A real-time adaptive data calibration method for a laser radar

Abstract

The application provides a real-time adaptive data calibration method of a laser radar. The steps of the application are as follows: step 1, initializing system parameters; step 2, acquiring Doppler frequency shift data of the laser radar; step 3, calculating system inherent deviation; step 4, judging inherent deviation constraint conditions; step 5, performing frequency shift calibration; and step 6, updating and outputting calibration results and wind speed inversion. The application is applied to the technical field of laser radars.

Description

Technical Field

[0001] This invention relates to the field of lidar technology, specifically to a real-time adaptive data calibration method for lidar, applicable to lidar systems for meteorological detection, wind farm wind measurement, environmental monitoring, and other applications requiring high-precision wind speed measurement. Background Technology

[0002] Laser Doppler radar, as a high-precision wind speed measurement device, has been widely used in meteorological detection, wind farm wind measurement, environmental monitoring, and other fields. Its working principle involves emitting modulated laser signals, receiving the echoes scattered by atmospheric particles, and measuring the Doppler frequency shift to invert the radial wind speed of the target area. However, existing laser Doppler radars are susceptible to factors such as temperature changes, light source aging, and system local oscillator drift during long-term operation, leading to systematic shifts in the echo signal frequency and ultimately causing wind speed inversion errors, affecting measurement accuracy.

[0003] To correct the aforementioned errors, traditional calibration methods primarily rely on external standard targets or offline calibration models. The former requires setting up a standard reference device in a specific environment, is highly dependent on environmental conditions (e.g., deployment is impossible in severe weather), and cannot be calibrated in real-time during radar operation. The latter builds a model based on preset static calibration parameters or periodic calibration data, failing to dynamically track real-time deviations caused by hardware drift and environmental disturbances, leading to continuous error accumulation over long-term observations. Furthermore, existing calibration algorithms lack constraint mechanisms: current technologies often fail to set constraints during system deviation estimation, making them prone to abnormal deviation estimations when noise is strong or signal fluctuations are large, affecting the reliability of calibration results. Simultaneously, the utilization of lidar blind zone signals is insufficient. Signals within the blind zone, formed by emitted light leakage and optical crosstalk (typically located within a short distance of tens to hundreds of meters from the radar), while lacking effective wind field information, stably reflect the system's local oscillator frequency characteristics. Existing technologies do not adequately extract the spectral characteristics of this signal, resulting in limited accuracy in estimating the system's inherent deviations, making it difficult to meet the requirements of high-precision wind measurement.

[0004] Therefore, how to utilize the characteristics of the blind zone signal of lidar to achieve real-time extraction and dynamic compensation of the inherent deviation of the system, and solve the problems of poor real-time performance, strong environmental dependence and inability to dynamically correct errors in traditional calibration, has become the key to improving the accuracy and stability of lidar Doppler radar wind measurement, and is also the purpose of this invention. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a real-time adaptive data calibration method for lidar. This method utilizes the characteristic that the lidar blind zone signal stably reflects the local oscillator frequency characteristics of the system, extracts the inherent deviation of the system in real time, and then performs dynamic calibration on the frequency shift data of the effective observation layer based on the deviation. At the same time, the validity of the deviation is ensured by judging the constraint conditions, and finally the high-precision inversion of wind speed is achieved.

[0006] The technical solution adopted in this invention is a real-time adaptive data calibration method for lidar, which includes the following steps:

[0007] Step 1: Initialize the lidar system parameters, including laser wavelength, laser pulse width, range layer resolution, sampling frequency, modulation intermediate frequency, and deviation upper limit constraint threshold.

[0008] Step 2: Control the lidar to emit modulated light signals and receive echo signals to obtain Doppler frequency shift observation values ​​for each range layer; divide the observation space into multiple range layers according to the range layer resolution, wherein the near-range layer is the blind zone layer and the far-range layer is the effective observation layer;

[0009] Step 3: Calculate the inherent bias of the system based on the difference between the observed Doppler frequency shift of the blind zone layer and the theoretical reference frequency shift;

[0010] Step 4: Determine whether the difference between the system inherent deviation at the current moment and the effective system inherent deviation at the previous moment exceeds the upper limit constraint threshold of the deviation. If it exceeds the threshold, the effective system inherent deviation at the previous moment is used as the effective system inherent deviation at the current moment; if it does not exceed the threshold, the system inherent deviation calculated at the current moment is the effective system inherent deviation.

[0011] Step 5: Use the inherent bias of the effective system at the current time to calibrate the Doppler frequency shift observations of the effective observation layer to obtain the calibrated frequency shift values;

[0012] Step 6: Store the current effective system inherent bias in the system buffer for subsequent data calibration; and based on the calibrated frequency shift value and the theoretical reference frequency shift, invert the radial wind speed of the effective observation layer using the Doppler frequency shift relation.

[0013] Repeat steps 2-6 to achieve real-time adaptive calibration of lidar data.

[0014] Further, in step 2, the blind zone layer is a near-range layer of 0-60m from the lidar, specifically including a range layer 1 of 0-30m and a range layer 2 of 30-60m; the effective observation layer is a far-range layer of more than 60m from the lidar, defined as range layer 3 and subsequent range layers.

[0015] Furthermore, in step 3, the formula for calculating the inherent bias of the system is:

[0016] ,

[0017] in, For t i The system's inherent bias at time t, where m is the number of blind zones used for estimation. f obs ( t i , r j ) for t i Doppler frequency shift observations at time j of the dead zone layer For t i The theoretical reference frequency shift or expected frequency shift value of the j-th blind zone layer at time j.

[0018] Furthermore, in step 1, the upper limit constraint threshold for deviation is set to Δ. f th In step 4, the judgment condition is: ,

[0019] in, For t i-1 The inherent bias of the effective system at any given time.

[0020] Furthermore, in step 5, the formula for calibrating the effective observation layer frequency shift data is:

[0021] ,

[0022] in, For t i The calibrated frequency shift value of the k-th effective observation layer at time k. f obs ( t i , r k ) for t i The Doppler shift observation value of the kth effective observation layer at time k.

[0023] Furthermore, in step 6, the Doppler frequency shift relationship for retrieving radial wind speed is:

[0024] ,

[0025] in, For t i The radial wind speed at the k-th effective observation layer at time k, where λ is the laser wavelength.

[0026] Furthermore, in step 1, the laser wavelength λ = 1550 nm.

[0027] Furthermore, in step 1, the sampling frequency is 5×10⁻⁶. 8 Hz.

[0028] Furthermore, in step 1, the modulation intermediate frequency is 80MHz.

[0029] Furthermore, in step 1, the distance layer resolution is 30m.

[0030] The beneficial effects of this invention are:

[0031] 1. This invention utilizes the blind zone signal of lidar (which does not contain wind field information but reflects local oscillator characteristics) to extract the inherent deviation of the system in real time. It does not require external standard targets or offline calibration models, thus eliminating the dependence of traditional calibration on environmental conditions and realizing the "online" calibration process. It solves the problem of poor real-time performance of traditional calibration and ensures that lidar can correct errors in real time during long-term operation.

[0032] 2. This invention continuously updates the inherent bias of the system through time series and dynamically calibrates the effective observation layer data based on the current effective bias. It can compensate for dynamic errors caused by temperature changes, light source aging, and system local oscillator drift in real time, avoid the accumulation of errors in long-term observation, and significantly improve the accuracy and stability of lidar wind measurement data.

[0033] 3. This invention sets an upper limit constraint threshold for deviation. By judging whether the difference between the current deviation and the deviation at the previous moment exceeds the limit, valid deviations are screened. This can avoid the impact of abnormal deviations (such as deviation abrupt changes caused by signal interference) on the calibration results and ensure the stability and continuity of the calibration process.

[0034] 4. The calibration logic of this invention relies only on the signal processing and parameter settings of the lidar itself, without being bound to the hardware structure of a specific lidar model. Furthermore, it can be adapted to different application scenarios (such as meteorological detection and wind farm wind measurement) by adjusting parameters such as range layer resolution and deviation threshold. It has good versatility and is suitable for various types of lidar Doppler radar systems. Attached Figure Description

[0035] Figure 1 This is a system framework diagram that matches the method of this invention;

[0036] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0037] like Figure 1As shown, the method of this invention is implemented through a real-time adaptive data calibration system framework of lidar. This system uses a lidar Doppler radar to emit a laser beam carrying a modulated signal to perform multi-level detection of the three-dimensional atmospheric space. During signal propagation and scattering, the echo is received and processed to obtain Doppler frequency shift data related to wind field motion. The system divides the detection space into several range layers according to the ranging resolution, where the near-range layer (blind zone) is used to extract the inherent bias of the system, and the far-range layer (effective observation area) is used for accurate wind field inversion and analysis.

[0038] At any time Where n is a constant, the system calculates the current inherent deviation based on the frequency shift characteristics of the blind zone layer. , where i is a constant, and this deviation characterizes the radar's internal error state at that moment. The system adaptively corrects the observation data of the outer layer of the blind zone based on this deviation, realizing dynamic compensation for hardware drift, environmental disturbances, and time-accumulated errors.

[0039] As the time series progresses, the system continuously updates its inherent bias parameters and dynamically adjusts them. Through interactive data processing, the system continuously senses and adaptively corrects the three-dimensional atmospheric field, thereby ensuring the stability and accuracy of the observation results.

[0040] The adaptive calibration algorithm proposed in this invention is based on the Doppler frequency shift measurement principle of laser Doppler radar. It achieves real-time adaptive calibration of radar wind measurement data by extracting the inherent bias of the blind zone signal and dynamically correcting the effective layer data. The main algorithm flow is as follows: Figure 2 As shown, the process includes steps such as inherent bias extraction, condition constraint judgment, data calibration, and wind speed inversion.

[0041] During the system observation process, the laser Doppler radar in time... The frequency shift observation obtained at time t can be expressed as:

[0042] ,

[0043] in, For distance layer The actual observed frequency shift at that location This represents the ideal frequency shift corresponding to a real atmospheric echo. This represents the system's inherent bias term that varies over time. To observe noise.

[0044] In multi-range gate structures, the range layers near the radar blind zone (such as range layer 1 and range layer 2) are prone to significant inherent biases due to signal reflection and system coupling. The inherent biases of multiple blind zones can be expressed as:

[0045] ,

[0046] in, The number of blind layers used for estimation. f obs ( t i , r j ) for t i Doppler frequency shift observations at time j of the dead zone layer This represents the theoretical reference frequency shift or the desired frequency shift value. The formula obtains the value at time [time value] by averaging observations from multiple blind zones. Effective inherent bias estimation under [the given conditions].

[0047] When the calculated inherent bias meets the set constraints (e.g., the change in inherent bias between consecutive time steps does not exceed a preset threshold, the bias stability meets statistical confidence requirements, etc.), the system determines the bias as the valid inherent bias at the current time and calibrates the data outside the blind zone (effective observation layer) accordingly. The calibration relationship for different distance layers can be expressed as:

[0048] ,

[0049] in, This is the calibrated frequency shift value. For any effective distance layer outside the blind zone, f obs ( t i , r k ) for t i The Doppler frequency shift observation at time k is the value of the effective observation layer. The calibrated frequency shift data is used for wind speed inversion and atmospheric motion feature extraction.

[0050] As the time series progresses, the system continuously updates the inherent bias parameters at each time point. Furthermore, by comparing the effective deviation between the current time and the previous time, the inherent deviation is dynamically adjusted, thereby ensuring the stability and real-time nature of the calibration results.

[0051] The present invention will be further described below with reference to specific embodiments.

[0052] This invention provides a real-time adaptive data calibration method for lidar, addressing the problem of Doppler frequency shift measurement error accumulation in lidar with blind zone signals, and achieving dynamic adaptive calibration based on blind zone signals. The specific implementation steps are as follows:

[0053] Step 1: Initialize system parameters

[0054] The laser wavelength is set to λ = 1550 nm, the laser pulse width is 400 ns, and the corresponding system blind zone is approximately 60 m; the distance layer resolution is 30 m; and the sampling frequency is 5 × 10⁻⁶. 8 The frequency is Hz, and the modulation intermediate frequency is 80MHz. The upper limit constraint threshold for deviation is set to Δf. th =0.15MHz, used to limit abnormal changes in inherent bias.

[0055] Step 2: Obtain Doppler frequency shift data from the lidar.

[0056] Step 201: The lidar transmits modulated optical signals and receives echo signals to obtain the observation frequency shift for each range layer. .

[0057] Step 202: Divide the observation space into multiple distance layers according to resolution. The blind zone layer includes distance layer 1 (0-30 m) and distance layer 2 (30-60 m). Outside the blind zone are distance layer 3 and subsequent layers.

[0058] Step 3: Calculate the inherent deviation of the system

[0059] Step 301: Calculate the inherent bias of the system based on the difference between the observed frequency shift and the theoretical reference frequency shift in the blind zone layer. The calculation formula is as follows:

[0060] ,

[0061] in, The number of blind layers used for estimation. This is the theoretical reference frequency shift.

[0062] For example, if the measured frequency shifts of blind zone 1 and blind zone 2 are 80.08MHz and 80.08MHz respectively, and the reference frequency shift is 80.00MHz, then:

[0063] .

[0064] Step 4: Determine the inherent bias constraints

[0065] Step 401: Compare the difference between the current inherent bias and the bias at the previous moment.

[0066] when At that time, it was considered that the deviation was unreasonable;

[0067] Step 402: If the constraints are not met, then the effective inherent deviation value from the previous time step is used. Replacement is performed to ensure the stability and continuity of calibration.

[0068] Step 5: Perform frequency shift calibration

[0069] Step 501: Effective distance layer outside the blind zone Perform calibration; the calibration formula is:

[0070] ,

[0071] in, This is the calibrated frequency shift value. For any effective distance layer outside the blind zone, f obs ( t i , r k ) for t i The Doppler shift observation value of the kth effective observation layer at time k.

[0072] For example, when the frequency shift of the outer layer observation in the blind zone is 86.45 MHz and the system bias is 0.08 MHz, then:

[0073] .

[0074] Step 6: Update and output calibration results and wind speed inversion.

[0075] Step 601: Set the inherent deviation at the current moment. Stored in the system cache;

[0076] Step 602: In subsequent data processing, this deviation parameter is called in real time for calibration;

[0077] Step 603: Based on the calibrated frequency shift value and theoretical reference frequency shift The radial wind speed component is obtained by inversion based on the Doppler frequency shift relation. The calculation formula is as follows:

[0078] ,

[0079] Where λ is the laser wavelength. Indicates distance layer Radial wind speed at the location;

[0080] Step 604: When environmental changes cause frequency shift characteristics to drift, the system automatically recalculates. This enables dynamic adaptive adjustment of inherent biases and real-time updates of frequency shift and wind speed inversion results after calibration.

[0081] This invention utilizes the spectral characteristics of the lidar blind zone signal to extract the inherent deviation of the system in real time and dynamically corrects the Doppler frequency shift outside the blind zone, achieving online adaptive compensation for system frequency drift. This method boasts advantages such as strong real-time performance, high adaptability, and good versatility, significantly improving the accuracy and stability of lidar Doppler wind measurement data, and is applicable to various types of lidar systems.

[0082] Finally, it should be emphasized that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A real-time adaptive data calibration method for lidar, characterized in that, The method includes the following steps: Step 1: Initialize the lidar system parameters, including laser wavelength, laser pulse width, range layer resolution, sampling frequency, modulation intermediate frequency, and deviation upper limit constraint threshold. Step 2: Control the lidar to emit modulated light signals and receive echo signals to obtain Doppler frequency shift observation values ​​for each range layer; divide the observation space into multiple range layers according to the range layer resolution, wherein the near-range layer is the blind zone layer and the far-range layer is the effective observation layer; Step 3: Calculate the inherent bias of the system based on the difference between the observed Doppler frequency shift of the blind zone layer and the theoretical reference frequency shift; Step 4: Determine whether the difference between the system inherent deviation at the current moment and the effective system inherent deviation at the previous moment exceeds the upper limit constraint threshold of the deviation. If it exceeds the threshold, the effective system inherent deviation at the previous moment is used as the effective system inherent deviation at the current moment; if it does not exceed the threshold, the system inherent deviation calculated at the current moment is the effective system inherent deviation. Step 5: Use the inherent bias of the effective system at the current time to calibrate the Doppler frequency shift observations of the effective observation layer to obtain the calibrated frequency shift values; Step 6: Store the current effective system inherent bias in the system buffer for subsequent data calibration; and based on the calibrated frequency shift value and the theoretical reference frequency shift, invert the radial wind speed of the effective observation layer using the Doppler frequency shift relation. Repeat steps 2-6 to achieve real-time adaptive calibration of lidar data.

2. The method according to claim 1, characterized in that, In step 2, the blind zone layer is a near-range layer with a distance of 0 to 60m from the lidar, specifically including a range layer 1 with a distance of 0 to 30m and a range layer 2 with a distance of 30 to 60m; the effective observation layer is a far-range layer with a distance greater than 60m from the lidar, defined as range layer 3 and subsequent range layers.

3. The method according to claim 1, characterized in that, In step 3, the formula for calculating the inherent deviation of the system is: ， in, For t i The system's inherent bias at time t, m is the number of blind layers used for estimation, and f obs (t i ,r j ) for t i Doppler frequency shift observations at time j of the dead zone layer For t i The theoretical reference frequency shift or expected frequency shift value of the j-th blind zone layer at time j.

4. The method according to claim 3, characterized in that, In step 1, the upper limit constraint threshold for deviation is set to Δf. th In step 4, the judgment condition is: ， in, For t i-1 The inherent bias of the effective system at any given time.

5. The method according to claim 3, characterized in that, In step 5, the formula for calibrating the effective observation layer frequency shift data is: ， in, For t i The calibrated frequency shift value of the k-th effective observation layer at time k. For any effective distance layer outside the blind zone, f obs (t i ,r k ) for t i The Doppler shift observation value of the k-th effective observation layer at time k.

6. The method according to claim 5, characterized in that, In step 6, the Doppler frequency shift relationship for inverting radial wind speed is: ， in, For t i The radial wind speed at the k-th effective observation layer at time k, where λ is the laser wavelength.

7. The method according to claim 6, characterized in that, In step 1, the laser wavelength λ = 1550 nm.

8. The method according to claim 1, characterized in that, In step 1, the sampling frequency is 5×10⁻⁶. 8 Hz.

9. The method according to claim 1, characterized in that, In step 1, the modulation intermediate frequency is 80MHz.

10. The method according to claim 1, characterized in that, In step 1, the distance layer resolution is 30m.

Images