An airborne wind lidar method and system

Abstract

The present application relates to the technical field of airborne wind measurement, and discloses an airborne wind measurement laser radar method and system.The method comprises the following steps: generating a scanning timing identifier and a path correction marker structure through scanning beat synchronization and stray light shielding processing; generating an observation wind speed initial value sequence through time-frequency demodulation and quality gating based on backscattering signals; performing time alignment and motion compensation in combination with inertial navigation data, generating a wind speed vector grid segment through spatial interpolation and scanning geometry mapping; and finally generating a configuration update instruction structure through coordinate alignment and grid splicing.The present application effectively improves the precision and anti-interference ability of wind field measurement in an airborne environment, significantly enhances the data reliability and system stability under complex weather conditions through multi-source data fusion and an adaptive compensation mechanism.

Description

Technical Field

[0001] This invention relates to the field of airborne wind measurement technology, and in particular to an airborne wind measurement lidar method and system. Background Technology

[0002] In the field of airborne wind measurement and meteorological monitoring, existing solutions for airborne wind-measuring lidar typically integrate the transmitting and receiving optics, scanning mechanism, window and shading structure, timing control and acquisition link, signal processing module, inertial navigation unit and avionics interface, and wind field reconstruction module into a single operational link. This approach suffers from limitations such as vibrations and attitude disturbances introduced by high-speed flight leading to timing misalignment of echo sampling and channels, echo contamination caused by window reflections and stray light entering the receiving path, and a lack of consistent rules for data organization and quality labeling across channels and beats. Existing methods largely rely on fixed thresholds and empirical parameter tables to complete time-frequency processing, peak retrieval, and channel aggregation. Under maneuvering flight and aerosol-unstable environments, these methods are prone to beat reference drift and window configuration mismatch, as well as spectral pseudo-peaks and sidelobe diffusion induced by non-target light, making it difficult to achieve stable reconstruction of wind speed sequences and wind speed vector grids. For the joint processing of scanning time sequence identifiers, backscatter signal sets, path correction marker structures, quality gating labels, observed wind speed initial value sequence structures, and inertial navigation unit attitude and velocity data, existing technologies generally lack the integration of common source key values ​​and status markers in time alignment, coordinate mapping, spectral peak determination, channel aggregation, segment splicing, and parameter write-back. This makes it difficult to form a continuous process of acquisition-alignment-determination-control-recording in airborne operation scenarios, resulting in breakpoints and inconsistencies in the organization and tracing of wind field data across segments, channels, and beats. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides an airborne wind-measuring lidar method, comprising:

[0004] From the multi-beam scanning parameters and optical path calibration data, scanning timing is synchronized and stray light is shielded to obtain scanning timing identifiers; channel indexes are extracted for transmit-receive matching, and finally, path correction marker structures and backscatter signal sets are generated through optical path correlation correction.

[0005] Based on the backscattered signal set and scanning time sequence identifier, time-frequency demodulation, quality gating labeling and channel aggregation processing are performed to generate the initial value sequence structure of observed wind speed;

[0006] The system acquires attitude and velocity data of the inertial navigation unit, including attitude angle sequence, angular velocity sequence, linear acceleration sequence, body velocity sequence, and altitude information, along with device timestamps, health status markers, and sensor temperature records. It then performs time alignment and coordinate system mapping, including a time service unit that establishes a master-slave association between the inertial navigation unit device timestamp, airborne mission control timestamp, and scan timing identifier reference time; a coordinate mapping unit that establishes a mapping chain between body coordinates, inertial navigation unit coordinates, and lidar coordinates and outputs a mapping matrix snapshot; obtaining a set of motion compensation parameters; applying adaptive filtering; and then, based on the angular step position and beat number in the scan timing identifier, using spatial interpolation and scan geometry mapping. This includes a spatial reconstruction unit constructing an observation ray set; and a geometry mapping unit that projects the observation ray set onto a geographic reference plane or flight mission reference plane to generate scan slices. Finally, it generates a wind speed vector grid fragment structure containing a grid node set organized by beat, a grid cell set, boundary descriptions, source records, neighborhood expansion records, sub-segment descriptions, and beat quality indicators.

[0007] Based on the wind speed vector grid segment structure and scanning time sequence identifier, segment coordinate alignment, grid stitching and confidence verification are performed to generate a configuration update instruction structure.

[0008] Furthermore, the process of generating the path correction marker structure and the backscattered signal set specifically includes:

[0009] From the multi-beam scanning parameters and optical path calibration data, based on the job start alignment pulse and beat resynchronization mechanism generated by the timing control unit, scanning beat synchronization and stray light shielding are performed. The scanning beat synchronization includes establishing a beat index within the same absolute time base by the transmit scheduler and the receive scheduler. The stray light shielding process includes geometrically limiting the non-target angular domain through the aperture and the light shield, and estimating the non-target light distribution based on the window transmittance and the residual reflectivity of the anti-reflective film to obtain the scanning timing identifier. The channel index, which is linked and mapped with the angular step position, beat number and time window, is extracted from the scanning timing identifier and the transmit and receive matching is performed. Finally, through optical path correlation correction, a path correction mark structure is generated, which includes the non-target light influence type, expected arrival time, affected angular domain, receive channel gain correction rule, data quality suggestion, time base transition neighborhood and limit neighborhood marker. A backscatter signal set is also generated, which is organized by the channel index and beat number and includes the original echo sample, sampling time label, acquisition threshold configuration, disturbance state and noise candidate time period reference.

[0010] Furthermore, the multi-beam scanning parameters include:

[0011] The multi-beam scanning parameters are a set of parameters indicating the number of beams, beam-to-beam angle, scanning coverage, scanning start and end angles, angular step, channel transmission sequence, single pulse width, pulse repetition structure, ranging window setting, and retrieval integration threshold for each transmission channel within a single operation.

[0012] Furthermore, the optical path calibration data includes:

[0013] The optical path calibration data system is a calibration record of the mounting orientation, aperture, off-axis deviation, reflective surface roughness, window transmittance, and residual reflectance of the anti-reflective coating of the optical components at the transmitting and receiving ends, as well as the geometric correspondence between the corresponding body mounting reference surface, inertial navigation reference surface, and attitude measurement reference surface.

[0014] Furthermore, the attitude and velocity data of the inertial navigation unit include:

[0015] The attitude and velocity data of the inertial navigation unit includes attitude angle sequence, angular velocity sequence, linear acceleration sequence, body velocity sequence and altitude information, along with device timestamps, health status markers and sensor temperature records.

[0016] Furthermore, the process of generating the initial sequence structure of observed wind speeds specifically includes:

[0017] Based on the backscattered signal set and scanning timing identifier, combined with the receiving channel gain correction rules and data quality suggestions in the path correction mark structure, time-frequency demodulation and path correction mark application processing are performed. This includes a window construction unit that generates time window groups based on the expected echo time window and transition switching point and binds path correction mark entries to each window; a time-frequency mapping unit that performs time-frequency transformation to output amplitude-frequency-time three-dimensional intermediate results, extracts frequency shift peaks and spectral stability indices, and uses a combination of segmented thresholding and neighborhood comparison methods to perform candidate peak retrieval and time consistency and frequency band coherence evaluation, performs quality gating, and then generates an initial wind speed sequence structure through channel aggregation, which includes an aggregated observation matrix arranged by angular domain and beat and an aggregated meta-information set storing participation entry summaries, weight allocation records, anomaly suppression records, fragment merging records, and low-confidence fragment markers.

[0018] Furthermore, the process of generating the wind speed vector mesh fragment structure includes:

[0019] The system acquires attitude and velocity data of the inertial navigation unit, including attitude angle sequence, angular velocity sequence, linear acceleration sequence, body velocity sequence, and altitude information, along with device timestamps, health status markers, and sensor temperature records. It then performs time alignment and coordinate system mapping processing. This includes a time service unit reading the inertial navigation unit device timestamp, airborne mission control timestamp, and reference time of the corresponding beat in the scan timing identifier and associating the three in a master-slave manner. A coordinate mapping unit establishes a mapping chain between body coordinates, inertial navigation unit coordinates, and lidar coordinates and outputs a mapping matrix snapshot. Finally, it obtains a motion compensation parameter set, which includes attitude and velocity descriptions for each beat, mapping snapshots, interpolation annotations, transient markers, and weighting strategy markers.

[0020] The channel wind speed components are extracted from the initial wind speed observation sequence. Based on the attitude and velocity descriptions and mapping snapshots in the motion compensation parameter set, the motion compensation parameter set is applied and adaptive filtering is performed. This includes the compensation application unit interpolating the radial observations given by the component analysis unit according to the multi-point spline description within the beat, and the filtering execution unit constructing the filtering scene description and configuring the filtering parameters within each beat to generate a compensated wind speed sequence. The compensated wind speed sequence is arranged by beat number and angular domain index, and processing records such as deferred participation, non-participation, backoff mode, robustness suppressor triggering, and trajectory disconnection are written in the metadata area of ​​the sequence.

[0021] Furthermore, the process of generating the wind speed vector mesh fragment structure also includes:

[0022] For the compensated wind speed sequence, spatial interpolation and scan geometric mapping are performed based on the angular step position and beat number in the scan time sequence identifier. This includes the spatial reconstruction unit constructing the observation ray set corresponding to the beat based on the angular domain index and beat number of the compensated wind speed sequence, and the geometric mapping unit projecting the observation ray set onto the geographic reference plane or flight mission reference plane and generating scan slices according to the beat, generating a wind speed vector grid fragment structure. The wind speed vector grid fragment structure includes a grid node set organized according to the beat, a grid cell set, boundary description, source record, neighborhood extension record, sub-segment description, and beat quality indication.

[0023] Furthermore, the process of generating the configuration update instruction structure also includes:

[0024] Based on the wind speed vector grid segment structure and scanning time sequence identifier, combined with the transition switching points and beat numbers in the scanning time sequence identifier, segment coordinate alignment and channel overlap processing are performed. This includes the coordinate reference management unit establishing a unified reference coordinate framework and splitting multi-segment description beats into micro-segments; the alignment execution unit establishing candidate alignment surfaces according to the grid node distribution and scanning geometry, extracting boundary overlapping areas; and based on the overlap weights and conflict point list in the 3D wind field block data, performing grid splicing and confidence verification. This includes the splicing execution unit processing adjacent grid blocks at the node and unit levels respectively, and the verification execution unit constructing a confidence assessment model to output confidence markers and classifications. Through data consistency verification, based on the consistency rule set of four dimensions—coverage time consistency, spatial consistency, cross-channel consistency, and boundary consistency—a configuration update instruction structure is generated, which includes an instruction header, instruction body, and evidence attachments. The instruction header specifies the target module and effective scope, the instruction body lists parameter items and suggested value ranges, the evidence attachments store consistency judgment logs, boundary overlapping area screenshot indexes, source difference grouping lists, and transition zone parameter trajectories.

[0025] Furthermore, an airborne wind-measuring lidar system, applied to any of the methods described above, includes:

[0026] The scanning clock synchronization and stray light shielding unit is used to acquire multi-beam scanning parameters and optical path calibration data, perform scanning clock synchronization and stray light shielding processing, and output scanning timing identifiers and noise annotation data;

[0027] The transmit and receive trigger matching unit is used to parse the channel index according to the scan timing identifier and perform transmit and receive trigger matching, and output the backscatter signal set;

[0028] An optical path correlation correction unit is used to combine noise labeling data and backscattering signals to generate a path correction mark structure.

[0029] The time-frequency demodulation and path correction mark application unit is used to invoke the path correction mark structure and output the demodulation spectrum sequence under the constraints of the backscattered signal set and the scan timing mark.

[0030] The frequency shift peak and spectral peak stability index extraction unit is used to extract the frequency shift peak and spectral peak stability index from the demodulated spectral sequence and output the quality gating label and Doppler frequency shift sequence;

[0031] The channel aggregation unit is used to perform channel aggregation based on the Doppler frequency shift sequence and channel index, and output the initial value sequence structure of the observed wind speed;

[0032] The inertial navigation data access and time-aligned coordinate mapping unit is used to receive attitude and velocity data from the inertial navigation unit, perform time alignment and coordinate system mapping processing, and output a set of motion compensation parameters.

[0033] The motion compensation and adaptive filtering unit is used to extract the channel wind speed components from the initial observation wind speed sequence structure and call the motion compensation parameter set to perform adaptive filtering processing, and output the compensated wind speed sequence.

[0034] Spatial interpolation and scan geometry mapping unit, used to perform spatial interpolation and scan geometry mapping on the compensated wind speed sequence and output wind speed vector grid fragment structure;

[0035] The fragment coordinate alignment and channel overlap processing unit is used to perform fragment coordinate alignment and channel overlap processing under the constraints of wind speed vector grid fragment structure and scan timing identifier, and output three-dimensional wind field block data;

[0036] The mesh stitching and confidence check unit is used to extract the boundary overlapping area on the 3D wind field block data and perform mesh stitching and confidence check, and output the 3D wind field data set;

[0037] The data consistency verification and parameter write-back unit is used to perform data consistency verification on the three-dimensional wind field data set and generate a configuration update instruction structure. The configuration update instruction structure is used to send parameter update instructions back to the scanning beat synchronization and stray light shielding unit, the time-frequency demodulation and path correction marking application unit, the frequency shift peak and spectral peak stability index extraction unit, the channel aggregation unit, the motion compensation and adaptive filtering unit, and the grid stitching and confidence verification unit, and register the source and effective range.

[0038] The key innovations of this invention include:

[0039] (1) Construct a channel index driven by the scanning timing identifier output by the scanning beat synchronization and stray light shielding unit, and fuse the noise label data and backscatter signal set to form a path correction mark structure. In the time-frequency demodulation and path correction mark application unit, the window, mark and time base are processed in an integrated manner, and the demodulated spectrum sequence is output and connected to the quality gate label and channel aggregation under the same key value.

[0040] (2) Receive attitude and velocity data from the inertial navigation unit, form a set of motion compensation parameters through time alignment and coordinate system mapping, and couple the initial value sequence of observed wind speed with the structure of the observation wind speed in the motion compensation and adaptive filtering unit. Generate a compensation wind speed sequence according to the beat and angular domain to form a dynamic compensation link consistent with the scanning geometry.

[0041] (3) Around the wind speed vector grid segment structure, set up segment coordinate alignment and channel overlap processing unit, establish candidate alignment surface and weight fusion strategy for adjacent segments, output three-dimensional wind field block data, and form a continuous spatial expression of the three-dimensional wind field data set by boundary overlap area processing and transition zone organization in the grid splicing and confidence verification unit.

[0042] (4) In the data consistency verification and parameter write-back unit, the three-dimensional wind field data set is subjected to joint consistency verification in terms of time, space, cross channels and boundaries. The configuration update instruction structure is generated and sent back to the scanning beat synchronization and stray light shielding unit, time frequency demodulation and path correction marking application unit, frequency shift peak and spectral peak stability index extraction unit, channel aggregation unit, motion compensation and adaptive filtering unit, and grid stitching and confidence verification unit to form a closed-loop update of parameters and strategies.

[0043] (5) Based on the demodulated spectrum sequence, extract the frequency shift peak and spectrum peak stability index, generate quality gating labels and work together with the channel index to aggregate the channel, output the initial value sequence structure of the observed wind speed carrying the participation summary and weight information, and establish a traceable data organization for subsequent motion compensation and spatial interpolation.

[0044] The following are its main beneficial effects:

[0045] (1) The processing link for backscattered signal set and scanning timing identifier maintains a unified key value and status mark between time-frequency demodulation, quality gating labeling and channel aggregation through path correction mark structure. During operation, it provides clear constraints on noise period, limit neighborhood and time base transition, and is suitable for continuous processing under multi-beam scanning and complex window reflection conditions.

[0046] (2) For dynamic correction of the initial wind speed sequence structure, adaptive filtering is performed in the unified coordinates after time alignment and coordinate mapping by the motion compensation parameter set. The operation link performs weight reduction and rollback management on transient markers and limited states, and maintains the temporal and angular domain consistency of the compensation wind speed sequence under high-speed flight and attitude change conditions.

[0047] (3) Spatial assembly of wind speed vector grid fragment structure, through channel overlap processing and boundary overlap area splicing, a continuous expression with weight control and transition zone constraint is formed in the assembly process from three-dimensional wind field block data to three-dimensional wind field data set. The running link retains verifiable marks for conflict points and verification zones, which is suitable for grid organization across beats, sub-segments and fragments.

[0048] (4) For parameter management throughout the scanning, demodulation, gating, aggregation, filtering and splicing stages, the rules and strategies are kept consistent under long-term operation and multi-scenario switching conditions by sending back and registering the configuration update instruction structure, thus realizing closed-loop operation from data consistency verification to parameter write-back.

[0049] (5) For the organization of demodulated spectrum sequences, quality gating annotation and channel index are used to participate in channel aggregation. The running link is deposited in the structure of the initial wind speed sequence to participate in the summary and weight record, providing stable input and boundary prompts for subsequent motion compensation and scanning geometry mapping. It is suitable for data filtering and aggregation across channels and across beats. Attached Figure Description

[0050] Figure 1 A flowchart illustrating an airborne wind-measuring lidar method provided in an embodiment of this application;

[0051] Figure 2 This is a structural block diagram of an airborne wind-measuring lidar system provided in an embodiment of this application. Detailed Implementation

[0052] Example 1: Refer to Figure 1 This is a flowchart illustrating an airborne wind-measuring lidar method according to an embodiment of the present invention. The flowchart may include at least steps S100-S400:

[0053] S100: From the multi-beam scanning parameters and optical path calibration data, perform scanning beat synchronization and stray light shielding processing to obtain scanning timing identifiers; extract channel indexes for transmit-receive matching, and finally generate path correction mark structures and backscatter signal sets through optical path correlation correction;

[0054] S200, based on the backscattered signal set and scanning time sequence identifier, performs time-frequency demodulation, quality gating labeling and channel aggregation processing to generate the initial value sequence structure of observed wind speed;

[0055] S300: Acquire attitude and velocity data of the inertial navigation unit, perform time alignment and coordinate system mapping to obtain a set of motion compensation parameters, apply adaptive filtering, and generate wind speed vector grid fragment structure through spatial interpolation and scanning geometry mapping.

[0056] S400, based on the wind speed vector grid segment structure and scanning time sequence identifier, performs segment coordinate alignment, grid stitching and confidence verification, and generates configuration update instruction structure.

[0057] Step S100 includes at least steps S110-S130:

[0058] S110. Acquire multi-beam scanning parameters and optical path calibration data, perform scanning clock synchronization and stray light shielding processing, and obtain scanning timing identifiers and noise labeling data.

[0059] Specifically, the multi-beam scanning parameters in the preceding system configuration file are used as input. These parameters are a set of parameters indicating the number of beams, beam-to-beam angle, scanning coverage, scanning start and end angles, angular step, channel transmission sequence, single pulse width, pulse repetition structure, ranging window setting, and retrieval integration threshold for each transmission channel within a single operation. Simultaneously, optical path calibration data is also input. This data is a calibration record of the mounting orientation, aperture, off-axis deviation, reflective surface roughness, window transmittance, and residual reflectivity of the anti-reflective coating for the transmitting and receiving optical elements, as well as the geometric correspondence between the corresponding airframe mounting reference surface, inertial navigation reference surface, and attitude measurement reference surface. The multi-beam scanning parameters are read by the scanning control module and loaded into the timing control unit. The optical path calibration data is read by the optomechanical component calibration unit and distributed to the transmitting end alignment, receiving end collimation, and stray light estimation units. These two inputs are used together in this section, and no new terminology is introduced. Furthermore, scan clock synchronization is performed in the scan control module. Scan clock synchronization is a time base alignment process carried out under the conditions of vibration and acceleration disturbance of the airborne flight platform to eliminate time base drift of transmission and reception actions. The implementation method is as follows: first, the timing control unit generates an operation start alignment pulse, and sends the alignment pulse to the transmit scheduler and the receive scheduler. The transmit scheduler assigns clock numbers to the channel transmission sequence in the multi-beam scan parameters, and the receive scheduler assigns clock numbers to the expected echo acquisition threshold and time window. The two establish a one-to-one corresponding clock index within the same absolute time base. When a transition occurs in the flight state, the timing control unit triggers clock resynchronization, records the switching point of the clock number before and after the transition, and includes the switching point in the scan timing identifier. The scan timing identifier is a structured record containing the clock number, channel transmission sequence, expected echo time window, angular step position, and transition switching point. To reduce background interference, stray light shielding is jointly performed by the optomechanical component calibration unit and the receiving front end. This stray light shielding process targets and marks non-target light power components introduced by factors such as window reflections, cockpit scattering, aircraft surface reflections, and strong external light. The implementation involves first geometrically limiting the non-target angular region entering the receiving optical path using an aperture and a light shield. Then, based on the window transmittance and residual reflectance of the anti-reflective film in the optical path calibration data, the arrival time and intensity distribution of non-target light on the detection surface are estimated. Simultaneously, the suppression threshold for non-target time periods is raised in the electrical threshold control of the receiving front end, forming a list of noise candidate time periods corresponding to the clock cycle. For angular region drift caused by rapid changes in aircraft attitude during flight, the scan control module sends new angular region limiting parameters to the stray light shielding unit after clock cycle resynchronization, triggering fine-tuning of the mechanical position of the light shielding component. When the light shielding component reaches the mechanical travel boundary under extreme attitude conditions, the system writes this boundary state along with the affected clock cycle number into the state description field of the noise marking data.Understandably, after the scanning clock synchronization is completed, the transmission and reception establish a correspondence between the clock number and the angular domain position on a unified time reference. After stray light shielding is completed, the system obtains a record of the noise candidate time period and noise source type aligned with the clock number. Regarding anomaly handling, when the read multi-beam scanning parameters and optical path calibration data have inconsistent versions or missing fields, the timing control unit enters a restricted mode, enabling only the verified channel transmission sequence and time window configuration, and writing a restricted mode mark into the scanning timing identifier. If the light-shielding component reports mechanical jamming, a manual verification mark is added to the noise annotation data while maintaining the electrical threshold suppression strategy. At this point, the output products are the scanning timing identifier and noise annotation data, both of which will be used in subsequent steps. The scanning timing identifier will be referenced as a "timing identifier" in the demodulation process of S210, and the noise annotation data will be used in S210 for path correction marking in non-target time periods. Both are also used in the segment coordinate alignment processing of S410 for initial screening of channel overlap, thus completing the cross-stage connection with subsequent main steps.

[0060] S120. Extract the channel index from the scanning timing identifier, perform transmit and receive trigger matching, and generate a backscatter signal set;

[0061] Specifically, the channel index is a retrieval key that uniquely numbers all transmission channels in the multi-beam scanning parameters and maps them in conjunction with the angular step position, cycle number, and time window. The channel index is generated by the scanning control module based on the scanning timing identifier. In this section, the system first reads the scanning timing identifier output by S110, unfolds each record sequentially according to the timing order of the cycle number, and parses out the channel number, angular domain position, expected echo time window, and transition switching point for each record. These four parsing results are then combined into the channel index. Subsequently, the transmission scheduler triggers the transmitting device according to the channel index. The transmitting device emits laser pulses for the corresponding channel at a given beam angle and angular step position. The receiving scheduler activates the receiving channel within the expected echo time window of that channel and configures the sampling threshold and integration threshold. The receiving front end converts and pre-amplifies the optical signal before transmitting it to the acquisition unit. The trigger matching phase is driven by the timing control unit. The trigger matching process is as follows: a two-way binding is established between the transmit trigger timestamp and the start and end timestamps of the receive window for each channel index; when there is a clock resynchronization switching point, the trigger matching logic calculates the transmit trigger time and receive window according to the two time bases before and after the switching point to avoid cross-segment mismatch; when there is angular domain drift caused by drastic changes in aircraft attitude, the trigger matching logic adjusts the receive window in terms of time shift and width according to the angular domain position update in the scanning timing identifier to ensure that the receive window still covers the main energy segment of the target echo. On abnormal branches, if the transmit trigger feedback of a certain channel is missing, the system marks the channel index as missing and skips the echo acquisition of that channel; if the receive window overlaps with the task switching of the flight control system, the acquisition unit marks the clock as disturbed and shortens the sampling duration of that clock, while recording the "disturbed" label in the backscatter signal set. The backscatter signal set refers to the collection of echo samples gathered across all channels and beats. Internally, it is organized according to channel index and beat number, and includes the original echo samples, sampling time stamps, acquisition threshold configuration, disturbance state, and noise candidate time period references. Furthermore, to reduce the transient impact of body vibration on electrical contact, the acquisition unit performs multiple short-segment samplings within each receiving window and performs intra-segment convergence at the end of the window, writing the intra-segment convergence result as the echo sample for that window into the backscatter signal set. When noise labeling data indicates the presence of strong background light during certain periods, the acquisition unit lowers the integration threshold or enables shorter sampling segments during these periods to avoid saturation. Understandably, the transmission and reception trigger matching process is repeated in each cycle, and the echo samples from all channels are aggregated into a structured backscattered signal set. The channel index and cycle number serve as a retrieval key to locate the processing parameters of each echo sample in the subsequent time-frequency demodulation of S210. At the same time, the backscattered signal set and the scan timing identifier are sent to the input of S210 and consumed as the "backscattered signal set" and "timing identifier". In the segment coordinate alignment process of S410, it serves as one of the basic data for the reconstruction of the three-dimensional wind field block.This concludes the complete chain from parsing the scan timing identifier to the channel index, from the channel index to the execution of trigger matching, and from trigger matching to the generation of the backscatter signal set, while maintaining the inherent consistency with the output of the previous section.

[0062] S130. Perform optical path correlation correction on the noise-labeled data and the backscattered signal set to generate a path correction marker structure.

[0063] Specifically, optical path-related correction refers to adding a label related to optical path characteristics to each echo sample based on optical path calibration data, noise labeling data, and channel geometry, without altering the acquired echo sample values. The label content includes the type of non-target light influence, expected arrival time, affected angular region, receiving channel gain correction rules, and data quality suggestions. The inputs to this process include the noise labeling data and optical path calibration data output from S110, and the backscatter signal set output from S120. In terms of implementation, the label generation unit first reads the candidate time periods from the noise labeling data and constructs a candidate influence list for each beat and channel index. This list includes influence types such as window reflection, cockpit scattering, aircraft surface reflection, and external strong light. Subsequently, the label generation unit maps the candidate influence list to the influence area on the channel angular region based on the off-axis deviation, window transmittance, and residual reflectivity of the anti-reflective coating in the optical path calibration data, corresponding to the sampling time label of each echo sample, forming an influence label that corresponds one-to-one with the sample. During the labeling refinement stage, the system adds data quality suggestions such as "recalibration suggestion" or "removal suggestion" to samples where the disturbed state intersects with the candidate time period. For samples near the beat resynchronization switching point, the label generation unit adds a "time base transition neighborhood" label to prompt subsequent processing steps to use the same time base as the switching point during demodulation. For states where the mechanical travel of the shading component reaches the boundary, the label generation unit adds a "limiting neighborhood" label to the relevant channel and provides a gain convergence strategy for this neighborhood in the gain correction rules. In terms of anomaly handling, if the noise labeling data lacks a record of a certain beat, the system does not add noise-related labels to the samples of that beat, but still retains the geometric-related labels generated based on the optical path calibration data. If there are missing fields in the optical path calibration data, the system only adds some labels that can be derived from the remaining fields and records "field missing" in the state description domain of the path correction label structure. To accommodate subsequent time-frequency demodulation, the path correction marker structure maintains the same key-value consistency as the backscattered signal set, using channel index and beat number as retrieval keys, with a complete marker list appended to each sample record. Specifically, the non-target light influence type and expected arrival time will be directly processed and consumed by the path correction markers in S210; the receive channel gain correction rule will guide the demodulation strategy selection within the time window of S210; data quality suggestions will correspond to the quality gating markers in S220; and the time base transition neighborhood and limit neighborhood markers will prompt the segment coordinate alignment stage in S410 to adjust neighborhood weights during channel overlap processing. Understandably, the path correction marker structure does not alter the echo samples themselves, but rather serves as a set of interpretive and constraining information bound to the samples, used by subsequent steps to select processing paths, set processing thresholds, and perform data filtering.At this point, the output product is a path correction mark structure. This "path correction mark structure" is directly consumed by the path correction mark application processing in S210, and together with the scan timing identifier and backscatter signal set, constitutes the three basic inputs of the main step in S200. Simultaneously, it serves as reference information for boundary neighborhood processing and overlap evaluation in the mesh stitching and confidence verification in S420. In summary, the technical effect of this step is that through the sequential processing of scan clock synchronization, trigger matching, and optical path correlation correction, a time-consistent, state-traceable input chain with optical path interpretation information is formed, providing a continuous, stable, and traceable data foundation for subsequent time-frequency demodulation, quality gating, and 3D wind field reconstruction.

[0064] Step S200 includes at least steps S210-S230:

[0065] S210. Obtain the backscattered signal set and scanning timing identifier, perform time-frequency demodulation and path correction mark application processing to obtain the demodulated spectrum sequence;

[0066] Specifically, the backscattered signal set is used as input. The backscattered signal set is the set of echo samples collected in the aforementioned steps according to the channel index and beat number. It contains fields such as the original echo sample, sampling time label, acquisition threshold configuration, and disturbance status. At the same time, the scanning timing identifier is used as parallel input. The scanning timing identifier is a structured record that records the beat number, channel transmission order, expected echo time window, angular step position, and transition switching point. The path correction mark structure generated in the previous steps is also loaded into the same processing task. The path correction mark structure is a set of optical path-related marks attached to each echo sample. It contains fields such as non-target light influence type, expected arrival time, affected angular domain, receiving channel gain correction rule, data quality suggestion, time base transition neighborhood and limit neighborhood mark. In this section, the processing link first involves the timing scheduling unit expanding the task queue according to the scan timing identifier, locking the corresponding channel index step by step, and calling the echo samples within the backscatter signal set that match the channel index. Subsequently, the window construction unit generates a time window group based on the expected echo time window and the transition switching point, and binds a corresponding path correction mark structure entry to each window, forming a time-frequency analysis preparation list with window-marker correspondence. Understandably, when the scan timing identifier indicates a transition in the beat resynchronization, the window construction unit divides the echo samples into two sub-time bases before and after the transition, independently setting the window and aligning them with the time axis to avoid aliasing caused by cross-segment superposition. When the path correction mark structure indicates that a window is in a limited neighborhood, the window construction unit adds a gain convergence prompt to the processing parameters of that window and retains this prompt for subsequent processing steps to read. Subsequently, the preprocessing unit performs detrending, threshold clipping, and intra-segment resampling on the original echo samples within each time window. Detrending is used to eliminate baseline rise caused by slow drift, threshold clipping is used to suppress the impact of occasional high-amplitude interference on short-time energy distribution, and intra-segment resampling is used to equalize local sampling points without changing the sampling label order. In the window where the disturbance state is "disturbed", the preprocessing unit sets the step size of intra-segment resampling to a smaller value and adds a buffer area at the end of the window to accommodate the main echo energy that may be compressed. Next, the time-frequency mapping unit performs time-frequency transformation on each window and outputs intermediate results of the amplitude-frequency-time three-dimensional structure. The transformation strategy of each window is jointly determined by the path correction mark structure and the receiver channel gain correction rule: if the mark is window reflection or cabin scattering, the time-frequency mapping unit enables a narrower analysis bandwidth and weighted suppression of the frequency band response corresponding to the non-target angular domain in this window; if the mark is strong external light, the time-frequency mapping unit introduces an initial suppression time slot in the first few sub-segments of this window to reduce the influence of strong background on the short time spectrum; if the mark is a time base transition neighborhood, each sub-time base is transformed and merged using its own reference clock, and the merge is spliced ​​according to the coverage ratio of the sub-time bases within the window.Subsequently, the spectral shaping unit performs spectral domain resampling and sidelobe suppression processing on the time-frequency mapping results, and selects differentiated sidelobe suppression intensities based on the perturbation state and influence type. Within the window where the confined neighborhood marker exists, the spectral shaping unit preferentially adopts a conservative sidelobe suppression scheme and records this selection for traceability. Furthermore, path correction marker application processing is integrated throughout the above stages: in the window construction stage, it is used to filter the window and marker binding relationship; in the preprocessing stage, it is used to configure the intra-segment resampling strategy and threshold peak clipping strategy; in the time-frequency mapping stage, it is used to determine the analysis bandwidth and multi-sub-timebase merging method; and in the spectral shaping stage, it is used to select the sidelobe suppression intensity and resampling step. In the anomaly handling branch, if a window lacks a path correction marker structure entry, the system places the window in the "default path" and records the missing event, while using the conventional analysis bandwidth and standard sidelobe suppression parameters; if the backscattered signal corresponding to a beat is missing, the system generates an empty entry at that beat position and marks the beat as "missing measurement." This marker is considered a skippable item in subsequent stages and does not participate in statistical aggregation. After all windows have been processed, the spectral segment splicing unit splices the time-frequency results of each window into a unified frequency-amplitude sequence according to the beat number and channel index order, and associates the angular step position in the scan timing identifier with the channel transmission order to form a demodulated spectral segment with a unified key value. Finally, the result archiving unit combines each demodulated spectral segment into a demodulated spectral sequence oriented towards the channel-beat two-dimensional index, and writes the path correction mark summary and anomaly handling record actually used in this subsection into each entry. Thus, the output product of this subsection is a demodulated spectral sequence, which will be directly consumed by the frequency shift peak extraction and spectral peak stability index calculation of S220 as a "demodulated spectral sequence". At the same time, the path correction mark summary in the demodulated spectral sequence will be implicitly associated with the quality gating annotation of S220. In addition, the demodulated spectral sequence can be used as reference information for frequency domain quality traceability in the subsequent grid splicing and confidence verification of S420, thereby forming data connection with the cross-step processing of S400.

[0067] S220. Extract the frequency shift peak and spectral peak stability index from the demodulated spectrum sequence, perform quality gating annotation, and generate the Doppler frequency shift sequence.

[0068] Specifically, the demodulated spectrum sequence is used as the sole input source. This sequence is organized by channel index and beat number, and each entry includes a frequency-amplitude sequence, path correction marker summary, and anomaly records. This section first describes how the peak candidate retrieval unit performs a candidate peak retrieval on each frequency-amplitude sequence. The retrieval employs a combination of segmented thresholding and neighborhood comparison: the segmented thresholding method divides the frequency domain into several segments and determines the retrieval threshold based on the background statistics within each segment; the neighborhood comparison method constructs a comparison window around the candidate peak and compares the energy difference between the main lobe and the neighboring area. Both methods together provide a candidate set. When a demodulated spectrum sequence entry has a "default path" marker, the peak candidate retrieval unit uses a higher segment threshold and stricter neighborhood comparison conditions to reduce false triggers. When an entry has a "limited neighborhood" marker, the peak candidate retrieval unit adds a candidate verification step to that entry, performs secondary screening on the candidate set, and retains a screening log. Subsequently, the peak refinement and localization unit performs peak position refinement processing on each candidate peak. This refinement process includes three actions: local interpolation, energy redistribution, and peak symmetry evaluation. Local interpolation constructs a finer frequency grid near the candidate peaks; energy redistribution corrects the peak position based on the energy patterns of the neighborhood; and peak symmetry evaluation identifies skewed peak shapes caused by background disturbances. After processing, the peak refinement and localization unit outputs a main peak position and several candidate peaks for each record, sorting the candidate peaks from highest to lowest correlation with the main peak. Next, the stability evaluation unit evaluates the temporal consistency and frequency band coherence of the main peak and its neighborhood. Temporal consistency evaluation is performed within the same channel index, calculated by beat number, calculating the rate and direction of change of the main peak position in adjacent beats, and generating a temporal consistency label accordingly. Frequency band coherence evaluation is performed within the same beat, calculated by channel index, checking whether the main peaks near the same angular step position exhibit continuous distribution, and generating a frequency band coherence label accordingly. If a record contains "missing" or "disturbed" abnormal records in the preceding steps, the stability assessment unit introduces a gap compensation strategy in the calculation of the record's temporal consistency and frequency band coherence. The gap compensation strategy is implemented by combining interpolation completion of adjacent valid records with weight attenuation, and the completion traces are written into the stability assessment notes. Furthermore, the quality gating unit generates quality gating labels based on the above peak value and stability assessment results.The quality gating annotation system is a set of tags used to classify and annotate the availability of each record. It includes three categories: Pass, Pending Review, and Rejection, with the triggering reason and reference information source recorded next to each tag. When the path correction tag summary indicates that "portal reflection" or "strong external light" has a significant impact, the quality gating unit places the record in "Pending Review" and writes the corresponding impact type in the triggering reason. When the time base transition neighborhood tag exists and the time consistency tag of the record does not meet the stability requirements, the quality gating unit places the record in "Rejection" and writes "Time base inconsistency" in the reference information source. When the peak height and local background difference under the segment threshold and neighborhood comparison are insufficient, the quality gating unit places the record in "Pending Review" and adds the annotation "Peak shape unreliable." Subsequently, the generation unit writes the main peak positions of the Pass and Pending Review records into the output structure according to the channel index and beat number, forming time-series-oriented Doppler frequency shift entries. For records marked as rejected, the generation unit leaves a blank in the output structure and writes the rejection reason so that subsequent steps can identify entries that do not participate in aggregation. Regarding anomaly handling, if a channel is marked as rejected for multiple consecutive beats, the quality gating unit marks the channel as "short-term failure" and triggers an internal recording mechanism to prompt subsequent main steps to reduce the participation weight of that channel during spatial interpolation. If most channels within a beat are marked as pending review, the quality gating unit adds an overall review process. This overall review compares peak distribution patterns and energy layouts across channels and attempts to promote some pending review entries to pass or downgrade them to rejection. After completing the above processing, the result archiving unit outputs a Doppler frequency shift sequence, which is a set of main peak positions organized by channel index and beat number, with quality gating annotations and stability assessment notes attached next to each entry. Therefore, the output of this section is a Doppler frequency shift sequence. This Doppler frequency shift sequence is directly consumed by the channel aggregation in S230. At the same time, the quality gating label is used by S230 for weight allocation and entry filtering during aggregation. In addition, the quality gating label and stability evaluation annotation will be read by the adaptive filtering process in S320 to allocate the robustness factor of the channel wind speed component, thereby establishing a cross-stage data relationship with the subsequent main steps.

[0069] S230. Channel aggregation is performed based on the Doppler frequency shift sequence and channel index to generate the initial value sequence structure of observed wind speed;

[0070] Specifically, the Doppler frequency shift sequence and the channel index are used as inputs. The Doppler frequency shift sequence contains the main peak position, quality gating label, and stability evaluation annotation of the channel-beat two-dimensional key value. The channel index is a unique number for the transmission channel and establishes a retrieval relationship with the angular step position, beat number, and time window. In this subsection, the aggregation preparation unit first loads the channel set under each angular step position according to the channel index and constructs the angular domain observation segment by the channel set within the same beat. Subsequently, the item filtering unit reads the quality gating label of the corresponding item in the Doppler frequency shift sequence, filters out the items marked as to be removed, and sets a participation mark with reduced weight for the items marked as to be reviewed. When the number of available channels at a certain angular step position is lower than a preset threshold, the item filtering unit triggers a gap filling strategy on the angular domain observation segment. The gap filling strategy is based on the available items at the adjacent angular step positions and fills the gap in relative azimuth order, while recording the source of the gap filling in the metadata of the observation segment. Subsequently, the geometric mapping unit maps the channel index and angular step position to a set of observation rays. This set of observation rays describes the orientation of each channel in the body coordinate system and its angular position in the scan plane. When the scan timing identifier records a beat resynchronization transition in the preceding step, the geometric mapping unit splits the set of observation rays into two groups before and after the transition within that beat, and processes them separately in subsequent aggregation. Next, the robust aggregation unit performs aggregation operations on the channel entries of the same angular domain observation segment within the same beat. The aggregation operation comprises three types of actions: First, weight allocation, which, referencing quality gating annotations and stability assessment notes, assigns standard weights to passing entries and reduced weights to entries awaiting review, while setting a weight decay band near consecutive "short-term failure" channels; second, anomaly suppression, which eliminates significantly deviating entries within a local neighborhood based on the relative dispersion between entries, recording the triggering reason and neighborhood range in the elimination record; and third, segment merging, which merges grouped observations at the same corner location before and after transitions, recording the transition information source in the merging record. After these three types of actions are executed, the robust aggregation unit outputs the aggregated observation value and a summary of participating entries for each corner observation segment. The summary of participating entries includes the number of participating entries, weight distribution, and a list of eliminated entries. Subsequently, the ray aggregation unit aggregates the observation segments from each corner domain within the same beat according to channel geometry at the ray level, and establishes an observation hierarchy structure with a triple index of ray-corner domain-beat. When a beat has a large concentration of "to be verified" tags, the ray aggregation unit marks the observation hierarchy structure of that beat as a "low-confidence segment." This marking serves as a reminder for subsequent main steps to process the segments using a robust strategy. Furthermore, the sequence construction unit connects the ray aggregation results of each beat in the time dimension, forming an observation time series organized by beat number, and attaches a summary of the participating entries and anomaly suppression records generated in this subsection at each time point.To facilitate subsequent main steps, the sequence construction unit structurally encapsulates the aforementioned observation time series, forming an initial wind speed sequence structure. This initial wind speed sequence structure comprises two parts: first, an aggregated observation matrix arranged by angle domain and beat, used to express the observation results corresponding to each angle domain position and beat; second, an aggregated metadata set, used to store traceability information such as participant entry summaries, weight allocation records, anomaly suppression records, fragment merging records, and low-confidence fragment markers. Regarding anomaly handling, if consecutive "low-confidence fragments" exist within a certain time period, the sequence construction unit adds stability indicators to that time period, which will be used as a reference for adaptive filtering configuration in subsequent main steps. If an angle domain position lacks any available entries in multiple consecutive beats, the sequence construction unit inserts a gap marker at that angle domain position and limits the influence range of that position when participating in subsequent spatial interpolation. After completing the above processing, the result archiving unit outputs the initial observation wind speed sequence structure and explicitly passes this output to the input of the subsequent main steps. The initial observation wind speed sequence structure is called by the application link after time alignment and coordinate system mapping in the subsequent processing link of S310 as the "initial observation wind speed sequence", and serves as the main data source for channel wind speed component extraction and adaptive filtering in S320. At the same time, the weight allocation record and anomaly suppression record in the aggregated meta-information set will be read by the robustness strategy of S320 and used to set different channel participation levels in the process of compensating for wind speed sequence generation. At the cross-step level, the initial observation wind speed sequence structure is also one of the time references for channel overlap processing in the segment coordinate alignment of S410, thus establishing a connection with the three-dimensional wind field data organization of S400. In summary, the technical effects of this step are as follows: by locating the peak value, evaluating the stability, and gating the quality of the demodulation results, and by combining channel geometry and angular domain organization to carry out robust aggregation, the key transition output of this invention from the data domain to the observation domain is formed, which establishes a clear, traceable, and spatiotemporally mapping initial value sequence structure for subsequent inertial information fusion and filtering.

[0071] Step S300 includes at least steps S310-S330:

[0072] S310. Acquire the attitude and velocity data of the inertial navigation unit, perform time alignment and coordinate system mapping processing, and obtain a set of motion compensation parameters;

[0073] Specifically, the raw data from the Inertial Measurement Unit (IMU), which is pushed in real time by the airborne avionics system, is used as input. The attitude and velocity data of the IMU includes attitude angle sequence, angular velocity sequence, linear acceleration sequence, airframe velocity sequence, and altitude information, along with device timestamps, health status markers, and sensor temperature records. At the same time, the initial wind speed sequence output from the previous step is used as a parallel reference. In this main step, the initial wind speed sequence is only used as a beat reference and data sparsity indicator, and does not participate in attitude estimation and velocity calculation. Time alignment is initiated by the time service unit. The time service unit first reads the timestamp of the inertial navigation unit device, the timestamp of the airborne mission control, and the reference time of the corresponding beat in the scan timing identifier, and establishes a master-slave association among the three. When there is a fixed offset between the inertial navigation unit device time and the mission control time, the time service unit establishes an offset compensation entry and performs alignment before the start of each beat. When there is a jump or jitter in the inertial navigation unit device time, the time service unit marks the data segment as an unstable interval according to the health status flag, and maps the unstable interval to the corresponding beat for subsequent calculations to avoid or reduce weight. Coordinate mapping is performed by the coordinate mapping unit. After receiving the attitude angle sequence, the coordinate mapping unit establishes a mapping chain between the body coordinates, inertial navigation unit coordinates, and lidar coordinates according to the geometric relationship between the body mounting reference plane and the optomechanical component mounting reference plane. This mapping chain is provided by optical path calibration data and installation calibration records during system installation and maintenance, including sensor installation offset, reference plane angle, installation tolerance, and temperature drift compensation coefficient. The coordinate mapping unit reads the above parameters and outputs a mapping matrix snapshot in each cycle, with each snapshot corresponding one-to-one with the cycle time provided by the time service unit. To handle transient changes caused by high-speed maneuvers, when the coordinate mapping unit detects that the rate of change of attitude angle exceeds a threshold, it triggers a high-frequency update channel, shortens the output interval of the mapping snapshot, and appends a transient marker field to the mapping snapshot to indicate that subsequent calculations should use a higher time resolution. Preprocessing of velocity and acceleration data is performed by the inertial navigation preprocessing unit. The inertial navigation preprocessing unit first excludes sensor saturation segments and obviously distorted segments based on health status markers, performs bias removal and temperature-related drift correction on the remaining segments, and resamples them to the beat time grid according to the alignment results provided by the time service unit. When a segment of data forms a hole due to communication loss, the inertial navigation preprocessing unit uses the effective segments before and after it for short-time interpolation, and marks the interpolation source and confidence level on the interpolation result. The interpolation mark will be included in the metadata area of ​​the motion compensation parameter set.Subsequently, under the mapping snapshot constraints of the coordinate mapping unit, the motion calculation unit transforms the attitude angle sequence, angular velocity sequence, and linear acceleration sequence to the lidar coordinate system, and combines the body velocity sequence and altitude information to obtain the attitude and velocity descriptions corresponding to each cycle. When the cycle span is long or the attitude change rate is high, the motion calculation unit establishes a multi-point spline description within a single cycle. The spline nodes are located at the start and end boundaries and midpoint of the window to characterize the attitude and velocity change trajectory within the cycle. Regarding anomaly handling, when the inertial navigation unit's health status is marked as unstable or the sensor temperature exceeds the operating range, the motion calculation unit marks the attitude and velocity description of that cycle as pending review and triggers a weighting strategy. When the installation bias record is missing or the mapping parameter version is mismatched, the coordinate mapping unit marks the output of that cycle as restricted and writes it into the list of parameter fields that need to be read back. Finally, the parameter construction unit summarizes the attitude and velocity descriptions, mapping snapshots, interpolation annotations, transient markers, and weight reduction strategy markers for each beat, encapsulates them into a motion compensation parameter set, and keeps the time key and beat number in the set consistent with the scan timing identifier. This motion compensation parameter set serves as the output field of this section and is directly used by the motion compensation parameter set application and adaptive filtering processing in S320. At the same time, the transient markers in the set will be used for segment neighborhood weight allocation in spatial interpolation and scan geometry mapping in S330, and as a time-varying reference for channel overlap processing in segment coordinate alignment in S410, thus forming a continuous data channel with the subsequent main steps.

[0074] S320. Extract the channel wind speed component from the initial observation wind speed sequence, apply motion compensation parameter set and adaptive filtering to generate a compensated wind speed sequence.

[0075] Specifically, the initial wind speed sequence output by S230 and the motion compensation parameter set output by S310 are input in parallel to the processing link in this section. Channel wind speed component extraction is performed by the component analysis unit. The component analysis unit first reads the angular domain organization structure and cycle number of the initial wind speed sequence, maps each angular domain observation value within the same cycle to the radial measurement direction of the lidar coordinate system according to the channel geometry, and applies pre-weighting to the angular domain positions with low-confidence segment markers. After reading the angular step position in the scan timing identifier, the component analysis unit performs neighborhood consistency judgment on adjacent angular domain observations. If an angular domain observation differs significantly from its neighbors and the angular domain position has a history of pending verification or removal in the quality gating annotation, a deferred participation mark is set for the observation during the component analysis stage. The deferred participation mark can be revoked or upgraded to a non-participation state as the cycle progresses. The motion compensation parameter set application is executed by the compensation application unit. Within each cycle, the compensation application unit reads the attitude and velocity descriptions and mapping snapshots from the motion compensation parameter set, interpolates the radial observations given by the component analysis unit according to the multi-point spline description within the cycle, and shortens the interpolation step size for sub-segments with transient markers. When a cycle is marked as a restricted state, the compensation application unit minimizes the compensation weight for that cycle and retains the original observations for that cycle in the output as a comparison. For observation entries inherited from previous steps in the perturbed state, the compensation application unit adds a suppression term during the projection process of the attitude and velocity descriptions. In this section, the suppression term is manifested as reducing the influence of the entry on the compensation result, and the reasons and sources of suppression are recorded. Subsequently, adaptive filtering is performed by the filtering execution unit. Within each cycle, the filtering execution unit constructs a filtering scenario description consisting of component observations, transient markers, interpolation annotations, low-confidence segment markers, and limited-use state markers, and configures filtering parameters accordingly. When the scenario description indicates a high rate of attitude change within the cycle, the filtering execution unit uses a higher update frequency and a shorter observation window. When the scenario description indicates a large proportion of observations requiring verification or missing measurements, the filtering execution unit introduces a robust suppressor within that cycle to prune abnormally discrete observation residuals. During the filtering process, the filtering execution unit maintains a data write-back channel with the component analysis unit: if the filtered residual shows significant and continuous deviations at a certain angular region position, the filtering execution unit upgrades the deferred participation marker for that angular region position to non-participation and writes the upgrade behavior into the verification record of the aggregate metadata; if the filtered residual tends to converge over multiple cycles, the filtering execution unit can revoke the previous deferred participation marker and record the basis for the revocation in the metadata.In terms of anomaly handling, when the set of motion compensation parameters is missing in a certain beat, the compensation application unit puts that beat into a rollback mode. In the rollback mode, only the component analysis results are applied and the attitude statistics of the beats with the nearest time are used as a temporary substitute, and the rollback mode is written into the output beat summary. When a certain angular region of the observed initial wind speed sequence has a persistently low confidence over a long period of time, the filtering execution unit triggers trajectory disconnection labeling. The trajectory disconnection label marks the position of the angular region in the output, prompting the subsequent spatial interpolation stage to adopt a conservative extension at that position. Finally, the results organization unit arranges the compensation results of each corner position within each beat according to the beat number and corner index to form a compensated wind speed sequence. The unit also writes the processing records such as deferred participation, non-participation, rollback mode, robustness suppressor triggering, and trajectory disconnection into the metadata area of ​​the sequence. The compensated wind speed sequence, as the output field of this subsection, is directly called by the spatial interpolation and scan geometry mapping in S330. At the same time, the metadata area of ​​this sequence will be used as a weight reference for the time dimension and corner dimension in the segment coordinate alignment and channel overlap processing in S410, and can be read by the confidence check in S420 for quality assessment of the boundary overlapping area, thereby establishing a continuous operation link between the main steps.

[0076] S330. Perform spatial interpolation and scanning geometric mapping on the compensated wind speed sequence to generate a wind speed vector grid fragment structure;

[0077] Specifically, the compensated wind speed sequence output by S320 is used as the sole numerical input, and the angular step position and beat number in the scan timing identifier are read simultaneously to establish a common reference for the time dimension and angular domain dimension. Spatial interpolation is performed by the spatial reconstruction unit. The spatial reconstruction unit first constructs the observation ray set for the corresponding beat based on the angular domain index and beat number of the compensated wind speed sequence. The observation ray set describes the observation direction and effective range of each angular domain position in the lidar coordinate system. When the sequence metadata area shows that some angular domain positions have trajectory disconnection markers or long-term low-confidence markers, the spatial reconstruction unit downgrades the participation weight of these positions and enables a neighborhood expansion strategy in the area that needs to be filled. The neighborhood expansion strategy is referenced by the surrounding angular domain positions and adjacent beats, and the source record is written during the expansion process. Subsequently, the scanning geometry mapping is performed by the geometry mapping unit. This unit reads the installation geometry relationship between the airframe and the lidar, as well as the scanning coverage area, projects the observation ray set onto the geographic reference plane or the flight mission reference plane, and generates scan slices according to a timeframe. In the presence of transient markers within a timeframe, the geometry mapping unit establishes multi-segment descriptions within that slice, ensuring that the attitude and velocity states corresponding to each segment are consistent with the mapping snapshot of S310, and inserts transition bands between segments. To form a regular data carrier within the slice space, the mesh generation unit constructs mesh nodes and mesh cells based on a preset mesh scheme of the flight mission reference plane. When the mesh resolution required by the mission does not perfectly match the scanning coverage, the mesh generation unit records the mismatched areas and performs adaptive sparsification or densification processing on the mesh. The result of the adaptive processing is written into the mesh description field for identification in subsequent main steps. After the grid is established, value assignment and fusion are performed by the value fusion unit. This unit assigns values ​​to the corresponding grid nodes or grid cells for each observation ray provided by the spatial reconstruction unit, and performs weighted fusion on overlapping observation contributions. The weighted fusion is based on the metadata region of the compensated wind speed sequence, the neighborhood extension record of the spatial reconstruction unit, and the segment description of the geometric mapping unit. If a grid node is contributed by multiple beats or multiple angular locations, the value fusion unit comprehensively allocates values ​​according to temporal proximity, angular proximity, and metadata weights, and adds side notes to contributions with extremely low weights or those marked with anomalies. Boundary processing is performed by the boundary tuning unit. This unit locates the intersection region of the scan slice's coverage boundary and the grid boundary, applies a smooth transition to sparse nodes within the intersection region, and adds extrapolation marks to the extrapolated regions. When a break occurs in the scan coverage, the boundary tuning unit establishes a soft connection region at the break point, allowing values ​​on both sides of the break to gradually transition within the soft connection region. The range and transition strength of the soft connection region are written into the boundary description for subsequent use by the S420 during grid stitching and confidence verification.In terms of anomaly handling, when the compensated wind speed sequence is in a large-scale retreat mode or a large proportion of entries are temporarily suspended in a certain beat, the assignment and fusion unit reduces the overall contribution weight of that beat and writes a beat quality prompt in the beat summary of the wind speed vector grid fragment structure. When the grid generation unit finds that the overlap between the scan geometry mapping and the preset grid scheme is lower than the threshold, the system triggers a sparse region alarm and sets a mark for the relevant grid units not to participate in subsequent splicing. After all processing is completed, the result construction unit encapsulates the gridded wind speed value of each beat along with the grid description field, boundary description, source record, neighborhood expansion record, sub-segment description, and beat quality prompt to form a wind speed vector grid fragment structure. The wind speed vector grid fragment structure is used as the output field of this subsection and is directly called by the fragment coordinate alignment and channel overlap processing of S410, and is used as the basic input in the grid splicing and confidence verification of S420. In addition, the grid description field and boundary description will serve as auxiliary information across the main steps, providing a positioning reference for quality traceability and three-dimensional wind field data organization within the S400 main step. In summary, the technical effects of this step are as follows: By coordinating the temporal and spatial dimensions, the compensated wind speed sequence is subjected to component-consistent spatial interpolation and scanning geometric mapping, forming a wind speed vector grid segment structure with source, weight, and boundary descriptions. This provides a stable, continuous, and traceable gridded foundation for subsequent segment alignment, splicing, and confidence verification.

[0078] Step S400 includes at least steps S410-S430:

[0079] S410. Obtain the wind speed vector grid segment structure and scanning time sequence identifier, perform segment coordinate alignment and channel overlap processing, and obtain three-dimensional wind field block data.

[0080] Specifically, the wind speed vector grid segment structure is used as input. This wind speed vector grid segment structure is a gridded carrier formed after spatial interpolation and scanning geometric mapping based on the compensated wind speed sequence in the previous step. It contains information such as grid node sets, grid cell sets, boundary descriptions, source records, neighborhood expansion records, sub-segment descriptions, and beat quality prompts organized by beat. At the same time, the scanning timing identifier is connected in parallel. The scanning timing identifier records key elements such as beat number, channel transmission sequence, expected echo time window, angular step position, and transition switching point, which are used to establish a consistent reference relationship in the time dimension and angular domain dimension. This section begins by establishing a unified reference coordinate framework using the coordinate reference management unit. It reads the sub-segment description corresponding to each beat in the wind speed vector grid segment structure and compares it with the transition switching points in the scan timing identifier. When multiple sub-segment descriptions are found for a particular beat, the system splits the grid segment of that beat into several micro-segments and binds each micro-segment to its corresponding sub-segment time range and attitude state. Subsequently, the time axis integration unit rearranges the order of each beat and its micro-segments on a unified time axis, eliminating overlaps or gaps across beats and sub-segments. The integration result is then written into a segment index table, organized by beat number, sub-segment order, and grid block number, serving as the retrieval key for subsequent channel overlap calculations and coordinate alignment.

[0081] After completing the above baseline preparation, fragment coordinate alignment is performed by the alignment execution unit. The alignment execution unit first reads the mesh description and boundary description, identifying the spatial coverage of each mesh fragment and its potential contact surfaces with adjacent fragments. For adjacent fragments within the same temporal neighborhood, the alignment execution unit establishes candidate alignment surfaces according to the mesh node distribution and scan geometry, and assigns initial weights based on the source record and neighborhood expansion record. When a candidate alignment surface involves a region marked as a low-confidence fragment in a previous step, the alignment execution unit reduces the participation weight of that region and retains the reason for the weight reduction in the alignment log. Within a beat with a transition point, the alignment execution unit uses the attitude state in the sub-segment description as an additional constraint, performing self-alignment of mesh fragments from different sub-segments within the same beat before splicing them across beats with fragments from adjacent beats, avoiding errors introduced by temporal base inconsistencies. Regarding anomaly handling, if a fragment is found to be missing necessary boundary descriptions or source records, the alignment execution unit lists that fragment as pending review, excluding it from this round of coordinate alignment, and marks it as "pending review" in the fragment index table for subsequent rounds.

[0082] Channel overlap processing is performed by the overlap assessment unit. After aligning the segment coordinates, the overlap assessment unit decomposes the channel contribution of the overlapping area of ​​adjacent segments. The channel contribution is determined by the source record of the wind speed vector grid segment structure and the metadata of the previous main step. If the same grid node has values ​​assigned from multiple channels and multiple beats, the overlap assessment unit calculates the channel coverage ratio, temporal proximity, and relative position difference in the angular domain to generate a comprehensive overlap index. This overlap index is then correlated with the segment quality indicator to form an overlap weight table. For nodes with long-distance neighborhood extension records, the overlap assessment unit introduces a reduction term for such nodes in the weight table and writes the reduction details into the weight notes. For sources showing contributions from fallback mode or deferred participation, the overlap assessment unit further reduces their participation coefficient. After completing the overlap assessment, the fragment fusion unit performs a fusion operation on each candidate alignment surface in the alignment coordinate system. The fusion operation is based on the overlap weight table and performs layered fusion on overlapping nodes and overlapping units: first, the vector values ​​at the same coordinate position are integrated at the node level, then the transition between adjacent nodes is filled at the unit level, and the boundary band is smoothed when necessary; if a situation is encountered in the fusion process where the node values ​​are significantly different and have no common source, the fragment fusion unit marks the point as a conflict point and retains the values ​​on both sides for subsequent confidence verification. Finally, the data construction unit generates 3D wind field block data. The 3D wind field block data consists of several grid blocks that have been processed for coordinate alignment and channel overlap. It contains fields such as grid block index, node value, cell topology, overlap weight, and conflict point list. The 3D wind field block data is the output field of this section and is directly consumed by the grid stitching and confidence check in S420. At the same time, the overlap weight and conflict point list in the 3D wind field block data will serve as key references in the boundary overlap area processing and confidence calculation in S420, and will echo the grid description and boundary description in the preceding S330 at the cross-main step level.

[0083] S420. Extract the boundary overlapping areas from the 3D wind field block data, perform mesh stitching and confidence verification, and generate a 3D wind field data set.

[0084] Specifically, the three-dimensional wind field block data is used as the sole input. This data is organized by grid block index and includes fields such as node value, cell topology, overlap weight, and conflict point list in each grid block. This section first uses a boundary identification unit to scan the spatial coverage and adjacency relationships of all grid blocks to determine the overlapping boundary areas that need to be stitched together. During the identification process, the boundary identification unit retrieves the boundary description of each grid block and the fragment index table from previous steps. For areas with multiple sub-segment descriptions or cross-beat contact, a time adjacency marker is added, and the above information is integrated into a boundary grouping list. Next, mesh stitching is performed by the stitching execution unit. This unit establishes stitching channels for adjacent mesh blocks in each boundary grouping list. These channels are processed at both the node and cell levels: at the node level, the unit reads the overlap weights, merges overlapping nodes according to their weights, and applies conflict decomposition strategies to nodes with conflict points, forming preliminary merged values ​​while preserving conflict source information; at the cell level, the unit performs topological corrections on the merged nodes, including cell index rearrangement, duplicate cell removal, and gap cell filling. Gap cell filling references the shape and orientation of adjacent cells, and records the filling source and applicable range. For stitching across beats or sub-segments, the unit establishes a transition zone inside the stitching seam based on temporal adjacency markers. Node values ​​within the transition zone are gradually changed according to both temporal and angular domain weights. The transition zone range and gradient weights are written into the stitching specification.

[0085] After the structural layer stitching is completed, confidence verification is carried out by the verification execution unit. The verification execution unit first constructs a confidence assessment model within the boundary overlap area based on overlap weights, a conflict point list, and stitching instructions. It outputs a confidence flag and confidence grade for each merged node. When the confidence grade in a certain area is consistently low, the verification execution unit designates that area as a check zone. Within the check zone, it traces back the conflict sources of the 3D wind field block data and performs secondary smoothing on nodes with consistent sources and controllable outliers. If the sources are inconsistent or the outlier level is too large, the check zone mark is retained in the annotation area of ​​the 3D wind field data set, prompting subsequent use to pay attention to the interpretability of this area. Subsequently, continuity verification is performed across the entire domain. Continuity verification reads the overall structure after mesh stitching and determines the flow connectivity across mesh blocks, the consistency of velocity vector directions, and the continuity of local extremum distributions. When non-physical breaks or abrupt changes are detected, continuity verification establishes a repair suggestion list for the problem area. This list does not directly modify node values ​​but only announces possible repair directions and ranges in the annotation area for reference in subsequent write-back phases or external strategies. Regarding anomaly handling, if a boundary group list has missing key fields or inconsistent mesh topology during stitching, the stitching execution unit activates a downgrade mode for that group, merging values ​​only at the node level while pausing topology correction at the cell level, and writing the reason for the downgrade mode in the stitching description. If confidence verification reveals a large area of ​​low confidence that cannot be improved through secondary smoothing, the verification execution unit marks this area as a delayed release area, which is presented as a mask in the final output. After completing the above process, the result integration unit aggregates all spliced ​​regions that have passed the confidence check into an integrated data carrier, forming a three-dimensional wind field data set. The three-dimensional wind field data set, as the output field of this section, will be directly consumed by the data consistency check and parameter write-back of S430. At the same time, the splicing description, transition zone range, and verification zone mark in the three-dimensional wind field data set will be used as the basis for judgment and write-back in the consistency strategy of S430, and will provide a reverse comparison reference for the mesh generation of the preceding S330 and the segment alignment of the preceding S410 in the cross-main step dimension.

[0086] S430: Perform data consistency verification and parameter write-back on the three-dimensional wind field data set, and generate a configuration update instruction structure;

[0087] Specifically, the three-dimensional wind field data set is used as input. This set includes fields such as integrated mesh, node values, stitching instructions, transition zone range, verification zone markers, and delayed release area masks. This section first loads the registered end-to-end consistency rule set by the consistency rule management unit. The consistency rule set originates from the collaborative constraints established in this invention during multi-beam scanning, optical path management, backscatter demodulation, inertial information fusion, and spatial mapping, covering four dimensions: temporal consistency, spatial consistency, cross-channel consistency, and boundary consistency. Subsequently, the consistency determination unit performs a comprehensive check on the entire 3D wind field data set: the temporal consistency check scans the temporal evolution trajectory of each grid node in beat order, identifies discontinuous change segments, and writes the time breakpoint location and associated beat in the determination log; the spatial consistency check evaluates the value difference between adjacent nodes and adjacent cells in terms of grid adjacency, marks abnormal gradient bands, and associates them with the splicing description and transition zone range; the cross-channel consistency check back the overlap weight and conflict point list in the 3D wind field block data, performs inter-source difference assessment on multi-source nodes retained in the set, and writes the difference level in groups according to source records; the boundary consistency check targets the verification zone markers and the delayed release area mask, determines the impact of these areas on grid coherence within the adjacency range, and forms a boundary revision suggestion table.

[0088] After the consistency determination is completed, the parameter write-back is carried out by the write-back execution unit. The write-back execution unit generates write-back instructions based on the consistency determination results: For time breakpoints located by time consistency checks, the write-back execution unit generates adjustment suggestions for the beat length and beat sampling window in the scan control direction, forming a write-back segment oriented towards the preceding scan time sequence identifier; for abnormal gradient bands located by spatial consistency checks, the write-back execution unit generates grid resolution adjustment and neighborhood expansion strategy update suggestions in the spatial mapping and interpolation direction, forming a write-back segment oriented towards the preceding wind speed vector grid segment structure generation strategy; for source differences identified by cross-channel consistency checks, the write-back execution unit generates weight revision and entry selection boundary update suggestions in the quality gating and robust aggregation direction, forming a write-back segment oriented towards the preceding Doppler shift sequence aggregation strategy; for boundary revision suggestions generated by boundary consistency checks, the write-back execution unit generates update suggestions for transition band width, gradient weight, and verification band handling methods in the splicing and confidence direction, forming a write-back segment oriented towards subsequent rounds within this main step. To ensure the traceability and closed-loop nature of the writeback, the writeback execution unit binds source evidence to each writeback suggestion, including the triggered rule entry, the grid area involved, the associated tick range, and the corresponding judgment log location. When conflicting suggestions occur, a priority arbitrator is used to provide a merging solution or a phased implementation strategy.

[0089] Finally, the configuration building unit integrates the aforementioned write-back fragments into a configuration update instruction structure. This structure consists of an instruction header, an instruction body, and supporting evidence: the instruction header specifies the target module and its effective scope, including multi-beam scanning parameters and optical path calibration configuration for preceding S110, demodulation and quality gating parameters for preceding S210 and S220, robust aggregation parameters for preceding S230, adaptive filtering configuration for preceding S320, and mesh stitching and confidence strategies within this main step; the instruction body lists specific parameter items and suggested value ranges, along with corresponding execution timing and rollback conditions; the supporting evidence contains consistency judgment logs, boundary overlap area screenshot indexes, source difference grouping lists, and transition zone parameter trajectories. Regarding anomaly handling, if a large area of ​​temporarily suspended release is found in the 3D wind field data set during consistency judgment and cannot be improved in the short term through parameter write-back, the configuration building unit will add a freeze entry to the configuration update instruction structure, temporarily blocking downstream releases and external interfaces in the relevant area, and declaring the freeze period and unfreezing conditions in the instruction header. At this point, the output product is a configuration update instruction structure. This structure, as the output field of this subsection, is read by the preceding module and the configuration management unit within this invention. Specifically, the write-back content oriented towards the scanning direction is read by the multi-beam scanning parameters and optical path calibration data of S110 and takes effect in the next round of scanning cycle synchronization and stray light shielding processing; the write-back content oriented towards the demodulation and gating direction is read by S210 and S220 and takes effect in time-frequency demodulation and quality gating annotation; the write-back content oriented towards the aggregation and filtering direction is read by S230 and S320 respectively and takes effect in robust aggregation and adaptive filtering processing; and the write-back content oriented towards the mesh stitching direction flows back to this main step for stitching and confidence verification in subsequent rounds, thereby completing the closed-loop linkage between the main step and within this step. In summary, the technical effects of this step are as follows: By performing multi-dimensional consistency verification and hierarchical parameter write-back on the three-dimensional wind field data set, an executable configuration update instruction structure is formed for each stage of scanning, demodulation, aggregation, filtering and stitching. This constructs a data-parameter two-way closed loop that connects the main steps before and after, supporting stable operation and traceable correction in consecutive rounds.

[0090] Example 2: Figure 2 A structural block diagram of an airborne wind-measuring lidar system according to an embodiment of the present invention is shown. Figure 2 As shown, the structure may include:

[0091] The scanning timing synchronization and stray light shielding unit 01 is used to acquire multi-beam scanning parameters and optical path calibration data, and perform scanning timing synchronization and stray light shielding processing, outputting scanning timing identifiers and noise labeling data. Specifically, it receives multi-beam scanning parameters and optical path calibration data from the installation and calibration process, reads the timing reference, channel transmission sequence, angle step setting, optical element assembly and adjustment record, and light-shielding component status, drives the timing control interface to send alignment pulses and establish a timing number, triggers the light-shielding component and electrical threshold to perform shielding during the noise candidate period, records the time period markers corresponding to the window transmittance and residual reflectance, and forms scanning timing identifiers and noise labeling data. The scanning timing synchronization and stray light shielding unit transmits the scanning timing identifiers and noise labeling data to the transmission and reception trigger matching unit as the basis for triggering and window configuration, and retains the timing switching position, light-shielding component pose, and threshold strategy entries in the log storage for subsequent traceability by the data consistency verification and parameter write-back unit.

[0092] The transmit and receive trigger matching unit 02 is used to parse the channel index based on the scan timing identifier and perform transmit and receive trigger matching, outputting a backscatter signal set. Specifically, it receives the scan timing identifier and noise label data from the scan clock synchronization and stray light shielding unit, parses the clock number and angular step position to construct the channel index, drives the transmitter to send pulses at the time corresponding to the channel index, schedules the receiver to start sampling in the expected window and applies the threshold associated with the noise label data, completes the binding of trigger and echo acquisition, and converges into a backscatter signal set according to the clock sequence. The backscatter signal set and the channel index are registered with time keys and status keys in the structured buffer and handed over to the optical path correlation correction unit for use. At the same time, the acquisition anomaly and missing measurement mark are written into the event log for reference by the subsequent grid stitching and confidence verification unit.

[0093] The optical path correlation correction unit 03 is used to merge noise-labeled data and backscattered signal sets to generate a path correction mark structure. Specifically, it receives noise-labeled data and backscattered signal sets, reads marks such as window reflection, cabin scattering, and external strong light, aligns echo samples according to channel index and beat, generates mark entries consisting of influence type, expected arrival time, affected angular domain, gain correction rules, and data quality suggestions, and binds them to the corresponding samples to form a path correction mark structure. The path correction mark structure and the backscattered signal set are jointly transmitted to the time-frequency demodulation and path correction mark application unit. The mark entries are simultaneously archived into the traceable storage area for data consistency verification and parameter write-back unit to read.

[0094] The time-frequency demodulation and path correction tagging application unit 04 is used to call the path correction tagging structure and output the demodulated spectrum sequence under the constraints of the backscattered signal set and the scan timing identifier. Specifically, it receives the backscattered signal set, the scan timing identifier and the path correction tagging structure, constructs a time window group and associates the tagging entries, performs preprocessing, short-time analysis and spectral shaping operations on each window, converges the analysis bandwidth and adjusts the sidelobe suppression intensity for non-target light tags, splits the time base transition tags into sub-segments and summarizes them separately, generates demodulated spectrum entries composed of frequency and amplitude sequences, and assembles them into a demodulated spectrum sequence according to the channel index and beat order. The demodulated spectrum sequence is transmitted to the frequency shift peak and spectral peak stability index extraction unit, and at the same time, the default path, the limited neighborhood and the disturbed record are registered in the status record for reference in subsequent quality gating and robust aggregation.

[0095] The frequency shift peak and spectral peak stability index extraction unit 05 is used to extract the frequency shift peak and spectral peak stability index from the demodulated spectrum sequence and output the quality gating label and Doppler frequency shift sequence. Specifically, it receives the demodulated spectrum sequence, performs candidate peak retrieval, peak position refinement and temporal coherence determination, generates time consistency and frequency band coherence tags based on the tag summary and abnormal records, forms quality gating labels, outputs the main peak position for each record and writes gating labels and comments next to the entry, and assembles them into a Doppler frequency shift sequence. The quality gating label and Doppler frequency shift sequence are synchronously transmitted to the channel aggregation unit, and the gating log and peak shape review record are entered into the archive area for use by the data consistency verification and parameter write-back unit.

[0096] Channel aggregation unit 06 is used to perform channel aggregation based on the Doppler frequency shift sequence and channel index, and output the initial value sequence structure of observed wind speed. Specifically, it receives the Doppler frequency shift sequence and channel index, constructs observation segments according to angular step positions, filters items according to quality gating labels and sets weights, removes or reduces the weight of abnormal discrete items, completes the aggregation of the same beat and angular domain, outputs the aggregated observation values ​​and participation summary, concatenates the results of each beat along the time direction, and encapsulates them into the initial value sequence structure of observed wind speed. The initial value sequence structure of observed wind speed is passed to the inertial navigation data access and time alignment coordinate mapping unit for beat reference, and passed to the motion compensation and adaptive filtering unit as the component analysis data source, and the participation summary is archived for data consistency verification and parameter write-back unit to read.

[0097] The inertial navigation data access and time alignment coordinate mapping unit 07 is used to receive attitude and velocity data from the inertial navigation unit and perform time alignment and coordinate system mapping processing, outputting a set of motion compensation parameters. Specifically, it receives attitude sequences, angular velocity sequences, linear acceleration sequences, velocity and altitude records, as well as equipment timestamps and health markers pushed by the avionics side, completes alignment by referencing the beat reference in the scan timing identifier, constructs a coordinate mapping chain according to the installation geometry and calibration records, and outputs a set of motion compensation parameters consisting of attitude and velocity descriptions corresponding to the beat, mapping snapshots, interpolation and weighting markers. The set of motion compensation parameters is transmitted to the motion compensation and adaptive filtering unit, and transient markers and limited state entries are registered in the state library for the segment coordinate alignment and channel overlap processing unit to establish weight references in the time dimension.

[0098] The motion compensation and adaptive filtering unit 08 is used to extract channel wind speed components from the initial observation wind speed sequence structure and call the motion compensation parameter set to perform adaptive filtering processing, outputting a compensated wind speed sequence. Specifically, it receives the initial observation wind speed sequence structure and the motion compensation parameter set, analyzes the radial component according to the channel geometry, performs time interpolation within the beat according to the attitude and velocity description, shortens the interpolation step size for transient marker segments, reduces the weight of the limited state beat, constructs a filtering scenario containing component observation, transient and confidence information, performs robust filtering and writes back the deferred participation and non-participation markers, and outputs the compensated wind speed sequence according to the angular domain and beat. The compensated wind speed sequence is transmitted to the spatial interpolation and scan geometry mapping unit, and at the same time, the rollback mode and trajectory disconnection record are registered in the metadata area for the grid splicing and confidence verification unit to read.

[0099] The spatial interpolation and scanning geometry mapping unit 09 is used to perform spatial interpolation and scanning geometry mapping on the compensated wind speed sequence and output the wind speed vector grid fragment structure. Specifically, it receives the angular step position in the compensated wind speed sequence and scanning time sequence identifier, constructs the observation ray set and projects it onto the task reference plane, generates scanning slices, establishes a preset grid and performs densification or sparse processing on sparse areas, completes node assignment and unit fusion according to weight and source record, processes the coverage boundary and sets soft connection regions at the break points, and summarizes to form the wind speed vector grid fragment structure. The wind speed vector grid fragment structure is passed to the fragment coordinate alignment and channel overlap processing unit, and the grid description and boundary description are archived for reference by the data consistency verification and parameter write-back unit.

[0100] The segment coordinate alignment and channel overlap processing unit 10 is used to perform segment coordinate alignment and channel overlap processing under the constraints of wind speed vector grid segment structure and scan timing identifier, and output three-dimensional wind field block data. Specifically, it receives the wind speed vector grid segment structure and scan timing identifier, reads the segment description and transition switching position, completes time axis integration and establishes a segment index table, establishes candidate alignment surfaces for adjacent segments in a unified coordinate framework, generates overlap weights based on source records and neighborhood expansion records, performs layered fusion of overlapping nodes and units and marks conflict points, and outputs three-dimensional wind field block data. The three-dimensional wind field block data is transmitted to the grid splicing and confidence verification unit, and the overlap weights and conflict point list are registered in the state database for subsequent reference by the data consistency verification and parameter write-back unit.

[0101] The mesh stitching and confidence verification unit 11 is used to extract boundary overlapping areas from the three-dimensional wind field block data and perform mesh stitching and confidence verification, outputting a three-dimensional wind field data set. Specifically, it receives three-dimensional wind field block data, scans the adjacent relationships of mesh blocks to identify boundary overlapping areas, completes value merging at the node level and performs decomposition strategies on conflict points, corrects the topology and fills gaps at the unit level, sets transition zones and registers gradual weights at cross-beat positions, constructs confidence assessment and continuity verification, generates verification zones and delayed release area markers, and summarizes and outputs a three-dimensional wind field data set. The three-dimensional wind field data set is delivered to the data consistency verification and parameter write-back unit, and the stitching instructions and transition zone range are synchronously stored in the database for reference during write-back.

[0102] The data consistency verification and parameter write-back unit 12 is used to perform data consistency verification on the three-dimensional wind field data set and generate a configuration update instruction structure. This configuration update instruction structure is used to send parameter update instructions back to the scanning beat synchronization and stray light shielding unit, the time-frequency demodulation and path correction marking application unit, the frequency shift peak and spectral peak stability index extraction unit, the channel aggregation unit, the motion compensation and adaptive filtering unit, and the grid stitching and confidence verification unit, and to register the source and effective range. Specifically, it receives the three-dimensional wind field data set, performs consistency inspection according to four categories of rules: time, space, cross-channel, and boundary, and generates suggestions for adjusting beat length and window settings, revising grid resolution and neighborhood expansion strategies, and quality gating and aggregation. The weight revision suggestions and transition band parameter update suggestions are assembled into a configuration update instruction structure and bound with source evidence. Updates related to the clock cycle are sent back to the scanning clock cycle synchronization and stray light shielding unit; updates on analysis bandwidth and sidelobe suppression strategies are sent back to the time-frequency demodulation and path correction marking application unit; updates on gating thresholds and coherence judgment boundaries are sent back to the frequency shift peak and spectral peak stability index extraction unit; updates on weight allocation and anomaly suppression boundaries are sent back to the channel aggregation unit; updates on filtering scene parameters and participation strategies are sent back to the motion compensation and adaptive filtering unit; and updates on transition band and verification band handling are sent back to the mesh stitching and confidence verification unit. The version and effective scope are registered in the configuration library, completing the parameter write-back closed loop.

Claims

1. An airborne wind-finding lidar method, characterized by, Comprise: S100, from the multi-beam scanning parameters and optical path calibration data, the scanning beat synchronization and stray light shielding processing are carried out, and the scanning time sequence mark and noise mark data are obtained; From the scanning time sequence mark, the channel index is extracted for emission-reception matching, and a backscattering signal set is generated; The noise mark data and the backscattering signal set are corrected to generate a path correction mark structure; Specifically comprising: The process of scanning beat synchronization and stray light shielding processing comprises: from the multi-beam scanning parameters and optical path calibration data, based on the job start alignment pulse generated by the time sequence control unit and the beat resynchronization mechanism, the scanning beat synchronization and stray light shielding processing are carried out, wherein the scanning beat synchronization comprises that the emission scheduler and the receiving scheduler establish the beat index in the same absolute time reference, and the stray light shielding processing comprises that the non-target angular domain is geometrically limited by the diaphragm and the light shield, and the non-target light distribution is estimated according to the window transmittance and the residual reflectivity of the anti-reflection film, to obtain the scanning time sequence mark and the noise mark data;From the scanning time sequence mark, the channel index is extracted for emission-reception matching, and a backscattering signal set is generated;The backscattering signal set contains original echo samples, sampling time labels, sampling threshold configurations, disturbed states and noise candidate period references; The noise mark data and the backscattering signal set are corrected to generate a path correction mark structure containing non-target light influence type, expected arrival period, affected angular domain, receiving channel gain correction rule, data quality suggestion, time base transition neighborhood and limit neighborhood mark; The multi-beam scanning parameters are a parameter set indicating the number of beams of each emission channel in a single job, the beam angle, the scanning coverage, the scanning start and end angle, the angular step, the channel emission sequence, the single pulse width, the pulse repetition structure, the ranging window setting and the back sampling integration threshold; The optical path calibration data are the calibration records of the installation pose, the light transmission aperture, the off-axis deviation, the reflection surface roughness, the window transmittance and the residual reflectivity of the anti-reflection film of the emission and receiving end optical elements, and the geometric correspondence of the corresponding body installation reference surface, inertial navigation reference surface and attitude measurement reference surface; S200, based on the backscattering signal set and the scanning time sequence mark, time-frequency demodulation, quality gating mark and channel aggregation processing are performed, and an observation wind speed initial value sequence structure is generated; S300, obtain inertial navigation unit attitude and speed data, including attitude angle sequence, angular velocity sequence, linear acceleration sequence, body speed sequence and height information, and attach device timestamp, health status marker and sensor temperature record, perform time alignment and coordinate system mapping, including time service unit master-slave association of inertial navigation unit device timestamp, airborne task control timestamp and scanning time sequence identifier reference time, coordinate mapping unit establishes mapping chain between body coordinates, inertial navigation unit coordinates and laser radar coordinates and outputs mapping matrix snapshot, obtains motion compensation parameter set, after adaptive filtering, based on angular step position and beat number in scanning time sequence identifier, through spatial interpolation and scanning geometry mapping, including space reconstruction unit constructs observation ray set, geometry mapping unit projects observation ray set to geographic reference plane or flight task reference plane to generate scanning slice, generates wind speed vector grid segment structure containing grid node set, grid cell set, boundary description, source record, neighborhood expansion record, subsection description and beat quality prompt organized by beat; S400, based on wind speed vector grid segment structure and scanning time sequence identifier, perform segment coordinate alignment, grid splicing and confidence check, generate configuration update instruction structure.

2. The method of claim 1, wherein, The process of generating the observation wind speed initial value sequence structure specifically includes: Based on the backscattering signal set and the scanning time sequence identifier, combined with the received channel gain correction rules and data quality suggestions in the path correction marker structure, perform time-frequency demodulation and path correction marker application processing, including a window construction unit generating time window groups according to the expected echo time window and the transition switching point and binding a path correction marker entry to each window, a time-frequency mapping unit performing time-frequency transformation to output an amplitude-frequency-time three-dimensional intermediate result, extracting frequency shift peak value and spectral stability index, using a combination of piecewise threshold method and neighborhood comparison method to implement candidate peak retrieval and perform time consistency and frequency band coherence evaluation, quality gating, and then generating an observation wind speed initial value sequence structure including an aggregated observation value matrix arranged by angular domain and beat and a aggregated meta-information set storing participation entry summary, weight distribution record, abnormal suppression record, segment merging record and low confidence segment marker.

3. The method of claim 1, wherein, The process of generating the wind speed vector grid segment structure includes: Obtain inertial navigation unit attitude and speed data, including attitude angle sequence, angular velocity sequence, linear acceleration sequence, body speed sequence and height information, and attach device timestamp, health status marker and sensor temperature record, perform time alignment and coordinate system mapping processing, including time service unit reading inertial navigation unit device timestamp, airborne task control timestamp and reference time of corresponding beat in scanning time sequence identifier and master-slave association of the three, coordinate mapping unit establishes mapping chain between body coordinates, inertial navigation unit coordinates and laser radar coordinates and outputs mapping matrix snapshot, obtains motion compensation parameter set, the motion compensation parameter set contains attitude and speed description, mapping snapshot, interpolation annotation, transient marker and weight reduction strategy marker of each beat; The channel wind speed component is extracted from the initial wind speed sequence, and the motion compensation parameter set is applied and adaptively filtered based on the attitude and speed description and mapping snapshot in the motion compensation parameter set. The compensation application unit performs time interpolation on the radial observation given by the component analysis unit according to the beat-in multi-point spline description, and the filter execution unit constructs a filter scene description and configures filter parameters in each beat to generate a compensated wind speed sequence. The compensated wind speed sequence is arranged according to the beat number and angular domain index.

4. The method of claim 3, wherein, The process of generating the compensated wind speed sequence also includes: For the compensated wind speed sequence, based on the angular step position and beat number in the scan timing identifier, spatial interpolation and scan geometry mapping are performed, including a spatial reconstruction unit constructing an observation ray set for the corresponding beat according to the angular domain index and beat number of the compensated wind speed sequence, and a geometry mapping unit projecting the observation ray set onto a geographic reference plane or a flight task reference plane and generating a scan slice according to the beat to generate a wind speed vector grid segment structure. The wind speed vector grid segment structure contains a grid node set organized by beat, a grid cell set, a boundary description, a source record, a neighborhood expansion record, a sub-segment description, and a beat quality prompt.

5. The method of claim 1, wherein, The process of generating the configuration update instruction structure also includes: Based on the wind speed vector grid segment structure and the scan timing identifier, combined with the transition switching point and beat number in the scan timing identifier, perform segment coordinate alignment and channel coincidence processing, including a coordinate reference management unit establishing a unified reference coordinate framework and splitting multiple sub-segment description beats into micro-segments, an alignment execution unit establishing a candidate alignment surface according to the grid node distribution and scan geometry relationship, extracting a boundary overlap area, and based on the coincidence weight and conflict point list in the three-dimensional wind field block data, performing grid splicing and confidence verification, including a splicing execution unit processing adjacent grid blocks at the node layer and the cell layer respectively, and a verification execution unit constructing a confidence evaluation model to output confidence markers and sub-grades. Through data consistency verification, based on the consistency rule set of four dimensions of coverage time consistency, spatial consistency, cross-channel consistency and boundary consistency, a configuration update instruction structure is generated, which contains an instruction header, an instruction body and evidence attachments, and the instruction header indicates the target module and the effective range, the instruction body lists the parameter items and the recommended value interval, the evidence attachments store the consistency judgment log, the boundary overlap area screenshot index, the source difference grouping list and the transition zone parameter trajectory.

6. An airborne wind lidar system for use in the method of any one of claims 1-5, characterized in that, It includes: A scan beat synchronization and stray light shielding unit is used to obtain multi-beam scanning parameters and optical path calibration data and perform scan beat synchronization and stray light shielding processing, outputting a scan timing identifier and noise annotation data; A transmission and reception trigger matching unit is used to analyze the channel index according to the scan timing identifier and perform transmission and reception trigger matching, outputting a set of backscatter signals; An optical path related correction unit is used to combine the noise annotation data and the backscatter signal set to generate a path correction marker structure; A time-frequency demodulation and path correction marker application unit is used to call the path correction marker structure under the constraints of the backscatter signal set and the scan timing identifier and output a demodulation spectrum sequence; The frequency shift peak and spectral peak stability index extraction unit is configured to extract the frequency shift peak and spectral peak stability index from the demodulation spectrum sequence and output the quality gating label and the Doppler frequency shift sequence; The channel aggregation unit is configured to perform channel aggregation according to the Doppler frequency shift sequence and the channel index and output the observed wind speed initial value sequence structure; The inertial navigation data access and time alignment coordinate mapping unit is configured to receive the inertial navigation unit attitude and velocity data and perform time alignment and coordinate system mapping processing, and output the motion compensation parameter set; The motion compensation and adaptive filtering unit is configured to extract the channel wind speed component from the observed wind speed initial value sequence structure and call the motion compensation parameter set to perform adaptive filtering processing, and output the compensated wind speed sequence; The spatial interpolation and scanning geometry mapping unit is configured to perform spatial interpolation and scanning geometry mapping on the compensated wind speed sequence and output the wind speed vector grid segment structure; The segment coordinate alignment and channel coincidence processing unit is configured to perform segment coordinate alignment and channel coincidence processing under the constraint of the wind speed vector grid segment structure and the scanning time sequence identifier, and output the three-dimensional wind field block data; The grid stitching and confidence check unit is configured to extract the boundary overlapping area on the three-dimensional wind field block data and perform grid stitching and confidence check, and output the three-dimensional wind field data total set; The data consistency check and parameter backwrite unit is configured to perform data consistency check on the three-dimensional wind field data total set and generate the configuration update instruction structure, which is used to send the parameter update instruction to the scanning beat synchronization and stray light shielding unit, the time-frequency demodulation and path correction label application unit, the frequency shift peak and spectral peak stability index extraction unit, the channel aggregation unit, the motion compensation and adaptive filtering unit, and the grid stitching and confidence check unit, and register the source and effective range.

Images