A route lightning environment early warning method and system based on multi-source data

Abstract

The application relates to a route lightning environment early warning method and system based on multi-source data, which comprises the following steps: collecting multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time; performing fusion processing on the collected multi-source heterogeneous data to generate corresponding data packets for each data source; evaluating each data packet through three-layer data verification to generate a real-time lightning occurrence probability based on the data packet of each data source; and providing a route lightning composite early warning by comprehensively analyzing the real-time lightning occurrence probability based on the data packet of each data source.

Description

Technical Field

[0001] This application relates to the field of lightning warning in civil aviation, specifically to a method and system for early warning of lightning environment along routes based on multi-source data, which can be widely applied to scenarios such as lightning environment monitoring and aircraft lightning warning. Background Technology

[0002] Lightning is a common atmospheric discharge phenomenon in nature. When the atmospheric temperature is unstable, the water vapor supply is sufficient, and the convective wind speed is strong, ice crystals in the atmosphere collide and bounce off graupel particles, causing charge separation. The electric field in the cloud is enhanced, resulting in lightning.

[0003] Civil aircraft encounter lightning in two ways: natural lightning strikes occur when a lightning leader from the cloud's charge center gets close enough to the aircraft to be transferred to it; and triggered lightning strikes occur when the aircraft interacts with a strong electric field while flying in such an area. On average, each civil aircraft encounters lightning at least once a year, and statistics show that the vast majority of lightning strikes encountered by aircraft are triggered lightning strikes.

[0004] Although modern aircraft are specifically designed to protect against lightning, the potential hazards remain significant. If protection fails, the damage to civil aircraft caused by lightning can be incalculable. Therefore, there is an urgent need in the industry to increase research into lightning warning methods and improve the accuracy of warnings.

[0005] Currently, civil aircraft mainly rely on airborne weather radar to identify dangerous weather such as thunderstorms and cumulonimbus clouds, helping pilots avoid areas of strong convection, assisting the crew in choosing safe detour routes, and improving operational capabilities under complex weather conditions. However, this approach has certain limitations.

[0006] First, rerouting is uneconomical. Even with specialized lightning protection designs, lightning strikes can still cause ablation damage during operation. Therefore, aircraft cannot take off in thunderstorms and must detour around thunderstorm areas, leading to flight delays, passenger travel disruptions, and financial losses for airlines. Detours also consume more fuel.

[0007] Secondly, lightning monitoring is ineffective. Current data sources are limited. Because aircraft speed and heading change during flight, and thunderstorm clouds vary significantly in location and center position, current methods cannot address these variations. They can only preliminarily identify some characteristics of thunderstorm clouds and cannot accurately identify the electric field in the aircraft's location, especially for triggered lightning warnings.

[0008] To facilitate aircraft in accurately assessing the probability of lightning occurrence along various flight paths and to provide accurate information for multi-decision-making, it is urgent to research a lightning warning method that can provide accurate lightning warning information to aircraft, reduce false alarm / missed alarm rates, and provide data support for new technologies. Summary of the Invention

[0009] This application discloses a method and system for early warning of road lightning based on multi-source data.

[0010] According to a first aspect of this application, a method for early warning of airborne lightning environments based on multi-source data is provided, comprising: Collect multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time; The collected multi-source heterogeneous data is fused and processed to generate corresponding data packets for each data source; Each data packet is evaluated using a three-layer data verification process to generate a real-time lightning occurrence probability based on the data packets from each data source; and A comprehensive early warning system for road lightning is provided by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources.

[0011] According to a second aspect of this application, a route lightning environmental early warning system based on multi-source data is provided, comprising: A three-dimensional air-space-ground sensing network is configured to collect multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time; A data fusion processing device is configured to fuse the collected multi-source heterogeneous data to generate corresponding data packets for each data source; A data evaluation device is configured to evaluate each data packet using a three-layer data verification process to generate a real-time lightning occurrence probability based on the data packets from each data source; and The lightning warning device is configured to provide a comprehensive warning of lightning along a flight path by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources.

[0012] This overview is provided to introduce, in a simplified form, some of the concepts further described in the detailed description below. This overview is not intended to identify key or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter. Attached Figure Description

[0013] To describe how the above and other advantages and features of the invention are obtained, a more detailed description of the invention, which has been briefly described above, will be presented with reference to specific embodiments of the invention shown in the accompanying drawings. It will be understood that these drawings depict only exemplary embodiments of the invention and are therefore not intended to limit its scope. The invention will be described and explained using the drawings and with the aid of additional features and details, in which: Figure 1 An example flowchart of a route lightning environmental early warning method based on multi-source data according to an embodiment of this application is shown.

[0014] Figure 2 An example structural diagram of a route lightning environmental early warning system based on multi-source data according to an embodiment of this application is shown.

[0015] Figure 3 A schematic diagram of an example lightning warning display interface obtained by a route lightning environment warning scheme based on multi-source data according to an embodiment of this application is shown. Detailed Implementation

[0016] To overcome the problems mentioned in the prior art, this application provides a method and system for early warning of road lightning environment based on multi-source data.

[0017] This application provides a spatiotemporal fusion processing architecture for multi-source heterogeneous data. It constructs an air-space-ground three-dimensional sensing network to form a multi-source heterogeneous data fusion processing architecture; based on a multimodal data conversion channel and an adaptive dynamic weight allocation method, it calculates the real-time lightning occurrence probability after three layers of data verification. Compared with existing technologies, this application can obtain more accurate measurement data, reduce the impact of equipment errors, and intelligently calculate the correlation coefficient between significant lightning influence factors and lightning in the flight path area, improving the spatial resolution of measurement data. It is adaptable to various provinces, cities, and terrains, and is not limited to specific areas, making it effectively applicable to various flights.

[0018] The following is combined Figure 1 This document provides an example flow of a route lightning environmental early warning method based on multi-source data, according to an embodiment of this application.

[0019] First, in step 102, multi-source heterogeneous data related to lightning environment early warning are collected in real time from multiple data sources.

[0020] The lightning early warning system of this application constructs a three-dimensional sensing network that includes multiple data sources to collect multi-source heterogeneous data, in order to obtain historical data as well as real-time location of thunderstorm clouds, lightning impact factors, and other data.

[0021] Specifically, the airborne equipment of the three-dimensional air-space-ground sensing network preferably includes dual-polarization radar.

[0022] Radar, due to its high spatiotemporal resolution and radar reflectivity factor, has a certain indicative effect on the occurrence of lightning. Therefore, combining radar data with lightning location data is the most effective technical means for near-term lightning warning. Among them, dual-polarization radar is one of the newest and most effective means of detecting severe convective weather. It can send and receive pulses in both horizontal and vertical directions, thereby obtaining richer information about the cloud environment. In this application, the dual-polarization radar is mainly used to collect data related to lightning environment warning, such as thunderstorm cloud sweep data (elevation angle, azimuth angle, and radial distance), wind field speed, radar echo intensity, radar echo top height, liquid water content, and graupel specific mass. Based on these data and using methods such as interpolation or gridding (the most commonly used method is bilinear interpolation), three-dimensional individual maps of each thunderstorm region can be constructed.

[0023] The space-ground three-dimensional sensing network preferably includes: a geostationary meteorological satellite and a lightning imager (LMI). The lightning imager (LMI) can be installed as an additional component on the geostationary meteorological satellite.

[0024] The geostationary meteorological satellite is primarily configured to collect corresponding blackbody brightness temperature (TBB) data. TBB data refers to the radiation temperature of cloud tops or the Earth's surface as observed by the satellite, based on the principle of blackbody radiation. TBB is crucial for determining the intensity, altitude, and potential hazards of thunderstorms. By quantifying the cloud top temperature, blackbody brightness temperature (TBB) indirectly reveals the intensity and development altitude of convective clouds—the lower the TBB value, the higher the cloud top and the more intense the convection, thus providing a macroscopic "thermometer"-like indicator for assessing the lightning activity potential of thunderstorms.

[0025] The Lightning Imager (LMI) is configured to acquire lightning images. Its most significant feature is that it records large-scale, continuous, and real-time lightning "videos" from space, rather than simply taking a structural "photograph." The Lightning Imager can monitor all types of lightning (total lightning) in clouds in real time from space, providing the earliest core signals for early warning and trend judgment of thunderstorms by analyzing the frequency and location changes of lightning activity.

[0026] The preferred ground-based equipment for the three-dimensional air-space-ground sensing network is a lightning locator and a multi-dimensional ground-based electric field sensing device.

[0027] The lightning locator is mainly configured to detect the electromagnetic pulse signals emitted by lightning and collect key parameters such as the time, location (latitude and longitude), density, intensity, and polarity of the lightning strike, so as to accurately locate the lightning strike point and judge the intensity and trend of thunderstorms in lightning early warning.

[0028] The multi-dimensional ground-based electric field sensing device can acquire multi-dimensional data from the electric field sensing array around the aircraft. Specifically, it can be used to accurately collect multi-dimensional data such as electric field amplitude, polarity, and rate of change near the aircraft within a small area. The core advantage of polarization radar lies in identifying the phase state of particles within clouds, such as distinguishing between raindrops, ice crystals, graupel, and hail. This is the basis for determining whether a thunderstorm has the conditions for electrification. However, it cannot directly tell you the stage or intensity of the electrification process. The multi-dimensional data, including the polarity and rate of change of the electric field near the aircraft, accurately collected by the multi-dimensional ground-based electric field sensing device, can serve as direct evidence of charge separation and accumulation within the cloud. This data fills the gap in radar's ability to "see the electrification conditions but not the electrification results." For example, a high electric field amplitude directly indicates that the collision electrification process in that area is exceptionally intense, and the thunderstorm is in its active phase.

[0029] Therefore, by fusing data from these five different types of devices, lightning observations from different perspectives and with different physical quantities can be aligned within the same spatiotemporal framework. This allows for comprehensive analysis to uncover the synergistic early warning value of "1+1>2". For example, combining data from dual-polarization radar and multi-dimensional ground-based electric field sensors can provide three-dimensional individual maps of various thunderstorm regions, rendered with intensity markers for "charge density" or "electric field strength distribution". With this three-dimensional distribution of charged particles, aircraft can not only bypass thunderstorm clouds but also actively avoid the most active core regions within the clouds where the electric field is strongest and lightning is most likely to occur.

[0030] It should be understood that the specific devices disclosed in the described air-space-ground three-dimensional sensing network are provided for illustrative purposes only and are not intended to limit the scope. Technicians can add more devices, remove some devices, or replace some devices to implement this air-space-ground three-dimensional sensing network architecture according to specific application scenarios.

[0031] Subsequently, in step 104, the collected multi-source heterogeneous data is fused to generate corresponding data packets for each data source.

[0032] The fusion process includes: S1) Data classification and transformation processing of multi-source heterogeneous data. The lightning early warning system classifies and transforms source data based on the multi-modal data transformation channel, obtaining radar, electric field, satellite, and positioning instrument data, as well as lightning image sets, which consist of location information and lightning impact factors.

[0033] For example, using the radar data processing channel, spherical volumetric scanning data is converted into 3D mesh data through interpolation; using the electric field data processing channel, the original electric field time-series data is filtered and denoised, and the real-time electric field amplitude, polarity, and rate of change are calculated; using the satellite data processing channel, the blackbody brightness temperature is calculated from infrared channel data, cold cloud clusters are identified, and their center temperature, gradient, and other characteristics are calculated. For Lightning Imager (LMI) data, the Lightning Clustering Filtering (LCFA) algorithm is used to aggregate the original pixel-level optical events into "groups" based on spatiotemporal proximity, and then further aggregate them into "flashes," thus realizing the conversion from raw pixels to a complete lightning image. Using the lightning locator data processing channel, the TDOA algorithm is used to calculate the 3D coordinates of lightning occurrence based on the time difference of arrival of electromagnetic pulses received from multiple stations.

[0034] After the above processing, each modal data is transformed into a feature field with location information. Subsequently, the data undergoes further spatiotemporal alignment. Spatial alignment refers to projecting all data onto the same grid coordinate system. Temporal alignment refers to aligning data with different time resolutions to a unified time window, which can be achieved using interpolation or nearest neighbor matching.

[0035] The lightning impact factor refers to data indirectly related to lightning; when it exceeds a certain threshold, lightning may occur. As mentioned earlier, the lightning impact factors in the examples include: thunderstorm cloud scan data (elevation angle, azimuth angle, and radial distance), wind field speed, radar echo intensity, radar echo top height, liquid water content, graupel specific mass, electric field amplitude and polarity near the aircraft, TBB data, and the time, location (latitude and longitude), density, and intensity of the electromagnetic pulse signal emitted by the detected lightning, etc.

[0036] The lightning image set includes: lightning channel images generated by lightning locator clustering, and lightning distribution images synthesized by lightning imager.

[0037] S2) Construct a 3D grid map of each thunderstorm region based on data acquired by dual-polarization radar. Specifically, the thunderstorm cloud scan data (including elevation, azimuth, and radial distance, generally in the form of irregularly distributed data points) in the acquired radar data is interpolated or gridded to transform it into a regular 3D grid. Based on a set threshold (e.g., echo intensity threshold), spatially continuous (adjacent) grid points are grouped into a "cluster," and each cluster represents a single thunderstorm region. After constructing the 3D grid map of each thunderstorm region, the number of thunderstorm regions can be counted.

[0038] S3) By projecting the three-dimensional volume scan data of each thunderstorm region onto a two-dimensional grid map, the three-dimensional grid map of the thunderstorm region is compressed into a two-dimensional planar image to more intuitively display the horizontal coverage, intensity distribution, and movement trend of the thunderstorm. After obtaining the corresponding two-dimensional planar image of the thunderstorm region, the vector coordinates of the thunderstorm center are calculated.

[0039] S4) Match corresponding thresholds to each lightning influencing factor to generate corresponding data packets. The threshold refers to the numerical boundary used in lightning early warning research to determine whether a lightning influencing factor can cause lightning to occur. If the threshold is exceeded, lightning is considered to have occurred.

[0040] For example, based on the province / city, current season, and whether the terrain is mountainous or marine (climate and geographical factors that significantly influence thresholds in current research) of each grid point along the aircraft's current flight path, corresponding thresholds are retrieved from a database containing these lightning impact factors and assigned to the grid points. This way, each lightning impact factor measured at each map grid point will be matched with a corresponding threshold. The thresholds in the database are obtained by researchers through statistical analysis or fitting of the relationship between historical monitoring data of these lightning impact factors and lightning occurrence rates in the region over many years.

[0041] In this way, radar data packets, electric field data packets, satellite data packets, and locator data packets can be obtained, consisting of lightning impact factors, corresponding location information, and corresponding thresholds.

[0042] Next, in step 106, each data packet is evaluated using a three-layer data verification process to generate a real-time lightning occurrence probability based on the data packets from each data source. The evaluation includes aspects such as data correctness, consistency, and relevance, wherein: 1) Data accuracy assessment Data accuracy assessment refers to the process of self-checking the source data.

[0043] For example, the accuracy assessment of radar-acquired data packets mainly involves removing outlier data that may be false alarms, such as small-scale thunderstorms that are far from a large-scale thunderstorm area.

[0044] The accuracy assessment of electric field data packets refers to the process of quality inspection and reliability verification of the raw data collected by electric field sensors and its processed characteristics (amplitude, polarity, rate of change). The purpose is to eliminate invalid or erroneous data caused by equipment failure, environmental interference or physical anomalies, and to ensure that the electric field information used for lightning warning truly reflects the actual state of the atmospheric electric field.

[0045] The accuracy assessment of satellite data packets refers to the process of systematically checking and verifying the reliability of raw data from meteorological satellites (including Lightning Imager (LMI)) and their inversion products (such as Blackbody Brightness Temperature (TBB), lightning events, cloud classification, etc.). Its core purpose is to eliminate invalid data caused by sensor noise, calibration errors, cloud detection errors, or spatiotemporal matching problems, ensuring that the satellite information used for lightning warnings accurately reflects the macroscopic characteristics of cloud clusters and the state of lightning activity.

[0046] The accuracy assessment of lightning locator data packets refers to the process of systematically inspecting and verifying the reliability of the structured data output by the lightning locator (including lightning occurrence time, latitude and longitude coordinates, current intensity, polarity, number of return strokes, etc.). Its core purpose is to eliminate invalid or erroneous data caused by detector anomalies, electromagnetic interference, positioning algorithm errors, or physical parameter inversion deviations, ensuring that the positioning information entering the lightning early warning system accurately reflects the actual discharge characteristics of lightning.

[0047] 2) Data consistency assessment

[0048] Data consistency assessment refers to the process of verifying data using data from other sources.

[0049] For example, at longitude 120°51′E and latitude 30°40′N, the radar echo intensity is 22 dBz, corresponding to a threshold of 45 dBz. From the perspective of a single radar source, this area is not a thunderstorm area. However, when reading the location data packet at this coordinate, it is found that the area is marked as a lightning cluster. Upon further inspection of the lightning image taken by the satellite at this coordinate, it is found that the cloud image after semantic segmentation is also marked as a lightning cluster area. Therefore, the map grid is corrected to a thunderstorm area.

[0050] Therefore, data from one source can be cross-verified with data from other sources to ensure consistency between data from different sources, thereby identifying accurate thunderstorm areas and lightning cluster areas.

[0051] 3) Data Relevance Assessment

[0052] Data correlation assessment refers to the process of dynamically weighting lightning impact factors based on their correlation with lightning occurrence and calculating the real-time probability of lightning occurrence. The core of this dynamic weighting is weighting based on the correlation coefficients of each lightning impact factor. These correlation coefficients are used in the study to characterize the relationship between lightning impact factors and lightning occurrence.

[0053] Overall, a dataset of correlation coefficients for lightning impact factors in different regions and seasons is established based on historical research as prior knowledge. This is then combined with a deep learning model to dynamically assign corresponding correlation coefficients to the lightning impact factors obtained from each map grid along the current seasonal flight path.

[0054] Specifically, the first step is to construct a correlation coefficient dataset to establish a prior knowledge base reflecting the statistical relationship between lightning influencing factors and lightning occurrence, providing a benchmark for subsequent dynamic weighting. For example, multi-year historical lightning data for the target area is collected, including historical lightning events and lightning influencing factors recorded by ground-based lightning locators and satellite lightning imagers. Then, lightning events and lightning influencing factors are matched within a spatiotemporal window (e.g., ±5 minutes, 5km × 5km grid) to construct feature-label sample pairs. Next, correlation coefficient calculations are performed, using methods such as Pearson correlation coefficient, mutual information, and random forest feature importance to calculate the correlation strength between each lightning influencing factor and lightning occurrence. Finally, the corresponding correlation coefficient dataset is constructed based on the above statistical results.

[0055] Secondly, the weights of each lightning influencing factor are dynamically adjusted based on a deep learning model, so that the weight allocation can adapt to different terrain and weather scenarios.

[0056] In the model building and training phases, the deep learning model (such as an attention mechanism network model) is first trained using the aforementioned correlation coefficient dataset to initially construct the weights of lightning influencing factors during lightning occurrences. Subsequently, by using seasonal encoding (such as one-hot encoding or seasonal sequence numbers) as input features to the deep learning model, the model can learn the weight preferences of each factor under different seasons, thereby dynamically adjusting it. Simultaneously, by using landform type encoding (such as mountain / plain / coastal classification) or high-precision geographic information (such as altitude, slope, and distance from the sea) as model input, the model can employ differentiated weight allocation strategies for different landform regions.

[0057] Subsequently, after the model is trained, in practical applications, real-time observed lightning influencing factors such as radar echoes and wind fields are input into the model. Combined with the corresponding seasonal and geomorphological features, the weights of each lightning influencing factor can be dynamically generated to adapt to the current weather and geomorphological scenario.

[0058] Finally, the real-time lightning occurrence probability of each data packet in each grid is calculated by weighted averaging of the lightning impact factors in each data packet, thereby realizing the adaptive adjustment of weights as meteorological, seasonal and topographical conditions change.

[0059] Taking radar data packets as an example, based on a trained deep learning model, the correlation coefficient of radar echo intensity can be calculated as R1, the correlation coefficient of radar reflectivity as R2, and so on, until the correlation coefficient of the last nth lightning influence factor in the radar data packet is calculated as R. n Then, using a weighted average formula, for example, (R1+R2+…+R…) nThe probability of lightning occurrence based on radar data packets is calculated by dividing the result by n. Similarly, after processing all map grids accordingly, the probability of lightning occurrence based on radar data packets at any grid point on the map can be obtained.

[0060] Similarly, the above dynamic weight allocation and calculation can be performed on other types of packets to obtain the real-time lightning occurrence probability based on these packets.

[0061] Finally, after processing all data packets as described above, any grid point on the map can obtain multiple real-time lightning occurrence probabilities based on different data sources (i.e., different types of data packets).

[0062] In this way, by using the above three-layer verification mechanism, accurate thunderstorm areas and lightning cluster areas can be calculated and identified. Furthermore, by combining deep learning to dynamically assign lightning correlation coefficients to different seasons, landforms, and climate routes, the real-time lightning occurrence probability based on data packets from various data sources can be obtained.

[0063] Subsequently, in step 108, a comprehensive early warning of lightning along the route is provided by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources.

[0064] First, for a given grid, a composite lightning occurrence probability P(x,y,t) is generated based on the real-time lightning occurrence probabilities of the data packets from each data source for that grid. Here, P represents the composite lightning occurrence probability, with a value between 0 and 100%, x and y represent the grid coordinates, and t represents the generation time. The composite lightning occurrence probability P can be a weighted sum of the real-time lightning occurrence probabilities from each data source at that grid location. P(x,y,t)=P 雷达 (x,y,t) ·ω 雷达 + P 电场 (x,y,t) ·ω 电场 +P 卫星 (x,y,t) ·ω 卫星 + P 定位仪 (x,y,t) ·ω 定位仪 Among them, P 雷达 (x,y,t) is the probability of lightning occurring at time t in the radar data packet, ω 雷达 That is its corresponding weight; P 电场 (x,y,t) is the probability of lightning occurring at time t in the electric field data packet, ω 电场 That is its corresponding weight; P 卫星 (x,y,t) is the probability of lightning occurring at time t in the satellite data packet, ω卫星 That is its corresponding weight; P 定位仪 (x,y,t) is the probability of lightning occurrence at time t in the lightning locator data packet, ω 定位仪 That is its corresponding weight; Where, ω 雷达 +ω 电场 +ω 卫星 +ω 定位仪 =1.

[0065] The weight ω of the real-time lightning occurrence probability from each data source in P(x,y,t) can be determined based on the importance of the data packets from each data source when predicting the lightning occurrence probability. In practical applications, the real-time lightning occurrence probability based on radar data packets plays a dominant role, while data packets from other data sources are generally used as auxiliary references. Generally speaking, if only one data source's data packet contradicts the judgment of the dual-polarization radar data packet, then the result of the dual-polarization radar is taken as the standard; only if data packets from other data sources contradict the judgment of the radar data packet will the dual-polarization radar data packet be re-examined. Therefore, its weight ω 雷达 Typically, it can occupy more than half of the weight. Additionally, if a set of lightning images of the grid (x,y) is captured at time t, then P(x,y,t) can be directly set to 100%.

[0066] After calculating the real-time composite probability of lightning occurrence at all grid points along the aircraft's flight path, the areas with high lightning occurrence probabilities can be identified and displayed to the flight crew based on these real-time composite probabilities.

[0067] Furthermore, by analyzing the temporal changes in historical lightning occurrence probability values ​​in areas where thunderstorms and lightning clusters are located along flight routes (for example, to determine the increasing or decreasing trend of lightning occurrence probability generated in the past 3 hours on an hourly basis, obtain the previously calculated P(x,y,t-3), P(x,y,t-2), and P(x,y,t-1)), and combining this with real-time monitoring data from other data sources, such as the rate of change of electric field (whether the electric field in these areas gradually increases or decreases in these three time series), it is possible to predict the development trend (increasing, maintaining, or decreasing) of lightning occurrence probability values ​​in the identified thunderstorms and lightning clusters at future times, thereby achieving early warning of lightning along flight routes.

[0068] Furthermore, for the map grid, based on the real-time composite probability of lightning occurrence in each grid, a certain area around the aircraft is divided into multiple zones, and corresponding lightning warning levels are generated. These levels are then transmitted to the aircraft after confirmation by ground personnel. This provides a more intuitive and clear lightning warning. For example, a lightning warning level of "low" can be set for the zone with a probability of 0%-20%, "medium" for the zone with a probability of 20%-60%, and "high" for the zone with a probability of 60%-100%. Moreover, when displaying the map grid, different colors can be used to visually distinguish these zones with different lightning warning levels, making it easier for the crew to view.

[0069] For example, in Figure 3 The diagram illustrates an example lightning warning display interface obtained from a multi-source data-based route lightning environment warning scheme according to an embodiment of this application. The diagram provides a clear overview of flight and route information, identifying current lightning-influencing factors, the probability and severity of lightning warnings, locations along the route with thunderstorms and lightning clusters, and system-recommended actions (such as advising pilots to activate lasers).

[0070] After understanding the example flow of the multi-source data-based route lightning environmental early warning method of this application, the following is based on... Figure 2 To further understand an example structural diagram of a route lightning environmental early warning system based on multi-source data according to an embodiment of this application.

[0071] like Figure 2 As shown, the airway lightning environment early warning system includes an air-ground integrated sensing network 202, a data fusion processing device 204, a data evaluation device 206, and a lightning early warning device 208.

[0072] The air-space-ground integrated sensing network 202 is configured to collect multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time.

[0073] As mentioned above, the air-space-ground integrated sensing network may include dual-polarization radar, used to collect data related to lightning environment early warning, such as thunderstorm cloud body scan data (elevation angle, azimuth angle and radial distance), wind field speed, radar echo intensity, radar echo top height, liquid water content, graupel specific mass, etc.

[0074] Geostationary weather satellites are used to collect corresponding blackbody brightness temperature (TBB) data.

[0075] A lightning imager is used to capture images of lightning.

[0076] A lightning locator is used to detect the electromagnetic pulse signals emitted by lightning and collect key parameters such as the time, location (latitude and longitude), density, intensity, and polarity of the lightning.

[0077] A multi-dimensional ground-based electric field sensing device is used to acquire multi-dimensional data of the electric field sensing array around the aircraft, including electric field amplitude, polarity, and rate of change.

[0078] The data fusion processing device 204 is configured to fuse collected multi-source heterogeneous data to generate corresponding data packets for each data source.

[0079] The fusion process includes: S1) classifying and transforming multi-source heterogeneous data; S2) constructing a three-dimensional grid map of each thunderstorm region based on data acquired by dual-polarization radar; S3) compressing the three-dimensional grid map of the thunderstorm region into a two-dimensional planar image by projecting the three-dimensional volume scan data of each thunderstorm region onto a two-dimensional grid map, so as to more intuitively display the horizontal coverage, intensity distribution and movement trend of the thunderstorm; and S4) matching corresponding thresholds for each lightning influence factor to form corresponding data packets.

[0080] The data evaluation device 206 is configured to evaluate each data packet through three-layer data verification to generate a real-time lightning occurrence probability based on each data source. The evaluation includes the correctness, consistency, and correlation of the data to obtain the real-time lightning occurrence probability based on each data source.

[0081] The lightning warning device 208 is configured to provide a comprehensive lightning warning along the flight path by comprehensively analyzing the real-time lightning occurrence probability based on data packets from various data sources. Specifically, the lightning warning device 208 first generates a gridded lightning occurrence probability P(x,y,t) covering the region based on the real-time lightning occurrence probability of data packets from various data sources for each grid in the region. Then, it identifies and displays areas with high lightning occurrence probabilities to the flight crew based on the lightning occurrence probability of each grid point. Furthermore, the lightning warning device 208 can also predict the future trend of lightning occurrence probabilities within the identified thunderstorms and lightning clusters by analyzing the temporal changes in historical lightning occurrence probability values ​​of the identified thunderstorms and lightning clusters along the flight path and combining this with real-time monitoring data from other data sources. Additionally, the lightning warning device 208 is further configured to determine the lightning warning level for each interval within a certain range around the aircraft based on the lightning occurrence probability.

[0082] Example application scenarios: Hardware environment: In an example scenario, the flight path lightning environment early warning system based on multi-source data described in this application consists of a host computer, communication equipment, satellite-based, air-based, and ground-based equipment, as well as a display panel. It utilizes a spatiotemporal fusion processing architecture based on multi-source heterogeneous data to achieve accurate early warning of natural lightning and triggered lightning along the aircraft's flight path.

[0083] Satellite-based, airborne, and ground-based equipment are used to acquire lightning-related data in real time, including location and lightning impact factors; a host computer is deployed in a ground data center to process the received lightning-related data and generate early warning results; communication equipment is used to transmit the data to the host computer and transmit early warning results to the aircraft; a display panel is installed in the aircraft's cockpit to facilitate pilots in viewing early warnings and operational suggestions.

[0084] The host computer includes a data fusion processing device. This device fuses and processes various types of collected data, removes abnormal or poorly matched data, and calculates the real-time lightning occurrence probability of the current thunderstorm cloud area, lightning cluster area, and each grid point on the map. Subsequently, it generates the real-time composite lightning occurrence probability of all grid points in the area traversed by the aircraft's flight path and predicts the thunderstorm area and lightning cluster area after a certain period of time. The data fusion processing device can also further define the corresponding lightning warning level for each area around the aircraft based on the composite lightning occurrence probability.

[0085] Operational Example: After takeoff, each device transmits lightning-related data to the host computer at the ground data center. The host computer calculates and displays in real time the map of the area the aircraft's flight path passes through, the aircraft's current flight path, the aircraft's speed and altitude, the current thunderstorm cloud area, the lightning cluster area, the real-time composite probability of lightning occurrence at each map grid point, the predicted thunderstorm cloud area, the predicted lightning cluster area, and the corresponding lightning warning level for each interval within a certain range around the aircraft.

[0086] When an aircraft is about to enter a thunderstorm area, ground data center staff will determine whether to detour or make other decisions based on the current lightning warning level. They will also upload a map of the area around the aircraft's current location, the lightning warning level, and operational suggestions to the cockpit display panel for the pilot to evaluate and review.

[0087] Compared with existing lightning warning schemes, the scheme in this application has the following advantages: 1. Significantly improve the accuracy of lightning warnings; 2. The data sources are abundant, and the calculation methods are intelligent, covering all provinces and cities along the flight route; 3. It can provide early warnings for both natural lightning and lightning triggered by aircraft.

[0088] Although the techniques have been described using language specific to structural features and / or methodological actions, it should be understood that the appended claims are not necessarily limited to the described features or actions. Rather, these features and actions are described as exemplary forms of implementing these techniques.

[0089] The operations of the example processes are shown in separate boxes and are summarized with reference to these boxes. These processes are shown as a flow of logical boxes, each of which may represent one or more operations that can be implemented using hardware, software, or a combination thereof. In the context of software, these operations represent computer-executable instructions stored on one or more computer-readable media that, when executed by one or more processors, cause one or more processors to perform a given operation. Generally, computer-executable instructions include routines, programs, objects, modules, components, data structures, etc., that perform a particular function or implement a particular abstract data type. The order in which the operations are described is not intended to be construed as limiting, and any number of the operations may be executed in any order, combined in any order, subdivided into multiple sub-operations, and / or executed in parallel to implement the described process. The described process may be executed by resources associated with one or more computing devices, such as one or more internal or external CPUs or GPUs, and / or one or more pieces of hardware logic, such as FPGAs, DSPs, or other types of accelerators.

[0090] All of the methods and processes described above can be embodied in software code modules executed by one or more general-purpose computers or processors, and can be fully automated via these software code modules. These code modules can be stored on any type of computer-executable storage medium or other computer storage device. This code can also be packaged into corresponding computer program products. Some or all of these methods can alternatively be embodied in dedicated computer hardware.

[0091] Any routine description, element, or box in the flowcharts described herein and / or in the accompanying drawings should be understood as potentially representing a module, segment, or portion of code comprising one or more executable instructions for implementing a specific logical function or element in that routine. Alternative implementations are included within the scope of the examples described herein, wherein elements or functions may be removed or performed inconsistently with the order shown or discussed, including substantially synchronous or reverse order execution, depending on the functionality involved, as will be understood by those skilled in the art.

[0092] While different embodiments have been described above, it should be understood that they are merely examples and not limitations. Those skilled in the art will appreciate that various modifications in form and detail may be made without departing from the spirit and scope of the invention as defined in the appended claims. Therefore, the breadth and scope of the invention disclosed herein should not be limited by the exemplary embodiments disclosed above, but should be defined solely by the appended claims and their equivalents.

Claims

1. A method for early warning of en-route lightning environment based on multi-source data, comprising: Collect multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time; The collected multi-source heterogeneous data is fused and processed to generate corresponding data packets for each data source; Each data packet is evaluated using a three-layer data verification to generate a real-time lightning occurrence probability based on the data packets from each data source. as well as A comprehensive early warning system for road lightning is provided by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources.

2. The en route lightning environment warning method of claim 1, wherein, Also includes: Construct a three-dimensional air-space-ground perception network including the multiple data sources, wherein: The airborne equipment of the three-dimensional air-space-ground sensing network includes dual-polarization radar; The space-ground three-dimensional sensing network includes geostationary meteorological satellites and lightning imagers, wherein the lightning imagers are installed as an additional component on the geostationary meteorological satellites. The ground-based equipment of the three-dimensional air-space-ground sensing network includes lightning locators and multi-dimensional ground-based electric field sensing devices.

3. The en route lightning environment alerting method of claim 2, wherein, The step of fusing the collected multi-source heterogeneous data to generate corresponding data packets for each data source includes: Perform data classification and transformation on multi-source heterogeneous data; A three-dimensional grid map of each thunderstorm region is constructed based on the radar data collected by the dual-polarization radar. By projecting the three-dimensional volume scan data of each thunderstorm region onto a two-dimensional grid map, the three-dimensional grid map of the thunderstorm region is compressed into a two-dimensional planar image; and Match appropriate thresholds to each lightning impact factor to generate corresponding data packets; The threshold refers to the numerical boundary used in lightning early warning research to determine whether the lightning influencing factor can cause lightning to occur.

4. The airway lightning environment early warning method as described in claim 3, characterized in that, The step of evaluating each data packet through three-layer data verification to generate a real-time lightning occurrence probability based on each data source includes: The correctness of the data is assessed by self-checking the source data; Use external data for verification to assess data consistency; Based on the correlation between the lightning impact factor and lightning occurrence, the lightning impact factor is dynamically weighted and the corresponding real-time lightning occurrence probability is calculated to evaluate the correlation of the data.

5. The method for early warning of lightning conditions along flight routes as described in claim 4, characterized in that, The step of dynamically weighting the lightning impact factor based on its correlation with lightning occurrence and calculating the corresponding real-time lightning occurrence probability to assess the data correlation includes: A dataset of correlation coefficients for the lightning influencing factors in different regions and seasons was established based on historical research. The weights of each lightning influencing factor are dynamically adjusted based on a deep learning model, so that the weight allocation can be adapted to different weather and terrain scenarios. The real-time observed lightning impact factors are input into the deep learning model, and the weights of each lightning impact factor are dynamically generated to adapt to the current weather and landform scenario, taking into account the corresponding seasonal and landform features. The real-time probability of lightning occurrence for each data packet is calculated by weighting the lightning impact factors in each data packet.

6. The airway lightning environment early warning method as described in claim 5, characterized in that, The step of providing a comprehensive early warning of airway lightning by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources includes: The composite probability P(x,y,t) of lightning occurrence for a grid is generated based on the real-time lightning occurrence probability of data packets from each data source. Here, P represents the composite probability of lightning occurrence, x and y represent the grid coordinates, and t represents the generation time. The composite probability P of lightning occurrence is calculated by weighted summation of the real-time lightning occurrence probabilities of data packets from each data source of the grid. Based on the real-time composite probability of lightning occurrence across all grids, identify areas where thunderstorms and lightning clusters with high lightning occurrence probabilities are located; and By analyzing the temporal changes in the historical lightning occurrence probability values ​​of the areas where the identified thunderstorms and lightning clusters are located along the flight path, and combining this with real-time monitoring data from other data sources, the future trend of the lightning occurrence probability values ​​within the identified thunderstorms and lightning clusters can be predicted.

7. The method for early warning of lightning conditions along flight routes as described in claim 6, characterized in that, Also includes: Based on the real-time composite probability of lightning occurrence, the area around the aircraft is divided into multiple intervals and corresponding lightning warning levels are generated.

8. A route lightning environment early warning system based on multi-source data, comprising: A three-dimensional air-space-ground sensing network is configured to collect multi-source heterogeneous data related to lightning environment early warning from multiple data sources in real time; A data fusion processing device is configured to fuse the collected multi-source heterogeneous data to generate corresponding data packets for each data source; The data evaluation device is configured to evaluate each data packet through three layers of data verification to generate a real-time lightning occurrence probability based on the data packets from each data source. as well as The lightning warning device is configured to provide a comprehensive warning of lightning along a flight path by comprehensively analyzing the real-time lightning occurrence probability of data packets based on various data sources.

9. The route lightning environment early warning system as described in claim 8, characterized in that, The lightning warning device is also configured to predict the future trend of the lightning occurrence probability value within the identified thunderstorms and lightning clusters by analyzing the temporal changes of historical lightning occurrence probability values ​​in the areas where the identified thunderstorms and lightning clusters are located along the flight path, combined with real-time monitoring data from other data sources.

10. The en-route lightning environment early warning system as described in claim 8, characterized in that, The lightning warning device is also configured to determine the lightning warning level of each interval within a certain range around the aircraft based on the real-time composite probability of lightning occurrence.

Images