Aerosol scattering coefficient and laser radar combined observation aerosol vertical profile correction method and device

Abstract

The invention relates to the technical field of atmospheric environment monitoring and optical detection, in particular to an aerosol vertical profile correction method and device for aerosol scattering coefficient and laser radar combined observation, and the method comprises the steps: S1, collecting echoes according to an integral window, and obtaining distance correction echoes; s2, collecting a height scattering coefficient greater than or equal to 2, and aligning the height scattering coefficient with a distance correction echo grid / time; s3, background deduction, resolution matching and invalid height elimination are carried out to obtain preprocessed echoes; s4, generating an initial scattering coefficient profile; s5, identifying layered boundary division layer intervals; s6, establishing a correction factor model for each layer, adding overlap compensation to the near earth, and introducing a system constant; s7, taking the consistency of the total height of the theoretical echo and the preprocessed echo as a target, taking the consistency of multi-height scattering coefficients as a constraint, and carrying out joint solving to obtain correction factor distribution; and S8, outputting a correction profile, boundary information and a parameter set. According to the method, system constants and overlap compensation are solved by preprocessing echo layering scale gradient correction factors through combination of full-height consistency and two-height actual measurement constraints, and a correction profile is stabilized.

Description

Technical Field

[0001] This invention relates to the field of atmospheric environment monitoring and optical detection technology, and in particular to a method and device for correcting the vertical profile of aerosols by combining aerosol scattering coefficient and lidar observation. Background Technology

[0002] The vertical distribution of aerosol scattering coefficients is a crucial parameter characterizing the optical properties of aerosols in the near-surface and boundary layers, commonly used in applications such as visibility assessment, pollution transport analysis, and atmospheric radiative transfer calculations. Existing ground-based lidar can acquire echo information varying with altitude, which can then be used to retrieve aerosol-related optical profiles. Simultaneously, ground-based scattering coefficient (or extinction coefficient) observation equipment can provide measured values ​​of aerosol optical parameters at specific observation heights. Effectively combining these two methods to obtain reliable aerosol vertical profiles is a common technical approach in this field.

[0003] For example, CN101504353A discloses a method and system for providing near-ground suspended particulate matter distribution. The idea includes: obtaining satellite aerosol optical thickness information, and combining the vertical profile of the extinction coefficient obtained from single-point lidar observations with the extinction coefficient data from ground observation stations to obtain information related to the vertical distribution of aerosols; and the technical solution description involves the processing flow of aerosol optical parameters obtained from observation / derivation for subsequent decomposition or regression estimation.

[0004] However, in actual atmospheric observations, aerosols often exhibit a distinct stratified structure, and the near-field of lidar is susceptible to factors such as insufficient geometric overlap, system constant drift, and signal-to-noise ratio variations. If existing technologies primarily rely on ground optical parameters at a single observation altitude as constraints, or employ relatively fixed distribution assumptions to connect ground observations with vertical profiles, insufficient constraint information can easily arise when the aforementioned uncertainties exist. This makes it difficult for the inverted or corrected vertical profile of the aerosol scattering coefficients to simultaneously maintain consistency with echo observations in the near-ground region, at abrupt changes between layers, and across the entire altitude range, thus affecting the stability and reliability of profile correction.

[0005] Therefore, a major technical problem that needs to be solved by existing technologies is: under the condition that there is uncertainty in the system constant and near-field geometric overlap of lidar and that aerosols have layered changes, how to use the joint observation of aerosol scattering coefficient and lidar echo to establish a sufficiently constrained and identifiable correction mechanism so that the obtained vertical profile of aerosol scattering coefficient can be consistent with the echo observation across the entire height range, and suppress the systematic deviations introduced by single-point constraints and system uncertainties. Summary of the Invention

[0006] To overcome the aforementioned technical deficiencies, the present invention aims to provide a method and apparatus for correcting the vertical profile of aerosols based on joint observation of aerosol scattering coefficients and lidar. The present invention first identifies aerosol layer boundaries based on preprocessed echoes and constructs a parameterized model of correction factors containing intra-layer scales / gradients. Then, under the constraints of "consistency target of forward echoes at all altitudes" and "consistency constraint of measured scattering coefficient values ​​at at least two different observation altitudes," it jointly solves for the correction factors, lidar system constants, and near-field geometric overlap compensation parameters, thereby achieving stable correction of the vertical profile of the aerosol scattering coefficients.

[0007] This invention discloses a method for correcting the vertical profile of aerosols based on joint observation of aerosol scattering coefficient and lidar, comprising: S1. During the observation period, the echo signal of the lidar to the atmosphere is acquired with a preset integration window, and a range-corrected echo signal is generated for profile processing. S2. Obtain the measured values ​​of aerosol scattering coefficient at at least two different observation heights within the observation period, and map the measured values ​​of aerosol scattering coefficient to the height grid of the distance correction echo signal and perform time alignment. S3. Preprocess the distance-corrected echo signal. The preprocessing includes at least background subtraction, resolution matching, and invalid height removal to obtain the preprocessed echo signal. S4. Generate the initial vertical profile of aerosol scattering coefficients based on the preprocessed echo signal; S5. Perform aerosol layer boundary identification based on the preprocessed echo signal to obtain at least one layer boundary and multiple height layer intervals divided by the layer boundary. S6. Construct a calibration factor parameterization model within each altitude layer interval. The calibration factor parameterization model includes at least intra-layer scale parameters and intra-layer gradient parameters. Furthermore, a geometric overlap compensation parameter is introduced for the low altitude range, and a constant parameter of the lidar system is also introduced. S7. Taking the consistency between the theoretical echo signal obtained by forward calculation of the vertical profile of the candidate aerosol scattering coefficient obtained by the vertical distribution of the candidate correction factor acting on the initial vertical profile of the aerosol scattering coefficient and the preprocessed echo signal in the full height range as the goal, and taking the consistency between the candidate aerosol scattering coefficient at at least two different observation heights and the corresponding measured values ​​of the aerosol scattering coefficient as the constraint, solve the parameters in the parameterization model of the correction factor, the constant parameters of the lidar system, and the geometric overlap compensation parameters to obtain the vertical distribution of the correction factor. S8. Apply the correction factor vertically to the initial vertical profile of the aerosol scattering coefficient, output the corrected vertical profile of the aerosol scattering coefficient, and output the layer boundary information and correction parameter set corresponding to the corrected vertical profile of the aerosol scattering coefficient.

[0008] Preferably, at least two different observation altitudes include a near-ground observation altitude and an upper-altitude observation altitude above the near-ground observation altitude. The upper-altitude observation altitude is provided by a tower, lifting platform, tethered platform, or UAV payload, and the observations are completed within the same preset integration window.

[0009] Preferably, the aerosol layer boundary identification is performed simultaneously based on the multi-scale edge features and polarization features of the preprocessed echo signal, and the layer boundary is only identified as a valid layer boundary when the height deviation of the layer boundary within multiple consecutive preset integration windows is less than a threshold.

[0010] Preferably, the correction factor parameterization model adopts a two-parameter form of intra-layer scale parameter and intra-layer gradient parameter within each height layer interval, and constrains the correction factor to change monotonically or piecewise monotonically with height within the same height layer interval.

[0011] Preferably, the correction factor between adjacent height layer intervals satisfies the interlayer continuity constraint; when the echo abrupt change intensity at the layer boundary exceeds the threshold, the interlayer continuity constraint is released and replaced with the interlayer restricted jump constraint; wherein, the echo abrupt change intensity is determined by the statistical difference of the preprocessed echo signal in the preset height bands adjacent to the layer boundary, and the statistical quantity includes at least the mean or median.

[0012] Preferably, the low altitude range is the altitude segment below the preset upper limit of near-ground altitude; the geometric overlap compensation parameter is obtained by performing a quadratic polynomial parameterization on the geometric overlap function in the low altitude range, and is jointly estimated with the parameter of the correction factor parameterization model during the solution process.

[0013] Preferably, the constant parameters of the lidar system are estimated by automatically selecting a clean reference height segment, which is a height segment that satisfies the conditions that "the echo gradient is below a threshold, the polarization characteristics are stable, and the rate of change of the initial vertical profile of the aerosol scattering coefficient is below a threshold".

[0014] Preferably, a robust loss function is used in combination with an iterative reweighting strategy to reduce the impact of the cloud base, strong reflective targets, or low signal-to-noise ratio altitude ranges on the overall altitude consistency target.

[0015] Preferably, cross-validation constraints are introduced during the solution process: at least one observation height is retained as a validation height among at least two different observation heights. The correction parameters are first solved based on the remaining observation heights, and then the deviation between the candidate aerosol scattering coefficient at the validation height and the corresponding measured aerosol scattering coefficient is verified. When the deviation exceeds the threshold, re-stratification or re-solution is triggered.

[0016] Preferably, when the observation wavelength of the measured aerosol scattering coefficient is inconsistent with the working wavelength of the lidar, the method converts the measured aerosol scattering coefficient to the working wavelength of the lidar based on a preset wavelength index model before solving, and the wavelength index is determined by multi-wavelength scattering observations at the same site or historical calibration data.

[0017] Preferably, when the measured value of the aerosol scattering coefficient is the measured value of the scattering coefficient under dry conditions, the method performs a humidity growth correction on the measured value of the scattering coefficient based on the relative humidity profile before solving. The humidity growth correction adopts a piecewise parameterized humidity growth curve and the parameters are determined by historical comparison test data at the same site.

[0018] Preferably, generating the initial vertical profile of the aerosol scattering coefficient includes: first, obtaining the backscattering correlation profile by inverting the preprocessed echo signal; and then, selecting the conversion parameter between scattering and backscattering from a preset type parameter table based on the aerosol type determined by polarization characteristics, in order to generate the initial vertical profile of the aerosol scattering coefficient.

[0019] Preferably, invalid height removal includes: generating a quality mark for each height, the quality mark being generated based at least on a signal-to-noise ratio threshold and an echo mutation threshold; when the lidar includes at least one polarization echo channel, the quality mark is also generated based on a polarization consistency threshold; and the marked heights are removed or constrained interpolation is performed during the solution process.

[0020] Preferably, the method further includes: performing sensor drift consistency verification on the measured values ​​of aerosol scattering coefficient, the consistency verification being obtained by comparing the variation trend of the measured values ​​of aerosol scattering coefficient at different observation heights with the intra-layer variation trend of the preprocessed echo signal; when drift is determined to exist, implementing an alternative strategy for the measured values ​​of aerosol scattering coefficient at the corresponding observation height, the alternative strategy including: using the statistical fusion value of the main channel and the co-located backup observation channel when there is a co-located backup observation channel, otherwise using the interpolated value of the adjacent observation height.

[0021] Preferably, the set of correction parameters includes at least the set of layer boundaries, the intra-layer scale parameters and intra-layer gradient parameters of each height layer interval, the geometric overlap compensation parameters, and the lidar system.

[0022] In view of this, the present invention also provides an aerosol vertical profile correction device for joint observation of aerosol scattering coefficient and lidar, characterized in that it includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described above.

[0023] Compared with existing technologies, the above technical solution has the following advantages: 1. The calibration results have stronger consistency and higher reliability across the entire altitude range: By jointly constraining the measured value of the aerosol scattering coefficient with the preprocessed echo signal of the lidar, and solving the problem with forward consistency across the entire altitude range as the objective, the vertical profile of the output aerosol scattering coefficient is made to maintain self-consistency with the echo observation across the entire altitude range, reducing the overall deviation caused by traditional single-point constraints.

[0024] 2. Better adaptability to stratification abrupt change scenarios: By identifying aerosol stratification boundaries and confirming cross-window stability, and establishing a stratification parameterized correction factor model based on stratification, it can correct for different height stratification intervals separately, and allows restricted inter-layer jumps when the abrupt change intensity is large, thereby improving the correction stability and effectiveness under strong stratification conditions such as boundary layer and transport layer.

[0025] 3. Enhanced near-ground error suppression capability: By introducing geometric overlap compensation parameters in the low-altitude range and estimating them together with correction factor parameters and lidar system constant parameters, the upward propagation of systematic errors caused by insufficient near-field geometric overlap is effectively reduced, thereby improving the reliability of the near-ground profile.

[0026] 4. More robust to system constant uncertainty and drift: The lidar system constant parameters are included in the joint solution and can be constrained based on the clean reference height segment, which reduces the impact of system constant uncertainty or drift on the inversion profile and improves stability in long-term operation scenarios.

[0027] 5. Improved anomaly and noise resistance, and more stable continuous operation: Invalid height is eliminated or restricted interpolation is performed by quality labeling, and a robust loss function and iterative reweighting strategy are adopted to reduce the impact of abnormal height segments. At the same time, a closed-loop mechanism for triggering re-layering or re-solution by cross-validation is introduced, so that the method still has high convergence stability under conditions of low signal-to-noise ratio, cloud bottom interference or strong reflective target interference.

[0028] 6. Enhanced engineering feasibility under multi-source observation conditions: Supports near-ground and high-altitude multi-altitude scattering coefficient observation (high tower, lifting platform, tethered platform or UAV payload, etc.), and provides optional processing such as wavelength conversion and humidity growth correction, so that data from different equipment and under different observation conditions can be unified into the same solution framework, which is convenient for engineering deployment and promotion.

[0029] 7. Results are traceable and easy to operate and evaluate: Output and associate the layer boundary information and the set of correction parameters, so that the vertical profile of the corrected aerosol scattering coefficient is traceable and reproducible, which facilitates subsequent quality control, long-term performance evaluation and model maintenance. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the preprocessed echo curve and layer boundary; Figure 2 A schematic diagram of the vertical distribution parameterization of the correction factor; Figure 3 A comparative diagram verifying high consistency; Figure 4 This is a schematic diagram of the steps of an aerosol vertical profile correction method and device based on joint observation of aerosol scattering coefficient and lidar according to the present invention. Detailed Implementation

[0031] The advantages of the present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments.

[0032] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0033] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0034] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0035] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0036] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0037] In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the convenience of the description of the invention and have no specific meaning in themselves. Therefore, "module" and "part" can be used interchangeably.

[0038] See Figure 4 As shown, this embodiment provides a method for aerosol vertical profile correction based on joint observation of aerosol scattering coefficient and lidar, including: S1, acquiring the lidar echo signal of the atmosphere within a preset integration window during the observation period, and generating a range-corrected echo signal for profile processing; S2, acquiring measured values ​​of aerosol scattering coefficient at at least two different observation heights during the observation period, mapping the measured values ​​of aerosol scattering coefficient to the height grid of the range-corrected echo signal, and performing time alignment; S3, preprocessing the range-corrected echo signal, including at least background subtraction, resolution matching, and invalid height removal, to obtain a preprocessed echo signal; S4, generating an initial vertical profile of aerosol scattering coefficient based on the preprocessed echo signal; S5, performing aerosol layer boundary identification based on the preprocessed echo signal to obtain at least one layer boundary and multiple height layer intervals divided by the layer boundary; S6, constructing a correction factor parameterization within each height layer interval. The model, the parameterized model of the correction factor, includes at least intra-layer scale parameters and intra-layer gradient parameters, and further introduces geometric overlap compensation parameters for the low-altitude range, while also introducing lidar system constant parameters; S7, with the goal of ensuring consistency between the theoretical echo signal and the preprocessed echo signal obtained by forward calculation of the vertical profile of the candidate aerosol scattering coefficient obtained by the vertical distribution of the candidate correction factor acting on the initial vertical profile of the aerosol scattering coefficient across the entire altitude range, and with the constraint that the candidate aerosol scattering coefficient at at least two different observation altitudes is consistent with the corresponding measured values ​​of the aerosol scattering coefficient, the parameters in the parameterized model of the correction factor, the lidar system constant parameters, and the geometric overlap compensation parameters are solved to obtain the vertical distribution of the correction factor; S8, the vertical distribution of the correction factor is applied to the initial vertical profile of the aerosol scattering coefficient, the corrected vertical profile of the aerosol scattering coefficient is output, and the layer boundary information and correction parameter set corresponding to the corrected vertical profile of the aerosol scattering coefficient are output.

[0039] This embodiment illustrates the implementation of an aerosol vertical profile correction method based on joint observation of aerosol scattering coefficient and lidar in a ground-based observation scenario. The implementation system includes a ground-based lidar observation device, an aerosol scattering coefficient observation device, and a data processing device. The ground-based lidar observation device emits laser pulses into the atmosphere and receives echoes during the observation period, outputting profile data showing the echo variation with distance or altitude. In one implementation, the ground-based lidar observation device further includes at least one polarization echo channel to output polarization-dependent profile data. The aerosol scattering coefficient observation device outputs measured aerosol scattering coefficient values ​​at at least two different observation altitudes, including at least a near-ground observation altitude and a higher-altitude observation altitude. The higher-altitude observation altitude can be provided by a tower, lifting platform, tethered platform, or UAV payload. The data processing device unifies the statistical caliber of the two types of observation data in the time and altitude dimensions, and completes preprocessing, hierarchical identification, parametric modeling, joint solution, and output storage.

[0040] To ensure that terms such as "accumulation," "average," "alignment," "threshold," and "stability" have clear engineering implications, this embodiment stipulates the following data organization method: Within each preset integration window, the ground-based lidar observation device generates multiple pulse transmission and reception processes. Each pulse reception yields an echo profile, which consists of multiple sampling points in the height direction. These sampling points can correspond to a range gate or a height gate. The data processing device uniformly maps these sampling points to height grid points. The height grid is generated by the data processing device from near-ground to the highest effective height according to a preset height step size. This preset height step size can be equal to the original height resolution or an integer multiple of the original height resolution, thus facilitating resampling and unified calculation. The data processing device also establishes a window identifier and a window timestamp for each preset integration window and performs "intra-window aggregation" on various types of data within the window to ensure that data from different sources have the same statistical meaning within the same window.

[0041] In addition, to facilitate the implementation of the "consistency" criterion, the data processing device maintains the following auxiliary quantities within each preset integration window: first, an effective height set, used to record the height grid points after removing invalid heights; second, a quality weight, used to represent the reliability of the data at each height point; third, a robust weight, used to suppress the influence of abnormal residual height points on the target during the solution process; and fourth, a threshold set, used to store various threshold calibers under the current site configuration. The threshold set includes at least a signal-to-noise ratio threshold, a boundary stability confirmation threshold, an echo mutation intensity threshold, a consistency deviation threshold, and a convergence criterion threshold. Each threshold can be obtained from equipment noise statistics, calibration uncertainty, or historical stability process statistics and is allowed to be updated with the season or equipment status.

[0042] In this embodiment, step S1 will be described in detail.

[0043] During the observation period, the ground-based lidar observation device continuously emits laser pulses and receives echoes. Each pulse reception yields a raw echo profile, which consists of multiple sampling points in the height direction. The data processing device organizes the raw echo profiles into a preset integration window. That is, within each preset integration window, all raw echo profiles within the window's time range are collected, and in-window aggregation is performed on the raw echo profiles within the window to obtain an echo profile representing that window.

[0044] The "accumulation" in this embodiment includes at least one of the following specific methods: The data processing device first performs the same height grid alignment on each original echo profile within the window, that is, maps the sampling points of each profile to a unified height grid point according to the height. Then, at each height grid point, the echo intensity of all profiles within the window at that height point is summed point by point to obtain the accumulated echo intensity at that height point, thereby forming a "window-in-accumulation echo profile". To avoid occasional spikes causing unreasonable amplification of the summation, the data processing device can perform outlier removal on the echo intensity of each profile within the window at that height point before point-by-point summation. The outlier removal includes at least: first calculating the median and dispersion of the echo intensity within the window at that height point, and then considering samples that deviate from the median by more than a preset multiple of dispersion as outliers and not participating in the summation. Through the above point-by-point summation method, the window-in-accumulation echo profile can improve the stability under random noise conditions and provide a basis for subsequent averaging or robust statistics.

[0045] In this embodiment, the "average" includes at least one of the following specific methods: After obtaining the accumulated echo profile within the window, the data processing device performs point-by-point division on the accumulated echo intensity using the number of valid profiles participating in the accumulation within the window, thereby obtaining the "arithmetic mean echo profile within the window." This average echo profile represents the expected level of echo intensity within the window. If the number of profiles within the window changes over time, the data processing device uses the actual number of profiles participating in the statistics as the divisor to ensure that the average value has a consistent caliber. To further enhance the anomaly resistance, the data processing device can also use the "median average within the window" instead of the arithmetic mean. That is, for each height point, the echo intensity of that height point within the window is sorted, and the median value is taken as the representative value of that height point, thereby forming the "median echo profile within the window." This median echo profile is more stable when there are a few strong anomalies. The data processing device can select either the arithmetic mean or the median mean as the representative echo profile of the window according to the site configuration, and maintain consistency under the same operating configuration at the same site to ensure the stability of the subsequent threshold caliber.

[0046] After obtaining the window-representative echo profile, the data processing device performs range-correlation correction to generate a range-corrected echo signal. Specifically, the data processing device multiplies the window-representative echo profile at each elevation point by a range square factor corresponding to that elevation, thereby compensating for echo attenuation with distance caused by propagation geometry diffusion, making the echo profile more suitable for profile processing. The data processing device outputs the range-corrected echo signal and records the window's timestamp, window duration, number of pulse profiles participating in aggregation, and elevation grid information. If the ground-based lidar observation device includes a polarization echo channel, the data processing device performs the same in-window aggregation and range-correlation correction on the polarization channel echo profile as described above, thereby obtaining the range-corrected echo signal for the polarization channel for subsequent calculation of polarization characteristics.

[0047] In this embodiment, step S2 will be described in detail.

[0048] During the observation period, the aerosol scattering coefficient observation device outputs measured aerosol scattering coefficient values ​​at at least two different observation altitudes, including near-ground observation altitudes and high-altitude observation altitudes above near-ground observation altitudes. To ensure consistency with the preset integration window, the data processing device performs "intra-window aggregation" on the measured scattering coefficient values. Specifically, for a certain observation altitude, if the scattering coefficient observation device at that altitude outputs multiple sample values ​​within a preset integration window, the data processing device aggregates these multiple sample values ​​into a single representative value for that window. The aggregation method includes at least the window mean, the window midpoint, or the window truncated mean; the window truncated mean means that the sample values ​​within the window are sorted, a small number of extreme values ​​at both ends are removed, and then the mean is calculated, thereby suppressing occasional anomalies. If the scattering coefficient observation device at that altitude outputs only one sample value within a window, that sample value is directly used as the representative value for that window. Through the above intra-window aggregation, the measured scattering coefficient values ​​and echo data have the same statistical meaning within the same window, thus achieving time alignment.

[0049] The "height correspondence" is implemented in this embodiment as follows: The data processing device obtains the nominal height value of each scattering coefficient observation height and maps the nominal height value to the nearest grid point in the height grid; when the nominal height falls between two adjacent height grid points, the data processing device maps the representative value of the scattering coefficient window to the equivalent value on the corresponding grid point using linear interpolation, and performs amplitude limiting processing on the interpolation result. The amplitude limiting processing includes at least ensuring that the interpolated value is within a reasonable range of the corresponding interpolation endpoint values ​​of adjacent height points, thereby avoiding unreasonable jumps in the interpolation. Based on this, the data processing device obtains the correspondence between "measured scattering coefficient value - height grid point - window identifier" and associates it one by one with the distance correction echo signal output in step S1 according to the window identifier.

[0050] When the observed wavelength of the measured scattering coefficient is inconsistent with the operating wavelength of the lidar, the data processing device performs wavelength conversion on the measured scattering coefficient before proceeding to the solution. In this embodiment, the wavelength conversion is achieved through a wavelength index model. Specifically, the data processing device obtains the wavelength index from co-located multi-wavelength scattering observations or historical calibration data, and uses this wavelength index to convert the measured scattering coefficient to the lidar's operating wavelength within the same window. Furthermore, the wavelength index can be grouped and maintained according to season or pollution type to reduce system bias.

[0051] When the measured scattering coefficient is under dry conditions, the data processing device performs a humidity growth correction on the measured scattering coefficient before proceeding to the solution. In this embodiment, the humidity growth correction is achieved through a piecewise parameterized humidity growth curve. The data processing device first acquires the relative humidity profile and interpolates it to a height grid. Then, based on the piecewise intervals into which the relative humidity at each height point falls, it selects the corresponding piecewise parameters to correct the measured scattering coefficient, ensuring that the corrected measured scattering coefficient reflects the humidity state consistent with that observed by the lidar. The relative humidity profile can originate from co-located sounding, assimilation products, or tower sensors, and is aligned with a preset integration window using an in-window aggregation method.

[0052] In this embodiment, step S3 will be described in detail.

[0053] The data processing device preprocesses the distance-corrected echo signal, and the preprocessing includes at least background subtraction, resolution matching, and invalid height removal to obtain a preprocessed echo signal.

[0054] The "background subtraction" is implemented as follows: The data processing device pre-sets a high-altitude background height range, which is selected within a long-term statistically significant range where echoes are primarily dominated by noise. For each preset integration window, the data processing device takes multiple height point samples of the distance-corrected echo signal within this high-altitude background height range, and calculates the background quantity and background fluctuation quantity. The background quantity is calculated using at least the mean or median, and the background fluctuation quantity is calculated using at least the standard deviation or median absolute deviation. The data processing device subtracts the background quantity from each height point of the distance-corrected echo signal to obtain the background-subtracted echo, and performs non-negative clipping on the negative values ​​resulting from the subtraction to maintain the physical reasonableness of the echo intensity. The background fluctuation quantity is used for subsequent signal-to-noise ratio calculations and threshold settings.

[0055] The "resolution matching" is implemented as follows: When the original height interval of the distance-corrected echo signal is inconsistent with the height grid step size used in this embodiment, the data processing device resamples the distance-corrected echo signal to the height grid. During resampling, the data processing device finds the two adjacent original height points of each target height grid point in the original height coordinates, uses linear interpolation to obtain the echo intensity of the target height point, and applies amplitude limiting to the interpolation result to prevent the interpolation from crossing abnormal points and causing spikes. If the site configuration requires further suppression of high-frequency noise, the data processing device performs height-direction smoothing on the echo profile after resampling. The smoothing method includes at least moving average or moving median. The width of the sliding window is set according to the height resolution and noise level and fixed in the site configuration to ensure consistency.

[0056] The "invalid altitude rejection" is implemented as follows: The data processing device generates a quality label for each altitude point and determines whether to reject, retain, or restrict interpolation accordingly. The quality label is generated based at least on a signal-to-noise ratio (SNR) threshold and an echo abrupt change threshold. In this embodiment, the SNR is calculated as follows: The data processing device uses the echo intensity after background subtraction as the signal quantity and the background fluctuation as the noise quantity, and calculates the ratio of the signal quantity to the noise quantity at each altitude point; when the ratio is lower than the SNR threshold, the altitude point is marked as a low-confidence altitude. The echo abrupt change threshold is used to identify strongly reflective targets, cloud bases, or anomalous peaks, and its calculation method includes at least one of the following: The data processing device calculates the absolute value of the difference in echo intensity between adjacent altitude points or calculates the abrupt change amplitude of the echo median within a fixed altitude band; when the abrupt change amplitude exceeds the abrupt change threshold, the altitude point or altitude band is marked as an anomalous altitude. When the ground-based lidar observation device includes a polarization echo channel, the quality marker is also generated based on the polarization consistency threshold. In this embodiment, polarization consistency is evaluated according to the following criteria: the data processing device calculates the change of polarization characteristics in adjacent time windows at the same height point or calculates the change of polarization characteristics in adjacent height points within the same window. When the change exceeds the polarization consistency threshold, it is marked as a low confidence height to suppress misjudgments caused by polarization channel anomalies.

[0057] For height points marked as low confidence or anomalous, the data processing device can perform either elimination or restricted interpolation. Elimination involves resetting the weight of the height point to zero in subsequent consistency evaluation and solution. Restricted interpolation involves finding the two nearest valid height points above and below the current height point in the height direction, using linear interpolation to fill the echo intensity of the current height point, and applying amplitude limiting to the interpolated value to ensure it falls within a reasonable range of the echo intensity of the above and below valid points. Simultaneously, a lower weight is assigned to the interpolated height point in the solution. Based on this, the data processing device generates a preprocessed echo signal and its corresponding set of valid heights and quality weights. Figure 1 The solid curve in the figure is used to illustrate the profile of the preprocessed echo signal.

[0058] In this embodiment, step S4 will be described in detail.

[0059] The data processing device generates an initial vertical profile of the aerosol scattering coefficient based on the preprocessed echo signal. Specifically, the data processing device first constructs the initial conditions and constraints required for inversion based on the preprocessed echo signal and outputs a "backscattering correlation profile." In this embodiment, the backscattering correlation profile represents a "height profile quantity monotonically correlated with backscattering intensity," and its acquisition method includes at least one of the following: the data processing device normalizes the preprocessed echo signal along the height direction and combines it with a molecular scattering reference profile or an empirical baseline profile to form a backscattering correlation profile, so that the profile can reflect the relative change trend of backscattering with height. The purpose of this step is to obtain an initial profile with a reasonable shape that can be used for conversion, rather than directly replacing the final correction result.

[0060] After obtaining the backscattering correlation profile, the data processing device converts it into an initial vertical profile of the aerosol scattering coefficient. To improve the consistency of the conversion, this embodiment adopts a type-driven parameter conversion mechanism: when a polarization channel exists, the data processing device calculates the polarization characteristics and determines the aerosol type accordingly, then selects a set of conversion parameters between scattering and backscattering from a preset type parameter table to complete the conversion. In this embodiment, the preset type parameter table is organized in the form of "polarization index interval - type label - conversion parameter set", where the polarization index interval is used to divide the typical polarization characteristic range of different aerosol types, the type label is used to identify the category, and the conversion parameter set is used to describe the conversion relationship from the backscattering correlation profile to the scattering coefficient profile. The calculation scope of the polarization index includes at least: taking the ratio or difference ratio of the polarization channel profile and the total echo profile at the same height point to obtain the polarization characterization quantity, and aggregating the characterization quantity within a window to suppress noise; when the polarization index falls into a certain interval, the conversion parameter set corresponding to that interval is selected. If the polarization channel is unavailable, the data processing device selects the default type of conversion parameter set and suppresses the influence of type error in subsequent solutions through robust weighting and cross-validation mechanisms.

[0061] To facilitate the subsequent use of the "rate of change is below the threshold" criterion, the data processing device calculates the rate of change of height after generating the initial vertical profile. In this embodiment, the rate of change is obtained by dividing the difference between adjacent height grid points by the height step size, and the difference result can be smoothed to reduce noise.

[0062] In this embodiment, step S5 will be described in detail.

[0063] The data processing device performs aerosol layer boundary identification based on the preprocessed echo signal, obtaining at least one layer boundary and dividing it into multiple height layer intervals. In this embodiment, layer boundary identification is achieved using a "candidate boundary detection + cross-window stability confirmation" approach. Candidate boundary detection specifically includes: the data processing device calculates first-order and second-order differences in the height direction of the preprocessed echo signal to obtain the slope and curvature changes of the echo with height, and repeatedly calculates the differences at multiple height scales; that is, the echo profile is first processed at different smoothing scales before calculating the differences, thus forming multi-scale difference features. When a certain height point simultaneously exhibits a local maximum of the absolute value of the difference or a significant curvature change at multiple scales, it is recorded as a candidate boundary height. If a polarization channel exists, the data processing device also performs the same difference detection on the polarization feature profile and performs height proximity matching between the echo candidate boundary and the polarization candidate boundary. Only when the two types of candidate boundaries are close to each other in height is their candidate confidence increased, thereby reducing false boundaries caused solely by noise.

[0064] Stability confirmation specifically includes: the data processing device tracks the height of candidate boundaries within multiple consecutive preset integration windows. If the height deviation of the same candidate boundary within each consecutive window is less than the stability confirmation threshold, it is confirmed as a valid layer boundary. In this embodiment, the height deviation is calculated as the absolute value of the difference between the heights of adjacent window boundaries. The number of consecutive windows and the stability confirmation threshold can be determined by site experience or historical statistics. The data processing device divides the height direction into multiple height layer intervals based on the valid layer boundaries and outputs the layer boundary heights for use in [further processing]. Figure 1 The dashed line markings in the text.

[0065] To facilitate subsequent switching of inter-layer constraints, the data processing device calculates the echo abrupt change intensity at the layer boundary. In this embodiment, the echo abrupt change intensity is calculated as follows: A preset height band is selected above and below the layer boundary height, centered on the boundary height. This preset height band can be defined by a fixed height and width or a fixed number of grid points. The data processing device calculates the statistics of the pre-processed echo signals within the upper and lower height bands respectively, and takes the absolute value of the difference between the two as the abrupt change intensity. The statistics include at least the mean or median. To enhance robustness, the median is preferably used as the statistics, and points marked as low confidence within the height band can be removed before calculation. The abrupt change intensity is compared with an abrupt change intensity threshold to determine the switching between subsequent inter-layer continuity constraints and inter-layer restricted jump constraints.

[0066] In this embodiment, step S6 will be described in detail.

[0067] The data processing device constructs a parameterized model of the correction factor within each altitude layer interval. This correction factor is used to correct the systematic deviation between the initial vertical profile of the aerosol scattering coefficient and the true scattering coefficient profile. To reduce the degrees of freedom and make the model identifiable, this embodiment uses a two-parameter form to describe the correction factor within each altitude layer interval: an intra-layer scale parameter to represent the overall amplitude correction of that layer, and an intra-layer gradient parameter to represent the linear trend of that layer with altitude. In its specific implementation, the data processing device calculates the correction factor value for each altitude grid point: first, it determines which altitude layer interval the altitude point belongs to, and then uses the intra-layer scale parameter and intra-layer gradient parameter of that layer to calculate the correction factor for that altitude point, thereby obtaining the vertical distribution of candidate correction factors for the complete altitude range.

[0068] To avoid unreasonable oscillations within layers, the data processing device applies monotonic or piecewise monotonic variation constraints to the correction factor within the same height layer interval. In this embodiment, monotonic variation constraints are implemented by limiting the sign of the gradient parameters within the layer or by limiting the magnitude relationship of the correction factor at the endpoints of the layer interval, allowing the correction factor to only increase or decrease overall within that layer interval. Piecewise monotonic variation constraints are implemented in this embodiment by setting a finite number of sub-segments within the same layer interval and applying monotonicity to each sub-segment separately. The number of sub-segments is fixed by the station configuration, thus allowing a finite number of trend changes within the layer without significantly increasing the degrees of freedom.

[0069] Regarding interlayer connectivity, the data processing device applies interlayer continuity constraints by default, ensuring that the correction factor values ​​between adjacent layers at the layer boundary are equal or the difference does not exceed the continuity tolerance. When the mutation intensity calculated in step S5 exceeds the mutation intensity threshold, the data processing device removes the continuity constraints and replaces them with interlayer restricted jump constraints. In this embodiment, the restricted jump constraints include at least limiting the difference between the correction factors above and below the boundary to not exceed a preset jump upper limit, thereby allowing interlayer discontinuity but avoiding unreasonable large jumps. Figure 2 This is used to illustrate the form of the correction factor that is parameterized by layer and has interlayer connections.

[0070] Regarding the low-altitude range, the data processing device defines the low-altitude range as the altitude segment below the preset upper limit of near-ground altitude, used to characterize the systematic deviation caused by insufficient near-field geometric overlap. The data processing device introduces geometric overlap compensation parameters for this low-altitude range, which are obtained by parameterizing the geometric overlap function in the form of a quadratic polynomial. To avoid introducing complex symbols, this embodiment defines the quadratic polynomial form as "a height function composed of a constant term, a first-order term, and a quadratic term." The data processing device outputs the overlap compensation amount at each altitude point using this function within the low-altitude range and uses it as one of the compensation factors when forward-calculating the theoretical echo signal. The coefficients corresponding to the constant term, the first-order term, and the quadratic term are the geometric overlap compensation parameters. These parameters are jointly estimated with the correction factor parameters during the solution process in step S7, thereby reducing the dependence on fixed empirical overlap curves.

[0071] At the system level, the data processing device introduces constant parameters of the lidar system and incorporates them into the set of parameters to be estimated. This is used to absorb the overall scale deviation caused by system gain, optical efficiency, and calibration errors, so that the consistency target between the theoretical echo signal and the preprocessed echo signal can be self-consistently satisfied at the system level.

[0072] In this embodiment, step S7 will be described in detail.

[0073] The data processing device aims to obtain the vertical profile of the candidate aerosol scattering coefficient by applying the vertical distribution of the candidate correction factor to the initial vertical profile of the aerosol scattering coefficient, and to calculate the theoretical echo signal from the vertical profile of the candidate aerosol scattering coefficient in the forward direction, with the theoretical echo signal being consistent with the preprocessed echo signal across the entire altitude range. At the same time, it is constrained to ensure that the candidate aerosol scattering coefficients at at least two different observation altitudes are consistent with the corresponding measured values ​​of the aerosol scattering coefficients. The device solves for the parameters of the parameterized model of the correction factor, the constant parameters of the lidar system, and the geometric overlap compensation parameters to obtain the vertical distribution of the correction factor.

[0074] To make "consistency / unity" feasible, this embodiment defines the consistency criteria as follows: First, the consistency constraint for multi-altitude scattering coefficients is implemented by "deviation not exceeding a threshold." That is, for each observation altitude, the data processing device calculates the absolute value of the difference between the candidate aerosol scattering coefficient at that altitude and the measured scattering coefficient, and requires that this absolute value does not exceed the scattering coefficient consistency threshold. The threshold can be determined by the calibration uncertainty of the scattering coefficient observation device, the fluctuation amplitude within the window, or the historical repeatability statistics, and can be set separately for each altitude. Second, the overall altitude consistency target is implemented by "weighted residuals not exceeding a threshold or satisfying the convergence criterion." That is, the data processing device calculates the difference between the theoretical echo signal and the preprocessed echo signal point by point within the effective altitude set, and performs a weighted summation on the difference to obtain the residual value. The weighted summation at least includes multiplying the absolute value or square of the difference by the altitude weight and summing them. The altitude weight is jointly determined by the quality weight in step S3 and the robust weight described later, thereby automatically reducing the contribution of low-confidence altitude points or abnormal residual altitude points to the residual. The residual value is compared with the full-height consistency threshold to determine whether the consistency requirement has been met, or it can be used as the target value for numerical optimization to minimize.

[0075] The data processing device performs the solution according to the following reproducible process. First, parameter initialization is performed. The data processing device initializes the intra-layer scale parameters and intra-layer gradient parameters for each height layer interval. The initialization method includes at least the following: using the height point where the measured scattering coefficient value is located as the anchor point, making the initial value of the correction factor of the initial profile near that height close to the proportion that matches the two, and setting the initial value of the intra-layer gradient parameter to zero or to a gradually varying trend estimated from the initial profile and the preprocessed echo morphology. The data processing device simultaneously initializes the geometric overlap compensation parameters and the system constant parameters. The initial value of the geometric overlap compensation parameters can be set as a smooth function that makes the compensation amount in the low height range close to one, and the initial value of the system constant parameters can be obtained from coarse matching of historical calibration or clean reference height segments. Then, an iterative loop is entered. In each iteration, the data processing device first generates a candidate correction factor vertical distribution using the current layer parameters, and applies this correction factor vertical distribution point by point to the initial vertical profile of the aerosol scattering coefficient to obtain the candidate aerosol scattering coefficient vertical profile. The data processing device then inputs the candidate aerosol scattering coefficient vertical profile into the lidar forward calculation module to obtain the theoretical echo signal. The forward calculation module includes at least combining the scattering coefficient profile with system constant parameters, geometric overlap compensation, and molecular scattering reference profile, outputting the theoretical echo intensity at each height point, and using the geometric overlap compensation as a multiplicative or equivalent compensation factor in the calculation within the low-height range. After completing the theoretical echo signal calculation, the data processing device calculates the difference between the theoretical echo and the preprocessed echo within the effective height set to form a residual profile, and performs robust weight updates on the residual profile.

[0076] In this embodiment, robust weight updates are implemented as follows: The data processing device compares the absolute value of the residual with a residual threshold at each altitude point. When the absolute value of the residual does not exceed the residual threshold, a higher robust weight is assigned to that altitude point. When the absolute value of the residual exceeds the residual threshold, the weight is reduced by the ratio of the residual threshold to the absolute value of the residual. This automatically reduces the impact on the target when abnormal residuals occur near the cloud base, near strongly reflective targets, or in low signal-to-noise ratio altitude ranges. Subsequently, the data processing device multiplies the robust weight with the quality weight from step S3 to obtain the final altitude weight, and uses this final altitude weight to calculate the weighted residual value.

[0077] In this embodiment, parameter updates employ an "alternating update + constraint projection" approach to balance simplicity and stability. Alternating updates specifically include: with fixed system constant parameters and geometric overlap compensation parameters, the data processing device updates the intra-layer scale parameters and intra-layer gradient parameters for each height layer interval. The update aims to reduce the weighted residual and satisfy the consistency constraint of multi-height scattering coefficients. If the update result violates the intra-layer monotonic constraint or the inter-layer continuous / restricted jump constraint, the update result is projected back to the constraint feasible region. Specifically, amplitude limiting or sign correction is applied to the intra-layer gradient parameters that violate the constraint, and boundary differences that violate the jump upper limit are truncated to bring them back to the allowable range. After completing the intra-layer parameter update, the data processing device updates the system constant parameters and geometric overlap compensation parameters while keeping the intra-layer parameters fixed, further reducing the weighted residual and improving consistency in the low-height range. This alternating update process is repeated until the convergence criterion is met.

[0078] The convergence criterion in this embodiment includes at least one of the following conditions: the change in the weighted residual value between two consecutive iterations is less than the convergence threshold; or the parameter update amount between two consecutive iterations is less than the parameter change threshold; or the number of iterations reaches a preset upper limit. To enhance the stability of the system constant parameters, the data processing device uses a clean reference height segment to anchor and estimate the system constant parameters during iteration. The clean reference height segment is determined as follows: the data processing device calculates the echo gradient, polarization stability, and initial profile change rate within the candidate height range. The echo gradient is calculated by dividing the echo difference between adjacent height points by the absolute value of the height step size; the initial profile change rate is calculated by dividing the initial profile difference between adjacent height points by the absolute value of the height step size; and the polarization stability is evaluated by the variance or maximum change of the polarization characteristics within the height segment. When all three indicators are below their respective thresholds, the height segment is selected as the clean reference height segment. Within this segment, the data processing device prioritizes matching theoretical echoes with preprocessed echoes to reduce the coupling drift between the system constant and the correction factor.

[0079] To improve the reliability of the solution, the data processing device introduces cross-validation constraints during the solution process. In this embodiment, cross-validation constraints are implemented as follows: at least one observation height is selected from at least two observation heights as a validation height. The measured scattering coefficient value at this validation height is not included in the hard constraints during parameter updates, but is only used for consistency checks after each or several iterations. The data processing device calculates the absolute value of the deviation between the candidate aerosol scattering coefficient at the validation height and the measured scattering coefficient at the validation height, and compares it with a validation threshold. When the deviation exceeds the validation threshold, a backoff mechanism is triggered. The backoff mechanism includes at least re-stratification or re-solving. Re-stratification obtains a new stratification boundary by adjusting the boundary stability confirmation threshold, multi-scale differential scale, or abrupt change intensity threshold. Re-solving restarts the iteration by resetting the initial parameter values, adjusting the residual threshold, or adjusting the upper limit of the low-height range, thus forming a closed-loop control of "solution-validation-backoff". Figure 3 This is a display method used to illustrate and verify a high degree of consistency.

[0080] To prevent abnormal measured values ​​of the scattering coefficient from violating constraints, the data processing device can also perform sensor drift consistency verification. This verification is implemented as follows: the data processing device generates a time series of representative values ​​for the scattering coefficient window at each observation height and calculates the trend of this time series. The trend can be obtained by differencing adjacent windows or by using a sliding regression slope. Simultaneously, the data processing device calculates the intra-layer trend of the preprocessed echo signal within the corresponding height layer interval and compares it with the scattering coefficient trend. If the scattering coefficient trend is significantly inconsistent with the intra-layer trend of the echo over a long period and exceeds the drift threshold, it is determined that the scattering coefficient at that observation height has drifted. After determining drift, the data processing device implements a substitution strategy for the scattering coefficient at that height. The substitution strategy includes: when a co-located backup observation channel exists, statistically fusing the main channel and the backup channel within the window to obtain a substitution value; otherwise, using the interpolated value from adjacent observation heights as the substitution value, and assigning a reduced weight to the substitution value in the solution to reduce the impact of substitution uncertainty.

[0081] In this embodiment, step S8 will be described in detail.

[0082] After completing step S7 and satisfying the convergence criterion, the data processing device applies the correction factor vertically distributed point-by-point to the initial vertical profile of the aerosol scattering coefficient, obtaining the corrected vertical profile of the aerosol scattering coefficient, and outputs the corrected vertical profile of the aerosol scattering coefficient. The data processing device simultaneously outputs the layer boundary information and correction parameter set corresponding to this profile. The correction parameter set includes at least the layer boundary set, intra-layer scale parameters and intra-layer gradient parameters for each height layer interval, geometric overlap compensation parameters, and lidar system constant parameters. For ease of reproduction and auditing, the data processing device also stores the following information: preset integration window identifier, window timestamp, number of pulse profiles participating in aggregation within the window, effective height set, quality weight, robustness weight, version number of the threshold set used, cross-validation results, and backoff trigger records. The data processing device associates and stores the above parameter set and the corrected profile according to the window identifier and timestamp, so that the output result of any window can be traced back to the corresponding input data, processing caliber, and solution process configuration.

[0083] It should be noted that in this embodiment, the following will be applied: Figure 1 , Figure 2 , Figure 3 A detailed explanation will be provided.

[0084] Figure 1 The horizontal axis represents the relative intensity of the preprocessed echo, which indicates the relative dimension of the preprocessed echo signal intensity obtained after in-window aggregation of multi-pulse echo profiles within the same preset integration window, and after background subtraction, resolution matching, and invalid height removal. Figure 1 The vertical axis represents the height. Figure 1 The solid curve in the figure is used to express the overall shape of the preprocessed echo signal in the height direction, and the horizontal dashed line is used to express the location of the layer boundary. Figure 1 This is primarily used to illustrate that the preprocessed echo profile can stably exhibit a morphology of "overall attenuation + local structural changes," where the local structural changes correspond to aerosol layering or transition layers; and the layer boundaries are obtained from this preprocessed echo profile through feature extraction and stability confirmation. Figure 1 The output forms corresponding to steps S3 and S5.

[0085] Figure 2 The horizontal axis represents the relative value of the correction factor, and the vertical axis represents the height. Figure 2 The solid curve in the figure is used to express the distribution of the correction factor obtained by joint solution in the height direction. Figure 2 The horizontal dashed line in the diagram is used to represent the layer boundary. Figure 2 This is primarily used to illustrate that the correction factor employs a low-degree-of-freedom structure parameterized by height layer intervals. Within a layer, it exhibits smooth and monotonic or piecewise monotonic variations, while between layers, connections can be made under either continuity constraints or restricted jump constraints. Figure 2The parameterized result forms obtained in steps S6 and S7 correspond to each other.

[0086] Figure 3 The horizontal axis represents the measured scattering coefficient at the verification height, and the vertical axis represents the estimated scattering coefficient at the verification height. Figure 3 The center dot represents the scatter points of high consistency in the verification of the method in this embodiment across multiple windows or multiple sample points, the cross represents the scatter points of the comparison method on the same sample points, and the dashed line represents the ideal reference line when the estimated value and the measured value are completely consistent. Figure 3 This is mainly used to illustrate that, under the condition that the verification height is not involved in solving the constraints, the scatter points of the method in this embodiment are closer to the reference line and have smaller dispersion, thus intuitively expressing "extrapolation consistency".

[0087] It should also be noted that, to verify the performance of the method in this embodiment under conditions of significant stratification, near-ground geometric overlap uncertainty, and the risk of system constant drift, comparative experiments were conducted using observational data from the same location. In the experiments, measured values ​​of the aerosol scattering coefficient from near-ground and high-altitude observation heights were used in the solution, and a third height was set as a verification height for cross-validation. The measured scattering coefficient values ​​at the verification height were not used in the solution; they were only used to evaluate the consistency of the correction results at heights not subject to constraint. Comparative method A uses a single-height optical parameter constraint and iterative inversion to obtain the vertical profile route, representing the stability differences of typical single-point constraint schemes under conditions of stratification variation and system uncertainty.

[0088] As shown in Table 1, Table 1 provides a comparison index of consistency and stability at the verification height. The mean absolute error and root mean square error at the verification height are used to characterize the magnitude of the consistency deviation between the correction result and the measured scattering coefficient at the verification height; the smaller the value, the better the consistency. The correlation coefficient at the verification height is used to characterize the synchronicity between the correction result and the measured scattering coefficient over time; the larger the value, the stronger the synchronicity. The deviation below the upper limit of the near-ground height is used to characterize the residual deviation level of the near-ground segment under conditions of insufficient geometric overlap and system uncertainty; the smaller the value, the less affected the near-ground segment is by overlap error. The cross-validation pass rate is used to characterize the proportion of correction results that meet the verification height consistency threshold during continuous window operation; the higher the value, the better the overall stability. The solution failure rate is used to characterize the proportion of divergence or triggering backtracking during the iteration process that still cannot converge; the lower the value, the more stable the solution.

[0089] Table 1 shows the comparison results of high consistency and stability.

[0090] Table 1 reflects the following: The method in this embodiment can still maintain small errors and high correlations at the verification heights where no constraints are involved, indicating that "full-height echo consistency target + multi-height scattering coefficient consistency constraint + hierarchical parameterized model + near-ground overlap compensation and joint estimation of system constants" can effectively propagate constraints to unconstrained heights and maintain overall self-consistency; at the same time, the near-ground deviation is smaller, reflecting that introducing geometric overlap compensation parameters in the low-height range and participating in joint estimation can effectively suppress near-field overlap errors; the cross-validation pass rate is higher and the solution failure rate is lower, reflecting that the hierarchical boundary stability confirmation, robust iterative reweighting and cross-validation backoff mechanism can improve the stability in continuous operation scenarios. Figure 3 The consistency differences in Table 1 are presented in a visual way: the scatter points in this embodiment are closer to the consistency reference line, and the errors and correlations in Table 1 are smaller.

[0091] As shown in Table 2, Table 2 illustrates that the key engineering configuration items and threshold calibers in this embodiment have clear and executable definitions, thereby ensuring that the layer identification, parameterized modeling, and joint solution processes are easy to reproduce and have consistent engineering meaning. Among them, the preset integral window is used to define the statistical range of echo accumulation / averaging and scattering coefficient time alignment; the preset near-ground height upper limit is used to define the low-height range, thereby limiting the effective height segment of geometric overlap compensation parameters; the background height interval is used to obtain the background amount and complete the background subtraction; the layer boundary stability confirmation is used to screen out false boundaries caused by instantaneous noise; the echo mutation intensity is used to trigger the switching of interlayer continuous / restricted jump constraints; the cross-validation trigger is used to start backoff when the verification height deviation exceeds the threshold; and the clean reference height segment is used to apply stable anchoring to the system constant parameters to reduce coupling drift.

[0092] Example Table 2 of Key Engineering Configuration and Threshold Scope

[0093] Table 2 reflects the following: the key concepts mentioned above all have clear calculation methods and engineering definitions, enabling the entire process of "layering—modeling—solving—verification—rollback—output" to be implemented by those skilled in the art, and providing a basis for Table 1 and... Figure 3 The comparative evaluation provides a consistent premise.

[0094] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for correcting the vertical profile of aerosols based on joint observation of aerosol scattering coefficient and lidar, characterized in that, include: S1. During the observation period, the echo signal of the lidar to the atmosphere is acquired with a preset integration window, and a range-corrected echo signal is generated for profile processing. S2. During the observation period, obtain at least two measured values ​​of aerosol scattering coefficient at different observation heights, and map the measured values ​​of aerosol scattering coefficient to the height grid of the distance-corrected echo signal and perform time alignment. S3. Preprocess the distance-corrected echo signal, the preprocessing including at least background subtraction, resolution matching and invalid height removal, to obtain a preprocessed echo signal; S4. Generate an initial vertical profile of the aerosol scattering coefficient based on the preprocessed echo signal; S5. Based on the preprocessed echo signal, perform aerosol layer boundary identification to obtain at least one layer boundary and multiple height layer intervals divided by the layer boundary. S6. Construct a correction factor parameterization model within each of the height layer intervals. The correction factor parameterization model includes at least intra-layer scale parameters and intra-layer gradient parameters. Furthermore, a geometric overlap compensation parameter is introduced for the low-height range, and a constant parameter of the lidar system is also introduced. S7. Taking the consistency between the theoretical echo signal obtained by forward calculation of the vertical profile of the candidate aerosol scattering coefficient obtained by the vertical distribution of the candidate correction factor acting on the initial vertical profile of the aerosol scattering coefficient and the preprocessed echo signal in the full height range as the goal, and taking the consistency between the candidate aerosol scattering coefficient at at least two different observation heights and the corresponding measured values ​​of the aerosol scattering coefficient as the constraint, solve the parameters in the parameterization model of the correction factor, the constant parameters of the lidar system, and the geometric overlap compensation parameters to obtain the vertical distribution of the correction factor; S8. Apply the correction factor vertically to the initial vertical profile of the aerosol scattering coefficient, output the corrected vertical profile of the aerosol scattering coefficient, and output the layer boundary information and correction parameter set corresponding to the corrected vertical profile of the aerosol scattering coefficient.

2. The method according to claim 1, characterized in that, The parameterization model of the correction factor adopts a two-parameter form of intra-layer scale parameter and intra-layer gradient parameter within each height layer interval, and constrains the correction factor to change monotonically or piecewise monotonically with height within the same height layer interval.

3. The method according to claim 1, characterized in that, The correction factor between adjacent height layer intervals satisfies the interlayer continuity constraint; when the echo abrupt change intensity at the layer boundary exceeds the threshold, the interlayer continuity constraint is released and replaced with an interlayer restricted jump constraint; wherein, the echo abrupt change intensity is determined by the statistical difference of the preprocessed echo signal in the preset height bands adjacent to the layer boundary, and the statistical quantity includes at least the mean or median.

4. The method according to claim 1, characterized in that, The low altitude range is the altitude segment below the preset upper limit of near-ground altitude; the geometric overlap compensation parameter is obtained by performing a quadratic polynomial parameterization on the geometric overlap function within the low altitude range, and is jointly estimated with the parameters of the correction factor parameterization model during the solution process.

5. The method according to claim 1, characterized in that, The constant parameters of the lidar system are estimated by automatically selecting a clean reference height segment, which is a height segment that satisfies the following conditions: "the echo gradient is below a threshold, the polarization characteristics are stable, and the rate of change of the initial vertical profile of the aerosol scattering coefficient is below a threshold".

6. The method according to claim 1, characterized in that, When the observation wavelength of the measured aerosol scattering coefficient is inconsistent with the operating wavelength of the lidar, the method converts the measured aerosol scattering coefficient to the operating wavelength of the lidar based on a preset wavelength index model before solving the problem, and the wavelength index is determined by multi-wavelength scattering observations or historical calibration data at the same site.

7. The method according to claim 1, characterized in that, When the measured value of the aerosol scattering coefficient is the measured value of the scattering coefficient under dry conditions, the method performs a humidity growth correction on the measured value of the scattering coefficient based on the relative humidity profile before the solution. The humidity growth correction adopts a piecewise parameterized humidity growth curve and the parameters are determined by historical comparison test data at the same site.

8. The method according to claim 1, characterized in that, The process of generating the initial vertical profile of the aerosol scattering coefficient includes: first, obtaining the backscattering correlation profile by inverting the preprocessed echo signal; and then, selecting the conversion parameter between scattering and backscattering from a preset type parameter table based on the aerosol type determined by polarization characteristics, in order to generate the initial vertical profile of the aerosol scattering coefficient.

9. The method according to claim 1, characterized in that, The invalid height removal includes: generating a quality mark for each height, the quality mark being generated based at least on a signal-to-noise ratio threshold and an echo mutation threshold; when the lidar includes at least one polarization echo channel, the quality mark is also generated based on a polarization consistency threshold; and the marked heights are removed or subjected to restricted interpolation in the solution.

10. An aerosol vertical profile correction device for joint observation of aerosol scattering coefficient and lidar, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method as claimed in any one of claims 1 to 9.

Images