An automatic message encoding method and system for airport thunderstorms

Abstract

This invention provides an automatic thunderstorm message encoding method and system for airports. The method includes: acquiring thunderstorm observation data of the airport area; identifying thunderstorms based on the observation data and extracting thunderstorm evolution features; performing trend prediction based on the thunderstorm evolution features to obtain a prediction result reflecting the degree of thunderstorm impact within a preset time period; and converting the prediction result into a discrete encoding result conforming to aviation meteorological message specifications and outputting it according to preset encoding rules and consistency constraints. Through standard logical mapping and stability filtering mechanisms, abnormal jumps in message content are effectively suppressed, significantly improving the consistency of airport thunderstorm messages.

Description

Technical Field

[0001] This invention relates to the field of aviation meteorological observation technology, and in particular to an automatic message coding method and system for airport thunderstorms. Background Technology

[0002] In civil aviation operations support, airport meteorological departments are required to compile and issue routine aviation meteorological reports (METAR), special aviation meteorological reports (SPECI), and short-term trend forecasts (TREND) regularly or as needed, in accordance with international and industry standards. Thunderstorms and their accompanying phenomena (such as lightning and heavy precipitation) are hazardous weather elements that must be closely monitored. In current operational processes, observers typically combine frequency and location data provided by lightning location systems with reflectivity images provided by weather radar for comprehensive interpretation. They use manual experience to determine the occurrence, duration, and end of thunderstorms and manually write discrete coded statements that conform to standards based on the movement trends of individual thunderstorm cells.

[0003] However, existing lightning and radar data are mostly descriptions of high-frequency, continuously changing physical processes, with outputs primarily in the form of numerical information such as probability and intensity distribution. Aviation meteorological reporting systems, on the other hand, require the generation of highly discrete, semantically constrained message statements (such as TS, VCTS, and BECMG). Due to a significant semantic gap between the output characteristics of multi-source automatic observation data and the operational logic of message encoding, current thunderstorm message compilation heavily relies on human experience. This can easily lead to inconsistent judgment standards among different observers, unstable message trend descriptions, and logical conflicts between different time-series messages during periods of intense thunderstorm evolution.

[0004] In summary, how to utilize continuous, probabilistic thunderstorm observation and forecast results to achieve automatic and stable conversion into discrete coded statements that conform to aviation meteorological report specifications, thereby eliminating the uncertainty caused by manual identification and improving the consistency of the reports, is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] This invention provides an automatic message encoding method and system for airport thunderstorms, which addresses the shortcomings of existing technologies in eliminating uncertainties caused by manual identification and improving message consistency.

[0006] This invention provides an automatic message encoding method for airport thunderstorms, comprising: Acquire thunderstorm observation data in the airport area; Thunderstorm identification is performed based on the thunderstorm observation data, and thunderstorm evolution characteristics are extracted. Based on the aforementioned thunderstorm evolution characteristics, trend prediction is performed to obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future. Based on preset coding rules and consistency constraints, the prediction results are converted into discrete coding results that conform to aviation meteorological message specifications and output.

[0007] According to the automatic message coding method for airport thunderstorms provided by the present invention, the thunderstorm observation data includes lightning location data and weather radar data; The step of identifying thunderstorms and extracting thunderstorm evolution features based on the thunderstorm observation data includes: The lightning location data and the weather radar data are subjected to spatiotemporal matching processing to construct a feature set characterizing the state of thunderstorm activity; Based on the feature set, thunderstorm cells are identified, and the spatiotemporal distribution attributes of the thunderstorm cells within the airport's preset spatial partitions and their evolution intensity attributes over time are extracted as the thunderstorm evolution features.

[0008] According to the automatic message encoding method for airport thunderstorms provided by the present invention, the step of extracting the spatiotemporal distribution attributes of the thunderstorm cell within a preset spatial partition of the airport includes: Based on the horizontal distance of lightning location data relative to the airport reference point in the thunderstorm observation data and a preset radius threshold, the location of thunderstorm activity is determined, and the location of thunderstorm activity is divided into a spatial region including the local area, the nearby area, and the external monitoring area; wherein, the radius threshold is set with the airport reference point as the center; Calculate the centroid location and spatial scale of the thunderstorm cell, and determine the spatial region to which the thunderstorm cell belongs; The step of extracting the evolution intensity attribute of the thunderstorm cell over time includes: Acquire lightning location data and construct a sliding time window containing lightning events at multiple consecutive moments; Spatial clustering of lightning events within the sliding time window is performed to determine the range of individual thunderstorm cells; The frequency and density change rate of lightning within the range of the thunderstorm cell are statistically analyzed and used as the evolution intensity attribute; The step of performing spatiotemporal matching processing on the lightning location data and the weather radar data to construct a feature set characterizing the thunderstorm activity state includes: Within a preset time window, the distribution of the lightning location data is statistically analyzed to generate a lightning density field, and the weather radar data is parsed into a radar reflectivity field. Map the lightning density field onto the same grid as the radar reflectivity field; The mapped lightning density field and radar reflectivity field are normalized and spatially superimposed to generate the feature set containing the superimposed composite intensity index and hail identification mark.

[0009] According to the automatic message coding method for airport thunderstorms provided by the present invention, the step of performing trend prediction based on the thunderstorm evolution characteristics to obtain a prediction result reflecting the degree of thunderstorm impact within a preset time period includes: Based on the movement trend of the thunderstorm cell, the predicted trajectory and predicted coverage area of ​​the thunderstorm cell within a preset time period are predicted. Determine the spatial overlap and transformation relationship between the predicted coverage area and the external monitoring area, the nearby area, and the local area; Based on the overlapping conversion relationship, the predicted time when the thunderstorm cell moves into or out of the external monitoring area, the nearby area and the local area is determined, and the prediction time window corresponding to each spatial partition is generated respectively. Based on the stability and evolution intensity attributes of the thunderstorm cells in their historical trajectories, the probability of thunderstorm activity occurring within the predicted time window of each spatial partition is calculated.

[0010] According to the automatic airport thunderstorm message coding method provided by the present invention, the step of converting the prediction result into a discrete coding result conforming to the aviation meteorological message specification according to preset coding rules and consistency constraints includes: According to the preset coding rules, the prediction time window, occurrence probability and evolution intensity attributes of each spatial partition are mapped to elements to generate preliminary discrete coding results. Based on the consistency constraints, the preliminary discrete coding results are subjected to time-series stability smoothing and cross-message logic verification to determine the final discrete coding results that conform to the specifications.

[0011] According to the automatic airport thunderstorm message coding method provided by the present invention, the step of mapping the prediction time window, occurrence probability, and evolution intensity attributes of each spatial partition to elements according to the preset coding rules to generate preliminary discrete coding results includes: For the standard message time interval, when the occurrence probability exceeds the preset trigger threshold, a thunderstorm weather phenomenon identifier is determined; Based on the spatial partition to which the prediction time window belongs, the identifier is matched with the corresponding location attribute information to distinguish the current thunderstorm from nearby thunderstorms; Based on the composite intensity index and hail identification identifier in the evolution intensity attribute, modifiers reflecting precipitation intensity level or special weather phenomena are added to the identifier with location attribute to form the preliminary discrete coding result.

[0012] According to the automatic message coding method for airport thunderstorms provided by the present invention, the step of performing time-series stability smoothing and cross-message logic verification on the preliminary discrete coding results based on the consistency constraint to determine the final discrete coding results that conform to the specification includes: Execution timing stability smoothing: The encoding result output at the previous moment is introduced as a reference state. If the change magnitude or duration of the preliminary discrete encoding result relative to the reference state does not reach a preset smoothing threshold, state transition is suppressed and the reference state is maintained. Perform cross-message logical verification: compare the smoothed encoding result with the actual situation elements in the airport routine weather report or airport special weather report released at the same time. If there is logical mutual exclusion, the encoding result is corrected or removed according to the preset business priority.

[0013] According to the automatic airport thunderstorm message encoding method provided by the present invention, adding modifiers reflecting precipitation intensity levels or special weather phenomena to the identifier having location attributes includes: Based on the reflectance peak value in the composite intensity index, a strong, medium or weak intensity level is matched in the precipitation intensity level, and then converted into the corresponding positive and negative intensity modifiers. When the hail identification identifier is detected and its location attribute belongs to this area, a hail phenomenon code is appended after the thunderstorm weather phenomenon identifier. Based on the start and end times of the predicted time window, add trend indicator time group codes representing the start, end, or change of weather phenomena to the preliminary discrete coding results.

[0014] According to the automatic airport thunderstorm message encoding method provided by the present invention, the step of correcting or eliminating the encoding result according to a preset service priority includes: Establish a priority ranking based on operational safety, setting the priority of manually issued correction messages and important weather warnings higher than that of automatically generated preliminary coding results; If the smoothed encoding result is inconsistent with the real-time data released at the same time in terms of the presence or absence of thunderstorms and intensity level, then the real-time data will be used as the standard, and the encoding result will be rolled back or deleted. When the change in the encoding result reaches the preset message release threshold, a special aviation meteorological message trigger command is generated, which includes the updated encoding content and release action suggestions.

[0015] The present invention also provides an automatic message encoding system for airport thunderstorms, comprising: The data acquisition module is used to acquire thunderstorm observation data in the airport area; The feature recognition module is used to identify thunderstorms based on the thunderstorm observation data and extract thunderstorm evolution features; The trend prediction module is used to predict trends based on the thunderstorm evolution characteristics and obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future. The encoding conversion module is used to convert the prediction results into discrete encoding results that conform to the aviation meteorological report specifications and output them according to preset encoding rules and consistency constraints.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the airport thunderstorm automatic message encoding method as described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the airport thunderstorm automatic message encoding method as described above.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the airport thunderstorm automatic message encoding method as described above.

[0019] The automatic airport thunderstorm message coding method and system provided by this invention identifies and extracts thunderstorm evolution features from thunderstorm observation data, transforming the original observations into objective and quantitative physical parameters, thus eliminating the subjectivity and uncertainty caused by manual identification at the source. Furthermore, based on the thunderstorm evolution features, trend prediction is performed to obtain prediction results, which are then converted and output as discrete coding results according to preset coding rules and consistency constraints. Through standard logical mapping and stability filtering mechanisms, abnormal jumps in message content are effectively suppressed, significantly improving the consistency of airport thunderstorm messages. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is one of the flowcharts of the automatic message encoding method for airport thunderstorms provided by the present invention.

[0022] Figure 2 This is a schematic diagram of the refined monitoring of the area surrounding the airport provided by the present invention.

[0023] Figure 3 This is a schematic diagram of the message consistency and conflict control logic provided by the present invention.

[0024] Figure 4 This is the second flowchart of the automatic message encoding method for airport thunderstorms provided by the present invention.

[0025] Figure 5 This is a schematic diagram of the structure of the airport thunderstorm automatic message coding system provided by the present invention.

[0026] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] Currently, although some automated auxiliary alarm methods have been introduced in the civil aviation meteorological field, the final message coding stage still heavily relies on manual identification and operation. This model has revealed significant technical shortcomings in actual operation: First, existing technologies cannot eliminate the uncertainty brought about by manual identification. Different observers have different professional backgrounds, observation experience and ability to capture the characteristics of complex convective weather. As a result, there are often different standards in determining the start and end time of thunderstorms, classifying intensity levels and determining the boundary between nearby thunderstorms (VCTS) and local thunderstorms (TS), which leads to the output message content having a strong subjective color.

[0029] Secondly, existing systems generally lack consistent message timing control mechanisms. Because radar echoes and lightning frequencies often exhibit high transient volatility during the evolution of thunderstorm systems, traditional automated logic, if based solely on thresholds at a single moment, is highly susceptible to pulse-like jumps in message content. For example, during the dissipation phase of a thunderstorm, sporadic flashes of lightning activity can cause the system to frequently switch between thunderstorm end and thunderstorm continuation within a few minutes. This unstable coded output severely violates aviation meteorological message compilation and issuance standards, and also causes significant confusion for air traffic controllers and pilots in their decision-making.

[0030] Furthermore, logical conflicts between messages are a challenge that existing technologies have not adequately addressed. Automatically generated thunderstorm codes are often isolated and lack logical loop verification with other routine or special reports released within the same time period. This can lead to physical and logical mismatches between thunderstorm identifiers and elements such as visibility, precipitation characteristics, or cloud base height in messages under certain complex scenarios, compromising the overall consistency and reliability of the message sequence.

[0031] The specific embodiments provided by this invention aim to effectively solve the uncertainty problem caused by high human intervention in the prior art through the following technical solutions, and to overcome the consistency problem of message jumps and logical inconsistencies in the automated coding process.

[0032] Before describing the technical solutions of the embodiments of the present invention, the terms and concepts involved in the embodiments of the present invention will be explained illustratively.

[0033] Thunderstorm evolution characteristics refer to a set of quantitative parameters extracted from observational data to describe the state of thunderstorm activity and its dynamic changes. These characteristics typically include the location, spatial extent, speed and direction of movement of individual thunderstorm cells (spatiotemporal distribution attributes), as well as their intensity and development or weakening trend (evolution intensity attributes). They are the core basis for the system to make trend predictions and coding decisions.

[0034] Lightning location data: Data acquired by lightning location detection networks records the precise time, geographical coordinates (latitude and longitude), current intensity, polarity, and other information of cloud-to-ground or cloud-to-cloud lightning events. Updated at a frequency of seconds or minutes, this data is a crucial real-time source for identifying the initial stages of convective activity and tracking the core areas of thunderstorms.

[0035] Weather radar data: This mainly refers to the basic reflectivity factor data obtained by weather radar. This data reflects the ability of precipitation particles (such as raindrops and hail) in the atmosphere to scatter electromagnetic waves emitted by radar. Its intensity value (usually expressed in dBZ) can be used to infer precipitation intensity, identify strong convective structures (such as hail storm cores), and determine precipitation type.

[0036] A thunderstorm cell refers to a spatially independent thunderstorm system or convective unit with a complete life cycle. In this invention, a continuous and concentrated area of ​​strong convective activity identified through spatiotemporal clustering analysis of lightning and radar data is considered an independent thunderstorm cell and serves as the basic object for tracking, prediction, and encoding.

[0037] Spatiotemporal distribution attributes: Describe the spatial and temporal existence of thunderstorm cells. These mainly include: the geographical location of the cell's centroid relative to the airport, the airport's pre-defined monitoring zone (local zone, nearby zone, etc.), the cell's spatial scale (e.g., equivalent radius), and its movement vector and historical trajectory.

[0038] Evolution intensity attributes: These describe the intensity of a thunderstorm cell itself and its characteristics over time. They mainly include: electrical activity intensity calculated based on lightning frequency and density, precipitation or hail intensity inferred from radar reflectivity, and the rate of change (intensification, maintenance, or weakening) of these intensity indicators over a short period.

[0039] Radar reflectivity field: A two-dimensional data field covering a certain area and with a regular grid distribution, formed by spatial interpolation of reflectivity data acquired by weather radar. This data field visually displays the scattering intensity levels of precipitation particles at different locations.

[0040] Lightning density field: Within a set time window (e.g., 10 minutes), lightning location data is statistically analyzed onto a regular geographic grid identical to the radar reflectivity field, and a two-dimensional data field is formed by calculating the number of lightning occurrences per unit area of ​​each grid point. This field is used to characterize the spatial distribution of convective activity and the intensity of electrical activity.

[0041] Discrete coding result: refers to a string composed of fixed characters and abbreviations that conforms to international aviation meteorological reporting specifications (such as WMO or ICAO standards). In this invention, it specifically refers to the report field suggestions automatically generated by the system to describe thunderstorm phenomena and their trends, such as TSRA (thunderstorm with moderate rain), VCTS (thunderstorm nearby), TEMPO 0006 TSGR (thunderstorm with hail in a short period of time), etc.

[0042] Modifiers: In aviation meteorological message coding, these are specific abbreviations added before or after the weather phenomenon code to describe its intensity or characteristics. For example, + indicates "strong," - indicates "weak," and VC indicates "nearby." In this embodiment of the invention, the system automatically matches appropriate modifiers to the thunderstorm code based on the predicted intensity attribute, such as encoding strong thunderstorm precipitation as +TSRA.

[0043] Aviation Meteorological Report (METAR): This is a standardized meteorological report issued by airports at regular intervals, describing the current weather conditions. It is typically issued hourly, or under special weather conditions. It includes core elements such as wind, visibility, weather phenomena, clouds, temperature, and air pressure.

[0044] Special Aviation Weather Report (SPECI): A special weather report issued immediately between two routine METAR reports when certain important meteorological elements (such as thunderstorms, wind shear, sudden changes in visibility, etc.) reach or exceed specific criteria. Its format is the same as METAR, and it is used to promptly notify of the onset, significant changes, or end of hazardous weather.

[0045] Short-term trend forecast (TREND): This is a brief forecast of weather trends over the next two hours, appended to the METAR or SPECI report. It uses fixed phrases such as BECMG (gradual change), TEMPO (short-term fluctuation), and NOSIG (no significant change) to describe the expected occurrence, duration, end, or change of weather phenomena.

[0046] The execution of the method in this invention relies on a dedicated information processing hardware system. The core of this system is one or more high-performance computing servers, equipped with central processing units (GPUs) and graphics processing units (GPUs) responsible for carrying and running software programs that implement core algorithms such as data fusion, feature extraction, trend prediction, and encoding mapping. To acquire real-time observation data, the system continuously receives raw data streams from lightning location networks and weather radars through deployed network communication equipment and data interfaces. Simultaneously, the system includes large-capacity storage devices for storing processing programs, a configurable encoding rule base, real-time data, and historical records. Finally, automatically generated message encoding suggestions are transmitted via the network to meteorological operational terminals, presented to observers on a display, and manually reviewed and confirmed through human-computer interaction devices such as keyboards and mice. This dedicated system, composed of computing, communication, storage, and interactive hardware, is the physical carrier and hardware entity that automatically executes the method described in this invention and realizes its functions.

[0047] Figure 1 This is a flowchart illustrating the automatic message encoding method for airport thunderstorms provided in an embodiment of the present invention. The method includes: Step 101: Obtain thunderstorm observation data for the airport area.

[0048] This step is the starting point and data foundation for the system's automated processing. The system establishes a real-time data link with a professional meteorological observation network deployed in and around the airport through its data acquisition module. This module continuously receives real-time observation data streams covering a radius of tens to hundreds of kilometers centered on the airport from external sources such as the lightning location system's data distribution server and the weather radar's data processing center, either through active subscription or periodic requests.

[0049] The core data acquired includes at least two categories: one is lightning location data reflecting the characteristics of thunderstorm activity, which records precise millisecond-level timestamps, geographic latitude and longitude coordinates, and current intensity for each lightning strike in the form of events; the other is weather radar data characterizing the distribution of precipitation particles in thunderstorm clouds, mainly referring to gridded field or image data from radar base data or generated product data that can describe the spatial distribution of reflectivity intensity. All of these data contain unified spatiotemporal reference information during transmission.

[0050] Simultaneously with data access, the system performs preliminary real-time preprocessing. This includes parsing the raw data stream and converting it into an internally unified spatiotemporal data format; validating the data, such as filtering out invalid records that clearly exceed physically reasonable ranges or have abnormal coordinates; and aligning and synchronizing data from different sources in time to provide standardized, high-quality, and continuous data input for subsequent spatiotemporal fusion analysis. The entire acquisition process incorporates status monitoring and error reconnection mechanisms to ensure the reliability and stability of the data source.

[0051] Step 102: Based on the thunderstorm observation data, identify thunderstorms and extract thunderstorm evolution features.

[0052] The core task of this step is to transform the raw observation data into a set of quantitative features that can structurally describe the state and dynamics of thunderstorm activity through the system's feature recognition module. The system first performs comprehensive integrated analysis on continuous observation data, and based on preset recognition logic, automatically identifies valid signals and active entities related to thunderstorms from the data field containing complex information. This process is not simply event detection, but rather, by analyzing the spatiotemporal correlation and statistical characteristics of the data, it aggregates discrete or continuous observation points and confirms them as independent thunderstorm activity objects, thus completing thunderstorm identification.

[0053] Building upon this foundation, the system further performs in-depth feature analysis on each identified thunderstorm object. It extracts and calculates a set of key parameters from the data to characterize the core behavioral patterns of the object; these parameters collectively constitute the thunderstorm evolution characteristics. This feature set mainly covers two aspects: first, the spatiotemporal distribution attributes describing the existence and location changes of the thunderstorm object in space, such as its geographical impact and the movement of its core area; and second, the energy intensity reflecting the thunderstorm object's own dynamic evolution over time, such as the rising and falling trends of its activity level. By generating this structured feature set, the system completes the transformation from raw data to feature indicators with clear physical and operational significance, providing directly usable and objective quantitative evidence for subsequent trend prediction and coding decisions.

[0054] Step 103: Based on the thunderstorm evolution characteristics, perform trend prediction to obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future.

[0055] After extracting the evolution characteristics of thunderstorms, the system uses feature information reflecting the history and current state of thunderstorms to predict trends. The prediction process is mainly based on the movement vector and intensity dynamic change trends in the evolution characteristics.

[0056] Specifically, the system combines the location and geometric scale of the thunderstorm cell with its displacement direction and velocity over a continuous time period to calculate the spatial path the thunderstorm cell may take within a preset future timeframe. Simultaneously, the system utilizes the intensity attribute from its evolutionary characteristics to analyze the development trend of thunderstorm activity and determine whether the thunderstorm energy will strengthen, remain stable, or weaken in the future. By establishing a correlation between the predicted trajectory and the airport's preset spatial zones, the system further calculates the coverage and potential impact of the thunderstorm cell on various areas of the airport at different future times.

[0057] This quantitative analysis, which integrates changes in spatial location and evolution of physical intensity, ultimately yields a prediction of the extent of thunderstorm impact within a predetermined timeframe. This prediction provides crucial data support for the subsequent transformation of the continuous physical evolution process into a compliant discrete code, ensuring the coding logic possesses sufficient predictability and objectivity.

[0058] Step 104: Based on the preset coding rules and consistency constraints, convert the prediction results into discrete coding results that conform to the aviation meteorological message specifications and output them.

[0059] This step is not a simple format conversion, but rather a dual mechanism of logical filtering and business rule judgment to ensure that the automatically generated messages can accurately capture weather changes while maintaining the stability required for business operations.

[0060] When executing the encoding mapping logic, the system uses the predicted probability, time, and evolution intensity attributes as input parameters. Through a pre-defined encoding rule base, the system retrieves the current message timeframe and filters elements based on the coverage of the prediction time window. For example, when the prediction shows a thunderstorm with an extremely high probability of impact within the airport's area, the rule base automatically activates the weather phenomenon code representing the thunderstorm and determines whether an intensity modifier needs to be added based on the individual storm's intensity evolution attribute. If the thunderstorm only affects a specific area around the airport, the rule base maps it to a location attribute representing nearby thunderstorms. This mapping mechanism transforms continuously changing physical quantitative predictions into discrete symbol sequences with a clear grammatical structure, achieving a cross-dimensional translation from meteorological scientific descriptions to civil aviation operational syntax.

[0061] To address the common data jitter issue in automated coding, the system introduces consistency constraints for secondary verification and smoothing after initially generating the coding sequence. The core of these consistency constraints is maintaining the temporal stability of the message state, preventing frequent switching of message content within a very short time due to instantaneous fluctuations in radar echoes or occasional drops in lightning frequency. The system retrieves message states from historical time periods as a benchmark, comparing the deviation between the current preliminary coding and historical states. Only when the change in the predicted result exceeds a preset stability threshold, and this change exhibits sufficient persistence over time, will the system allow the coding state to change. Furthermore, the consistency constraints also include cross-element logical conflict detection. By retrieving the release information of other meteorological elements within the same timeframe in real time, it ensures that thunderstorm coding is physically and logically consistent with elements such as precipitation, clouds, and visibility.

[0062] Finally, the discrete coding results, determined after rule transformation and constraint filtering, will be distributed in real time through the system's output interface. This result can be presented directly as a standard text message conforming to civil aviation regulations, or pushed as a structured data stream to forecaster review terminals or air traffic control decision support systems. In this way, step 104 effectively distills complex thunderstorm detection conclusions into concise, professional, and logically rigorous safeguards, eliminating the blind spots inherent in manual judgment and providing highly consistent and forward-looking information support for safe airport operations.

[0063] The automatic airport thunderstorm message coding method provided in this invention identifies and extracts thunderstorm evolution features from thunderstorm observation data, transforming the original observations into objective and quantitative physical parameters, thus eliminating the subjectivity and uncertainty caused by manual identification at the source. Furthermore, it performs trend prediction based on thunderstorm evolution features to obtain prediction results, which are then converted and output as discrete coding results according to preset coding rules and consistency constraints. Through standard logical mapping and stability filtering mechanisms, it effectively suppresses abnormal jumps in message content and significantly improves the consistency of airport thunderstorm messages.

[0064] Furthermore, during implementation, the system first receives thunderstorm observation data, which consists of lightning location data and weather radar data. The lightning location data includes the latitude and longitude coordinates of the lightning occurrence and a timestamp; the weather radar data includes the reflectivity factor grid values ​​and their corresponding spatial coordinates.

[0065] The system performs spatiotemporal matching processing on these two types of data. Spatially, the system establishes a unified geographic coordinate system, mapping discrete lightning location data onto a spatial grid of weather radar reflectivity factors. Temporally, the system aggregates lightning location data according to preset time windows, aligning it with the radar scanning cycle. Through these processes, the system constructs a feature set characterizing thunderstorm activity, where each grid cell contains both reflectivity intensity values ​​and lightning activity frequency.

[0066] Based on this feature set, the system uses a clustering algorithm to identify thunderstorm cells. The system extracts spatially continuous grid regions from the feature set whose intensity exceeds a preset threshold and defines them as independent thunderstorm cell entities. Subsequently, the system extracts spatiotemporal distribution attributes for each identified thunderstorm cell. These attributes include the centroid latitude and longitude of the thunderstorm cell, its geometric boundaries, and its coverage area. The system correlates the spatial location of the thunderstorm cell with preset spatial partitions of the airport to determine whether the thunderstorm cell is currently located in the local area, the nearby area, or the external monitoring area.

[0067] Simultaneously, the system extracts the evolution intensity attributes of thunderstorm cells over time. The system compares the highest reflectivity factor, total lightning frequency, and lightning density within the same thunderstorm cell over continuous observation periods. By calculating the rate of change of these indicators over time, intensity data reflecting the development, maintenance, or dissipation trend of thunderstorm cells is obtained. The extracted spatiotemporal distribution attributes and evolution intensity attributes together constitute the thunderstorm evolution characteristics, serving as the input basis for subsequent automatic coding logic.

[0068] In its implementation, to achieve a more robust and comprehensive assessment of thunderstorm conditions, the system performs deep fusion processing on two heterogeneous data sets: lightning and radar data, constructing a feature set characterizing thunderstorm activity. Within a preset common time window (e.g., the past 6 minutes), the system statistically analyzes the distribution of lightning location data to generate a spatial lightning density field. Simultaneously, it resolves the baseline data from weather radar into a radar reflectivity field reflecting precipitation particle intensity. To ensure comparability and fusion, the system maps the lightning density field to a regular geographic grid identical to the radar reflectivity field using spatial interpolation. Next, the system normalizes the two fields to eliminate dimensional differences, and then spatially superimposes them according to preset weights, generating a feature set containing a composite intensity index and hail identification markers. This feature set integrates the electrical and microphysical characteristics of convective activity. Based on this feature set, the system can identify individual thunderstorms.

[0069] After identifying individual thunderstorm cells, a crucial step is to determine the spatial location and zonify thunderstorm activity based on lightning location data. The system uses the airport runway center or a designated meteorological observation point as a fixed airport reference point. Using this point as the center, multiple distance radius thresholds are preset to spatially divide the area into different monitoring zones. These zones typically include the local zone adjacent to the airport, the surrounding area, and the outermost external monitoring zone. For example, 0 to 8 kilometers can be set as the local zone, 8 to 16 kilometers as the surrounding area, and 16 to 50 kilometers as the external monitoring zone. The system calculates the horizontal distance between each lightning event and the airport reference point in real time and automatically assigns it to the corresponding zone based on the preset radius thresholds. Furthermore, for each identified thunderstorm cell, the system calculates the geographic coordinates and spatial scale (such as equivalent radius) of its centroid, thus accurately determining which spatial zone the cell is currently primarily located within. This zoning mechanism provides direct quantitative evidence for distinguishing between local thunderstorms (TS) and nearby thunderstorms (VCTS) in subsequent message coding.

[0070] To accurately capture the intensity variation trend of thunderstorms, the system needs to extract their evolutionary intensity attributes over time. The system first constructs a sliding time window containing lightning events across multiple consecutive moments, such as a 10-minute window. Within this time window, the system uses spatial clustering algorithms (such as DBSCAN) to analyze all lightning events, aggregating spatially concentrated lightning points into clusters, thereby defining the spatial extent of each thunderstorm cell. Subsequently, the system counts the total number of lightning strikes within this cell's extent and calculates the lightning density per unit time and per unit area. By analyzing the rate of change of lightning frequency and density (e.g., increasing, decreasing, or remaining constant) across multiple sliding time windows, the system can quantify whether the thunderstorm cell is in a state of recent enhancement, maintenance, or weakening. These dynamic intensity indicators constitute the core intensity attributes of thunderstorm evolution.

[0071] Furthermore, based on the extracted evolution characteristics of thunderstorm cells, the system performs short-term trend prediction to quantify their future impact on the airport. First, based on the current movement speed, direction, and other movement trends of the thunderstorm cells, the system uses linear extrapolation, Kalman filtering, or methods combined with environmental wind field advection to extrapolate the predicted trajectory of its centroid over a future period (e.g., the next 2 hours), and estimates the predicted coverage area that its echoes and lightning activity may cover.

[0072] Subsequently, the system compares the spatiotemporal contour of this predicted coverage area with pre-defined spatial partitions such as the local area, nearby area, and external monitoring area, and analyzes the spatial overlap and transformation relationship between the two. Specifically, the system determines when the predicted coverage area will begin to touch, enter, or completely leave a certain partition.

[0073] Based on the dynamic analysis of this spatial overlap, the system can calculate the key time points when thunderstorm cells are expected to enter and leave each zone. For example, the system can determine that a thunderstorm will enter the nearby zone from the external monitoring area in about 15 minutes and further intrude into the local area in about 40 minutes. Based on these time points, the system generates a specific predicted impact time window for each relevant zone, such as "impact time window for the nearby zone: next 15 to 40 minutes".

[0074] Finally, to assess the uncertainty of the impact, the system comprehensively considers the stability of thunderstorm cells in their historical trajectories and the current and historical evolution intensity attributes (such as whether lightning activity is continuously increasing or rapidly decreasing). Through an internal calculation model, it assigns a quantified probability of occurrence to each predicted impact time window. This probability value characterizes the likelihood of significant thunderstorm activity occurring in the corresponding time period and region, providing core decision parameters for subsequently converting continuous predictions into deterministic business message coding.

[0075] In its implementation, the system employs a multi-layered concentric circle spatial partitioning model for refined monitoring of the area surrounding the airport. For example... Figure 2 As shown, the model uses the airport reference point (such as the runway center) as the center and sets three typical monitoring radii: R1 (0-8 km) is defined as the "local area" that directly affects flight take-off and landing; R2 (8-16 km) is defined as the "nearby area" that may affect arrival and departure routes; and R3 (16-50 km or more) is defined as the "external monitoring area" for early warning and tracking.

[0076] Once the system identifies and tracks a thunderstorm cell using lightning and radar data, it determines the zone in real time where its centroid is located. As the thunderstorm cell moves, the system monitors its dynamic process of "moving in" or "moving out" of each zone. This spatial overlap and transformation relationship directly drives differentiated message linkage suggestions: for example, when a thunderstorm moves from zone R3 to zone R2 (nearby zone), the system may suggest coding "VCTS" (thunderstorm nearby) in the routine METAR message and suggest "TEMPO TS..." in the TREND section to predict its potential further impact on the local area; when the thunderstorm moves further into zone R1 (local area), the system will immediately suggest a more severe code such as "TS" or "TSRA" in the METAR body, and is likely to trigger a SPECI (Special Weather Report) suggestion. Through this mechanism of "zone monitoring, dynamic determination, and linkage triggering," the system achieves spatial classification of thunderstorm impact and precise mapping of message coding types, ensuring that the automatically generated suggestions not only conform to the spatiotemporal logic of operations but also reflect the progressive nature of hazardous weather.

[0077] Furthermore, after obtaining the prediction time windows, occurrence probabilities, and evolution intensity attributes of each spatial partition obtained from trend prediction, the system performs the operation of converting the above prediction results into discrete coding results that conform to the aviation meteorological report specifications. This process is divided into two stages: first, preliminary discrete coding results are generated using preset coding rules, and then the final discrete coding results are determined using consistency constraints.

[0078] In the initial discrete coding stage, the system first maps the prediction results to the standard message timeframe, which corresponds to the release timeframe of aviation meteorological messages. The system determines whether the probability of occurrence of the corresponding spatial partition within the timeframe of the prediction time exceeds a preset trigger threshold; if it exceeds the threshold, a thunderstorm weather phenomenon identifier is determined. Subsequently, the system matches location attribute information to the identifier based on the spatial partition to which the prediction time window belongs: if the individual cell is located in the local area, the location attribute corresponds to a local thunderstorm; if the individual cell is located in a nearby area, the location attribute corresponds to a nearby thunderstorm. Furthermore, the system extracts the composite intensity index and hail identification identifier from the evolution intensity attributes, adding intensity modifiers or special weather phenomenon modifiers to the identifier, thus forming the initial discrete coding result.

[0079] In one specific embodiment, the process of adding modifiers reflecting precipitation intensity levels or special weather phenomena to the identifier having location attributes includes the following three dimensions: First, the system performs a refined mapping of precipitation intensity levels. After acquiring the basic identifier of thunderstorm phenomena, it further extracts composite intensity indicators from the composite feature field, particularly the peak radar reflectivity. The system internally pre-sets intensity classification thresholds corresponding to civil aviation meteorological observation standards. For example, when the peak radar reflectivity of the core area of ​​a thunderstorm cell is below the first threshold (e.g., 39 dBZ), the system classifies it as a weak intensity level and automatically matches a mild modifier (i.e., the "-" sign in the standard); when the peak reflectivity is between the first and second thresholds (e.g., 49 dBZ), it is classified as a moderate intensity level, and no positive or negative modifier is added; when the peak reflectivity exceeds the second threshold, it is classified as a strong intensity level and automatically converted to a heavy modifier (i.e., the "+" sign in the standard). Through this direct mapping from physical quantities to grammatical modifiers, the system can automatically convert basic thunderstorm identifiers into precise codes such as light thunderstorms, moderate thunderstorms, or severe thunderstorms.

[0080] Secondly, the system handles complex judgments and additions for special weather phenomena. Thunderstorms are often accompanied by severe weather such as hail, which pose a serious threat to aircraft operational safety. The system reads the feature set generated by multi-source data fusion in real time and checks whether it contains a hail identification identifier (this identifier is usually calculated based on radar vertical integral liquid water content or dual polarization parameters). After confirming the presence of hail features, the system further performs a strict spatial location verification: only when the centroid or coverage area of ​​the hail feature falls within the preset local area will the system add a specific code indicating hail after the original thunderstorm weather phenomenon identifier. If hail exists only in the nearby area or external monitoring area, based on the principles of message conciseness and standardization, the system may only retain the thunderstorm identifier without adding a hail code. This spatially constrained complex coding logic effectively avoids over-warning and ensures the accuracy of message indication.

[0081] Finally, the system assembles time-based data for the evolution of weather phenomena. Aviation meteorological reports not only need to describe "what is happening," but also accurately predict "when it will happen" and "when it will end." The system retrieves the prediction time windows generated by the trend prediction module and extracts the start and end times. For the coding requirements of short-term trend forecasts or special weather reports, the system converts these specific times into trend indication time groups conforming to meteorological code specifications. For example, if a prediction time window indicates that thunderstorm activity will begin moving into the airfield at a specific time in the future, the system extracts the "start time" and adds a trend indication time group code indicating "from a certain time"; if the prediction indicates that thunderstorm activity will move out or dissipate at a certain time, the system extracts the "end time" and adds a time group code indicating "until a certain time." By logically concatenating these time group codes with the aforementioned weather phenomenon identifiers containing intensity and special weather modifiers, the system ultimately generates a complete and preliminary discrete coding result that includes both a phenomenon description and a precise time evolution trajectory.

[0082] In the final discrete coding result determination stage, the system performs consistency constraint processing on the preliminary coding results. First, the system performs temporal stability smoothing, obtaining the coding result output at the previous time step as a reference state and comparing it with the preliminary discrete coding result. If the change between the two does not reach a preset smoothing threshold, or the duration of the preliminary coding result does not reach a preset time threshold, the system suppresses state changes and maintains the reference state to avoid coding jumps caused by data fluctuations. After smoothing, the system performs cross-message logical verification, comparing the current coding result with the real-time elements in the airport's routine weather report or special weather report published at the same time. If a logical mutual exclusion is detected, the system corrects or removes the current coding result according to preset business priority rules, ensuring content consistency of different types of aviation meteorological messages within the same time period, and finally determining and outputting discrete coding results that conform to aviation meteorological message specifications.

[0083] In the actual operation of automatic thunderstorm message encoding at airports, to ensure the authority and logical rigor of the message content, the system executes a correction and verification mechanism based on business priorities, specifically including: First, the system establishes a priority ranking based on operational safety. Within the civil aviation meteorological operational system, manually issued corrected reports and critical weather warnings represent the forecasters' highest comprehensive assessment of complex weather conditions. Therefore, the system explicitly prioritizes manually issued corrected reports and critical weather warnings over automatically generated preliminary coding results. This mechanism, from a fundamental logical perspective, avoids potential misjudgments by automated algorithms under extreme weather conditions, ensuring the uniqueness of safety decision-making information.

[0084] Secondly, the system performs real-time logical consistency comparisons across messages. If the smoothed encoding result is inconsistent with the actual weather data, such as the airport's routine weather report released at the same time, in terms of the presence or intensity of thunderstorms, the system will trigger an automatic correction procedure. Since the actual weather data is based on direct feedback from the airport's physical detection equipment, it has a very high degree of objectivity and authenticity. In this case, the system will force the actual weather data to be the standard and roll back or delete the encoding result. For example, if the automated algorithm predicts that there will still be thunderstorms at the airport, but the actual weather data released at the same time has consistently confirmed that the thunderstorms have disappeared, the system will automatically discard the initially generated thunderstorm code to prevent the output of contradictory meteorological information.

[0085] Finally, in response to the sudden and severe nature of thunderstorms, the system possesses agile early warning triggering capabilities. When the change in the coding result reaches the preset message release threshold, it signifies a significant change in weather conditions that could affect flight safety. At this point, the system immediately generates a special aviation meteorological message trigger command containing updated coding content and release action recommendations. This command is rapidly pushed to the operational terminal, reminding the on-duty forecaster to immediately compile and sign a special report, thereby ensuring that critical meteorological information such as the start and end times of thunderstorms and dramatic changes in intensity can be transmitted to air traffic control and flight support departments in a timely manner, maximizing the operational safety and efficiency of the airport.

[0086] See Figure 3 The schematic diagram of message consistency and conflict control logic shown in this embodiment simulates a typical scenario where the predicted probability fluctuates around a critical value. Figure 3 The document demonstrates the interaction process between the real-time observation / prediction module, the historical state database (previous message), the consistency control module, and the final discrete coding result.

[0087] Suppose that the trend forecast (TREND) code suggested by the system in the previous processing cycle (time T-1) is "NOSIG", indicating that there will be no significant weather changes in the next two hours. This code has been recorded in the system's historical state database as the baseline state (State_Ref) for judgment in the current cycle (time T).

[0088] At time T, the real-time observation / prediction module calculates, based on the latest data, that the current predicted probability (P_curr) of a thunderstorm occurring at the airport within the next 30 minutes is 52%. Meanwhile, the reference probability (P_ref) for the previous period corresponding to the current prediction time, stored in the historical state database, is 48%. The system then triggers the consistency control module to make a decision.

[0089] The module first calculates the deviation Δ, i.e., |52% - 48%| = 4%. Then, it compares Δ with a preset compatibility threshold (e.g., 5%). Since 4% < 5%, the system determines that the change in the current predicted probability is still within the "dead zone," which is a normal fluctuation of the prediction model near the critical value (e.g., 50%), rather than representing a definite and significant change in the actual weather conditions.

[0090] Therefore, the system suppresses the impulse to generate new coding suggestions (such as "TEMPO TS") based on the current instantaneous probability. Its decision is to maintain the original baseline state, that is, the final discrete coding result is still output as "NOSIG". This mechanism effectively avoids the problem of message content frequently switching between "with thunderstorm" and "without thunderstorm" (i.e., "flickering") due to slight fluctuations in probability values ​​around the threshold, thus ensuring the stability of the message sequence.

[0091] Afterward, the system will update the currently maintained "NOSIG" state to a new baseline state for use in the next cycle.

[0092] Conversely, if P_curr calculated at time T is 58%, then the deviation Δ is 10% (|58% - 48%|). Since 10% > 5%, the system's judgment has changed significantly. At this point, the consistency control module will allow the new encoded content (such as "TEMPOTSRA") to pass and determine whether to trigger the SPECI recommendation or update TREND. Before outputting the new recommendation, the module will also quickly compare it with the published live message to complete the final logical verification before outputting it as the discrete encoding result.

[0093] pass Figure 3 By calculating the deviation and determining the compatibility threshold, this embodiment of the invention achieves quantitative control over the timing stability of thunderstorm messages, ensuring that the output encoding results can not only sensitively reflect significant weather evolution, but also filter out invalid instantaneous fluctuations, thereby improving message consistency.

[0094] After outputting discrete coding results that conform to the specifications, the system executes an observation consistency check process as the final quality control step of automated coding, ensuring that the output results do not significantly deviate from the actual ground conditions. This includes: acquiring ground meteorological element data collected in real time by the airport's automatic weather observation system; checking whether the discrete coding results and the current weather conditions reflected by the ground meteorological element data meet preset consistency criteria; if the consistency criteria are not met, a manual review prompt message is generated.

[0095] Specifically, the system acquires real-time ground meteorological data collected by the airport's automatic weather observation system through a data interface. This real-time ground data includes at least instantaneous precipitation, minute precipitation, ground air temperature, dew point temperature, and ground wind direction and speed. Because the airport's automatic weather observation system is deployed at fixed locations around the runway, the physical quantities it observes can most directly reflect the real-time weather conditions of the airport area.

[0096] Subsequently, the system executes a crucial logical hedging step, namely, checking whether the discrete coding result and the current weather conditions reflected by the ground meteorological element data meet a preset consistency criterion. This criterion is based on meteorological logical consistency. For example, if the final determined discrete coding result indicates the presence of moderate or strong thunderstorms at the airport, but the airport's automatic weather observation system does not record any precipitation within a preset continuous monitoring time step, and the ground temperature does not exhibit the typical decreasing characteristics of convective weather, then the system determines that it does not meet the consistency criterion. Similarly, if ground sensors have detected significant precipitation and convective wind shear, but the automatically generated coding result still indicates no thunderstorms, it is also determined to be logically inconsistent.

[0097] When the system identifies a situation that does not meet the consistency criteria, in order to ensure clear attribution of responsibility for the message and prevent a single sensor failure from interfering with the coding logic, the system will not perform mandatory automatic modification on the generated discrete coding results. Instead, it will immediately generate a manual review prompt. This prompt is pushed in high brightness through the graphical interface of the business terminal, reminding meteorological observers that there is a significant conflict between the current prediction conclusion of the automated algorithm and the actual situation of the ground sensors, requiring manual intervention for on-site observation and verification. Through this closed-loop verification mechanism that introduces real-time feedback, this embodiment effectively avoids coding errors caused by occasional abnormalities in detection equipment or algorithm failures in extreme weather scenarios while ensuring the efficiency of automatic coding.

[0098] To further understand the solutions of the embodiments of the present invention, Figure 4 The complete process of the automatic message encoding method for airport thunderstorms according to an embodiment of the present invention is shown, mainly including: S1: Multi-source data acquisition and preprocessing.

[0099] The system acquires real-time lightning location network data, weather radar reflectivity data, and airport automatic weather observation system (AWOS) data for the target airport area via a dedicated data interface. After acquisition, the system performs instant parsing, time synchronization, and validity verification of the raw data, and converts it into an internally unified standardized format to provide reliable and standardized input for subsequent analysis.

[0100] S2: Lightning spatial partitioning based on multi-radius thresholds.

[0101] The system uses airport meteorological observation points or runway center points as reference points and pre-sets multiple concentric circular areas as monitoring zones, typically divided into the local zone (e.g., 0-8 km), the nearby zone (e.g., 8-16 km), and the external monitoring zone (e.g., 16-50 km). The system calculates the distance between each lightning event and the reference point in real time and assigns it to the corresponding zone, thereby achieving second-level perception and qualitative analysis of the spatial location of thunderstorm activity (local, nearby, or peripheral).

[0102] S3: Thunderstorm identification and evolution feature extraction.

[0103] The system performs spatiotemporal fusion of lightning and radar data to construct a composite feature field that integrates electrical activity characteristics (lightning density field) and cloud microphysical characteristics (radar reflectivity field). Based on this field, a clustering algorithm is used to identify independent thunderstorm cells and extract their core evolutionary features, including spatiotemporal distribution attributes (such as centroid location, movement speed, and region) and evolution intensity attributes (such as lightning frequency variation trend and reflectivity intensity level).

[0104] S4: Future trend prediction and probabilistic representation.

[0105] Based on the historical trajectory and current characteristics of thunderstorm cells, the system uses an extrapolation algorithm to predict their movement path (predicted trajectory) and impact range (predicted coverage area) for the next 1-2 hours. By analyzing the overlap between the predicted coverage area and preset spatial partitions, the prediction time window for thunderstorms entering or leaving each partition is determined. Simultaneously, considering the stability and intensity evolution trends of the individual cells, the probability (P) of thunderstorms affecting airports and the reliability (C) of the prediction results are calculated within each prediction time window.

[0106] S5: Probabilistic-semantic mapping generates preliminary coding suggestions.

[0107] The system calls the built-in coding rule library to map the prediction time window, probability of occurrence (P), confidence level (C), and intensity attributes generated in step S4 into discrete coded statements that conform to the syntax specifications of aviation meteorological reports (METAR / SPECI / TREND). For example, "high probability and high confidence of strong thunderstorm precipitation in this area within the next 30 minutes" is mapped to the preliminary coding suggestion "TEMPO 0030 +TSRA".

[0108] S6: Time Consistency and Conflict Control.

[0109] The system executes its core consistency control logic. This step compares the initial coding suggestions generated by S5 with the valid coding suggestions (as the baseline state) stored in the historical state database from the previous processing cycle, calculating the deviation between the two. The system then compares this deviation with a preset compatibility threshold (or dead zone) and makes a decision accordingly. If the deviation does not exceed the threshold: the system determines that the current change is a normal fluctuation or temporary rise and fall of the predicted value near the critical point, and does not represent a definite turning point in the weather. Therefore, the system will suppress the output of this preliminary suggestion and decide to maintain the baseline state of the previous period (e.g., continue to output "NOSIG" or maintain the original trend description) as the valid output result for this period.

[0110] If the deviation exceeds the threshold: the system confirms a significant, business-significant weather change. In this case, the system will adopt the preliminary coding suggestion generated by S5 and allow it to enter the output process. Simultaneously, the mechanism will perform a rapid logical check on this new suggestion against the already published weather reports to resolve potential timing discrepancies.

[0111] S7: Output the final encoding suggestion to the human-machine interface.

[0112] The system outputs the final coded recommendation generated from the decision-making process (which may maintain the previous state or confirm a new, updated recommendation) to the human-computer interaction interface of the meteorological operational terminal. This recommendation is presented in a clear format, such as highlighting, a structured list, or visual prompts, serving as a direct decision-making reference for observers or forecasters in compiling and issuing formal reports. At this stage, the system only provides auxiliary suggestions and does not automatically publish the report.

[0113] Simultaneously or subsequently, the system can further acquire real-time ground weather data collected by the Airport Automated Weather Observation System (AWOS). The system automatically compares the weather phenomena described in the final coding suggestion (especially the trend forecast portion) with the ground weather data. If a pre-defined logical inconsistency is detected (e.g., the TREND forecast predicts the end of a thunderstorm, but the actual data shows heavy precipitation is underway), the system will generate a clear manual review prompt, reminding staff to pay attention to and comprehensively assess the situation, thus adding a layer of real-time quality verification to the automated process.

[0114] The beneficial effects achieved by the method in this embodiment of the invention include: (1) It realizes the automation and standardization of thunderstorm identification and message encoding, significantly reducing reliance on manual labor and subjective differences.

[0115] By establishing a technological chain that automatically identifies thunderstorms from multi-source observation data (lightning, radar), extracts features, predicts trends, and generates coding suggestions, this invention transforms a business decision-making process that heavily relies on personal experience into an objective processing flow driven by explicit algorithms and rules. This eliminates the problem of inconsistent judgment standards among different observers and improves the objectivity, repeatability, and efficiency of message generation.

[0116] (2) It effectively bridges the gap between continuous observation data and discrete message syntax, and solves the problem of operational application of probability prediction results.

[0117] This invention creatively proposes a mapping rule base mechanism of "probability / credibility (P / C) → discrete coding". By automatically and stably converting continuous, probabilistic thunderstorm prediction results into discrete coding suggestions that conform to the strict syntax of METAR, SPECI, and TREND messages, based on preset time segmentation and threshold rules, this core breakthrough enables advanced, high-frequency automatic observation and nowcasting data to directly and reliably serve responsible message compilation operations.

[0118] (3) By introducing time consistency and conflict control mechanisms, the logical consistency and business stability of message sequences are greatly improved.

[0119] This invention employs an intelligent smoothing control logic based on historical state references and compatibility thresholds. This mechanism effectively suppresses frequent jumps in coding suggestions ("flickering") caused by normal fluctuations in data near critical values, ensuring a smooth transition in weather trend descriptions within adjacent message cycles. Simultaneously, a systematic cross-message logic verification function automatically detects and alerts to potential contradictions between new suggestions and previously published real-time messages, technically preventing the release of contradictory information and enhancing the inherent logical reliability and user trust of the entire message product.

[0120] (4) A clear, explainable and traceable decision support process has been established, which has strengthened business quality control and personnel training capabilities.

[0121] Each coding suggestion generated by the system has a clear data source, characteristic basis, and rule mapping path, making the process transparent and explainable. Simultaneously, the system fully records the entire chain of information from automatic suggestions, manual modifications, and final decisions. This not only meets the stringent traceability requirements of aviation meteorological reports as responsible information but also provides a valuable data foundation and technical means for post-event analysis, case review, rule optimization, and personnel training.

[0122] (5) A flexible and configurable system architecture that supports human-machine collaboration was designed, which enhanced business adaptability and practicality.

[0123] In this embodiment of the invention, the encoding rules, trigger thresholds, spatial partitioning parameters, etc., are all stored in an updatable rule base, which can be flexibly adjusted according to changes in civil aviation business specifications or the localization needs of different airports without modifying the core algorithm. The system is positioned as an auxiliary decision-making tool, outputting suggestions in the form of recommendations, which are then finalized and published by business personnel. This aligns with the current principle of "human responsibility" in business operations, achieving efficient collaboration between automated intelligence and human experience, and ensuring the practicality and feasibility of the solution.

[0124] The automatic airport thunderstorm message coding system provided in the embodiments of the present invention is described below. The automatic airport thunderstorm message coding system described below can be referred to in correspondence with the automatic airport thunderstorm message coding method described above.

[0125] This invention provides an automatic message encoding system for airport thunderstorms, see [link to relevant documentation]. Figure 5 ,include: Data acquisition module 510 is used to acquire thunderstorm observation data in the airport area; Feature recognition module 520 is used to identify thunderstorms based on the thunderstorm observation data and extract thunderstorm evolution features; The trend prediction module 530 is used to make trend predictions based on the thunderstorm evolution characteristics to obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future. The encoding conversion module 540 is used to convert the prediction results into discrete encoding results that conform to the aviation meteorological report specifications and output them according to preset encoding rules and consistency constraints.

[0126] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840. The processor 810, communications interface 820, and memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute an automatic airport thunderstorm message coding method. This method includes: acquiring thunderstorm observation data of the airport area; identifying thunderstorms based on the observation data and extracting thunderstorm evolution features; performing trend prediction based on the thunderstorm evolution features to obtain a prediction result reflecting the degree of thunderstorm impact within a preset future time period; and converting the prediction result into a discrete coding result conforming to aviation meteorological message specifications and outputting it according to preset coding rules and consistency constraints.

[0127] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0128] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the airport thunderstorm automatic message coding method provided by the above methods. The method includes: acquiring thunderstorm observation data of the airport area; identifying thunderstorms based on the thunderstorm observation data and extracting thunderstorm evolution features; performing trend prediction based on the thunderstorm evolution features to obtain a prediction result reflecting the degree of thunderstorm impact within a preset time period; and converting the prediction result into a discrete coding result conforming to the aviation meteorological message specification and outputting it according to preset coding rules and consistency constraints.

[0129] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the automatic airport thunderstorm message coding method provided by the above methods. The method includes: acquiring thunderstorm observation data of an airport area; identifying thunderstorms based on the thunderstorm observation data and extracting thunderstorm evolution features; performing trend prediction based on the thunderstorm evolution features to obtain a prediction result reflecting the degree of thunderstorm impact within a preset time period; and converting the prediction result into a discrete coding result conforming to the aviation meteorological message specification and outputting it according to preset coding rules and consistency constraints.

[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic message encoding method for airport thunderstorms, characterized in that, include: Acquire thunderstorm observation data in the airport area; Thunderstorm identification is performed based on the thunderstorm observation data, and thunderstorm evolution characteristics are extracted. Based on the aforementioned thunderstorm evolution characteristics, trend prediction is performed to obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future. Based on preset coding rules and consistency constraints, the prediction results are converted into discrete coding results that conform to aviation meteorological message specifications and output.

2. The method according to claim 1, characterized in that, The thunderstorm observation data includes lightning location data and weather radar data; The step of identifying thunderstorms and extracting thunderstorm evolution features based on the thunderstorm observation data includes: The lightning location data and the weather radar data are subjected to spatiotemporal matching processing to construct a feature set characterizing the state of thunderstorm activity; Based on the feature set, thunderstorm cells are identified, and the spatiotemporal distribution attributes of the thunderstorm cells within the airport's preset spatial partitions and their evolution intensity attributes over time are extracted as the thunderstorm evolution features.

3. The method according to claim 2, characterized in that, The step of extracting the spatiotemporal distribution attributes of the thunderstorm cells within a preset spatial partition of the airport includes: Based on the horizontal distance of lightning location data relative to the airport reference point in the thunderstorm observation data and a preset radius threshold, the location of thunderstorm activity is determined, and the location of thunderstorm activity is divided into a spatial region including the local area, the nearby area, and the external monitoring area; wherein, the radius threshold is set with the airport reference point as the center; Calculate the centroid location and spatial scale of the thunderstorm cell, and determine the spatial region to which the thunderstorm cell belongs; The step of extracting the evolution intensity attribute of the thunderstorm cell over time includes: Acquire lightning location data and construct a sliding time window containing lightning events at multiple consecutive moments; Spatial clustering of lightning events within the sliding time window is performed to determine the range of individual thunderstorm cells; The frequency and density change rate of lightning within the range of the thunderstorm cell are statistically analyzed and used as the evolution intensity attribute; The step of performing spatiotemporal matching processing on the lightning location data and the weather radar data to construct a feature set characterizing the thunderstorm activity state includes: Within a preset time window, the distribution of the lightning location data is statistically analyzed to generate a lightning density field, and the weather radar data is parsed into a radar reflectivity field. Map the lightning density field onto the same grid as the radar reflectivity field; The mapped lightning density field and radar reflectivity field are normalized and spatially superimposed to generate the feature set containing the superimposed composite intensity index and hail identification mark.

4. The method according to claim 3, characterized in that, The trend prediction based on the thunderstorm evolution characteristics, to obtain a prediction result reflecting the degree of thunderstorm impact within a preset time period, includes: Based on the movement trend of the thunderstorm cell, the predicted trajectory and predicted coverage area of ​​the thunderstorm cell within a preset time period are predicted. Determine the spatial overlap and transformation relationship between the predicted coverage area and the external monitoring area, the nearby area, and the local area; Based on the overlapping conversion relationship, the predicted time when the thunderstorm cell moves into or out of the external monitoring area, the nearby area and the local area is determined, and the prediction time window corresponding to each spatial partition is generated respectively. Based on the stability and evolution intensity attributes of the thunderstorm cells in their historical trajectories, the probability of thunderstorm activity occurring within the predicted time window of each spatial partition is calculated.

5. The method according to claim 4, characterized in that, The step of converting the prediction result into a discrete coded result conforming to the aviation meteorological message specification according to preset coding rules and consistency constraints includes: According to the preset coding rules, the prediction time window, occurrence probability and evolution intensity attributes of each spatial partition are mapped to elements to generate preliminary discrete coding results. Based on the consistency constraints, the preliminary discrete coding results are subjected to time-series stability smoothing and cross-message logic verification to determine the final discrete coding results that conform to the specifications.

6. The method according to claim 5, characterized in that, The step involves mapping the prediction time window, occurrence probability, and evolution intensity attributes of each spatial partition to elements according to the preset coding rules, generating preliminary discrete coding results, including: For the standard message time interval, when the occurrence probability exceeds the preset trigger threshold, a thunderstorm weather phenomenon identifier is determined; Based on the spatial partition to which the prediction time window belongs, the identifier is matched with the corresponding location attribute information to distinguish the current thunderstorm from nearby thunderstorms; Based on the composite intensity index and hail identification identifier in the evolution intensity attribute, modifiers reflecting precipitation intensity level or special weather phenomena are added to the identifier with location attribute to form the preliminary discrete coding result.

7. The method according to claim 5, characterized in that, The step of performing time-series stability smoothing and cross-message logic verification on the preliminary discrete coding results based on the consistency constraints to determine the final discrete coding results that conform to the specifications includes: Execution timing stability smoothing: The encoding result output at the previous moment is introduced as a reference state. If the change magnitude or duration of the preliminary discrete encoding result relative to the reference state does not reach a preset smoothing threshold, state transition is suppressed and the reference state is maintained. Perform cross-message logical verification: compare the smoothed encoding result with the actual situation elements in the airport routine weather report or airport special weather report released at the same time. If there is logical mutual exclusion, the encoding result is corrected or removed according to the preset business priority.

8. The method according to claim 6, characterized in that, The addition of modifiers reflecting precipitation intensity levels or special weather phenomena to the identifiers with location attributes includes: Based on the reflectance peak value in the composite intensity index, a strong, medium or weak intensity level is matched in the precipitation intensity level, and then converted into the corresponding positive and negative intensity modifiers. When the hail identification identifier is detected and its location attribute belongs to this area, the hail phenomenon code is appended after the thunderstorm weather phenomenon identifier. Based on the start and end times of the predicted time window, add trend indicator time group codes representing the start, end, or change of weather phenomena to the preliminary discrete coding results.

9. The method according to claim 7, characterized in that, The step of correcting or eliminating the encoding results according to preset service priorities includes: Establish a priority ranking based on operational safety, setting the priority of manually issued correction messages and important weather warnings higher than that of automatically generated preliminary coding results; If the smoothed encoding result is inconsistent with the real-time data released at the same time in terms of the presence or absence of thunderstorms and intensity level, the real-time data will be used as the standard, and the encoding result will be rolled back or deleted. When the change in the encoding result reaches the preset message release threshold, a special aviation meteorological message trigger command is generated, which includes the updated encoding content and release action suggestions.

10. An automatic message coding system for airport thunderstorms, characterized in that, include: The data acquisition module is used to acquire thunderstorm observation data in the airport area; The feature recognition module is used to identify thunderstorms based on the thunderstorm observation data and extract thunderstorm evolution features; The trend prediction module is used to predict trends based on the thunderstorm evolution characteristics and obtain prediction results that reflect the degree of thunderstorm impact within a preset time period in the future. The encoding conversion module is used to convert the prediction results into discrete encoding results that conform to the aviation meteorological report specifications and output them according to preset encoding rules and consistency constraints.

Images