Aircraft icing early warning method

Abstract

The invention discloses an airplane icing early warning method, and relates to the technical field of icing early warning. According to the airplane icing early warning method, cloud picture data and cloud product file data of a set airplane are obtained, a cloud field operation state feature set of the set airplane is extracted based on the cloud product data of the set airplane, and a cloud liquid water path value of the set airplane is analyzed; based on a pre-trained cloud picture error mapping model and in combination with cloud picture data of the set airplane, analyzing a cloud type temperature correction factor of the set airplane, and in combination with the cloud field operation state feature set, analyzing an optimized cloud top temperature value of the set airplane; based on the cloud liquid water path value of the set airplane, the cloud top temperature value is optimized, and the ice accretion index of the set airplane is analyzed, ice accretion early warning processing is performed on the set airplane based on the ice accretion index, so that the finally generated ice accretion index result has higher environmental adaptability, dynamic correction and high-precision characterization of the ice accretion risk are realized, and the method is suitable for large-scale popularization and application. And thus, the accuracy of ice accumulation early warning is improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of ice accretion early warning, in particular to an aircraft ice accretion early warning method. BACKGROUND

[0002] When an aircraft passes through a convective cloud zone or a humid meteorological environment such as cumulus and stratocumulus clouds, a large amount of supercooled water droplets are often contained in the external air. When these water droplets freeze rapidly after contacting the aircraft body, ice layers are formed on the surface of the wings, air intakes or measuring devices, which leads to reduced lift, increased resistance and instrument errors. In severe cases, it may even cause engine flameout or stall accidents.

[0003] In recent years, with the development of multi-spectral remote sensing technology of meteorological satellites, the microphysical properties of cloud layers can be obtained through inversion. However, the existing ice accretion identification method based on satellite data fails to fully consider the radiation deviation between cloud top brightness temperature and true cloud temperature under different cloud types, and lacks a mechanism for quantitatively correcting satellite inversion errors, resulting in an underestimation of temperature, which makes ice accretion early warning prone to misjudgment.

[0004] The prior art such as the patent application with the publication number CN114927009A discloses an aviation flight dangerous weather diagnosis and analysis system, which comprises: an aircraft ice accretion diagnosis module and a cloud layer analysis module. The aircraft ice accretion diagnosis module is used to obtain the ice accretion condition of the aircraft according to an ice accretion algorithm. The ice accretion algorithm includes at least one of the following ice accretion algorithms: IC ice accretion algorithm, RAP ice accretion algorithm and RAOB ice accretion algorithm. The cloud layer analysis module is used to determine the cloud layer condition of the sounding layer according to the environmental temperature and dew point temperature of the sounding layer. The aviation flight dangerous weather diagnosis and analysis system provided by the application can comprehensively judge the ice accretion condition and cloud layer condition that may occur during aviation flight by using the ice accretion algorithm and the cloud layer analysis algorithm, so as to warn the main factors of dangerous aviation flight and improve the warning level of aviation flight.

[0005] Based on the above scheme, the limitations of the prior art at least include the following problems: the prior art does not fully utilize high spatiotemporal resolution cloud product file data, lacks comprehensive analysis of cloud top temperature, cloud optical thickness, effective particle radius and other parameters, and is difficult to accurately characterize the thermal state and ice accretion environment of the cloud field, and is difficult to adaptively correct according to different cloud types, resulting in large cloud temperature estimation deviation and insufficient ice accretion index calculation precision. In addition, the prior art lacks multi-source joint analysis of cloud type temperature correction system, cloud product file data and flight parameters, and is difficult to realize dynamic response analysis of cloud liquid water path, cloud top temperature and flight height difference, which leads to insufficient accuracy of ice accretion early warning. SUMMARY

[0006] In view of the defects of the prior art, the aircraft icing early warning method is provided to solve the problem of insufficient icing early warning accuracy caused by the lack of multi-source fusion in the prior art.

[0007] To achieve the above object, the present application is implemented by the following technical scheme: An aircraft icing early warning method, comprising the following steps: obtaining cloud image data of a set aircraft, cloud product file data, extracting cloud field running state feature set of the set aircraft based on the cloud product file data of the set aircraft, and analyzing cloud liquid water path value of the set aircraft; analyzing cloud type temperature correction factor of the set aircraft based on a pre-trained cloud image error mapping model and in combination with the cloud image data of the set aircraft, and analyzing optimized cloud top temperature value of the set aircraft in combination with the cloud field running state feature set; analyzing icing index of the set aircraft based on the cloud liquid water path value and the optimized cloud top temperature value of the set aircraft; and performing icing early warning processing on the set aircraft based on the icing index.

[0008] Further, the specific steps of extracting the cloud field running state feature set of the set aircraft are as follows: performing analysis processing on the cloud product file data of the set aircraft; obtaining the longitude and latitude of the set aircraft and performing matching processing on the cloud product file data of the set aircraft after the analysis processing to obtain grid data of the set aircraft; and outputting the cloud field running state feature set of the set aircraft including cloud top temperature value, effective particle radius value, cloud optical thickness value and cloud top height value based on the grid data of the set aircraft.

[0009] Further, the specific steps of analyzing the cloud liquid water path value of the set aircraft are as follows: reading the effective particle radius value and optical thickness value of the set aircraft; obtaining water density value stored in a pre-established database and analyzing the cloud liquid water path value of the set aircraft in combination with the effective particle radius value and optical thickness value of the set aircraft.

[0010] Further, the cloud image data is specifically the albedo value and channel brightness temperature set of each pixel in the cloud image and the corresponding two-dimensional coordinates, and the cloud image error mapping model is an input layer, a cloud image deconstruction layer and a mapping output layer.

[0011] Further, the specific steps of analyzing the cloud type temperature correction factor of the set aircraft are as follows: inputting the cloud image data of the set aircraft into the pre-trained cloud image error mapping model to analyze the temperature correction mapping feature set of the set aircraft including spectral difference temperature mapping value, cloud body fluctuation temperature mapping value and absorption scattering cooperative offset temperature mapping value; and analyzing the cloud type temperature correction value of the set aircraft based on the temperature correction mapping feature set of the set aircraft.

[0012] Further, the specific steps of analyzing the temperature correction mapping feature set of the setting aircraft are as follows: in the input layer of the cloud image error mapping model, the cloud image data of the setting aircraft is received and preprocessed; in the cloud image disassembly layer of the cloud image error mapping model, the preprocessed cloud image data of the setting aircraft is subjected to feature extraction processing to obtain the cloud image radiation feature vector of the setting aircraft; and in the mapping output layer of the cloud image error mapping model, the cloud image radiation feature vector of the setting aircraft is used to output the temperature correction mapping feature set of the setting aircraft.

[0013] Further, the specific steps of analyzing the optimized cloud top temperature value of the setting aircraft are as follows: reading the cloud top temperature value and the cloud top height value of the setting aircraft and the cloud type temperature correction value; obtaining the flight height value of the setting aircraft, and comprehensively processing the cloud top temperature value and the cloud top height value of the setting aircraft and the cloud type temperature correction value to obtain the optimized cloud top temperature value of the setting aircraft.

[0014] Further, the specific formula for calculating the optimized cloud top temperature value of the setting aircraft is as follows: ; wherein, is the optimized cloud top temperature value of the setting aircraft, is the cloud top temperature value of the setting aircraft, is the cloud top height value of the setting aircraft, is the flight height value of the setting aircraft, is the temperature decay adjustment coefficient stored in the database, is the cloud type temperature correction value of the setting aircraft.

[0015] Further, the specific formula for calculating the icing index of the setting aircraft is as follows: ; wherein, is the icing index of the setting aircraft, is the cloud liquid water path value of the setting aircraft, is the icing amplification factor stored in the database, is the optimized cloud top temperature value of the setting aircraft, is the temperature response normalization coefficient stored in the database.

[0016] Further, the specific steps of the icing warning processing of the setting aircraft based on the icing index are as follows: comparing the icing index of the setting aircraft with the preset warning interval set; and sending the corresponding warning level to the setting aircraft based on the comparison result.

[0017] The present application has the following advantages:

[0018] (1) The aircraft ice accretion early warning method fuses cloud image data and cloud product file data to construct an ice accretion index analysis system based on multiple information sources. The system not only uses the cloud field running state feature set obtained from the cloud product file, but also uses a pre-trained cloud image error mapping model to adaptively output a cloud type temperature correction factor for different cloud types, so as to obtain an optimized cloud top temperature value. Then, the cloud liquid water path value generated by the cloud field running state feature set is fused to generate an ice accretion index, thereby effectively solving the cloud temperature estimation deviation caused by a single data source, realizing dynamic correction and high-precision quantification of the ice accretion risk, and significantly improving the accuracy of ice accretion early warning.

[0019] (2) The aircraft ice accretion early warning method jointly analyzes the cloud liquid water path generated by the cloud field running state feature and the optimized cloud top temperature, thereby avoiding the neglect of the complexity of the cloud field and the uneven thermal force, making the finally generated ice accretion index result have higher environmental adaptability. The cloud product file data is analyzed and processed to automatically extract the cloud field running state feature set, and the cloud liquid water path value is calculated based on the feature set. Then, based on the optimized cloud top temperature value, the coupling effect of the liquid water content and the temperature condition on the ice accretion formation is comprehensively analyzed to obtain the ice accretion index, thereby realizing dynamic evaluation of the ice accretion risk and effectively improving the safety of aircraft flight.

[0020] (3) The aircraft ice accretion early warning method realizes an intelligent ice accretion early warning feedback mechanism through automatic comparison of the ice accretion index and the preset early warning interval, thereby realizing quantitative and graded early warning of the ice accretion risk according to the numerical value and the corresponding early warning interval after analyzing the ice accretion index of the set aircraft. When the ice accretion index is in different intervals, the corresponding early warning level information is automatically output, thereby responding to the cloud field change during flight, continuously maintaining high sensitivity and high reliability under complex weather conditions, and significantly improving the safety guarantee level of aviation operation.

[0021] Of course, implementing any product of the present application does not necessarily require all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A flowchart of the aircraft ice accretion early warning method of the present application.

[0023] Figure 2 A flowchart of the specific steps of analyzing the cloud type temperature correction factor of the set aircraft in the aircraft ice accretion early warning method of the present application.

[0024] Figure 3 A flowchart of the temperature correction mapping feature set analysis of the set aircraft in the aircraft ice accretion early warning method of the present application. DETAILED DESCRIPTION

[0025] Please refer toFigure 1 , an embodiment of the present invention provides a technical solution: an aircraft icing warning method, including the following steps: obtaining cloud map data and cloud product file data of a specified aircraft, extracting a cloud field operation state feature set of the specified aircraft based on the cloud product file data of the specified aircraft, and analyzing the cloud liquid water path value of the specified aircraft; based on a pre-trained cloud map error mapping model, and combining the cloud map data of the specified aircraft, analyzing the cloud type temperature correction factor of the specified aircraft, and combining the cloud field operation state feature set, analyzing the optimized cloud top temperature value of the specified aircraft; based on the cloud liquid water path value and the optimized cloud top temperature value of the specified aircraft, analyzing the icing index of the specified aircraft; performing icing warning processing on the specified aircraft based on the icing index.

[0026] The specific formula for calculating the icing index of the specified aircraft is as follows: ; where is the icing index of the specified aircraft, is the cloud liquid water path value of the specified aircraft, is the icing amplification factor stored in the database (when the cloud liquid water path value is low, a higher value is used to improve sensitivity, and when the cloud liquid water path value is high, a lower value is used to maintain stability), is the optimized cloud top temperature value of the specified aircraft, [[ID=2O]]is the temperature response normalization coefficient stored in the database (and in this embodiment, it can take the value of ).

[0027] It should be noted that the icing amplification factor stored in the database is obtained as follows: obtaining the cloud liquid water path values of several historical times, and extracting its 25th percentile (Q25) and 75th percentile (Q75), and using these as interval boundaries, and dividing them into a low liquid water interval (cloud liquid water path value < Q25), a medium liquid water interval (Q25 ≤ cloud liquid water path value < Q75), and a high liquid water interval (cloud liquid water path value ≥ Q75), where Q25 and Q75 respectively correspond to the 25% and 75% distribution positions of the cloud liquid water path value, to reflect the physical stratification characteristics of different cloud water contents. If the current cloud liquid water path value is within the low liquid water interval, the icing amplification factor can take 6.2849. If the current cloud liquid water path value is within the medium liquid water interval, the icing amplification factor can take 4.1899. If the current cloud liquid water path value is within the high liquid water interval, the icing amplification factor can take 3.3519.

[0028] The temperature response normalization coefficient stored in the database The acquisition step is as follows: a sample data set of several times of icing environment in history is acquired, the sample data set includes multi-dimensional meteorological feature data and physical observation data related to icing index calculation, and the sample data is stored in the form of time sequence, and a time sequence division strategy is adopted to divide the sample data into a training set and a verification set in a ratio of 8:2, wherein the training set is used for model training and parameter search, and the verification set is used for model performance evaluation;

[0029] In the training stage, the initial value of the temperature response normalization coefficient is set, and automatic optimization search is performed in the parameter space of [-80, -5], and a target function is constructed based on the Bayesian optimization framework, taking the icing index calculation error as the optimization constraint, and introducing a sliding time window cross-validation mechanism, dividing the later time series data into multiple continuous time windows (such as 5), each window is taken as a verification sample in turn, and the remaining windows are taken as training samples. After each round of training is completed, the following training indicators are extracted based on the prediction results of the verification set and the corresponding measured label values: critical success index (indicating the matching degree of the model prediction of icing events and the measured icing events), false alarm rate (the omission rate of the model to the real icing sample), false alarm rate (the probability of misjudging non-icing samples as icing), and overall accuracy (evaluating the overall prediction consistency). Through window-by-window iterative training and verification, the corresponding comprehensive performance score of each window is calculated.

[0030] The comprehensive performance score is obtained by multi-index weighted calculation, and the evaluation function is: ; wherein, the critical success index, with a weight of 40%, reflecting the core accuracy of icing prediction; the false alarm rate, with a weight of 30%, emphasizing the identification ability of real icing; the false alarm rate, with a weight of 20%, controlling the false alarm of non-icing scene; the overall accuracy, with a weight of 10%, measuring the overall prediction consistency.

[0031] In the Bayesian optimization process, the performance of the value generated by each iteration is evaluated, and the corresponding value with the highest score is determined as the final temperature response normalization coefficient, and the target value obtained after optimization is = .

[0032] ​The specific steps of the ice accretion index-based ice accretion early warning processing of the set aircraft are as follows: comparing the ice accretion index of the set aircraft with a preset early warning interval set, and the early warning interval set includes a first early warning interval: when the ice accretion index < 0, it is determined that there is no ice accretion risk area, corresponding to a clear sky or a low cloud thin layer area; a second early warning interval: when 20≤ ice accretion index < 50, it is determined that there is a light ice accretion risk area, corresponding to an area where the cloud liquid water content is low and local supercooled water droplets appear; a third early warning interval: when 50≤ ice accretion index < 80, it is determined that there is a moderate ice accretion risk area, corresponding to an area where the cloud thickness is large, the liquid water path is moderate, and the temperature is low; and a fourth early warning interval: when the ice accretion index is ≥ 80, it is determined that there is a strong ice accretion risk area, corresponding to a strong icing area where the cloud liquid water content is high, the temperature is extremely low, and the ice phase particles are active.

[0033] Based on the comparison processing result, the corresponding early warning level is sent to the set aircraft, which is specifically: if the ice accretion index of the set aircraft is within the first early warning interval, the normal monitoring state is maintained, and no alarm is issued; if the ice accretion index of the set aircraft is within the second early warning interval, an ice accretion attention signal is generated, prompting the pilot to pay attention to the cloud layer change and continuously monitor the ice accretion index; if the ice accretion index of the set aircraft is within the third early warning interval, an ice accretion warning signal is generated, an image early warning prompt is sent to the flight control terminal, and the pilot is advised to take height adjustment or heading avoidance measures; and if the ice accretion index of the set aircraft is within the fourth early warning interval, a serious ice accretion early warning signal is generated, a forced alarm mechanism is triggered, and is sent to the ground control center through the flight communication link, and the early warning time, position coordinates and related parameters are recorded for subsequent analysis and emergency response.

[0034] Specifically, the cloud product file data is the original field data read from the L2 level cloud product file obtained from the Himawari-8 / 9 satellite, which is stored in the NetCDF format and internally contains multiple cloud microphysical parameter fields, and each grid is attached with corresponding latitude and longitude coordinate information and multiple cloud microphysical parameters (and it needs to be noted that the L2 cloud product file is automatically generated by the meteorological satellite ground inversion system based on L1 level satellite observation data, and the system directly obtains the inversion result by calling a meteorological data interface or a satellite data sharing service without performing radiation inversion operation in the design), and the specific steps of extracting the cloud field running state feature set of the set aircraft are as follows:

[0035] The cloud product file data of the set aircraft is parsed, that is, the structure of the file is identified and the field is decoded through a data parsing interface, the (original binary) field data in the cloud product file is parsed into a processable numerical matrix form, and preprocessing operations such as outlier rejection, missing point interpolation and data format unification are performed, after the parsing is completed, a standardized data set containing cloud microphysical fields in multiple latitude and longitude grids is obtained;

[0036] The latitude and longitude of the set aircraft (which can be obtained by onboard GNSS or BDS) is obtained, and is matched with the parsed cloud product file data of the set aircraft to obtain grid data of the set aircraft, that is, the latitude and longitude of the set aircraft are compared with the latitude and longitude coordinate information in each latitude and longitude grid to obtain the corresponding latitude and longitude grid of the set aircraft, and the cloud microphysical fields in the grid, such as cloud top temperature field, effective particle radius field, cloud optical thickness field, cloud top height field, etc., are read as grid data;

[0037] Based on the grid data of the set aircraft, the cloud field running state feature set of the set aircraft is output, including cloud top temperature value (temperature value at the upper boundary of the cloud layer), effective particle radius value (equivalent average radius of liquid droplets or ice crystals in the cloud body), cloud optical thickness value (overall extinction ability of the cloud layer to solar radiation, reflecting the shielding strength of the cloud layer to incident radiation, the thicker the cloud layer, the higher the particle concentration, the larger the optical thickness value, the thinner the cloud layer, the sparser the particles, the smaller the optical thickness value), and cloud top height value (geometric height of the upper boundary of the cloud layer from the ground), that is, the cloud microphysical fields in the grid data, such as cloud top temperature field, effective particle radius field, cloud optical thickness, cloud top height field, etc., are output with their corresponding numerical values.

[0038] The specific steps of analyzing the cloud liquid water path value of the set aircraft are as follows: reading the effective particle radius value and optical thickness value in the cloud field running state feature set of the set aircraft; obtaining the water density value stored in the pre-established database, and combining the effective particle radius value and optical thickness value of the set aircraft to analyze the cloud liquid water path value (total mass of liquid water contained per unit area when vertically passing through the entire cloud layer) of the set aircraft, the specific formula is as follows: ; wherein, is the cloud liquid water path value of the set aircraft, is the water density value stored in the pre-established database, is the effective particle radius value of the set aircraft, is the optical thickness value of the set aircraft.

[0039] It needs to be explained that the calculation of the cloud liquid water path value is based on the radiation transmission and optical thickness theory of the cloud body. In the cloud body microphysical model, the radius distribution function of the cloud droplet and its volume distribution determine the relationship between the optical thickness and the liquid water content. When the effective radius of the cloud droplet is approximately constant in the vertical direction, the quantitative mapping relationship between the optical thickness of the cloud body, the average radius of the liquid droplet and the liquid water density can be established by integration simplification, so as to obtain the formula, which reflects the coupling law between the total mass of the cloud liquid water and the cloud droplet size, optical extinction ability. The effective particle radius value and the optical thickness value are obtained by analyzing the L2 level cloud product file of the Himawari-8 / 9 satellite. The parameters are inverted based on the multi-channel radiation observation by the meteorological satellite ground inversion system in the file. Therefore, it is not necessary to repeat the inversion calculation. Only the physical formula is used to calculate the cloud liquid water path value of the aircraft.

[0040] In the embodiment, the step can realize the structured analysis and quantitative utilization of the satellite inversion cloud product file data, significantly improve the availability and accuracy of the cloud field microphysical parameters, and avoid the complex radiation inversion calculation in the local area by directly calling the L2 level cloud product file generated by the meteorological satellite ground inversion system. The calculation burden is reduced, the data reliability is ensured, the grid data corresponding to the position of the aircraft is extracted and set through the latitude and longitude matching mode, the accurate correspondence between the cloud field spatial distribution and the flight position is realized, so that the cloud field thermal state and structural characteristics of the area where the aircraft is located can be reflected in real time. Secondly, the cloud liquid water path is calculated by using the effective particle radius value and the optical thickness value, the dynamic quantitative expression from the micro-particle characteristics of the cloud layer to the macro water content is realized, so that the water vapor content and the potential icing condition of the cloud layer can be reflected, and the accuracy of the whole ice identification and risk assessment is improved.

[0041] Specifically, as shown in Figures 2-3 The cloud image data (the cloud image is a local cloud image area with a radius of 10 km centered on the latitude and longitude position of the set aircraft) is specifically the albedo value and channel brightness temperature set (including water vapor channel brightness temperature value, infrared window region brightness temperature value, short wave infrared brightness temperature) of each pixel in the cloud image and the corresponding two-dimensional coordinates. The cloud image error mapping model is an input layer, a cloud image deconstruction layer and a mapping output layer.

[0042] The specific steps of analyzing the cloud type temperature correction factor of the setting aircraft are as follows: inputting the cloud image data of the setting aircraft into the pre-trained cloud image error mapping model, analyzing the temperature correction mapping feature set of the setting aircraft, including the spectral difference temperature mapping value, the cloud fluctuation temperature mapping value and the absorption scattering cooperative offset temperature mapping value; based on the temperature correction mapping feature set of the setting aircraft, analyzing the cloud type temperature correction value of the setting aircraft, which is specifically: performing weighted processing on the spectral difference temperature mapping value, the cloud fluctuation temperature mapping value and the absorption scattering cooperative offset temperature mapping value of the setting aircraft to obtain the cloud type temperature correction value of the setting aircraft.

[0043] And it should be noted that in the weighting process, the weighting coefficients corresponding to the spectral difference temperature mapping value, the cloud fluctuation temperature mapping value and the absorption scattering cooperative offset temperature mapping value can be obtained by the following steps: obtaining historical spectral difference temperature mapping values, historical cloud fluctuation temperature mapping values and historical absorption scattering cooperative offset temperature mapping values (and it should be noted that the above values are absolute values), and extracting the historical spectral difference temperature mapping mean value, the historical cloud fluctuation temperature mapping mean value and the historical absorption scattering cooperative offset temperature mapping mean value respectively, and performing sum processing to obtain a correction sum value, and performing ratio processing on the historical spectral difference temperature mapping mean value, the historical cloud fluctuation temperature mapping mean value and the historical absorption scattering cooperative offset temperature mapping mean value respectively with the correction sum value, and taking the corresponding results as the corresponding weighting coefficients.

[0044] The specific steps of analyzing the temperature correction mapping feature set of the setting aircraft are as follows: in the input layer of the cloud image error mapping model, receiving the cloud image data of the setting aircraft and performing preprocessing, which is specifically: performing range standardization processing on the albedo value and the brightness temperature value of each channel to normalize them to the same numerical interval, such as 0-1, to facilitate the uniformity of the feature scale of the model input, and for missing pixels in the cloud image caused by shielding, noise or data loss, compensating by neighborhood interpolation or spatial smoothing method to maintain the continuity and integrity of the data, after which 3x3 neighborhood mean convolution can be used to perform spatial filtering on the pixels in the local area to reduce random noise and enhance the spatial consistency of the cloud brightness temperature and albedo distribution;

[0045] In the cloud image disassembly layer of the cloud image error mapping model, the cloud image data of the set aircraft after preprocessing is subjected to feature extraction processing to obtain a cloud image radiation feature vector of the set aircraft, which is specifically: reading the water vapor channel brightness temperature value, the infrared window region brightness temperature value, and the short-wave infrared brightness temperature value of each pixel in the cloud image, respectively extracting the water vapor channel and infrared window region brightness temperature difference, the short-wave infrared and infrared window region brightness temperature difference of each pixel, and performing ratio processing to obtain the brightness temperature difference ratio of each pixel, and reading the two-dimensional coordinates of each pixel, performing mean value processing to obtain the center coordinates of the center pixel, respectively extracting the Euclidean distance value of the center pixel and each pixel based on the Euclidean distance method, and representing it in inverse normalization, such as 1 / (1+Euclidean distance value), and performing summation processing to obtain the Euclidean distance sum value, and respectively performing ratio processing on the Euclidean distance value of the center pixel and each pixel with the Euclidean distance sum value to obtain the weight proportion value of each pixel, and performing weighted product operation with the brightness temperature difference ratio of the corresponding pixel to extract the spectral difference ratio feature for representing the thickness distribution characteristics of the cloud above the set aircraft in different spectral channels. When it is higher, it indicates that the cloud layer is thicker, the water vapor content is higher, and the radiation attenuation is significant, and the cloud top brightness temperature exists a low trend;

[0046] Reading the infrared window region brightness temperature value of each pixel, and for each pixel, respectively calculating the infrared window region brightness temperature difference value in the horizontal and vertical directions with the adjacent pixels, and performing summation on the square of the infrared window region brightness temperature difference value in the horizontal and vertical directions, and performing root processing to obtain the brightness temperature gradient value of each pixel, and reading the weight proportion value of each pixel, and performing weighted product operation with the brightness temperature gradient value of the corresponding pixel to extract the cloud fluctuation energy feature for representing the spatial fluctuation degree of the cloud above the set aircraft. When it is higher, it indicates that the cloud top height in this region changes dramatically, the cloud layer structure is complex, and the uneven distribution of infrared radiation leads to a low trend of the cloud top brightness temperature;

[0047] Reading the albedo value and the infrared window region brightness temperature value of each pixel, respectively extracting the albedo mean value, the albedo standard deviation value, the infrared window region brightness temperature mean value, and the infrared window region brightness temperature standard deviation value, and using the correlation coefficient method, i.e., based on the albedo value and the infrared window region brightness temperature value of each pixel, as well as the albedo mean value and the infrared window region brightness temperature mean value, calculating the covariance value, and performing ratio processing with the product of the albedo standard deviation value and the infrared window region brightness temperature standard deviation value to extract the absorption and scattering collaborative shift feature for adjusting the absorption and scattering collaborative performance of the cloud above the set aircraft in the phase change process. When it is smaller, it indicates that the visible light albedo is higher and the infrared brightness temperature is lower, there is ice-water mixing or uneven phase in the cloud, the infrared radiation absorption is weakened, and the cloud top temperature value is biased low. The spectral difference ratio feature, the cloud fluctuation energy feature, and the absorption and scattering collaborative shift feature are spliced into a cloud image radiation feature vector;

[0048] In the mapping output layer of the cloud image error mapping model, based on the cloud image radiation feature vector of the set aircraft, the temperature correction mapping feature set of the set aircraft is output, which is specifically: calling the feature and temperature correction corresponding relationship established in the training stage, mapping the current values of the spectral difference feature, the cloud body fluctuation energy feature and the absorption scattering cooperative offset feature into specific spectral difference temperature mapping values, cloud body fluctuation temperature mapping values and absorption scattering cooperative offset temperature mapping values based on the spectral difference feature, the cloud body fluctuation energy feature and the absorption scattering cooperative offset feature, and the spectral difference temperature mapping values, the cloud body fluctuation temperature mapping values and the absorption scattering cooperative offset temperature mapping values respectively represent the specific correction values of the corresponding features to the current cloud top temperature value.

[0049] The pre-training step of the cloud image error mapping model is as follows:

[0050] The labeled data set is obtained, which is composed of historical multi-channel cloud image product file data of Himawari-8 / 9 and historical meteorological sounding observation temperature data of the same time and same area. Each sample is labeled by meteorological observation experts according to the deviation between historical satellite inversion cloud top brightness temperature and measured cloud top temperature, and each group of samples includes visible light albedo value, water vapor channel brightness temperature value, short-wave infrared brightness temperature value, infrared window region brightness temperature value and corresponding real cloud top temperature label value of each pixel in the set region.

[0051] The labeled data set is preprocessed, including range standardization processing, missing point compensation and outlier removal of each channel brightness temperature and albedo value, to maintain the continuity and spatial consistency of the data. The preprocessed historical data set is divided into training set, validation set and test set in proportion, and the samples are ensured to cover different cloud types, different climate zones and different height distributions to enhance the generalization ability of the model.

[0052] In the training stage, the spectral difference feature, the cloud body fluctuation energy feature and the absorption scattering cooperative offset feature of each sample are extracted, and the three features are sequentially spliced into a cloud image comprehensive radiation feature vector. At the same time, the deviation between the satellite inversion cloud top brightness temperature and the measured cloud top temperature in the sample is calculated as the training label input model, and the cloud image error mapping model is supervised trained.

[0053] The model adopts a regression learning mechanism based on historical data, with the optimization objective of minimizing the mean square error (MSE) between the predicted temperature correction value and the historical measured temperature error. The mapping function parameters are iteratively updated through the back propagation algorithm, and the adaptive optimization algorithm (such as Adam optimizer) is used for convergence control. The model performance is evaluated by the validation set, and the hyperparameters are adjusted to ensure that the model can accurately learn the relationship between different cloud image features and temperature deviation.

[0054] After the training is completed, the feature vectors of all historical samples and their corresponding correction amounts are statistically analyzed to form a historical correspondence table of feature value intervals and temperature correction amounts, that is, a historical mapping relationship between the spectral difference feature, the cloud body fluctuation energy feature, the absorption and scattering cooperative shift feature, and the cloud top temperature correction value is established, and the relationship is solidified as a model parameter file.

[0055] Finally, the trained cloud image error mapping model is saved, and the historical correspondence table in the model will be called in the running stage for the rapid temperature correction calculation of the real-time input cloud image feature vector, so as to realize the adaptive correction of the cloud top temperature based on historical experience data.

[0056] In the embodiment, the cloud image data is deeply analyzed by introducing the pre-trained cloud image error mapping model, so as to realize the intelligent correction of the cloud top temperature deviation, thereby significantly improving the accuracy and adaptability of the cloud field temperature. Secondly, the pre-trained cloud image error mapping model is used to jointly deconstruct the albedo, water vapor channel brightness temperature, short-wave infrared brightness temperature, and infrared window brightness temperature, extract the spectral difference feature, cloud body fluctuation energy feature, and absorption and scattering cooperative shift feature, and realize the quantitative correction of the cloud top temperature through the feature and temperature correction correspondence relationship established in the training stage. The model can adaptively output correction values according to different cloud types and cloud layer thicknesses. Finally, the historical data is introduced for supervised learning, so that the model learns the brightness temperature deviation law under multiple cloud types, so that it has the ability to quickly and accurately correct the real-time cloud image in the running stage, thereby realizing the automatic mapping from features to temperature correction values. Not only the physical reality and regional adaptability of the cloud top temperature inversion are improved, but also the prediction accuracy of the ice accretion is significantly enhanced.

[0057] Specifically, the specific steps of analyzing and setting the optimized cloud top temperature value of the aircraft are as follows: reading the cloud top temperature value and the cloud top height value of the set aircraft and the cloud type temperature correction value; and obtaining the flight height value of the set aircraft (which can be obtained by a barometric height sensor, that is, the barometric height sensor detects the static pressure outside the aircraft in real time, and converts the relationship between the barometric height equation and the standard atmospheric pressure height to obtain the flight height value), and the cloud top temperature value and the cloud top height value of the set aircraft and the cloud type temperature correction value are combined and processed to obtain the optimized cloud top temperature value of the set aircraft.

[0058] The specific formula for calculating the optimized cloud top temperature value of the set aircraft is as follows: ; wherein, is the optimized cloud top temperature value of the set aircraft, is the cloud top temperature value of the set aircraft, is the cloud top height value of the set aircraft, is the flight height value of the set aircraft, The temperature attenuation adjustment coefficient stored in the database (in this embodiment, it can be 0.65), The cloud type temperature correction value of the aircraft is set.

[0059] It is necessary to explain that the temperature attenuation adjustment coefficient stored in the database The acquisition steps are as follows: obtaining historical cloud top height values, historical flight height values and corresponding measured temperature values in the past, extracting the measured temperature difference value (i.e. the difference between the measured temperature value corresponding to the historical cloud top height value and the historical flight height value) of each time in the past, the height difference value (i.e. the absolute value of the difference between the historical cloud top height value and the historical flight height value), and performing ratio processing to obtain the temperature attenuation value of each time in the past, and performing sliding average processing, and taking the result as the temperature attenuation adjustment coefficient .

[0060] In this embodiment, on the basis of the cloud top temperature, the flight height of the aircraft and the cloud type temperature correction value are combined to realize dynamic optimization correction of the cloud top temperature, thereby significantly improving the accuracy of the cloud field thermal parameter, and the flight height is obtained in real time by using the air pressure height sensor, the accurate spatial position is obtained through the air pressure and height conversion relationship, and the difference value analysis is performed on the cloud top height, thereby establishing the thermal coupling relationship between the height difference and the temperature difference, and then the temperature attenuation adjustment coefficient is calculated by statistically analyzing the historical multiple groups of cloud top height, flight height and measured temperature data, thereby reflecting the change characteristics of the temperature gradient between different height layers, so that the temperature correction can be dynamically adjusted according to the actual atmospheric vertical structure, and finally, this step realizes the adaptive correction of the cloud top temperature, thereby avoiding the deviation problem of the fixed temperature attenuation rate model under different climate conditions, so that the optimized cloud top temperature value can more truly reflect the thermal state of the environment where the aircraft is located, and the regional adaptability of the ice accumulation identification is significantly enhanced.

[0061] Although the preferred embodiments of the present application have been described, those skilled in the art can make further changes and modifications to the embodiments once they know the basic inventive concept. Therefore, the appended claims are intended to be interpreted as including all changes and modifications falling within the scope of the present application.

[0062] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A method for early warning of aircraft icing, characterized in that, Includes the following steps: Obtain cloud image data and cloud product file data for the designated aircraft. The cloud product file data is the raw field data read from the L2 level cloud product file obtained from Himawari-8 / 9 satellites. Based on the cloud product file data of the designated aircraft, the cloud field operation status feature set of the designated aircraft is extracted, and the cloud liquid water path value of the designated aircraft is analyzed. Based on the pre-trained cloud map error mapping model, which consists of an input layer, a cloud map deconstruction layer, and a mapping output layer, and combined with the cloud map data of the set aircraft, the cloud shape temperature correction factor of the set aircraft is analyzed, and combined with the cloud field operation state feature set, the optimized cloud top temperature value of the set aircraft is analyzed. Based on the set cloud liquid water path value and optimized cloud top temperature value, the icing index of the set aircraft is analyzed. Icing warnings are issued for designated aircraft based on the icing index.

2. The aircraft icing early warning method according to claim 1, characterized in that, The specific steps for extracting the cloud field operation status feature set of a given aircraft are as follows: The cloud product file data for the aircraft is parsed and processed. Obtain the latitude and longitude of the designated aircraft and match it with the parsed cloud product file data of the designated aircraft to obtain the grid data of the designated aircraft. Based on the grid data of the specified aircraft, the cloud field operation status feature set of the specified aircraft is output, including cloud top temperature value, effective particle radius value, cloud optical thickness value, and cloud top height value.

3. The aircraft icing early warning method according to claim 2, characterized in that, The specific steps for analyzing and setting the cloud liquid water path values ​​for an aircraft are as follows: Read the effective particle radius and optical thickness values ​​of the set aircraft; Obtain the water density value stored in the pre-established database, and combine it with the effective particle radius value and optical thickness value of the set aircraft to analyze the cloud liquid water path value of the set aircraft.

4. The aircraft icing early warning method according to claim 1, characterized in that, The cloud image data specifically includes the albedo value and channel brightness temperature set of each pixel in the cloud image, as well as the corresponding two-dimensional coordinates.

5. The aircraft icing early warning method according to claim 4, characterized in that, The specific steps for analyzing and setting the cloud type temperature correction factor for an aircraft are as follows: The cloud image data of the set aircraft is input into the pre-trained cloud image error mapping model, and the temperature correction mapping feature set of the set aircraft is analyzed, including spectral difference temperature mapping value, cloud undulation temperature mapping value, and absorption and scattering co-shift temperature mapping value. Based on the temperature correction mapping feature set of the set aircraft, the cloud temperature correction value of the set aircraft is analyzed.

6. The aircraft icing early warning method according to claim 5, characterized in that, The specific steps for analyzing and setting the temperature correction mapping feature set of the aircraft are as follows: In the input layer of the cloud image error mapping model, cloud image data of the specified aircraft is received and preprocessed. In the cloud map deconstruction layer of the cloud map error mapping model, feature extraction processing is performed on the preprocessed cloud map data of the specified aircraft to obtain the cloud map radiation feature vector of the specified aircraft. In the mapping output layer of the cloud map error mapping model, the temperature correction mapping feature set of the set aircraft is output based on the cloud map radiation feature vector of the set aircraft.

7. The aircraft icing early warning method according to claim 5, characterized in that, The specific steps for analyzing and setting the optimal cloud top temperature value for the aircraft are as follows: Read the cloud top temperature value, cloud top altitude value, and cloud shape temperature correction value of the set aircraft; The system obtains the flight altitude of the set aircraft and combines it with the cloud top temperature value, cloud top altitude value, and cloud shape temperature correction value to obtain the optimized cloud top temperature value of the set aircraft.

8. The aircraft icing early warning method according to claim 7, characterized in that, The specific formula for calculating the optimal cloud top temperature value for the aircraft is as follows: ； in, To set the optimal cloud top temperature value for the aircraft, To set the cloud top temperature value for the aircraft, To set the cloud top altitude value for the aircraft, To set the aircraft's flight altitude, The temperature decay adjustment coefficient is stored in the database. To set the cloud cover temperature correction value for the aircraft.

9. The aircraft icing early warning method according to claim 1, characterized in that, The specific formula for calculating the icing index of a given aircraft is as follows: ； in, To set the aircraft icing index, To set the cloud liquid water path value for the aircraft, This refers to the ice accumulation amplification factor stored in the database. To set the optimal cloud top temperature value for the aircraft, The temperature response normalization coefficient is stored in the database.

10. The aircraft icing early warning method according to claim 1, characterized in that, The specific steps for issuing icing warnings for designated aircraft based on the icing index are as follows: The icing index of the aircraft is set and compared with the preset warning interval set. Based on the comparison and processing results, the corresponding early warning level is sent to the designated aircraft.

Images