A CNN-based dynamic control method for laser beam parameters of three-dimensional laser wind radar

Abstract

The present invention belongs to the field of laser wind measurement technology. A method for dynamically controlling laser beam parameters of a three-dimensional laser wind measurement radar based on CNN is provided, comprising: controlling the three-dimensional laser wind measurement radar to emit a first group of laser beams into the wind measurement area where it is arranged, and predicting the dust particle size grade; obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located, and using a CNN-based deep processing model to analyze and process them, and predicting the wind speed change curve and the dust particle variation coefficient; determining the laser beam adjustment timing according to the wind speed change curve, and matching the dust particle variation coefficient and the dust particle size grade to obtain the target diameter of the laser beam; when the laser beam adjustment timing is reached, controlling the three-dimensional laser wind measurement radar to emit a second group of laser beams into the wind measurement area where it is arranged, and measuring the wind speed data in the wind measurement area. The present invention realizes more accurate laser wind measurement.

Description

Technical Field

[0001] The invention belongs to the technical field of laser wind measurement, and in particular relates to a CNN-based three-dimensional laser wind measurement radar laser beam parameter dynamic control method. Background Art

[0002] Laser wind measurement technology is an advanced method that uses laser to measure wind speed and direction. It is mainly used in meteorological and climate monitoring, weather detection, air pollution tracking, atmospheric research and wind energy utilization.

[0003] The working principle of the laser wind radar is as follows: the laser wind radar uses the Doppler heterodyne method to measure the wind speed and direction by measuring the Doppler frequency shift of the laser backscatter echo of the particles in the air. When the laser passes through the atmosphere, the aerosol particles in the atmosphere will produce a scattering effect on the incident light, and the running aerosol particles will cause the frequency of the scattered light to produce a Doppler frequency shift effect. By comparing the frequency difference between the reference light (unscattered emitted light) and the scattered light in the receiver, the airflow speed carrying the aerosol particles can be determined.

[0004] Laser wind radar measures wind speed and direction by emitting laser pulses and receiving backscattered signals from particles in the atmosphere. The size of dust particles directly affects the degree of scattering and absorption of the laser beam. The diameter of the laser beam of laser wind radar in the prior art is almost always fixed, but the size of dust particles in the area where the laser wind radar is located changes dynamically. Obviously, the change in the size of dust particles will affect the wind measurement accuracy of the laser wind radar.

[0005] How to further improve the wind measurement accuracy of laser wind radar is a technical issue that currently needs further improvement. Summary of the invention

[0006] In order to solve the technical problems existing in the above background technology, the present invention provides a method for dynamically controlling laser beam parameters of a three-dimensional laser wind radar based on CNN, comprising the following steps:

[0007] Controlling the three-dimensional laser wind measurement radar to emit a first group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and predicting the dust particle size grade according to the scattered light angle and scattered light intensity corresponding to the first group of laser beams;

[0008] Obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located, using a CNN-based deep processing model to analyze and process the real-time wind measurement data, the geographical area big data, and the meteorological big data, and predicting a wind speed change curve and a dust particle change coefficient;

[0009] Determine the laser beam adjustment timing according to the wind speed variation curve, and determine the target diameter of the laser beam according to the dust particle variation coefficient and the dust particle size grade matching;

[0010] When the laser beam adjustment timing is reached, the three-dimensional laser wind measurement radar is controlled to emit a second group of laser beams into the wind measurement area where it is arranged, and the wind speed data in the wind measurement area is measured according to the second group of laser beams; wherein the second group of laser beams includes a laser beam of the target diameter.

[0011] In some embodiments, the step of obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located includes:

[0012] Calculating the geographical area and the aspect ratio of the wind measurement area where the three-dimensional laser wind measurement radar is located, and calculating the interference assessment value according to the geographical area and the aspect ratio;

[0013] If the interference assessment value is higher than the interference threshold, a domain expansion coefficient is calculated according to the difference between the interference assessment value and the interference threshold, and the domain expansion coefficient is used to perform domain expansion processing on the wind measurement area to obtain the geographical area big data and the meteorological big data of the wind measurement area after the domain expansion processing;

[0014] If the interference assessment value is not higher than the interference threshold, the geographical area big data and the meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located are obtained.

[0015] In some embodiments, a CNN-based deep processing model is used to analyze and process the real-time wind measurement data, the geographic area big data, and the meteorological big data to predict the wind speed change curve and the dust particle change coefficient, including:

[0016] Performing statistical analysis and fitting processing on the meteorological big data to obtain the wind speed change curve and output it;

[0017] The wind speed variation curve, the real-time wind measurement data and the geographical area big data are input into a CNN-based deep processing model, and the dust particle variation coefficient is calculated according to the output result of the deep processing model.

[0018] In some embodiments, the calculating the dust particle variation coefficient according to the output result of the deep processing model includes:

[0019] The deep processing model outputs the predicted preliminary dust particle variation coefficient;

[0020] Identify several target time periods in the wind speed change curve that are within the analysis time period, calculate the overall time span of these target time periods, and calculate the optimization coefficient based on the overall time span; wherein the target time period refers to a set time window in the wind speed change curve where the wind speed change amplitude is higher than a preset value;

[0021] The optimization coefficient is multiplied by the preliminary dust particle variation coefficient to obtain the dust particle variation coefficient.

[0022] In some embodiments, obtaining a target diameter of the laser beam according to the matching of the dust particle variation coefficient and the dust particle size grade includes:

[0023] Multiplying the dust particle variation coefficient by the dust particle size grade to calculate a target dust particle size grade;

[0024] The target dust particle size grade is matched with a comparison table obtained in advance to obtain the target diameter, and the target diameter is evenly divided to obtain multiple target sub-diameters; wherein the largest target sub-diameter is the target diameter, and the smallest target sub-diameter is the diameter of the first group of laser beams.

[0025] In some embodiments, controlling the three-dimensional laser wind measurement radar to emit a second group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and measuring the wind speed data in the wind measurement area according to the second group of laser beams includes:

[0026] Controlling the three-dimensional laser wind measurement radar to respectively emit laser beams in the second group of laser beams that meet the diameters of the target sub-diameters into the wind measurement area where the three-dimensional laser wind measurement radar is arranged;

[0027] Preliminary wind speed data are calculated based on the scattered light beams of the laser light beams of each target sub-diameter, and weighted calculation is performed on each of the preliminary wind speed data to obtain the wind speed data in the wind measurement area.

[0028] In some embodiments, the weighted weights of the preliminary wind speed data corresponding to each target sub-diameter decrease in sequence from the largest target sub-diameter to the target sub-diameter.

[0029] The beneficial effects of the present invention are as follows: the present invention conducts a comprehensive analysis of geographical area big data and meteorological big data to obtain a target diameter of a laser beam that is more compatible with the predicted dust particles, and a laser beam adjustment timing for switching to use the laser beam of the target diameter. The wind speed data measured thereby is more accurate, which is beneficial for management personnel to accurately make relevant analysis decisions and disposal plans based on the wind speed data. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0031] Figure 1 It is a flow chart of a method for dynamically controlling laser beam parameters of a three-dimensional laser wind measuring radar based on CNN disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0033] Laser wind radar measures wind speed and direction by emitting laser pulses and receiving backscattered signals from particles in the atmosphere. The size of dust particles directly affects the degree of scattering and absorption of the laser beam. The diameter of the laser beam of laser wind radar in the prior art is almost always fixed, but the size of dust particles in the area where the laser wind radar is located changes dynamically. Obviously, the change in the size of dust particles will affect the wind measurement accuracy of the laser wind radar. How to further improve the wind measurement accuracy of laser wind radar is a technical problem that currently needs further improvement.

[0034] like Figure 1 As shown, the embodiment of the present invention discloses a method for dynamically controlling laser beam parameters of a three-dimensional laser wind radar based on CNN, comprising the following steps:

[0035] Controlling the three-dimensional laser wind measurement radar to emit a first group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and predicting the dust particle size grade according to the scattered light angle and scattered light intensity corresponding to the first group of laser beams;

[0036] Obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located, using a CNN-based deep processing model to analyze and process the real-time wind measurement data, the geographical area big data, and the meteorological big data, and predicting a wind speed change curve and a dust particle change coefficient;

[0037] Determine the laser beam adjustment timing according to the wind speed variation curve, and determine the target diameter of the laser beam according to the dust particle variation coefficient and the dust particle size grade matching;

[0038] When the laser beam adjustment timing is reached, the three-dimensional laser wind measurement radar is controlled to emit a second group of laser beams into the wind measurement area where it is arranged, and the wind speed data in the wind measurement area is measured according to the second group of laser beams; wherein the second group of laser beams includes a laser beam of the target diameter.

[0039] In order to solve the above technical problems, the present invention sets a three-dimensional laser wind measurement radar to first emit a first group of laser beams into the wind measurement area where it is arranged. The first group of laser beams will be scattered when encountering dust particles in the air. The size grade of the dust particles in the wind measurement area in the current period can be predicted based on the angle and intensity of the backscattered echo. The method of obtaining the dust particle size grade in this step can be referred to another prior invention patent application of the applicant (application number: 202411125047.0, a method and system for intelligent control of the laser beam of a three-dimensional laser wind measurement radar). The technical content of the prior invention patent application, at least related to the determination of the above-mentioned dust particle size grade, is introduced into the technical solution of the present invention, and its technical content will not be repeated here.

[0040] Then, the geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located are further obtained, wherein the geographical area big data mainly refers to the ground particle situation in the corresponding wind measurement area, such as the average size grade of ground particles, vegetation coverage, etc.; the meteorological big data mainly refers to the wind speed change law data of the corresponding wind measurement area in different periods of the day under each season attribute, such as the wind speed in the morning, the wind speed at noon, the wind speed in the evening, etc. The real-time wind measurement data, geographical area big data, and meteorological big data obtained are analyzed and processed using a CNN-based deep processing model to predict the wind speed change curve and the dust particle variation coefficient. Among them, the real-time wind measurement data is the wind speed data of the current location measured autonomously by the three-dimensional laser wind measurement radar. The dust particle variation coefficient predicted by the deep processing model is used to characterize the variation range of the size grade of dust particles in the area where the three-dimensional laser wind measurement radar is located in the next period of time, that is, the larger the absolute value of the variation range, the larger the corresponding dust particle variation coefficient, and vice versa, the smaller the absolute value of the variation range, the smaller the corresponding dust particle variation coefficient. It should be noted that the variation range in the present invention refers to positive variation range and negative variation range, that is, the size level of dust particles in the area where the three-dimensional laser wind measurement radar is located may become larger or smaller.

[0041] Next, the wind speed change curve can be used to intuitively analyze the change of wind speed over time in the wind measurement area. According to the wind speed change curve, the timing of laser beam adjustment can be determined. For example, when the wind speed change amplitude in a certain period is higher than the first value and the duration is higher than the second value, it is determined that the diameter of the laser beam needs to be adjusted for this period. Furthermore, the target diameter of the laser beam is obtained according to the dust particle variation coefficient and the dust particle size grade matching.

[0042] Finally, when the laser beam adjustment timing is reached, the three-dimensional laser wind measurement radar is controlled to emit a second group of laser beams including the laser beams of the target diameter into the wind measurement area where it is arranged, so as to measure the wind speed data in the wind measurement area according to the second group of laser beams. Since the second group of laser beams includes the laser beams of the target diameter, the adaptability thereof to the size grade of the dust particles in the time period is higher, so the accuracy of the measured wind speed data is higher.

[0043] Therefore, the present invention conducts a comprehensive analysis of geographical area big data and meteorological big data to obtain a target diameter of the laser beam that is more compatible with the predicted dust particles, and the laser beam adjustment timing for switching to use the laser beam of the target diameter. The wind speed data measured thereby is more accurate, which is conducive to managers making accurate relevant analysis decisions and disposal plans based on the wind speed data.

[0044] In some embodiments, the step of obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located includes:

[0045] Calculating the geographical area and the aspect ratio of the wind measurement area where the three-dimensional laser wind measurement radar is located, and calculating the interference assessment value according to the geographical area and the aspect ratio;

[0046] If the interference assessment value is higher than the interference threshold, a domain expansion coefficient is calculated according to the difference between the interference assessment value and the interference threshold, and the domain expansion coefficient is used to perform domain expansion processing on the wind measurement area to obtain the geographical area big data and the meteorological big data of the wind measurement area after the domain expansion processing;

[0047] If the interference assessment value is not higher than the interference threshold, the geographical area big data and the meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located are obtained.

[0048] In this embodiment, since the size, contour distribution, etc. of the geographical area of ​​the wind measurement area where the three-dimensional laser wind measurement radar is located are different, this leads to different degrees of influence of the dust particle level in the wind measurement area on the meteorological fluctuations in the adjacent area. In this regard, the present invention is configured to first calculate the geographical area and aspect ratio of the wind measurement area where the three-dimensional laser wind measurement radar is located, and calculate the interference assessment value based on the geographical area and aspect ratio. The interference assessment value is used to characterize the degree to which the dust particle level in the wind measurement area is affected by the meteorological fluctuations in the adjacent area due to its geographical distribution. The higher the interference assessment value, the greater the corresponding degree of influence, and vice versa. The calculation formula of the interference assessment value is, for example:

[0049] Interference assessment value = f1 (geographic area) / f2 (aspect ratio)

[0050] In the above formula, f1 and f2 are corresponding conversion functions, which are used to normalize the geographical area and aspect ratio respectively, so as to use the ratio between the dimensionless values ​​after normalization to express the magnitude of the above influence. Obviously, when the geographical area of ​​the wind measurement area is smaller and the aspect ratio is larger, it indicates that the narrowness of the wind measurement area is higher, and it is more susceptible to the influence of the adjacent area; when the geographical area of ​​the wind measurement area is larger and the aspect ratio is smaller, it indicates that the circularity of the wind measurement area is higher, and it is less susceptible to the influence of the adjacent area.

[0051] Furthermore, when the interference assessment value is higher than the interference threshold, the present invention further calculates the expansion coefficient (for example, 1.5, 2.0) based on the difference between the interference assessment value and the interference threshold, and then expands the wind measurement area based on the expansion coefficient, and obtains the geographical area big data and meteorological big data of the wind measurement area after the expansion. The higher the accuracy of the dust particle variation coefficient obtained by subsequent analysis based on the geographical area big data and meteorological big data of a larger range. When the interference assessment value is not higher than the interference threshold, it indicates that the wind measurement area is less affected by the adjacent area. At this time, there is no need to expand the wind measurement area, and the geographical area big data and meteorological big data of the wind measurement area can be directly obtained.

[0052] In some embodiments, a CNN-based deep processing model is used to analyze and process the real-time wind measurement data, the geographic area big data, and the meteorological big data to predict the wind speed change curve and the dust particle change coefficient, including:

[0053] Performing statistical analysis and fitting processing on the meteorological big data to obtain the wind speed change curve and output it;

[0054] The wind speed variation curve, the real-time wind measurement data and the geographical area big data are input into a CNN-based deep processing model, and the dust particle variation coefficient is calculated according to the output result of the deep processing model.

[0055] In this embodiment, the meteorological big data includes the wind speed variation regularity data of the wind measurement area at different time periods within a day under each season attribute, such as the wind speed in the morning, the wind speed at noon, the wind speed in the evening, etc. The above-mentioned wind speed variation curve can be obtained by statistically analyzing these wind speed data and performing curve fitting processing.

[0056] The obtained wind speed change curve, together with the real-time wind measurement data and geographic area big data, is input into the pre-built deep processing model. The deep processing model first extracts the convolution features from the wind speed change curve, real-time wind measurement data and geographic area big data, and then performs pooling, full connection and other processing on the convolution features to obtain the predicted dust particle variation coefficient for the next period compared to the current period. The dust particle variation coefficient is a value around 1. When the variation range is positive, the dust particle variation coefficient is a value greater than 1. When the variation range is negative, the dust particle variation coefficient is a value less than 1. When the variation range is close to 0, the dust particle variation coefficient is equal to 1. It should be noted that due to the change in wind speed, the ground particles in the wind measurement area are blown up or some of the blown particles can no longer be blown up, which causes the size grade of dust particles in the air to change. The deep processing model can derive the degree of this change based on the above data analysis, that is, the dust particle variation coefficient; moreover, whether the particles can be blown up should also consider the vegetation coverage rate. The higher the vegetation coverage rate, the lower the probability of the particles in the wind measurement area being blown up and the smaller the number of particles blown up. Conversely, the higher the probability of the particles in the wind measurement area being blown up and the larger the number of particles blown up.

[0057] The deep processing model includes an input layer, multiple convolutional layers, a pooling layer, a fully connected layer, and an output layer connected in sequence. The activation function used in the convolutional layer is ReLu; the pooling layer uses maximum pooling; the fully connected layer is located in the last N layers of CNN (N>=2), uses Sigmoid as the activation function, and uses dropout technology to randomly discard some neurons to reduce overfitting. In addition, the format of the training data used for training the deep processing model is [wind speed change curve, real-time wind measurement data, geographic area big data, label], where the label is the measured dust particle variation coefficient.

[0058] In some embodiments, the calculating the dust particle variation coefficient according to the output result of the deep processing model includes:

[0059] The deep processing model outputs the predicted preliminary dust particle variation coefficient;

[0060] Identify several target time periods in the wind speed change curve that are within the analysis time period, calculate the overall time span of these target time periods, and calculate the optimization coefficient based on the overall time span; wherein the target time period refers to a set time window in the wind speed change curve where the wind speed change amplitude is higher than a preset value;

[0061] The optimization coefficient is multiplied by the preliminary dust particle variation coefficient to obtain the dust particle variation coefficient.

[0062] In this embodiment, the wind speed change curve obtained above represents the predicted wind speed change of the wind measurement area in the next period of time, but it is not necessarily equivalent to the actual wind speed change. In order to ensure the confidence of the obtained dust particle variation coefficient, the present invention further analyzes the wind speed change curve by time window with a preset time window (for example, the time window is 0.5 hours, the analysis period is 6 hours, and the number of time windows is 12). If the wind speed change amplitude corresponding to the wind speed change curve segment in the time window is higher than the preset value, it is used as the target period. Then calculate the overall time span of these target periods, and calculate the optimization coefficient based on the overall time span. The optimization coefficient and the overall time span are in a negative correlation. In other words, the larger the overall time span of the target period, the lower the wind speed fluctuation frequency within the analysis time, and vice versa. When the wind speed fluctuation frequency is higher, the uncertainty of the wind speed in the corresponding time window is higher, and the corresponding optimization coefficient is set larger, and vice versa, the optimization coefficient is set smaller.

[0063] For example, the optimization coefficient = a + (b / overall time span), where a is the basic value of the optimization coefficient, and b is the overall time span of the wind speed variation curve for all target periods, which can be regarded as a constant.

[0064] It should be noted that the start time of the target period can be used to determine the subsequent laser beam adjustment timing.

[0065] In some embodiments, obtaining a target diameter of the laser beam according to the matching of the dust particle variation coefficient and the dust particle size grade includes:

[0066] Multiplying the dust particle variation coefficient by the dust particle size grade to calculate a target dust particle size grade;

[0067] The target dust particle size grade is matched with a comparison table obtained in advance to obtain the target diameter, and the target diameter is evenly divided to obtain multiple target sub-diameters; wherein the largest target sub-diameter is the target diameter, and the smallest target sub-diameter is the diameter of the first group of laser beams.

[0068] In this embodiment, after the dust particle variation coefficient is calculated in the above manner, it is multiplied by the predicted dust particle size grade to obtain the final target dust particle size grade.

[0069] At the same time, the diameter of the best laser beam adapted to different dust particle size classes is determined in advance, that is, the laser beam of this diameter is used to irradiate the atmosphere of the dust particle size class, and the wind measurement data obtained is closest to the real wind speed data. The above-mentioned comparison table is obtained through multiple measurements. The determination process of the comparison table is also based on the above-mentioned prior invention patent application. The technical content of the prior invention patent application at least related to the determination of the above-mentioned comparison table is introduced into the technical solution of the present invention, and its technical content will not be repeated here.

[0070] In some embodiments, controlling the three-dimensional laser wind measurement radar to emit a second group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and measuring the wind speed data in the wind measurement area according to the second group of laser beams includes:

[0071] Controlling the three-dimensional laser wind measurement radar to respectively emit laser beams in the second group of laser beams that meet the diameters of the target sub-diameters into the wind measurement area where the three-dimensional laser wind measurement radar is arranged;

[0072] Preliminary wind speed data are calculated based on the scattered light beams of the laser light beams of each target sub-diameter, and weighted calculation is performed on each of the preliminary wind speed data to obtain the wind speed data in the wind measurement area.

[0073] In this embodiment, the three-dimensional laser wind measurement radar is controlled to emit laser beams that meet multiple target sub-diameters into the wind measurement area, that is, a laser beam of one target sub-diameter is emitted each time, and preliminary wind speed data is calculated based on the scattered beam of the laser beam of each target sub-diameter. Finally, the preliminary wind speed data are weighted and fused to obtain the wind speed data in the wind measurement area.

[0074] In some embodiments, the weighted weights of the preliminary wind speed data corresponding to each target sub-diameter decrease in sequence from the largest target sub-diameter to the target sub-diameter.

[0075] In this embodiment, considering that the largest target sub-diameter (i.e., the target diameter) best matches the predicted target dust particle size grade, the weighted weights of the preliminary wind speed data corresponding to each target sub-diameter are set to decrease in sequence, that is, the weighted weight of the preliminary wind speed data corresponding to the largest target sub-diameter is the largest, and the weighted weight of the preliminary wind speed data corresponding to the smallest target sub-diameter is the smallest.

[0076] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of this disclosure can be achieved, and this document is not limited here.

[0077] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A CNN-based three-dimensional laser wind radar laser beam parameter dynamic control method, characterized in that: The steps include: Controlling the three-dimensional laser wind measurement radar to emit a first group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and predicting the dust particle size grade according to the scattered light angle and scattered light intensity corresponding to the first group of laser beams; Obtaining geographical area big data and meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located, using a CNN-based deep processing model to analyze and process the real-time wind measurement data, the geographical area big data, and the meteorological big data, and predicting a wind speed change curve and a dust particle change coefficient; Determine the laser beam adjustment timing according to the wind speed variation curve, and determine the target diameter of the laser beam according to the dust particle variation coefficient and the dust particle size grade matching; When the laser beam adjustment timing is reached, the three-dimensional laser wind measurement radar is controlled to emit a second group of laser beams into the wind measurement area where it is arranged, and the wind speed data in the wind measurement area is measured according to the second group of laser beams; wherein the second group of laser beams includes a laser beam of the target diameter.

2. The method for dynamic control of laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 1 is characterized in that: Obtain geographical area big data and meteorological big data of the wind measurement area where the 3D laser wind measurement radar is located, including: Calculating the geographical area and the aspect ratio of the wind measurement area where the three-dimensional laser wind measurement radar is located, and calculating the interference assessment value according to the geographical area and the aspect ratio; If the interference assessment value is higher than the interference threshold, a domain expansion coefficient is calculated according to the difference between the interference assessment value and the interference threshold, and the domain expansion coefficient is used to perform domain expansion processing on the wind measurement area to obtain the geographical area big data and the meteorological big data of the wind measurement area after the domain expansion processing; If the interference assessment value is not higher than the interference threshold, the geographical area big data and the meteorological big data of the wind measurement area where the three-dimensional laser wind measurement radar is located are obtained.

3. The method for dynamically controlling laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 2 is characterized in that: The real-time wind measurement data, the geographical area big data, and the meteorological big data are analyzed and processed using a CNN-based deep processing model to predict the wind speed change curve and the dust particle change coefficient, including: Performing statistical analysis and fitting processing on the meteorological big data to obtain the wind speed change curve and output it; The wind speed variation curve, the real-time wind measurement data and the geographical area big data are input into a CNN-based deep processing model, and the dust particle variation coefficient is calculated according to the output result of the deep processing model.

4. The method for dynamic control of laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 3 is characterized in that: The dust particle variation coefficient is calculated based on the output result of the deep processing model, including: The deep processing model outputs the predicted preliminary dust particle variation coefficient; Identify several target time periods in the wind speed change curve that are within the analysis time period, calculate the overall time span of these target time periods, and calculate the optimization coefficient based on the overall time span; wherein the target time period refers to a set time window in the wind speed change curve where the wind speed change amplitude is higher than a preset value; The optimization coefficient is multiplied by the preliminary dust particle variation coefficient to obtain the dust particle variation coefficient.

5. The method for dynamic control of laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 4 is characterized in that: The target diameter of the laser beam is obtained according to the matching of the dust particle variation coefficient and the dust particle size grade, including: Multiplying the dust particle variation coefficient by the dust particle size grade to calculate a target dust particle size grade; The target dust particle size grade is matched with a comparison table obtained in advance to obtain the target diameter, and the target diameter is evenly divided to obtain multiple target sub-diameters; wherein the largest target sub-diameter is the target diameter, and the smallest target sub-diameter is the diameter of the first group of laser beams.

6. The method for dynamically controlling laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 5 is characterized in that: Controlling the three-dimensional laser wind measurement radar to emit a second group of laser beams into the wind measurement area where the three-dimensional laser wind measurement radar is arranged, and measuring the wind speed data in the wind measurement area according to the second group of laser beams, including: Controlling the three-dimensional laser wind measurement radar to respectively emit laser beams in the second group of laser beams that meet the diameters of the target sub-dimensions into the wind measurement area where the three-dimensional laser wind measurement radar is arranged; Preliminary wind speed data are calculated based on the scattered light beams of the laser light beams of each target sub-diameter, and weighted calculation is performed on each of the preliminary wind speed data to obtain the wind speed data in the wind measurement area.

7. The method for dynamic control of laser beam parameters of a three-dimensional laser wind radar based on CNN according to claim 6 is characterized in that: The weighted weights of the preliminary wind speed data corresponding to each target sub-diameter decrease in sequence from the largest target sub-diameter to the target sub-diameter.

Images