Plateau weak-signal cloud and multi-layer cloud bottom identification method based on single-beam wind lidar

Abstract

The application discloses a kind of highland weak signal cloud and multilayer cloud bottom identification method based on single laser wind-radar, belong to the technical field of weather radar cloud bottom identification, including: obtaining single laser wind-radar data, and construct clear sky background reference profile;Calculate the sudden increase of spectral intensity relative to corresponding clear sky background reference profile, and preliminary extraction candidate cloud bottom;Through cloud layer minimum spacing constraint, background constraint and time continuity correction, the candidate cloud bottom is verified and layer position is matched in layers;Under the condition of multilayer cloud, the upper cloud cloud bottom is judged in detectability and missed detection risk, and the judgment result is output.The application extracts cloud bottom candidate point stably under weak signal background, reduces misidentification caused by cloud local enhancement and layer position mismatch under the condition of multilayer cloud, and further gives the reliability or detectability identification to the upper cloud cloud bottom identification result, so as to improve the accuracy, continuity and availability of cloud bottom height identification under complex highland cloud condition.

Description

Technical Field

[0001] This invention belongs to the technical field of meteorological radar cloud base identification, specifically relating to a method for identifying the cloud base of weak signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar. Background Technology

[0002] Cloud base height is a crucial parameter characterizing cloud vertical structure, cloud evolution, and cloud radiation effects, holding significant value in climate research, weather analysis, aviation safety, and weather modification operations. Accurate and continuous acquisition of cloud base height is essential for understanding cloud formation, development, and dissipation, as well as for monitoring cloud precipitation in complex terrain areas. Furthermore, with the development of the low-altitude economy and low-altitude flight activities such as unmanned aerial vehicles (UAVs), the demand for continuous cloud base height monitoring of low-altitude routes and takeoff / landing areas in complex terrain regions is further increasing. Cloud base height identification results can provide fundamental support for low-altitude flight meteorological support, route planning, and operational safety assessments.

[0003] Currently, the main methods for detecting cloud base height include microwave radiometers, radiosondes, millimeter-wave cloud radar, laser altimeters, and lidar. Microwave radiometers and radiosondes typically derive cloud base height based on temperature and humidity profiles, but microwave radiometers have relatively limited vertical resolution, and radiosonde data lacks temporal continuity, making it difficult to meet the continuous monitoring needs of rapidly changing cloud layers. Millimeter-wave cloud radar offers strong vertical cloud detection capabilities and is a commonly used device for observing cloud vertical structure; however, its equipment cost is high, and cloud base height identification can still be affected by precipitation, strong near-surface echoes, range sidelobes, and complex cloud conditions.

[0004] With the development of coherent lidar technology, the application of lidar in atmospheric boundary layer, wind field profile, and aerosol structure detection is constantly expanding. These devices typically output observations such as spectral intensity and signal-to-noise ratio that vary with time and altitude. Since cloud particles generally have stronger backscattering capabilities than clear-sky background aerosols, the spectral intensity or related signal quantity will significantly increase when the laser beam reaches the vicinity of the cloud base. Therefore, a single lidar unit has certain application potential in cloud base height identification. Compared with dedicated cloud radar or multi-device joint observation schemes, lidar typically features high miniaturization, simple equipment configuration, ease of deployment and maintenance, and good observation continuity. It also exhibits high response sensitivity to the enhanced backscattering near the base of thin clouds, fragmented clouds, and weak-signal clouds. Therefore, cloud base identification based on a single lidar unit has good operational application potential.

[0005] Existing lidar-based cloud base identification methods typically employ signal strength thresholds, signal gradients, slope abrupt changes, wavelet transforms, differential zero-crossing, differential enhancement, or statistical classification to identify cloud boundaries. These methods achieve good results under conditions of strong cloud signals, low background noise, and relatively simple cloud structures. However, in weak signal clouds, thin clouds, fragmented clouds, and multi-layered clouds, the enhancement of the cloud base signal is small and easily affected by background noise, aerosol echoes, local scattering enhancement, and signal attenuation with altitude, leading to unstable extraction of cloud base candidate points. In particular, methods using fixed thresholds or single gradient features struggle to simultaneously adapt to differences in background intensity and variations in cloud signal strength at different altitudes.

[0006] Plateau regions have complex terrain and high altitudes, resulting in a wide range of cloud heights and frequent occurrences of thin clouds, weak-signal clouds, fragmented clouds, and multi-layered cloud processes. Since laser signals gradually attenuate as their propagation path increases, the scattering and absorption of laser signals by lower-layer clouds in the presence of multi-layered clouds can affect the effective detection of the upper cloud base, weakening or even masking the signal surge characteristics of the upper cloud base. Existing methods mostly focus on the inversion of cloud base height itself, without addressing the detectability and missed detection risks of upper cloud bases under multi-layered cloud conditions in plateau regions. Furthermore, as multi-layered clouds move, form, or dissipate over time, the number of cloud layers may change between adjacent moments, easily causing layer mismatches between different cloud layers and affecting the continuity and reliability of the cloud base height time series. Summary of the Invention

[0007] This invention provides a method for identifying the cloud base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar. This method addresses the shortcomings of existing cloud base height identification methods based on single laser wind-measuring radar, which are characterized by complex cloud structures, frequent occurrences of thin and weak-signal clouds, and the prevalence of multi-layered cloud processes in high-altitude regions. First, the stability of cloud base candidate point extraction is insufficient under weak signal cloud and thin cloud conditions. Existing methods mostly rely directly on echo signal intensity, fixed thresholds, or single gradient features for cloud base identification, which is easily affected by the attenuation of lidar signals with altitude, differences in background intensity at different altitude layers, and fluctuations in background noise. When the enhancement of the cloud base signal is weak, it is easy to miss or misidentify, resulting in insufficient stability of cloud base identification under thin, fragmented, and weak signal cloud conditions.

[0008] Second, the detectability and reliability of the upper cloud base are difficult to determine. When a lower cloud layer exists, the lidar signal is cumulatively attenuated during upward propagation due to scattering and absorption by cloud particles, potentially weakening or even masking the abrupt increase in spectral intensity at the upper cloud base. Existing methods typically focus on outputting the cloud base height itself, making it difficult to determine whether the upper cloud base is reliably detectable or to identify the risk of missed detections of upper clouds, resulting in insufficient reliability of upper cloud identification results under multi-cloud conditions.

[0009] Third, under multi-layered cloud conditions, problems such as misidentification of cloud layers and poor continuity of time series are prone to occur. Loose cloud systems, fragmented clouds, and localized scattering enhancement are common in plateau regions. Changes in the phase state of particles within clouds or localized echo enhancement can easily produce signal surges resembling cloud bases, leading to misidentification as new cloud bases and causing misidentification of multi-layered clouds. Simultaneously, as multi-layered clouds move, form, or dissipate over time, the number of cloud layers at adjacent moments may change, easily causing misalignment of the layer correspondence between different cloud layers, affecting the continuity and reliability of the cloud base height time series. To address the following problems existing in current technologies...

[0010] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for identifying weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar includes the following steps: S1. Acquire single-laser wind radar data, preprocess the data, and construct a clear sky background reference profile; S2. Calculate the sudden increase in spectral intensity relative to the corresponding clear sky background reference profile at each observation time, and preliminarily extract candidate cloud bases; S3. By using the minimum spacing constraint of cloud layers, the background constraint of the distance library below the candidate cloud base, and the time continuity correction, perform layered verification and layer matching on the candidate cloud base, and output the upper-layer candidate cloud base. S4. Based on the interval integral spectrum intensity index and the signal-to-noise ratio of the candidate cloud base neighborhood, the detectability and missed detection risk of the upper cloud base under multi-layer cloud conditions are judged, and the judgment results are output.

[0011] This invention uses the spectral intensity (SI) output by a laser wind-measuring radar as the primary identification parameter. It leverages the abrupt increase in spectral intensity at the cloud base relative to a clear-sky background reference profile to construct a weak-signal cloud base candidate identification mechanism based on a background reference. Simultaneously, it incorporates the signal-to-noise ratio (SNR) to supplement the validity of the observation data and the credibility of the candidate cloud base, thereby improving the stability of cloud base candidate point extraction under weak-signal and thin cloud conditions. Furthermore, it introduces minimum cloud spacing constraints, background constraints below the candidate cloud base, and a time continuity correction mechanism to perform layered verification and layer matching for candidate cloud bases under multi-layer cloud conditions, reducing misidentification caused by localized enhancement within the cloud, fragmented clouds, and layer mismatch. For upper-layer cloud bases, this invention uses the interval integral spectral intensity index calculated from the spectral intensity data of a single laser wind-measuring radar, combined with the SNR characteristics near the upper-layer candidate cloud base, to determine the detectability, credibility, or risk of missed detection of the upper-layer cloud base.

[0012] The purpose of this invention is to achieve stable identification of the cloud base height of weak signal clouds, thin clouds, and multi-layered clouds under complex cloud conditions in plateau regions by fully utilizing the spectral intensity and signal-to-noise ratio data output by a single laser wind-measuring radar without relying on external detection equipment such as millimeter-wave cloud radar for identification calculation. This improves the accuracy, continuity, and reliability of the cloud base height identification results, and provides corresponding reliability or detectability indicators while outputting the cloud base height results. This provides technical support for cloud monitoring, meteorological support, low-altitude flight safety, and artificial weather modification operations in plateau regions.

[0013] Furthermore, S1 specifically includes: Acquire spectral intensity data and signal-to-noise ratio data output by a single-laser wind-measuring radar during a continuous observation period; Preprocessing includes: performing two-dimensional smoothing on the spectral intensity data to obtain smoothed spectral intensity; In addition, the effectiveness of the time-altitude distance database is screened based on the signal-to-noise ratio data. Specifically, for distance databases with signal-to-noise ratio data below the preset effective detection threshold, they are marked as low-confidence data points and are not used as cloud bottom candidate points alone. After preprocessing, an observation period under cloudless weather conditions was selected, and a clear sky background reference profile was constructed based on the smoothed spectral intensity data.

[0014] Furthermore, in S1, the clear sky background reference profile is a dynamic clear sky background reference profile constructed over a time period, and its construction process is as follows: The day is divided into multiple time periods. Smoothed spectral intensity data corresponding to cloudless samples within a time period are extracted, and the average value is calculated for each altitude distance library to serve as the clear sky background reference value for that altitude. The clear sky background reference values ​​of each altitude distance library are arranged by altitude to form the clear sky background reference profile corresponding to that time period.

[0015] Furthermore, S2 specifically includes: The sudden increase in spectral intensity relative to clear sky background is calculated by subtracting the clear sky background reference value for the corresponding time period from the smoothed spectral intensity data at each observation time. The initial screening criteria are as follows: when the sudden increase at a certain height exceeds the preset sudden increase threshold, and the height is not lower than the minimum effective detection height, and the signal-to-noise ratio data at or in the neighborhood of that height is not lower than the preset threshold, or when it has synchronous enhancement characteristics relative to the background distance library below it, the height is marked as a candidate cloud base location. At the same observation time, candidate cloud bases that meet the preliminary screening conditions are searched from bottom to top along a certain height direction, thus obtaining a set of candidate cloud bases.

[0016] Furthermore, in S3, layered filtering is performed using cloud minimum spacing constraints, including: For multiple candidate cloud bases identified at the same time, if the vertical distance between two adjacent candidate cloud bases is less than the minimum cloud spacing threshold, they are considered to belong to the same cloud layer's signal fluctuations, local enhancements, or fragmented cloud disturbances, and are merged into the same cloud layer candidate result, instead of being output as two independent cloud layers; if the vertical distance between two adjacent candidate cloud bases is not less than the minimum cloud spacing threshold, they are retained as candidate cloud bases of different cloud layers.

[0017] Furthermore, in S3, background constraint judgment is performed on the spectral intensity of the continuous distance library below the candidate cloud base, including: Based on cloudless samples, the mean and standard deviation of background spectral intensity at various altitudes were statistically analyzed, and the upper limit of normal fluctuation of background spectral intensity was calculated. Two consecutive distance libraries below the candidate cloud base height are selected. If the smoothed spectral intensity of both distance libraries is lower than the normal fluctuation upper limit of the corresponding height, the candidate cloud base is retained; otherwise, it is considered that there is an abnormal enhancement below the candidate cloud base, and it is removed or marked as a low-confidence candidate cloud base.

[0018] Furthermore, in S3, the layer correction of the multi-layer cloud candidate results based on temporal continuity includes: For candidate cloud base identification results at adjacent observation times, the cloud layers are numbered from low to high altitude, and the deviation of the candidate cloud base at the current time from the corresponding cloud base at the previous and next adjacent times is compared. If the height deviation between the current candidate cloud base and the corresponding layer cloud base does not exceed the continuity threshold, they are considered to belong to the same layer cloud. If the height deviation exceeds the continuity threshold, the current candidate cloud base is matched with cloud bases of other layers. If the height deviation with cloud bases of other layers is less than the continuity threshold, the candidate cloud base is assigned to the corresponding layer. If it cannot be matched with any layer, the candidate cloud base is marked as a low-confidence candidate point or is removed.

[0019] Furthermore, S4 specifically includes: The interval integral spectral intensity index is calculated based on the smoothed spectral intensity. When the interval integral spectral intensity index is lower than the preset attenuation risk threshold, and the target upper candidate cloud base meets the requirements of spectral intensity surge, signal-to-noise ratio effectiveness and time continuity, the upper cloud base is determined to be detectable, and a high-confidence label is output. When the interval integral spectral intensity index is higher than the preset attenuation risk threshold, or when the signal-to-noise ratio of the target upper candidate cloud base is lower than the effective detection condition and the spectral intensity surge characteristics are unstable, the detectability of the upper cloud base is determined to be limited, and a low confidence or missed detection risk indicator is output.

[0020] Furthermore, S4 specifically includes: Calculate the interval integral spectral intensity index based on the smoothed spectral intensity; The spectral intensity abrupt change, candidate cloud base neighborhood signal-to-noise ratio, temporal continuity deviation, and interval integral spectral intensity index are normalized and fused to obtain a comprehensive credibility scoring function; Based on the relationship between the comprehensive credibility scoring function and the preset level threshold, the identification results are divided into categories of high credibility, medium credibility, low credibility, or limited detectability.

[0021] Furthermore, in S4, the smoothed spectral intensity calculation interval integral spectral intensity exponent is expressed as: In the formula, The interval integral spectral intensity index; Indicates the height of the candidate cloud base above the target; To represent the height distance library, This represents the current height of the first cloud base. This represents the smoothed spectral intensity; Time period The outline of a clear sky background.

[0022] The method for identifying weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar provided by this invention has the following beneficial effects: (1) This invention improves the stability of cloud base candidate identification under weak signal clouds and thin clouds in plateau regions. Existing methods mostly use the absolute value of echo signal intensity, fixed threshold, or single gradient features for cloud base identification, which are easily affected by the attenuation of laser signal with altitude, background differences at different altitudes, and fluctuations in background noise. This invention constructs a clear sky background reference profile and calculates the sudden increase in spectral intensity relative to the corresponding clear sky background, transforming cloud base identification from "absolute intensity discrimination" to "background reference enhancement discrimination," which can reduce the impact of altitude attenuation and background intensity differences on the identification results. Furthermore, this invention can construct a dynamic clear sky background reference profile over time periods such as hours, so that the background reference value can adapt to the diurnal variation characteristics of the boundary layer, aerosols, and background noise in plateau regions, thereby improving the stability of cloud base candidate point extraction under weak signal clouds, thin clouds, and fragmented clouds.

[0023] (2) This invention reduces misidentification caused by noise disturbances and low-reliability data. Based on the spectral intensity surge criterion, this invention introduces signal-to-noise ratio (SNR) data to effectively screen the time-altitude distance database and assist in judging the observation validity of candidate cloud bases or their neighborhoods. For distance databases with low SNR, they are not considered as cloud base candidates, thereby reducing the risk of random noise peaks and low SNR anomalies being misidentified as cloud bases. This processing ensures that cloud base candidate identification not only focuses on whether the signal is enhanced but also on whether the enhancement has sufficient observation reliability.

[0024] (3) This invention reduces the identification of false cloud bases caused by local cloud enhancement and fragmented cloud disturbance. Loose cloud systems, fragmented clouds, and localized scattering enhancement within clouds are common in plateau regions. Changes in particle phase state or localized echo enhancement within the cloud layer may produce abrupt features similar to cloud bases. To address this issue, this invention sets a background constraint below the candidate cloud base: the spectral intensity of two consecutive distance libraries below the candidate cloud base is judged. If the spectral intensity of two consecutive distance libraries below the candidate cloud base is not lower than the upper limit of normal background fluctuation at the corresponding altitude, it is considered that there is an abnormal enhancement below the candidate point, which may belong to localized scattering enhancement within the cloud or a false cloud base, and it is removed or marked as a low-confidence candidate point. Through this constraint, the situation of misjudging localized cloud enhancement as a new cloud base can be effectively reduced.

[0025] (4) This invention improves the continuity and reliability of cloud base position identification under multi-layer cloud conditions; it introduces a minimum cloud layer spacing constraint and a time continuity correction mechanism. For adjacent candidate points with a vertical spacing less than the preset minimum cloud layer spacing threshold, they are merged into the same cloud layer candidate result to avoid misjudging fluctuations within the same cloud layer as multi-layer clouds. For cases where the number of cloud layers changes between adjacent times, this invention corrects potentially misaligned cloud layers by matching layer numbers with adjacent time heights, thereby reducing layer mismatches between different cloud layers and enhancing the continuity and physical rationality of the cloud base height time series.

[0026] (5) This invention can provide detectability or missed detection risk indicators for the upper cloud base identification results. Existing methods typically only output the cloud base height itself, making it difficult to determine whether the upper cloud base is reliable and detectable. This invention addresses the cumulative attenuation effect of lower clouds on laser signals under multi-cloud conditions by introducing an interval integral spectral intensity index to characterize the cumulative laser echo enhancement between the first cloud base and the target upper candidate cloud base and its impact on upper cloud detection. Combining the signal-to-noise ratio and temporal continuity information of the upper candidate cloud base neighborhood, this invention can determine whether the upper cloud base has high detectability or whether there is limited detectability or missed detection risk. Thus, this invention not only outputs the cloud base height but also the corresponding reliability or detectability indicator, improving the interpretability and operational usability of multi-cloud base identification results.

[0027] (6) This invention reduces the reliance on joint observations by multiple devices, facilitating operational deployment in plateau regions. The cloud base identification and reliability assessment process of this invention relies solely on the spectral intensity and signal-to-noise ratio data output by a single laser wind-measuring radar, without requiring millimeter-wave cloud radar, laser ceilometer, or all-sky imager to participate in the identification calculation. Other observation equipment can be used for result verification or sample annotation, but are not necessary conditions for the operation of this method. Therefore, this invention is suitable for plateau regions with relatively limited equipment conditions, scattered site distribution, and high deployment and maintenance costs, and is conducive to achieving continuous, low-cost, and easily deployable cloud base height monitoring.

[0028] (7) This invention enhances the application value of cloud base identification results under complex plateau cloud conditions. This invention can output cloud base height, cloud layer number, and reliability or detectability indicators under conditions of weak signal clouds, thin clouds, fragmented clouds, and multi-layered clouds, providing more stable, continuous, and reliable cloud base height data support for cloud monitoring in plateau areas, aviation meteorological support, low-altitude flight safety, UAV operations, artificial weather modification operations, and cloud precipitation process research. Attached Figure Description

[0029] Figure 1 This is a flowchart of the technical route in Example 1.

[0030] Figure 2 This is an exterior view of the test location and equipment in Example 2; Figure 2 (a) in the figure represents the terrain surrounding the experimental site; Figure 2 (b) in the image is an exterior view of the laser wind-measuring radar; Figure 2 (c) in the image is an exterior view of a Ka-band millimeter-wave cloud radar. Figure 2 (d) in the image is an exterior view of the All-Sky Imager.

[0031] Figure 3 The spectral intensity profiles under clear sky conditions in Example 2 are shown. Figure 3(a) in the image is an example of the spectral intensity profile of a laser wind measuring radar under cloudless conditions; Figure 3 (b) in the figure is an example of the spectral intensity profile of a laser wind measuring radar under low cloud conditions; Figure 3 (c) in the figure is an example of the spectral intensity profile of a laser wind measuring radar under high cloud conditions.

[0032] Figure 4 The images show the actual results after each step in Example 2. Figure 4 (a) in the text represents the original observation data; Figure 4 (b) in the figure shows the results of data preprocessing; Figure 4 (c) in the diagram represents the initial identification; Figure 4 (d) in the figure represents the result of secondary verification and multi-layer cloud level correction of cloud base height.

[0033] Figure 5 The figures show the results of identifying the height of the first layer of clouds under a typical multi-layer cloud structure in Example 2, as well as the results of judging the reliability of identifying the height of the cloud base of non-first layer clouds.

[0034] Figure 6 This is a diagram illustrating the applicability and identification performance in high-altitude areas in Example 2; Figure 6 (a) is a two-dimensional frequency comparison of the cloud base height identified by laser wind radar and the cloud base height inverted by Ka-band millimeter-wave cloud radar under all valid samples of the same cloud layer; Figure 6 (b) in the figure is a violin plot showing the difference in cloud base height between the two under different cloud height (low cloud, middle cloud, high cloud) ranges.

[0035] Figure 7 This is a comparison chart of the cloud base height results identified by the laser wind-measuring radar in a typical single-layer convolutional cloud case in Example 2 and the cloud base height results retrieved by the Ka-band millimeter-wave cloud radar. Figure 7 (a) in the image is the cloud base height map retrieved by superimposing the spectral intensity of the laser wind measurement radar. Figure 7 (b) in the image shows the cloud conditions at various times using the all-sky imager; Figure 7 (c) in the figure represents the reflectivity factor of the Ka-band millimeter-wave cloud radar and the cloud base height map obtained by superimposing it. Figure 7 (d) in the figure is a comparison of the cloud base heights of the two devices.

[0036] Figure 8 This is a comparison chart of the cloud base height results identified by the laser wind measurement radar in a typical multi-layer cloud case in Example 2 and the cloud base height results retrieved by the Ka-band millimeter-wave cloud radar. Figure 8 (a) in the image is the cloud base height map retrieved by superimposing the spectral intensity of the laser wind measurement radar. Figure 8 (b) in the image shows the cloud conditions at various times using the all-sky imager; Figure 8 (c) in the figure represents the reflectivity factor of the Ka-band millimeter-wave cloud radar and the cloud base height map obtained by superimposing it. Figure 8 (d) in the figure is a comparison of the cloud base heights of the two devices.

[0037] Figure 9 This is an overview of the comparison between cloud base heights identified by all laser wind-measuring radars and Ka-band millimeter-wave cloud radars during the case study in Example 2. Figure 10 This is a flowchart of the plateau weak signal cloud and multi-layer cloud base identification method based on a single laser wind measuring radar in Example 1. Detailed Implementation

[0038] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0039] Example 1 This embodiment provides a method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single-unit laser wind-measuring radar. The method uses the spectral intensity (SI) and signal-to-noise ratio (SNR) data output by the single-unit laser wind-measuring radar as input. The spectral intensity (SI) characterizes the relative strength of the laser echo signal with time and altitude, while the SNR represents the relative noise intensity of the signal in the corresponding time-altitude distance-based observation data, characterizing the validity and reliability of the data. The method uses the abrupt increase in spectral intensity (SI) relative to the clear-sky background reference profile as the primary criterion for candidate cloud base identification, and uses the SNR as an auxiliary criterion for effective data screening, candidate cloud base verification, and upper-layer cloud detectability assessment. Figure 1 and Figure 10 Specifically, it includes the following: S1. Acquire single-laser wind radar data, preprocess the data, and construct a clear sky background reference profile; In some embodiments, the spectral intensity data output by a single laser wind-measuring radar during a continuous observation period is first acquired. and signal-to-noise ratio data .in, Indicates the observation time. Indicates the height distance library, This represents the original spectral intensity at the corresponding time and altitude position. This indicates the signal-to-noise ratio at the corresponding location.

[0040] In the preprocessing stage, in order to reduce the impact of random noise, local anomalies and short-term fluctuations on cloud bottom candidate identification, the original spectral intensity data is subjected to two-dimensional smoothing.

[0041] In this embodiment, a two-dimensional Gaussian filter is preferably used to obtain the smoothed spectral intensity. : (1) (2) In the formula, It is a two-dimensional Gaussian function. and These are the standard deviations in the time and height directions, respectively, with the following signs. This represents a two-dimensional convolution operation. In this embodiment, Take 1 minute. Take 30 m; filter uses The window is implemented using discrete convolution, taking four sampling points on each side of the center point in both the time and height directions, corresponding to half-widths of 4 min and 120 m, respectively. These parameters can be adjusted according to the temporal and range resolutions of the laser wind radar.

[0042] In the preprocessing stage, based on signal-to-noise ratio data The time-altitude distance database is screened for validity. Distances with a signal-to-noise ratio (SNR) below a preset effective detection threshold are marked as low-confidence data points and are not considered as separate candidate cloud base points. The SNR threshold can be determined based on the instrument model, observation altitude range, and background noise statistics.

[0043] After the preprocessing stage, observation periods under cloudless weather conditions were selected, and the smoothed spectral intensity data was used as the basis for the analysis. The clear sky background reference value is statistically analyzed at each altitude to construct a clear sky background reference profile that varies with altitude. The clear sky background reference profile can be a uniform background reference profile or a dynamic background reference profile constructed according to time periods.

[0044] In the unified background reference profile scheme, the cloudless observation period is defined as... For each altitude distance database, smoothed spectral intensity samples of all corresponding altitudes within the cloudless observation period are extracted, and the average value is used as the clear sky background reference value for that altitude. The clear sky background reference values ​​of each altitude distance database are arranged sequentially to form a unified clear sky background reference profile. .

[0045] In one specific embodiment, considering the significant diurnal variation characteristics of the boundary layer, aerosols, and background noise in plateau regions, this invention employs a dynamic clear-sky background reference profile constructed over time periods, specifically as follows: Divide the day into multiple time periods, such as by hour; for the first... For each time period, smoothed spectral intensity data corresponding to cloudless samples are extracted, and the average value is calculated for each altitude distance database to serve as the clear sky background reference value for that altitude. The clear sky background reference values ​​of each altitude distance database are arranged by altitude to form the clear sky background reference profile corresponding to that time period. .

[0046] For any observation time Determine the time period number to which it belongs, and record it as... And select the clear sky background reference outline corresponding to that time period. It participates in subsequent burst calculations. If the number of cloudless samples is insufficient within a certain time period, the background reference profile for that time period can be determined by merging adjacent time periods, using sliding time window statistics, or supplementing with historical clear-sky sample databases.

[0047] The method in this embodiment S1 can incorporate the differences in clear sky background at different altitudes and at different times into the reference benchmark, thereby reducing the impact of laser signal attenuation with altitude, diurnal variation in the vertical distribution of background aerosols, and fluctuations in background noise on weak signal cloud base identification.

[0048] S2. Calculate the sudden increase in spectral intensity relative to the corresponding clear sky background reference profile at each observation time, and preliminarily extract candidate cloud bases; In some embodiments, after obtaining the clear-sky background reference profile, the smoothed spectral intensity at each observation time is... Clear sky background reference value for the corresponding time period at that moment Compare and calculate the sudden increase in spectral intensity relative to a clear-sky background; The burst increment can be represented by a difference, a normalized difference, or a decibel difference; this embodiment uses the difference form to calculate the background reference burst increment. : (3) Alternatively, normalized mutation increments can be used. : (4) In practice, the appropriate method can be selected based on the dimensions, dynamic range, and high dependence of the spectral intensity data. , Alternatively, the decibel difference can be used as the candidate cloud base identification value. When using a uniform background reference profile, the above formula... Can be replaced with .

[0049] When the background reference increment at a certain altitude exceeds a preset increment threshold, and this altitude is not lower than the minimum effective detection altitude, this altitude is marked as a candidate cloud base location. To reduce false identification caused by noise peaks, the candidate cloud base must also meet the signal-to-noise ratio (SNR) validity condition, i.e., the SNR at or around the candidate altitude. It is not lower than a preset threshold, or has synchronous enhancement features relative to the background distance library below it.

[0050] At the same observation time, candidate cloud bases satisfying the above conditions are searched from bottom to top along the altitude direction. If a region satisfying the conditions of sudden increase in spectral intensity and effective signal-to-noise ratio reappears at a higher altitude, it is marked as an upper-layer candidate cloud base. The preliminary set of candidate cloud bases is then used for subsequent secondary verification and stratified screening steps.

[0051] S3. By using the minimum spacing constraint of cloud layers, the background constraint of the distance library below the candidate cloud base, and the time continuity correction, perform layered verification and layer matching on the candidate cloud base, and output the upper-layer candidate cloud base. In some embodiments, firstly, a layered screening is performed using a minimum cloud spacing constraint. Specifically, for multiple candidate cloud bases identified at the same time, if the vertical spacing between two adjacent candidate cloud bases is less than the minimum cloud spacing threshold, they are considered to belong to signal fluctuations, local enhancements, or fragmented cloud disturbances within the same cloud layer, and are merged into the same cloud layer candidate result, not output as two independent cloud layers; if their vertical spacing is not less than the threshold, they are retained as candidate cloud bases of different cloud layers. The minimum cloud spacing threshold can be 200-500 m, preferably 300 m.

[0052] In some embodiments, secondly, a background constraint judgment is performed on the spectral intensity of the continuous distance database below the candidate cloud base to reduce the interference of particle phase changes or local scattering enhancement within the cloud layer on cloud base identification. Specifically, based on cloudless samples, the mean and standard deviation of the background spectral intensity of each altitude distance database are statistically analyzed, and the upper limit of normal fluctuation of the background spectral intensity is determined: (5) In the formula, For the first altitude within a time period The upper limit of normal fluctuations in background spectral intensity at that location. and These represent the mean and standard deviation of the spectral intensity for the corresponding cloudless samples. This is the fluctuation range coefficient. In this embodiment, Taking 2, the normal fluctuation range of the background can be expressed as: For any candidate cloud base height Select two consecutive distance libraries below it. and If the smoothed spectral intensity of both distance libraries is lower than the upper limit of normal background fluctuations at the corresponding altitude, the candidate cloud base is retained; otherwise, it is considered that there is already abnormal enhancement below the candidate cloud base, which may be due to local scattering enhancement within the cloud, fragmented cloud disturbance, or a false cloud base, and it is removed or marked as a low-confidence candidate point. The signal-to-noise ratio is mainly used in this step to confirm the validity of the observation data at the candidate cloud base; the background constraint of the lower distance library is mainly based on whether the spectral intensity is within the normal fluctuation range of a clear sky background.

[0053] In some embodiments, layer correction is performed on the candidate cloud layers based on temporal continuity. Specifically, for cloud base identification results at adjacent observation times, cloud layers are numbered from low to high altitude, and the height deviation of the candidate cloud base at the current time is compared with that of the corresponding cloud base at the previous and next adjacent times. If the height deviation between the current candidate cloud base and the corresponding cloud base does not exceed the continuity threshold, it is considered to belong to the same cloud layer; if the height deviation exceeds the continuity threshold, it is further attempted to match it with cloud bases at other layers. If the height deviation with cloud bases at other layers is less than the continuity threshold, the candidate cloud base is assigned to the corresponding layer; if it cannot be matched with any layer, the candidate cloud base is marked as a low-confidence candidate point or is removed.

[0054] The continuity threshold can be set between 300 and 600 m, with 450 m being preferred. When multiple matching layers exist, the matching result with a larger sudden increase in spectral intensity, higher signal-to-noise ratio, and better temporal continuity is given priority.

[0055] This embodiment, through the above-mentioned secondary verification and layer correction, can reduce problems such as local enhancements within the cloud being misjudged as new cloud bases, mismatch between different cloud layers at adjacent times, and jumps in the time series of cloud base height.

[0056] S4. Based on the interval integral spectrum intensity index and the signal-to-noise ratio of the candidate cloud base neighborhood, the detectability and missed detection risk of the upper cloud base under multi-layer cloud conditions are judged, and the judgment results are output. In some embodiments, under multi-layered cloud conditions, the laser signal is affected by scattering and absorption by cloud particles when passing through the lower layer of cloud, resulting in cumulative attenuation. When the lower cloud echo is strong or has a large vertical range, the sudden increase in spectral intensity at the bottom of the upper cloud may be weakened or even submerged, leading to unreliable identification results of the upper cloud bottom or a risk of missed detection.

[0057] To characterize the potential occlusion and attenuation effects of lower-layer clouds on the identification of upper-layer cloud bases, this invention introduces an interval integral spectral intensity index. This index is calculated from the spectral intensity data of a single laser wind radar and is used to characterize the cumulative laser echo enhancement from the first cloud base to the target upper-layer candidate cloud base and its impact on upper-layer cloud detection. For the first... Candidate cloud base, its interval integral spectral intensity index Represented as: (6) In the formula, This indicates the height of the first cloud base at the current moment. Indicates the height of the candidate cloud base above the target. This represents the height distance resolution. Using positive augmentation integrals avoids noise or attenuation ranges below the background value from offsetting the cumulative cloud augmentation features. When using a uniform background reference profile, the above equation... Can be replaced with .

[0058] When the interval integral spectral intensity index is lower than the preset attenuation risk threshold, and the target upper candidate cloud base meets the requirements of spectral intensity surge, signal-to-noise ratio effectiveness, and time continuity, the upper cloud base is determined to have high detectability, and a high-confidence flag is output. When the interval integral spectral intensity index is higher than the preset attenuation risk threshold, or the signal-to-noise ratio near the target upper candidate cloud base is lower than the effective detection condition, or the spectral intensity surge characteristic is unstable, the upper cloud base is determined to have limited detectability, and a low-confidence or missed detection risk flag is output.

[0059] In some embodiments, a comprehensive credibility scoring function can also be constructed to normalize and fuse the sudden increase in spectral intensity, the signal-to-noise ratio of the candidate cloud-bottom neighborhood, the temporal continuity bias, and the interval integral spectral intensity index. For example, it can be expressed as: (7) In the formula, For the first Candidate cloud base at any time The overall credibility score, , , ,and These represent the normalized scores for spectral intensity surge characteristics, signal-to-noise ratio effectiveness, temporal continuity, and attenuation risk, respectively. , , and These are the corresponding weighting coefficients.

[0060] Then, based on the overall credibility score Based on the relationship with preset level thresholds, the recognition results are divided into categories such as high confidence, medium confidence, low confidence, or limited detectability.

[0061] Finally, this invention outputs the cloud base height, cloud layer number, and corresponding reliability or detectability indicators for each observation time. This output includes not only the cloud base height and location but also the reliability of the upper cloud base identification results or a risk indication of missed detections. It can be used for continuous cloud monitoring and meteorological support applications under complex cloud conditions in plateau regions.

[0062] Example 2 This implementation case uses a comprehensive cloud precipitation characteristic observation experiment conducted in Hongyuan County, Aba Prefecture, Sichuan Province from July 11 to 30, 2024, as an example to illustrate the method in Example 1. Utilizing observation data from a single laser wind-measuring radar deployed during the experiment, cloud base height identification was carried out in a typical plateau region. Hongyuan County (102.513 °E, 32.411 °N, 3493 m) is located on the eastern edge of the Qinghai-Tibet Plateau, with a high altitude, open terrain, and frequent cloud activity in summer. Thin clouds, fragmented clouds, weak signal clouds, and multi-layered cloud processes are relatively common, making it a good representative of the plateau region. Therefore, selecting this region for implementation case verification can effectively test the applicability of the present invention under complex cloud conditions on the plateau.

[0063] It should be noted that the cloud base height identification in this invention is based solely on the spectral intensity data (SI) and signal-to-noise ratio (SNR) output from a single laser wind-measuring radar. To verify the identification results and assist in cloud condition interpretation, this embodiment simultaneously incorporates inversion results from a co-located Ka-band millimeter-wave cloud radar (KaCR) and images from an all-sky imager (ASI). The Ka-band millimeter-wave cloud radar and the all-sky imager are used only as verification equipment and auxiliary interpretation data for comparative analysis of the laser wind-measuring radar cloud base height identification results; they are not used as input data for this invention, nor do they participate in calculations such as cloud base candidate point identification, secondary verification, multi-layer cloud level correction, and detectability discrimination.

[0064] The Doppler lidar used in this implementation case is the Wind3D 6000 Coherent Doppler lidar (CDL), with an operating wavelength of 1.55 μm, pulse energy of 150 μJ, pulse width of 100–400 ns, a maximum detection range of 6000 m, adjustable spatial resolution between 5 / 30 / 60 / 150 m, and temporal resolution set between 1 s and 10 min. The Ka-band millimeter-wave cloud radar is a solid-state, Doppler-based cloud-measuring radar that acquires cloud echo information by emitting pulsed electromagnetic waves and receiving backscattered signals from cloud particles. This radar has an 8.6 mm emission wavelength, an antenna pointing vertically towards the zenith, a beamwidth of 0.3°, a temporal resolution of 1 min, a spatial resolution of 30 m, and a detection altitude range of 0.12–15.3 km. The all-sky imager is primarily used for cloud observation. Equipped with a 180° wide-angle lens and a high-resolution image sensor, it can automatically record the distribution of cloud cover and the evolution of cloud images across the entire sky, with a time resolution of 5 minutes. Figure 2 As shown Figure 2 .

[0065] During the results verification process, the cloud base height results identified by the laser wind-measuring radar of this invention were compared with the inversion results from Ka-band millimeter-wave cloud radar, and typical cloud conditions were further interpreted using all-sky imager images. The evaluation metrics uniformly adopted were correlation coefficient (CC), root mean square error (RMSE), mean absolute error (MAE), and mean error (ME). The specific calculation formulas are as follows: (8) (9) (10) (11) In the formula, Indicates the first The cloud base height is obtained by laser wind radar at each matching time point. This indicates the cloud base height obtained from Ka-band millimeter-wave cloud radar inversion at the corresponding time. and These are the average values ​​of the two types of cloud base height results; To effectively match the number of samples.

[0066] When identifying cloud base height, it is first necessary to construct a clear sky background reference profile hourly so that the background reference value can adapt to the diurnal variation characteristics of the boundary layer, aerosols and background noise in the plateau region. Figure 3 Hourly resolution background reference profiles for clear skies under cloudless conditions are presented, along with abrupt increases in spectral intensity (SI) relative to the background profiles for the corresponding time periods under low and high cloud conditions. For ease of demonstration, Figure 3 In (a) of the figure, the hourly background profile is further averaged over four time periods: 00–06, 06–12, 12–18, and 18–24. As shown in the figure, the clear-sky background reference profile exhibits similar distribution characteristics with altitude across different time periods. The spectral intensity is relatively high in the near-surface layer below 0.5 km, gradually weakening with increasing altitude and approaching background noise levels in the mid-to-high altitudes. However, certain differences exist between the near-surface layer and the lower atmosphere. For example, the overall spectral intensity in the lower atmosphere is lower during the nighttime period of 00–06, while during the daytime, due to boundary layer development and enhanced vertical aerosol mixing, the background profile from the near-surface layer to the lower atmosphere is relatively increased. This difference indicates that using an hourly clear-sky background reference profile can better adapt to changes in background conditions at different times and reduce identification biases that may result from a fixed background profile. Figure 3 (b) and Figure 3 Figure (c) shows the comparison results of the spectral intensity profiles with the corresponding clear-sky background reference profiles under two typical cases: low clouds and high clouds. As can be seen from the figure, under both low and high cloud conditions, the spectral intensity profiles show a significant abrupt increase near the cloud base compared to the clear-sky background profile for the corresponding time period. This abrupt increase location is basically consistent with the cloud base height marked in the figure, indicating that the abrupt increase in spectral intensity relative to the clear-sky background can effectively indicate the height of the cloud base. This result demonstrates that the abrupt increase in spectral intensity SI relative to the clear-sky background reference profile can effectively indicate the height of the cloud base and can be used as the main criterion for the preliminary identification of cloud base candidate points in this invention.

[0067] Figure 4 The processing results of each main step in the cloud base height identification method of this invention are presented, including the original spectral intensity data SI, preprocessing results, preliminary identification results of cloud base candidate points, and the final cloud base identification results after secondary verification and multi-layer cloud level correction. As shown in the figure, in addition to the signal surge near the cloud base, the original spectral intensity data also contains certain random noise, strong near-surface echoes, and local strong scattering structures within the cloud, such as... Figure 4 In (a), after two-dimensional smoothing, short-term fluctuations in the spectral intensity are weakened, and the continuous enhancement features near the cloud base relative to the background are clearer, as shown in (a). Figure 4(b) Based on this, the background reference burst increment is calculated by the difference between the smoothed spectral intensity and the corresponding clear-sky background reference profile. This can effectively capture burst signals near the base of low-level and upper-level clouds and extract preliminary candidate cloud base points. However, due to factors such as strong local scattering within clouds and cloud fragmentation, some non-true cloud base locations may still be identified as candidate cloud bases. Figure 4 (c) Further, by applying minimum cloud spacing constraints, background constraints below the candidate cloud base, and temporal continuity constraints to the initial candidate points, false candidate points are significantly eliminated. For multi-layer cloud processes, layer correction can reduce problems such as mismatched matching of different cloud layers at adjacent times and abrupt changes in cloud base height sequences, making the candidate cloud base results closer to the actual cloud base position. Figure 4 (d) In the end, the results show that the present invention can output the cloud base height and the corresponding cloud layer number at each observation time.

[0068] Figure 5 The results of identifying the cloud base height of the first layer under a typical multi-layer cloud structure, as well as the reliability judgment results of the cloud base height of non-first layer clouds, are presented. As shown in the figure, for the identification of the first layer cloud, this invention can effectively detect its cloud base height in the vast majority of samples, demonstrating high identification stability and reliability. This indicates that the first layer cloud is less affected by the occlusion of the lower layer cloud, and the abrupt increase in spectral intensity at the cloud base is relatively clear, which is beneficial for stable identification by the algorithm. For non-first layer clouds, the identification difficulty increases significantly due to the scattering and attenuation of the laser signal by the lower layer cloud, and in some cases, the cloud base height cannot be successfully identified. After classifying the identification results of non-first layer clouds into high, medium, low, and limited detectability categories, it can be seen that high reliability results mainly occur when the first layer cloud thickness is thin or the echo enhancement is weak. In this case, the attenuation effect of the lower layer cloud on the laser signal is limited, and the abrupt increase in intensity near the cloud base of the upper layer cloud can still be effectively captured. Conversely, low reliability results mostly correspond to cases where the first layer cloud thickness is large or the spectral intensity SI is strong. At this point, the lower cloud layer significantly attenuates the laser signal, weakening or even masking the signal surge at the base of the upper cloud layer, thus reducing the reliability of determining the cloud base height of non-first-layer clouds. This indicates that the present invention can not only identify the cloud base height of different layers under multi-layer cloud conditions, but also provide a reliability or detectability limitation indicator for the upper cloud base identification results, thereby supporting continuous cloud base monitoring and quality interpretation of identification results under complex cloud conditions in plateau regions.

[0069] To verify the applicability and identification performance of this invention in high-altitude areas, Figure 6Two-dimensional frequency comparison charts of cloud base height inversion by laser wind-measuring radar and Ka-band millimeter-wave cloud radar are presented for all valid samples of the same cloud layer. Violin plots of the cloud base height difference (ΔCBH) between the two methods are also provided for three cloud height ranges: low, middle, and high clouds. As shown in the figures, the cloud base heights obtained by laser wind-measuring radar and Ka-band millimeter-wave cloud radar show good overall consistency, with a correlation coefficient of 0.99. The root mean square error, mean absolute error, and mean error are 251 m, 182 m, and 38 m, respectively, indicating that the cloud base height results obtained by this invention based on laser wind-measuring radar are generally very close to the KaCR inversion results. Looking at the ΔCBH distribution across different cloud height ranges, the ΔCBH distribution is relatively discrete in the low cloud range with larger fluctuations; the middle cloud range is next; and the high cloud range has the most concentrated ΔCBH distribution with the smallest overall difference. These results demonstrate that the identification results of this invention maintain good consistency with the KaCR inversion results under different cloud height conditions.

[0070] Figure 7 and Figure 8 The results of cloud base height identification by laser wind-measuring radar and the results of cloud base height inversion by Ka-band millimeter-wave cloud radar are compared for single-layer convolutional cloud cases and multi-layer cloud cases, respectively, and are further verified by all-sky imager images. As can be seen from the figures, in the stage where the cloud layer is thicker and the echo structure is more continuous, both the identification results of this invention and the KaCR inversion results can identify the cloud base height relatively stably; however, in the stage where the cloud layer is thinner or the cloud fragmentation is more obvious (…), the results of this invention and the results of KaCR inversion can identify the cloud base height relatively stably. Figure 7 T1, T2, T4 and Figure 8 During T1 and T3, the laser wind radar can identify the cloud base height, but the corresponding cloud base signal in the KaCR results is relatively discontinuous or missing. All-sky imager images show that small-scale fragmented or thin clouds are indeed present near the zenith at the above times, indicating that this invention has good cloud base identification capabilities under thin, fragmented, and weak-signal cloud conditions. During precipitation ( Figure 8 From 16:30 to 18:30, the laser wind radar could still identify the low-level cloud base, while the Ka-band millimeter-wave cloud radar was affected by precipitation echoes, resulting in a significant enhancement of near-surface echoes and making it difficult to accurately determine the cloud base height. After the precipitation ended, the KaCR gradually recovered its ability to identify upper-level clouds, and its inversion results gradually became consistent with those of the laser wind radar. On the other hand, when the cloud layer was thick and there were multiple layers of cloud cover (such as...), the cloud base height could not be accurately determined. Figure 8 At time T5, the laser wind radar failed to identify high clouds at approximately 8 km, while KaCR provided the corresponding cloud layer results, indicating that under conditions of thick cloud layers and multiple layers of cloud cover, a single laser wind radar still faces limitations in detectability and the risk of missed detections in upper-level clouds.

[0071] Figure 9This paper presents a comprehensive comparison of the cloud base height identification results of laser wind-measuring radar and Ka-band millimeter-wave cloud radar during the case study period. This comparison illustrates the detectability of upper-layer clouds by laser wind-measuring radar under multi-layer cloud conditions and the potential risk of missed detections. As shown in the figure, under single-layer cloud conditions, the number of samples where both laser wind-measuring radar and Ka-band millimeter-wave cloud radar simultaneously identified the cloud base was 1110, while the number of missed samples by laser wind-measuring radar was 384. However, under multi-layer cloud conditions, the number of missed samples by laser wind-measuring radar increased to 4129, significantly higher than in the single-layer cloud case. This result indicates that the upper-layer cloud identification capability of laser wind-measuring radar is more severely limited under multi-layer cloud conditions. When the lower-layer cloud is thicker, has a higher particle concentration, or a greater optical thickness, the laser beam energy attenuates rapidly during upward propagation, resulting in a significant reduction in the spectral intensity (SI) upon reaching the upper-layer cloud, thus weakening the ability to identify the upper-layer cloud base. Therefore, under multi-layer cloud conditions, outputting only the cloud base height is insufficient to reflect the reliability of the upper-layer cloud identification results. To address this issue, this invention proposes a method for identifying the detectability and missed detection risk of upper cloud bases. This method provides a reliability or detectability indicator for the identification results of cloud bases other than the first cloud layer, thus offering risk warnings when upper cloud signals are affected by obstruction or attenuation from lower clouds. Compared to simply providing the cloud base height, this method better reflects the actual detection capabilities of a single laser wind-measuring radar under multi-layered cloud conditions, and also improves the interpretability and operational application value of the identification results.

[0072] Although specific embodiments of the invention have been described in detail with reference to the accompanying drawings, this should not be construed as limiting the scope of protection of this patent. Various modifications and variations that can be made by a person skilled in the art without inventive effort within the scope described in the claims still fall within the scope of protection of this patent.

Claims

1. A method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar, characterized in that, Includes the following steps: S1. Acquire single-laser wind radar data, preprocess the data, and construct a clear sky background reference profile; S2. Calculate the sudden increase in spectral intensity relative to the corresponding clear sky background reference profile at each observation time, and preliminarily extract candidate cloud bases; S3. By using the minimum spacing constraint of cloud layers, the background constraint of the distance library below the candidate cloud base, and the time continuity correction, perform layered verification and layer matching on the candidate cloud base, and output the upper-layer candidate cloud base. S4. Based on the interval integral spectrum intensity index and the signal-to-noise ratio of the candidate cloud base neighborhood, the detectability and missed detection risk of the upper cloud base under multi-layer cloud conditions are judged, and the judgment results are output.

2. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, S1 specifically includes: Acquire spectral intensity data and signal-to-noise ratio data output by a single-laser wind-measuring radar during a continuous observation period; Preprocessing includes: performing two-dimensional smoothing on the spectral intensity data to obtain smoothed spectral intensity; In addition, the effectiveness of the time-altitude distance database is screened based on the signal-to-noise ratio data. Specifically, for distance databases with signal-to-noise ratio data below the preset effective detection threshold, they are marked as low-confidence data points and are not used as cloud bottom candidate points alone. After preprocessing, an observation period under cloudless weather conditions was selected, and a clear sky background reference profile was constructed based on the smoothed spectral intensity data.

3. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 2, characterized in that, In S1, the clear sky background reference profile is a dynamic clear sky background reference profile constructed over a time period, and its construction process is as follows: The day is divided into multiple time periods. Smoothed spectral intensity data corresponding to cloudless samples within a time period are extracted, and the average value is calculated for each altitude distance library to serve as the clear sky background reference value for that altitude. The clear sky background reference values ​​of each altitude distance library are arranged by altitude to form the clear sky background reference profile corresponding to that time period.

4. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, S2 specifically includes: The sudden increase in spectral intensity relative to clear sky background is calculated by subtracting the clear sky background reference value for the corresponding time period from the smoothed spectral intensity data at each observation time. The initial screening criteria are as follows: when the sudden increase at a certain height exceeds the preset sudden increase threshold, and the height is not lower than the minimum effective detection height, and the signal-to-noise ratio data at or in the neighborhood of that height is not lower than the preset threshold, or when it has synchronous enhancement characteristics relative to the background distance library below it, the height is marked as a candidate cloud base location. At the same observation time, candidate cloud bases that meet the preliminary screening conditions are searched from bottom to top along a certain height direction, thus obtaining a set of candidate cloud bases.

5. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, In S3, layered filtering is performed using cloud minimum spacing constraints, including: For multiple candidate cloud bases identified at the same time, if the vertical distance between two adjacent candidate cloud bases is less than the minimum cloud spacing threshold, they are considered to belong to the same cloud layer's signal fluctuations, local enhancements, or fragmented cloud disturbances, and are merged into the same cloud layer candidate result, instead of being output as two independent cloud layers; if the vertical distance between two adjacent candidate cloud bases is not less than the minimum cloud spacing threshold, they are retained as candidate cloud bases of different cloud layers.

6. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, In step S3, background constraint judgment is performed on the spectral intensity of the continuous distance library below the candidate cloud base, including: Based on cloudless samples, the mean and standard deviation of background spectral intensity at various altitudes were statistically analyzed, and the upper limit of normal fluctuation of background spectral intensity was calculated. Two consecutive distance libraries below the candidate cloud base height are selected. If the smoothed spectral intensity of both distance libraries is lower than the normal fluctuation upper limit of the corresponding height, the candidate cloud base is retained; otherwise, it is considered that there is an abnormal enhancement below the candidate cloud base, and it is removed or marked as a low-confidence candidate cloud base.

7. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, In step S3, the layer-level correction of the multi-layer cloud candidate results based on temporal continuity includes: For candidate cloud base identification results at adjacent observation times, the cloud layers are numbered from low to high altitude, and the deviation of the candidate cloud base at the current time from the corresponding cloud base at the previous and next adjacent times is compared. If the height deviation between the current candidate cloud base and the corresponding layer cloud base does not exceed the continuity threshold, they are considered to belong to the same layer cloud. If the height deviation exceeds the continuity threshold, the current candidate cloud base is matched with cloud bases of other layers. If the height deviation with cloud bases of other layers is less than the continuity threshold, the candidate cloud base is assigned to the corresponding layer. If it cannot be matched with any layer, the candidate cloud base is marked as a low-confidence candidate point or is removed.

8. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, S4 specifically includes: The interval integral spectral intensity index is calculated based on the smoothed spectral intensity. When the interval integral spectral intensity index is lower than the preset attenuation risk threshold, and the target upper candidate cloud base meets the requirements of spectral intensity surge, signal-to-noise ratio effectiveness and time continuity, the upper cloud base is determined to be detectable, and a high-confidence label is output. When the interval integral spectral intensity index is higher than the preset attenuation risk threshold, or when the signal-to-noise ratio of the target upper candidate cloud base is lower than the effective detection condition and the spectral intensity surge characteristics are unstable, the detectability of the upper cloud base is determined to be limited, and a low confidence or missed detection risk indicator is output.

9. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 1, characterized in that, S4 specifically includes: Calculate the interval integral spectral intensity index based on the smoothed spectral intensity; The spectral intensity abrupt change, candidate cloud base neighborhood signal-to-noise ratio, temporal continuity deviation, and interval integral spectral intensity index are normalized and fused to obtain a comprehensive credibility scoring function; Based on the relationship between the comprehensive credibility scoring function and the preset level threshold, the identification results are divided into categories of high credibility, medium credibility, low credibility, or limited detectability.

10. The method for identifying the base of weak-signal clouds and multi-layered clouds in high-altitude areas based on a single laser wind-measuring radar according to claim 8 or 9, characterized in that, In step S4, the smoothed spectral intensity calculation interval integral spectral intensity exponent is expressed as: In the formula, The interval integral spectral intensity index; Indicates the height of the candidate cloud base above the target; To represent the height distance library, This represents the height of the first cloud base at the current moment. This represents the smoothed spectral intensity; Time period The outline of a clear sky background.

Images