasonable data weights, improve the rationality and accuracy of data fusion, enhance the motion compensation effect, and facilitate obtaining high-precision SAR images;
[0025] 2. Obtain weight calibration data through experiments, which is convenient to be used as preset data for the actual flight of the UAV, enabling the UAV SAR imaging to have quasi-real-time imaging ability, and without increasing the hardware burden, being suitable for application on UAVs with small load capacity.
[0026] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination schemes. Other features and advantages of the present invention will be described in the subsequent specification, and some advantages can be made obvious from the specification, or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained through the content specifically pointed out in the specification and the drawings. Description of the Drawings
[0027] The drawings are only for the purpose of showing specific embodiments, and are not considered as a limitation to the present invention. Throughout the drawings, the same reference signs represent the same components.
[0028] Figure 1 It is a block diagram of a position data weight calibration system for UAV SAR imaging in an embodiment of the present invention;
[0029] Figure 2 It is a schematic processing flow diagram of a weight calculation module in an embodiment of the present invention. Detailed Embodiments
[0030] The following will specifically describe the preferred embodiments of the present invention in conjunction with the drawings, where the drawings form a part of this application and are used together with the embodiments of the present invention to explain the principles of the present invention, and are not used to limit the scope of the present invention.
[0031] A specific embodiment of the present invention discloses a position data weight calibration system for UAV SAR imaging, as Figure 1 shown, including: a first linear guide rail, a second linear guide rail, a lifting guide rail, a motion sensor, a synthetic aperture radar, a data acquisition module, and a weight calculation module; the second linear guide rail is installed on the slider of the first linear guide rail and is perpendicular to the first linear guide rail on the horizontal plane; the lifting guide rail is installed on the slider of the second linear guide rail and is perpendicular to the second linear guide rail on the vertical plane; the motion sensor and the synthetic aperture radar are installed on the slider of the lifting guide rail;
[0032] The data acquisition module is used to obtain the first position data collected by the motion sensor and the radar echo data collected by the synthetic aperture radar when the sliders of the three guide rails move simultaneously in each experiment, and send them to the weight calculation module;
[0033] The weight calculation module is used to receive and convert the first position data in each experiment, obtain the second position data according to the received radar echo data, and obtain the third position data according to the acceleration of the sliders of the three guide rails; and calculate the three position weights using a global optimization algorithm based on the errors between the three position data and the expected position data in multiple experiments.
[0034] It should be noted that in this embodiment, the experimental scene is constructed by three guide rails, and the positional relationship of the three guide rails is similar to the three axes of the space coordinate system, with the movement direction of the first linear guide rail as the azimuth.
[0035] During the experiment, the sliders of the three guide rails were controlled to move in three directions with acceleration to simulate the movement of the UAV, that is, to ensure that the running trajectories of the motion sensors and synthetic aperture radar installed on the sliders of the lifting guide rails were close to the uniform linear motion of the UAV, thereby reducing the experimental cost while being close to the actual scene.
[0036] It should be noted that the accelerations in the three directions are measured in advance by the mounted inertial navigation equipment when the UAV flies at a constant speed in the azimuth direction at a set speed.
[0037] Furthermore, the motion sensor can use an inertial navigation device with cumulative errors but high short-term relative accuracy and fast update rate; it can also use GPS, Beidou and other positioning devices with a slower update rate but no cumulative errors; it can also use a combined navigation device that combines the above two.
[0038] It should be noted that the data acquisition module and the weight calculation module are both deployed in the host computer. The data collected by the motion sensor and the synthetic aperture radar can be first stored in their respective memories, and after the experiment is over, the data acquisition module in the host computer reads the memory to obtain the corresponding data. Alternatively, it is connected to the host computer by wired or wireless means, and the data acquisition module in the host computer receives the first position data collected by the motion sensor and the radar echo data collected by the synthetic aperture radar in real time, and then sends them to the weight calculation module.
[0039] In this embodiment, the radar sampling point is used as the standard, that is, three types of position data and expected position data at each radar sampling point are obtained according to the time of the radar sampling point in each experiment; the position data includes position data in two dimensions: line of sight and azimuth.
[0040] like Figure 2 As shown, the processing flow of the weight calculation module includes the following steps:
[0041] (1) Obtain the first position data from the motion sensor at each radar sampling point.
[0042] It should be noted that since the sampling rate of the motion sensor is generally lower than that of the synthetic aperture radar, the number of sampling points of the motion sensor within the same time period is less than the number of radar sampling points of the radar echo data. Moreover, the first position data collected by the motion sensor is the three-dimensional coordinate values based on the geodetic coordinate system: longitude, latitude, and altitude, and needs to be converted to the SAR image coordinate system.
[0043] Therefore, after receiving the first position data collected by the motion sensor, the weight calculation module needs to perform conversion processing, including:
[0044] Synchronize the first position data with the radar sampling points according to the time stamp. Among two adjacent first position data, linearly interpolate to fill in the missing first position data at the radar sampling points, and then convert the first position data at each radar sampling point to the first line-of-sight position data and the first azimuth position data in the SAR image coordinate system according to the azimuth direction.
[0045] (2) Obtain the second position data from the synthetic aperture radar at each radar sampling point.
[0046] It should be noted that the second position data obtained by the weight calculation module from the radar echo data includes the second line-of-sight position data and the second azimuth position data, which are obtained by calculating the displacement data of the received radar echo data in the line-of-sight direction and the azimuth direction.
[0047] Since the radar echo data obtained by the synthetic aperture radar is collected during the continuous movement of the radar, the radar echo data contains the motion information of the radar. In this embodiment, the following steps are used to calculate the displacement data of the radar echo data in the line-of-sight direction and the azimuth direction:
[0048] ① Take half of the synthetic aperture length as the segmentation length and segment the radar echo data in the azimuth direction; calculate the azimuth spectrum of each segment of data according to the fast Fourier transform, that is, the azimuth frequency data.
[0049] ② Calculate the correlation peak of two adjacent azimuth spectra according to the cross-correlation function to obtain the frequency shift amount.
[0050] It should be noted that the frequency shift amount consists of two parts: low-frequency information and high-frequency information. Among them, the frequency change in the azimuth direction is usually located in the low-frequency band, while the frequency change in the line-of-sight direction usually occurs in the high-frequency band.
[0051] ③ Use a high-pass filter and a low-pass filter to extract the frequency change values in the line-of-sight direction and the azimuth direction from the frequency shift amount, and then multiply them by the corresponding acceleration coefficients in the line-of-sight direction and the azimuth direction to obtain two accelerations. Integrate the two accelerations twice respectively to obtain the displacement data of the radar echo data in the line-of-sight direction and the azimuth direction.
[0052] Specifically, calculate the acceleration coefficient a1 in the line-of-sight direction and the acceleration coefficient a2 in the azimuth direction through the following formula:
[0053]
[0054] Where λ is the wavelength of the radar echo signal, Δτ is the time interval between two adjacent non-overlapping data segments in the azimuth direction, and R0 is the slant range of the radar from the target center.
[0055] (3) Obtain the third position data from the simulated flight of the UAV at each radar sampling point.
[0056] It should be noted that in this embodiment, the flight of the UAV is simulated by the movement trajectories of three sliders. Therefore, obtaining the third position data according to the accelerations of the sliders of the three guide rails includes: integrating the accelerations of the sliders of the first linear guide rail, the second linear guide rail, and the lifting guide rail twice respectively to obtain the first distance, the second distance, and the third distance; the first distance is used as the third azimuth position data in the third position data; calculate the square root of the sum of the squares of the second distance and the third distance to obtain the third line-of-sight position data in the third position data.
[0057] (4) Obtain the desired position data.
[0058] It should be noted that it is expected that the UAV flies in a uniform linear motion along the azimuth direction. Therefore, the desired azimuth position data in the desired position data is calculated according to the set speed of the UAV and the time of each radar sampling point, and the desired line-of-sight position data in the desired position data is 0.
[0059] (5) Calculate the three position errors and variances at each radar sampling point.
[0060] In each experiment, obtain the three position data at each radar sampling point through (1), (2), and (3), and compare them with the desired position data in (4) to obtain the three position errors at each radar sampling point, that is, obtain the position errors in two dimensions of the line-of-sight direction and the azimuth direction.
[0061] After multiple experiments, according to the position errors in two dimensions at each radar sampling point in multiple experiments, calculate the variances of the position errors in two dimensions at each radar sampling point respectively.
[0062] (6) Construct an objective function and calculate three position weights.
[0063] It should be noted that, in order to accurately fuse the three types of position data and improve the accuracy of motion data, position weights are set for each type of position data, and the position weights are obtained by constructing an objective function and solving the optimal solution.
[0064] Specifically, the objective function represents that the sum of the errors after fusing the three types of position errors at each radar sampling point is minimized, and the objective function is constructed by the following formula:
[0065]
[0066] where M represents the number of radar sampling points, Q1, Q2, and Q3 respectively represent the first position weight from the motion sensor, the second position weight from the synthetic aperture radar, and the third position weight from the simulated UAV flight, and Δp ni , Δq ni , and Δs ni respectively represent the position errors from the motion sensor, the synthetic aperture radar, and the simulated UAV at the i-th radar sampling point in a set of experimental data, and R ci , R di , and R ei respectively represent the variances of the position errors from the motion sensor, the synthetic aperture radar, and the simulated UAV flight at the i-th radar sampling point.
[0067] It can be understood that since the position data includes two dimensions, the line-of-sight direction and the azimuth direction, in the SAR image coordinate system, the objective function is established and solved separately for the two dimensions according to the position errors and variances corresponding to each dimension to obtain the optimal solutions of the position weights for the corresponding dimensions.
[0068] It should be noted that after solving the optimal solutions of the three types of position weights for multiple sets of experimental data respectively using the quantum particle swarm optimization algorithm or the genetic algorithm according to formula (2), the average value is taken as the final position weight.
[0069] After obtaining the calibration data of the three types of position weights according to this embodiment, it can be directly applied to the SAR imaging scenario during the UAV flight: during the UAV flight, the three types of position data in the two dimensions of the line-of-sight direction and the azimuth direction are obtained at each radar sampling point, and according to the preset three types of position weights, the weighted average of the first position data, the second position data, and the third position data corresponding to each radar sampling point in the corresponding dimension is calculated to obtain the corrected position data in the two dimensions; then, according to the corrected position data and the received radar echo data, the backscattering algorithm is called to generate the SAR image.
[0070] Compared with the prior art, the position data weight calibration system for UAV SAR imaging provided in this embodiment constructs an experimental scenario through a guide rail to simulate the movement mode of the UAV, reducing the experimental cost; considering the influence of data from multiple sources on the UAV flight, through the statistical analysis of multiple test data, reasonable data weights are obtained, improving the rationality and accuracy of data fusion, enhancing the motion compensation effect, and facilitating the acquisition of high-precision SAR images; weight calibration data is obtained through experiments, which is convenient to be used as preset data in the actual flight of the UAV, enabling the UAV SAR imaging to have quasi-real-time imaging capabilities, and without increasing the hardware burden, being suitable for application on UAVs with relatively small load capacities.
[0071] Those skilled in the art can understand that all or part of the processes for implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.
[0072] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.
Claims
1. A position data weight calibration system for UAV SAR imaging, characterized in that, Including: A first linear guide rail, a second linear guide rail, a lifting guide rail, a motion sensor, a synthetic aperture radar, a data acquisition module, and a weight calculation module; the second linear guide rail is installed on the slider of the first linear guide rail and is perpendicular to the first linear guide rail on the horizontal plane; the lifting guide rail is installed on the slider of the second linear guide rail and is perpendicular to the second linear guide rail on the vertical plane; the motion sensor and the synthetic aperture radar are installed on the slider of the lifting guide rail; The data acquisition module is used to obtain the first position data collected by the motion sensor and the radar echo data collected by the synthetic aperture radar when the sliders of the three guide rails move simultaneously in each experiment, and send them to the weight calculation module; The weight calculation module is used to receive and convert the first position data in each experiment, obtain the second position data according to the received radar echo data, and obtain the third position data according to the accelerations of the sliders of the three guide rails; according to the errors between the three kinds of position data and the expected position data in multiple experiments, use the global optimization algorithm to calculate the three position weights.
2. The position data weight calibration system for UAV SAR imaging according to claim 1, wherein The sliders of the three guide rails move along the accelerations in three directions respectively, simulating the movement mode of the unmanned aerial vehicle. The accelerations in the three directions are pre-measured by the inertial navigation equipment mounted when the unmanned aerial vehicle flies uniformly along the azimuth direction at a set speed.
3. The position data weight calibration system for UAV SAR imaging according to claim 2, characterized in that Taking the movement direction of the first linear guide rail as the azimuth direction; in each experiment, obtain the three kinds of position data and the expected position data at each radar sampling point according to the time of the radar sampling points; the position data all include the position data in two dimensions of the line-of-sight direction and the azimuth direction.
4. The position data weight calibration system for UAV SAR imaging according to claim 3, characterized in that The weight calculation module calculates the three position weights by using the global optimization algorithm according to the errors between the three kinds of position data and the expected position data in multiple experiments, including: Obtaining the variances of the three position errors at each radar sampling point according to the three position errors at each radar sampling point in multiple experiments; Constructing an objective function according to the three position weights, the three position errors, and the variances of the position errors, using the global optimization algorithm to calculate the optimal solutions of the three position weights in the multiple experimental data respectively, and taking the average to obtain the final three position weights.
5. The position data weight calibration system for UAV SAR imaging according to claim 4, wherein The objective function is constructed according to the three position weights, the three position errors, and the variances of the position errors through the following formula: Where M represents the number of radar sampling points, Q1, Q2, and Q3 represent the first position weight from the motion sensor, the second position weight from the synthetic aperture radar, and the third position weight from the simulated UAV flight respectively, and Δp ni , Δq ni , and Δs ni represent the position errors from the motion sensor, the synthetic aperture radar, and the simulated UAV at the i-th radar sampling point in a single experimental data respectively. R ci , R di , and R ei represent the variances of the position errors from the motion sensor, the synthetic aperture radar, and the simulated UAV flight at the i-th radar sampling point respectively.
6. The position data weight calibration system for UAV SAR imaging according to claim 3, characterized in that The expected azimuth position data in the expected position data is calculated according to the set speed of the unmanned aerial vehicle and the time of each radar sampling point, and the expected line-of-sight position data in the expected position data is 0.
7. The position data weight calibration system for UAV SAR imaging according to claim 3, characterized in that, The weight calculation module converts the first position data, including: synchronizing the first position data with the radar sampling points according to the time stamps, and filling in the data of the radar sampling points lacking the first position data by linear interpolation method between two adjacent first position data, and then converting the first position data at each radar sampling point into the first line-of-sight position data and the first azimuth position data in the SAR image coordinate system according to the azimuth direction.
8. The position data weight calibration system for UAV SAR imaging according to claim 3, characterized in that The second position data obtained by the weight calculation module according to the received radar echo data includes the second line-of-sight position data and the second azimuth position data, and is obtained by calculating the displacement data of the received radar echo data in the line-of-sight direction and the azimuth direction.
9. The position data weight calibration system for UAV SAR imaging according to claim 8, characterized in that Calculating the displacement data of the received radar echo data in the line-of-sight direction and the azimuth direction, including: Segmenting the radar echo data in the azimuth direction, calculating the correlation peak of adjacent two-segment azimuth spectra to obtain the frequency shift amount; using a high-pass filter and a low-pass filter to extract the frequency change values in the line-of-sight direction and the azimuth direction from the frequency shift amount, and then multiplying by the corresponding acceleration coefficients in the line-of-sight direction and the azimuth direction to obtain two accelerations, and respectively performing double integration on the two accelerations to obtain the displacement data of the radar echo data in the line-of-sight direction and the azimuth direction.
10. The position data weight calibration system for UAV SAR imaging according to claim 3, characterized in that, Obtaining the third position data according to the accelerations of the sliders of the three guide rails, including: respectively performing double integration on the accelerations of the sliders of the first linear guide rail, the second linear guide rail and the lifting guide rail to obtain the first distance, the second distance and the third distance; using the first distance as the third azimuth position data in the third position data; calculating the square root of the sum of the squares of the second distance and the third distance to obtain the third line-of-sight position data in the third position data.