An airship route identification method and device

Abstract

The application provides a method and device for identifying airway of a floating object, and belongs to the technical field of data processing. The method comprises: aggregating and averaging multi-year same-period wind field data by hour to obtain a climatic wind field sequence; calculating forward FTLE and backward FTLE of each discrete spatial grid per day; for any discrete spatial grid of any day, if the forward FTLE is low and the backward FTLE is low, the discrete spatial grid is determined as a candidate area of an airway on the day; if a high-value area of the backward FTLE surrounds a closed area in which the forward FTLE is low, the discrete spatial grid is determined as a candidate area of a retention zone on the day; if the forward FTLE is high and the backward FTLE is low, the discrete spatial grid is determined as a dangerous area on the day; if the forward FTLE is high and the backward FTLE is high, the discrete spatial grid is determined as a forbidden area on the day; and the airway is identified based on the candidate area of the airway. The application improves the reliability and safety of a long-time task of the floating object.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and apparatus for identifying airship routes. Background Technology

[0002] Zero-thrust aerostats rely entirely on atmospheric winds for passive transport in high-altitude environments such as the stratosphere. Their flight path planning essentially involves finding transport channels that conform to the Lagrange particle trajectory. In long-endurance missions, accurately identifying flight paths that ensure the aerostat's safe and stable navigation, as well as avoiding hidden or dangerous areas that could lead to mission failure, is a crucial prerequisite for efficient aerostat flight and mission success. Therefore, the reliability and accuracy of aerostat flight path identification technology directly determine the success or failure of long-endurance mission planning and flight safety.

[0003] Existing airship route identification methods are mainly based on Eulerian instantaneous or average wind field data, and typically employ wind speed threshold methods to identify suitable areas for airships to stay or navigate, such as using wind speeds below 5 m / s as a criterion. However, on the one hand, methods based on Eulerian wind fields cannot accurately reflect the transport laws of Lagrange particles, leading to easily broken planned routes and low reliability of long-duration planning; on the other hand, relying solely on wind speed thresholds cannot identify concealed stagnation areas such as Lagrangian Coherent Structures (LCS) where wind speeds are not low but are bound by vortices. In the stratosphere, the edge wind speeds of many vortex structures can reach above 15 m / s, and the internal wind speeds are often between 5 and 10 m / s. Traditional wind speed threshold methods will misjudge these areas as safe navigation areas, but in reality, according to Lagrange transport theory, once particles enter such vortices, they are difficult to escape. If an airship mistakenly enters such a vortex, it will remain there for a long time, potentially leading to mission failure. Existing technologies have an extremely high false negative rate for such concealed stagnation areas.

[0004] Therefore, how to accurately identify real waterways and effectively avoid hidden stagnation areas based on Lagrange transport theory has become an urgent technical problem to be solved. Summary of the Invention

[0005] This application provides a method and apparatus for airship channel identification, which addresses the shortcomings of low reliability in channel planning in the prior art.

[0006] This application provides a method for identifying the airship's flight path, including the following steps: By aggregating and averaging the wind field data from the same period over many years on an hourly basis, we can obtain the climatological wind field sequence for each hour of the year. Based on the climatological wind field sequence, the finite-time Lyapunov exponent FTLE is calculated for each discrete spatial grid on a daily basis. The FTLE includes forward FTLE and backward FTLE. For any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. Based on the candidate waterway areas, waterways are identified; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0007] According to the airship channel identification method provided in this application, the channel candidate areas are screened for each day based on the following conditions: First condition: The forward FTLE of the candidate channel area is lower than the first preset quantile of the global FTLE on that day; Second condition: The average wind speed of the day reaches the wind speed threshold. Third condition: For connected components that satisfy the first and second conditions, calculate the aspect ratio of the minimum area bounding rectangle, where the aspect ratio is greater than the aspect ratio threshold. Fourth condition: For a connected region composed of channel candidate areas whose backward FTLE is higher than the second preset quantile of the global FTLE of the day, if the geometric closure of the connected region is greater than the closure threshold, the connected region is removed.

[0008] According to the airship channel identification method provided in this application, after screening the channel candidate areas, the method further includes: The candidate waterway areas that meet the conditions are marked with 8-connected components, and the connected components with an area greater than the area threshold are retained to obtain the daily connected component set of the climate state for the whole year.

[0009] According to the airship channel identification method provided in this application, the identification of the channel based on the channel candidate area includes: For the i-th connected component on day d and the j-th connected component on day d+1, calculate the area intersection-union ratio and centroid drift distance of the two connected components; If the area intersection-union ratio is greater than the spatial overlap threshold and the centroid drift distance is less than the drift distance threshold, then the i-th connected component on day d and the j-th connected component on day d+1 are determined to be two adjacent days of the same climatological channel, and a cross-day correspondence is established. After obtaining the cross-day correspondence throughout the year, a time-limited waterway is determined based on the life cycle of each climatological channel.

[0010] According to the airship channel identification method provided in this application, the airship channel identification method further includes: The following attributes are extracted for each climatological channel: start date, number of consecutive days, climatological average wind speed, climatological area, climatological centroid location, climatological geometric profile, and Lagrange coherence structure intensity.

[0011] According to the airship channel identification method provided in this application, the airship channel identification method further includes: Based on the number of consecutive days of each climatic channel, the persistence of each climatic channel is classified. Based on the climatological average wind speed of each of the aforementioned climatological channels, the channel speed is classified for each of the aforementioned climatological channels; Based on the persistence classification results and channel speed classification results of each of the aforementioned climatic channels, channel labels are generated for each of the aforementioned channels.

[0012] According to the airship channel identification method provided in this application, the stagnation candidate area is extracted based on the following steps to obtain the stagnation area: For the connected regions composed of the candidate regions whose backward FTLE is higher than the third preset quantile of the global FTLE of the day, connected regions with a geometric closure degree greater than the closure degree threshold are selected. For the closed ridge contours of each of the connected domains, a polygon filling algorithm is used to generate their internal regions; The intersection of the candidate regions where the forward FTLE is lower than the fourth preset quantile of the global FTLE for that day, and the internal regions, is taken as the candidate connected region; If the area of ​​the candidate connected region is greater than the area threshold, and the geometric closure of the candidate connected region is greater than the closure threshold, then the candidate connected region is determined to be a stagnant region.

[0013] According to the airship channel identification method provided in this application, the stagnation zone is verified based on the following steps: n test particles are uniformly released using the Monte Carlo method. The forward Lagrange trajectory of the test particles is integrated over D days, and the number of escaped particles that successfully cross the boundary of the retention area after D days is counted. The escape rate is determined based on the ratio of the number of escaping particles to the number of test particles; Based on the escape rate, the risk level of the retention area is determined.

[0014] According to the airship channel identification method provided in this application, the airship channel identification method further includes: Based on the number of consecutive days each candidate connected component is maintained, the candidate connected components are classified according to their persistence. Based on the climatological average wind speed of each candidate connected domain, the channel speed is classified for each candidate connected domain. Based on the persistence classification results and channel speed classification results of each candidate connected domain, a stagnation zone label is generated for each stagnation zone.

[0015] This application also provides an airship navigation path identification device, including the following modules: The data preprocessing module is used to: aggregate and average the wind field data from the same period over many years on an hourly basis to obtain the climatological wind field sequence for each hour of the year; The FTLE field construction module is used to: calculate the finite-time Lyapunov exponent FTLE for each discrete spatial grid on a daily basis, based on the climatological wind field sequence, wherein the FTLE includes forward FTLE and backward FTLE; The region division module is used for: for any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. The channel determination module is used to: identify a channel based on the channel candidate area; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0016] This application 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 any of the above-described airship route identification methods.

[0017] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the airship channel identification method as described above.

[0018] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described airship channel identification methods.

[0019] The airship flight path identification method and apparatus provided in this application constructs an hourly climatological wind field sequence for the entire year by aggregating and averaging wind field data from multiple years of the same period on an hourly basis. This overcomes the shortcomings of existing technologies that rely on Eulerian instantaneous or average wind fields, which cannot reflect the true transport laws of particles, and provides highly reliable underlying data support for long-endurance mission planning. Furthermore, it incorporates the Finite-Time Lyapunov exponent... Using the Exponent Forced Leakage (FTLE) as the core identification indicator, the properties of the spatial grid are precisely defined by calculating the combined state of forward and backward FTLE. When a grid exhibits both low forward and backward FTLE, it indicates that the particles in this region will not diverge rapidly over time, nor are there any strong attractive ridges from the outside. This conforms to the Lagrange transport characteristics of a real waterway and is thus accurately identified as a waterway candidate area. When a grid exhibits a low forward FTLE and is located in a closed region surrounded by a high backward FTLE ridge, it indicates that although the particles in this region are not easily diverged, there are strong backward attractive ridges, i.e., vortex closed boundaries. This is a typical dynamic characteristic of LCS-type concealed stagnation areas in the stratosphere where the wind speed is not low but the vortex is confined. Through this dynamic boundary identification mechanism based on the Lagrange coherent structure, those concealed stagnation areas covered by high wind speeds can be accurately identified, significantly reducing the risk of airships accidentally entering vortices and causing long-term stagnation or mission failure, thus improving the scientific accuracy and safety of airship waterway identification. In summary, this application, by constructing a climatological wind field sequence and combining it with forward and backward feature combinations of the finite-time Lyapunov index, achieves accurate planning of real waterways and effective avoidance of hidden vortex stagnation zones, significantly improving the reliability and safety of long-endurance airship missions. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this application 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 application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is one of the flowcharts illustrating the airship channel identification method provided in this application; Figure 2 This is the second flowchart of the airship channel identification method provided in this application; Figure 3 This is a schematic diagram of the airship navigation identification device provided in this application; Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] It should be noted that in the description of the embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, they can be fixed connections, detachable connections, or integral connections; they can be mechanical connections or electrical connections; they can be direct connections or indirect connections through an intermediate medium; and they can be internal connections between two elements. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0024] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0025] The following is combined Figures 1-4 This application describes the airship channel identification method and apparatus provided in its embodiments.

[0026] Figure 1 This is one of the flowcharts illustrating the airship path identification method provided in this application, such as... Figure 1 As shown, the method includes the following: S110, aggregate and average the wind field data of the same period over many years by hour to obtain the climatological wind field sequence for each hour of the year; S120, Based on the climatological wind field sequence, calculate the finite-time Lyapunov exponent FTLE for each discrete spatial grid on a daily basis, wherein the FTLE includes forward FTLE and backward FTLE; S130, for any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. S140, Based on the channel candidate area, the channel is identified.

[0027] Here, "FTLE low" means that the FTLE is lower than the first threshold, and "FTLE high" means that the FTLE is higher than the second threshold.

[0028] It should be noted that the execution subject of the airship channel identification method provided in this application embodiment can be a server, computer equipment, such as mobile phone, tablet computer, laptop computer, handheld computer, vehicle electronic equipment, wearable device, ultra-mobile personal computer (UMPC), netbook or personal digital assistant (PDA), etc.

[0029] Here, an airship is an aircraft that relies on the static buoyancy of air to take off and fly in the atmosphere, such as a stratospheric airship or a weather balloon. A zero-thrust airship does not provide its own power and relies entirely on the atmospheric wind field for passive transport.

[0030] In S110, wind field data for the same period over many years in the target airspace (such as the stratosphere) are acquired, including wind speed and wind direction, and aggregated and averaged at hourly resolution. For example, the wind field data at 08:00 on January 1st of the past 15 years are averaged to obtain the wind field data for the 8th hour of the first day of the year. The wind field data for each hour of the year are averaged to finally generate an hourly climatological wind field sequence covering 365×24=8760 hours of the year.

[0031] In the specific implementation process, ECMWF ERA5 global multi-year (N) yearHourly reanalysis data (≥15 years). Horizontal resolution is 0.25° × 0.25°, and temporal resolution is 1 hour. Required variables include: horizontal wind field components. and , Longitude Latitude. The covering pressure layer is selected based on the typical operating altitude of the airship, with the default layers being 30hPa, 50hPa, 70hPa, 100hPa, 200hPa, 250hPa, and 300hPa, corresponding to the stratosphere at altitudes of 10–25 km. Synthetic wind speed .

[0032] Here, N year ≥15 is the preferred age range for climatological statistics, based on: (1) The stratospheric quasi-two-year oscillation (QBO) cycle is about 28 months. The 15-year sample covers about 6 complete QBO cycles, which can effectively smooth out this dominant low-frequency variability. (2) The solar activity cycle is about 11 years, and the 15-year sample covers more than one complete cycle, which can effectively reduce the systematic bias of solar forcing; (3) The ENSO cycle is about 2 to 7 years, and the 15-year sample covers 2 to 7 complete events, with sufficient representativeness of interannual variability.

[0033] For missing values, spatial bilinear interpolation and temporal linear interpolation are used to fill them in: (1) First, use bilinear interpolation with four valid grid points around the missing measurement point; (2) If there are fewer than 4 grid points, the number of grid points will be increased to 8. (3) If it is still insufficient, use linear interpolation of time 1 time before and after; (4) If the time neighborhood is also missing, it is marked as permanently missing and will not be included in subsequent statistics. If the proportion of permanently missing grid points on a certain day exceeds 5% (this threshold can be flexibly adjusted according to data quality), the data on that day is marked as incomplete, the product on that day is marked with a data quality warning, the missing grid points are filled with null values ​​NaN, and are skipped in subsequent FTLE calculations.

[0034] Furthermore, for the original hourly wind field over 15 years, according to d (d=1,2,…,365)+UTC Hourly (0–23) time periods were averaged over many years to obtain a continuous hourly climatological wind field sequence: , , ; in, Longitude Latitude For Heaven, For a time; This sequence is a continuous time-varying wind field for 365 days × 24 hours throughout the year. It filters out short-term weather disturbances and interannual anomalies, preserves the large-scale climatological circulation patterns in the stratosphere, and meets the basic requirements of the Lagrange trajectory integral for time series. It significantly reduces the amount of computation while ensuring physical rationality.

[0035] In S120, the Finite-Time Lyapunov Exponent (FTLE) is a mathematical index used to quantify the divergence or convergence rate of trajectories of adjacent particles in a fluid or atmosphere over a finite time interval. A larger FTLE value indicates faster particle separation, typically corresponding to transport barriers or boundaries of coherent structures in the fluid. A smaller FTLE value indicates relatively stable particle trajectories. The forward FTLE is calculated by integrating backward over a period of time from the current moment, characterizing the divergence trend of particles over a future time period. The backward FTLE is calculated by integrating backward over a period of time from the current moment, characterizing the convergence trend of particles over a past time period, and is often used to identify attractive Lagrange coherent structures.

[0036] In this embodiment, based on the aforementioned climatological wind field sequence, the hourly wind field for each day is spatially divided into discrete grids. Virtual particles are released for each grid point, and forward and backward trajectory integrations are performed to calculate the forward and backward FTLE values ​​for each discrete spatial grid on that day.

[0037] Preferably, the integration time is fixed at |T|=7 days. |T|<5 days will lead to insufficient FTLE structure resolution, and small-scale transport channels will not be fully evolved and will be missed, reducing the ability to capture dynamic features. |T|=10 days will result in a linear increase in computation, and at the same time, the excessively long integration will cause FTLE to become overly smooth, weakening local separation and mixed boundary recognition. In engineering, |T|=7 days ( (t=1 hour) is a trade-off between accuracy, computational cost, and structural fidelity.

[0038] In the specific implementation process, the initial time ℎ of calendar day d+ is used as the starting point for integration. Hourly wind fields from d-7 to d+7 (a total of 15 days) are extracted from the aforementioned climatological wind field sequence and used as input data for Lagrange integration, specifically for integrating the Lagrange trajectory over the previous 7 days and the next 7 days. For boundary days, i.e., d<8 or d>358, a cyclic modulo method is used to map dates outside the range of 1 to 365 into the valid range. For example, when d=1, d-7=359, and the corresponding wind field data is obtained using the annual cyclical characteristics of the climatological wind field.

[0039] Furthermore, for each grid point, particles are released centered on that grid point, while simultaneously applying four tiny perturbations in the longitudinal and latitudinal directions, causing the particles to deflect by a distance... δx Δ is the grid spacing xThe fourth-order Runge-Kutta method (RK4) is used for trajectory integration, and the integration step size is dynamically adjusted according to the severity of the flow field changes. The calculation is denser in complex regions such as jet streams and sparser in stable regions, balancing accuracy and efficiency. Calculate the deformation gradient tensor and the right Cauchy-Green tensor . Approximated by finite difference: ; FTLE is defined as: ; in, λ max The maximum eigenvalue of c represents the maximum stretching rate experienced by the fluid micro-element over 7 days; Forward FTLE ( Physical meaning: T=7 days, a high value indicates that the particle trajectory is diverging (repulsive boundary), and a low value indicates that the particle trajectory is stable and not easily dispersed; Backward FTLE ( Physical meaning: T = -7 days. A high value indicates that particles are attracted and converged (attraction boundary), while a low value indicates that particles are trapped (vortex core).

[0040] In S130, a first threshold (low threshold) and a second threshold (high threshold) are set to divide each discrete spatial grid into four quadrants: First quadrant: Channel candidate area, forward FTLE < first threshold and backward FTLE < first threshold, indicating that particles in this area are not likely to diverge in the future and there has been no strong attraction source in the past, making it a stable and safe transport channel; Second quadrant: Retention candidate region, a closed region surrounded by a high-value ridge line where the forward FTLE is less than the first threshold and the backward FTLE is greater than the second threshold. This indicates that there is a strong attractive boundary (such as a vortex) in this region, and particles are difficult to escape after entering. It belongs to the hidden retention region. Third quadrant: Danger zone, forward FTLE > second threshold and backward FTLE < first threshold, indicating that particles in this region will diverge violently in the future and may encounter strong wind shear or turbulence, posing a threat to the attitude control of the airship; Quadrant 4: Absolutely forbidden zone. Forward FTLE > second threshold and backward FTLE > second threshold, indicating that the dynamic environment in this region is extremely complex, with both strong attraction and strong repulsion. Aircraft should absolutely avoid this region.

[0041] In practice, the third and fourth quadrants are considered danger zones and are not considered as product output, but they can be marked as avoidable areas in the verification module.

[0042] Here, the first threshold and the second threshold can be adaptively set according to actual usage needs, with the first threshold being less than the second threshold.

[0043] In S140, connected domain analysis or topology extraction is performed on the continuously distributed channel candidate area grid in both spatial and temporal dimensions to ultimately identify continuous and smooth airship channels.

[0044] The airship route identification method provided in this application constructs an hourly climatological wind field sequence throughout the year by aggregating and averaging wind field data from multiple years of the same period on an hourly basis. This overcomes the shortcomings of existing technologies that rely on Eulerian instantaneous or average wind fields, which cannot reflect the true transport laws of particles, and provides highly reliable underlying data support for long-endurance mission planning. Furthermore, it incorporates the Finite-Time Lyapunov exponent... Using the Exponent Forced Leakage (FTLE) as the core identification indicator, the properties of the spatial grid are precisely defined by calculating the combined state of forward and backward FTLE. When a grid exhibits both low forward and backward FTLE, it indicates that the particles in this region will not diverge rapidly over time, nor are there any strong attractive ridges from the outside. This conforms to the Lagrange transport characteristics of a real waterway and is thus accurately identified as a waterway candidate area. When a grid exhibits a low forward FTLE and is located in a closed region surrounded by a high backward FTLE ridge, it indicates that although the particles in this region are not easily diverged, there are strong backward attractive ridges, i.e., vortex closed boundaries. This is a typical dynamic characteristic of LCS-type concealed stagnation areas in the stratosphere where the wind speed is not low but the vortex is confined. Through this dynamic boundary identification mechanism based on the Lagrange coherent structure, those concealed stagnation areas covered by high wind speeds can be accurately identified, significantly reducing the risk of airships accidentally entering vortices and causing long-term stagnation or mission failure, thus improving the scientific accuracy and safety of airship waterway identification. In summary, this application, by constructing a climatological wind field sequence and combining it with forward and backward feature combinations of the finite-time Lyapunov index, achieves accurate planning of real waterways and effective avoidance of hidden vortex stagnation zones, significantly improving the reliability and safety of long-endurance airship missions.

[0045] In an optional embodiment, the candidate waterway areas are screened for each day based on the following criteria: First condition: The forward FTLE of the candidate channel area is lower than the first preset quantile of the global FTLE on that day; Second condition: The average wind speed of the day reaches the wind speed threshold. Third condition: For connected components that satisfy the first and second conditions, calculate the aspect ratio of the minimum area bounding rectangle, where the aspect ratio is greater than the aspect ratio threshold. Fourth condition: For a connected region composed of channel candidate areas whose backward FTLE is higher than the second preset quantile of the global FTLE of the day, if the geometric closure of the connected region is greater than the closure threshold, the connected region is removed.

[0046] In this embodiment, the forward FTLE values ​​of all discrete spatial grids globally on a given day are calculated, and a first preset quantile, such as 30%, is extracted. Only candidate channels with forward FTLE values ​​below this quantile are retained. This step ensures that the selected channels not only have low absolute values ​​but also belong to the top 30% of the most stable regions globally on that day, excluding regions that are relatively volatile globally but are below a fixed threshold. Further, it is determined whether the average wind speed of the retained regions on that day reaches a preset wind speed threshold. Since zero-thrust aerostats rely on wind transport, if the wind speed is too low, the aerostat will not be able to obtain sufficient power for long-duration maneuvers. This condition eliminates invalid channels that are stable but stagnant. For channels that simultaneously meet the first... For the connected regions under the second condition, calculate the aspect ratio of the minimum area bounding rectangle. If the aspect ratio is greater than the set aspect ratio threshold, it is retained. This step ensures from a geometric perspective that the retained area is a narrow channel, which conforms to the basic physical definition of a waterway. Finally, for the above-retained connected regions, check whether there are grids with a backward FTLE higher than the second preset quantile (e.g., 70%). If so, it indicates that there is a strong attracting ridge inside or at the edge of the connected region. Further calculate the geometric closure of the connected region. If the closure is greater than the closure threshold, it indicates that this is a highly closed vortex structure. At this time, even if the forward FTLE of the area is low and the wind speed meets the standard, it must be removed to prevent the airship from accidentally entering the concealed and stagnant area.

[0047] In the specific implementation process, for each day d∈[1,365], identification is performed based on the climatological FTLE field: (1) Condition 1, LCS stability: forward FTLE low value region <P 30 ( ),Right now Below the 30th percentile globally on that day, a low FTLE indicates small fluid deformation and stable particle trajectories, which is a necessary condition for the existence of a transport channel; (2) Condition 2, transport capacity: average wind speed over the 24 hours of the day: ; High wind speeds are a necessary condition for transport channels. 10 m / s can effectively filter out weak wind areas in the background while preserving usable channels at the edge of the jet stream. This threshold can be adjusted between 10 and 20 m / s. ; (3) Condition 3, geometric shape, aspect ratio of the minimum area bounding rectangle: For the connected components that satisfy conditions (1) and (2), extract the latitude and longitude coordinates of their boundary grid points, and use the rotating caliper algorithm to calculate the minimum area bounding rectangle (MABR): Calculate the convex hull of the boundary grid points of a connected domain; For each edge of the convex hull, treat it as an edge of a rectangle, and calculate the smallest rectangle that contains the convex hull; Traverse all edges and select the rectangle with the smallest area as the MABR; Calculate the aspect ratio AR=L long / L short >3, where L long and L short These are the long and short side lengths (spherical distance) of the MABR, respectively. The strip structure ensures that the channel has long-distance transport capabilities, rather than closed circulation. This threshold range can be adjusted between 2.5 and 3.5. (4) Condition 4, exclude vortices: exclude the internal region (vortex core) surrounded by the high-value ridge line of the backward FTLE closure, and avoid misjudging the structure located in the vortex core as a channel; ; right Grid points above the 70th percentile of the global total for that day are labeled as 8-connected components, and the geometric closure of each connected component is calculated: c = 4πA / P 2 ; Where A is the area and P is the perimeter, the threshold corresponds to a shape that is close to a circle or ellipse, with closed circulation characteristics. The preferred threshold c>0.6 is used to determine that the connected domain is a closed structure; otherwise, it is determined to be an open structure. This can effectively screen vortex structures with long-term swirling and binding potential, and exclude narrow convergence zones and broken non-closed structures.

[0048] Furthermore, on the global 0.25° grid, the longitude direction is cyclic (0°=360°), and the 8-connected domain markers at the longitude boundaries (such as 0.25°E and 359.75°E) adopt periodic extension, treating the longitude grid as a ring topology. The eastern neighbor of the boundary grid point is the 0° longitude grid point, and the western neighbor is the 359.75° longitude grid point, ensuring that cross-domain connected domains are correctly merged.

[0049] For regions with latitude ≥ 87.5°, the conventional 8-neighborhood connected domain marking method fails due to meridian convergence. Polar stereo projection is used to assist in the determination. The region from 87.5° to 90° is projected onto the polar stereo plane, and 8-neighborhood connected domain marking is performed in this plane. The area calculation also needs to be switched to the polar stereo projection coordinate system simultaneously to avoid systematic errors in the spherical area formula at high latitudes.

[0050] The airship channel identification method provided in this application converts absolute thresholds into relative thresholds through a first condition, accurately extracting the most stable preceding regions globally on that day, eliminating evaluation biases caused by seasons and weather cycles, and improving the robustness of channel identification. The second condition couples Lagrange stability with Eulerian dynamics effectiveness, ensuring that only regions simultaneously satisfying trajectory non-divergence and possessing transport kinetic energy are truly suitable physical channels for airships to perform long-endurance missions, avoiding the planning of safe but immobile invalid paths. The third condition imposes topological constraints from a geometrical morphology perspective, filtering out linear structures with clear flow directions and eliminating planar regions lacking channel attributes, making the identification results more consistent with the physical morphology of real channels. The fourth condition constructs a dual immune mechanism against LCS-type vortices, reducing the false negative rate for high-wind-speed concealed stagnation areas and improving the safety of long-endurance airship missions.

[0051] In an optional embodiment, after screening the candidate waterway areas, the method further includes: The candidate waterway areas that meet the conditions are marked with 8-connected components, and the connected components with an area greater than the area threshold are retained to obtain the daily connected component set of the climate state for the whole year.

[0052] In this embodiment, grid points that satisfy all four conditions are labeled as 8-connected components, and the area A > 5000 km² is retained. 2 The connected components. Area threshold 5000km. 2 It can preserve a narrow, elongated rapid flow path while filtering out minute debris caused by noise. This threshold can be adjusted from 3000-10000 km depending on mission requirements. 2 Range adjustment.

[0053] This yields the daily connected domain set of climatological conditions for the entire year (365 days). ,in Let md represent the j-th climatological connected component on day d, and md be the number of connected components identified on day d. Each channel is a spatial connected component with independent area, centroid, and geometric profile.

[0054] The airship channel identification method provided in this application restores discrete qualified points into a continuous topological structure by marking 8-connected components, ensuring the physical continuity of the identified channel in spatial morphology; by filtering by area threshold, connected components with an area smaller than the threshold are eliminated, reducing the invalid search space of subsequent route planning algorithms and improving computational efficiency.

[0055] In an optional embodiment, identifying the waterway based on the waterway candidate area includes: For the i-th connected component on day d and the j-th connected component on day d+1, calculate the area intersection-union ratio and centroid drift distance of the two connected components; If the area intersection-union ratio is greater than the spatial overlap threshold and the centroid drift distance is less than the drift distance threshold, then the i-th connected component on day d and the j-th connected component on day d+1 are determined to be two adjacent days of the same climatological channel, and a cross-day correspondence is established. After obtaining the cross-day correspondence throughout the year, a time-limited waterway is determined based on the life cycle of each climatological channel.

[0056] In this embodiment of the application, for the first d The first day i The connected component and the first d +1 day j For each connected component, calculate its spatial overlap and positional offset. Set spatial overlap and drift distance thresholds. If the area intersection-union ratio is greater than the spatial overlap threshold and the centroid drift distance is less than the drift distance threshold, then the two connected components are considered to have high physical continuity and inheritance, belonging to the adjacent two-day performance of the same climatic channel, thus establishing a cross-day correspondence. Traverse the daily connected component set throughout the year, and use the path search algorithm in graph theory to connect all connected components that have established cross-day correspondences to form complete climatic channels. The start and end dates of each channel constitute the life cycle of the channel, and finally output a channel with a clear time limit.

[0057] In the specific implementation process, for the daily masking sequence of 365 climatological conditions throughout the year, the IoU and centroid drift distance of adjacent daily masking pairs are calculated: (1) Spatial overlap IoU ; Where Area(·) represents the spherical area of ​​the mask coverage area. Each grid cell is considered as a spherical quadrilateral, and its area is precisely calculated using the formula A. cell =R2·Δλ·|sin Where R = 6371 km is the Earth's radius; Δλ is the network's longitude span, in radians. Let A be the latitude of the north and south boundaries of the grid, and let A be the total area of ​​the connected domains. cell The sum of these values ​​automatically compensates for the east-west contraction effect of high-latitude grid points. In polar regions ≥87.5°, the area calculation switches to the polar stereo projection coordinate system.

[0058] (2) Distance of centroid drift ; Where Centroid(·) is the area-weighted centroid of the mask, and Distance(·) is the distance to the great circle of the sphere.

[0059] (3) Reference length (average of two-day equivalent radius) ; L ref This is a reference length based on the area equivalent radius (assuming the area is equivalent to a circle), used as an order of magnitude constraint for the centroid drift criterion.

[0060] (4) Joint judgment based on two criteria Criterion 1 (morphological stability): This criterion ensures that two adjacent sun masks maintain significant spatial overlap, with a preferred threshold. ; The preferred threshold is 0.55, which is determined based on statistical analysis of the daily drift distance of typical stratospheric channels. It can be adjusted in the range of 0.55 to 0.65 based on sensitivity analysis of historical data.

[0061] Criterion 2 (Position Stability): This criterion ensures that the centroid drift of the channel between two adjacent days does not exceed half of the equivalent radius, in order to capture the overall translational channel. ; If both criteria are met, then it is considered that... and For two adjacent days of the same channel, establish a cross-day correspondence. If a channel i on day d cannot find a matching object that meets the criteria on day d+1, the lifecycle of that channel is interrupted on the next day and is no longer tracked; if a similar channel appears on subsequent days, it is considered a new independent channel.

[0062] In the specific implementation process, for the M candidate channels on day d and the N candidate channels on day d+1, an M×N IoU matrix and a centroid drift matrix are constructed. A greedy matching strategy is adopted: all candidate pairs that satisfy the two criteria are sorted in descending order of IoU, and the pair with the largest IoU is matched first. The candidate regions that have already been matched will not participate in subsequent matching. If a candidate channel i on day d cannot find a matching object that satisfies the criteria on day d+1, the life cycle of the channel is interrupted on the next day and will no longer be tracked. If a similar channel appears on a subsequent date, it will be regarded as a new independent channel. If a candidate channel on day d+1 satisfies the criteria with multiple candidate channels on day d, the best match is selected according to the principle of the largest IoU, and the life cycle of the unmatched candidate channels on day d is interrupted.

[0063] The airship channel identification method provided in this application embodiment ensures the continuity of the channel in spatial topology by area intersection-union ratio, while the centroid drift distance ensures the rationality of the channel in dynamic transport. By eliminating pseudo-continuous matching caused by accidental similarity of local wind fields through dual verification, the absolute spatiotemporal coherence of the extracted climatological channel is ensured.

[0064] In an optional embodiment, the airship path identification method further includes: The following attributes are extracted for each climatological channel: start date, number of consecutive days, climatological average wind speed, climatological area, climatological centroid location, climatological geometric profile, and Lagrange coherence structure intensity.

[0065] Here, the start date refers to the specific date on the timeline when a climatological channel is first identified and a cross-day correspondence is established, marking the generation of the transport channel; the continuous maintenance days refer to the total number of days the climatological channel exists continuously on the timeline from the start date, reflecting the life cycle length and stability of the channel; the climatological average wind speed refers to the statistical average of all hourly wind field data within the coverage area of ​​the climatological channel throughout its entire life cycle, characterizing the channel's average transport capacity for airships; the climatological area refers to the statistical mean or integral value of the spatial coverage area (connected domain area) of the climatological channel during its life cycle, reflecting the channel's physical width and capacity. Capabilities; Climatic centroid location refers to the average position of the geometric center of a climatic channel in three-dimensional space (longitude, latitude, altitude) or two-dimensional space during its life cycle, used to characterize the macro-geographical location of the channel; Climatic geometric profile refers to the statistical characteristics of the spatial morphology of a climatic channel during its life cycle (such as average aspect ratio, shape factor, etc.), reflecting whether the channel is straight, curved, or patchy; Lagrange coherent structure strength refers to the dynamic index characterizing the ability of the channel boundary (composed of FTLE ridges) to bind or transport particles, usually quantified by the mean or extreme value of FTLE within the channel or at the boundary. The higher the value, the clearer and more stable the dynamic boundary of the channel.

[0066] In the specific implementation process, from the set of connected domains for 365 days a year, all continuous and stable climatological channels are tracked through the above-mentioned cross-day matching relationship, and the following attributes are extracted for each climatological channel: Start date; Number of consecutive days maintained (D) count Climatic average wind speed Take the average wind speed at all grid points within the channel. The area-weighted average, i.e. Climatic Area: For each day within the channel's lifecycle, calculate the total area of ​​the connected domain for that day, then take the median of all days within the lifecycle; Climatic Centroid Location: For each day within the channel's lifecycle, calculate the centroid of the connected domain for that day using area weighting, then take the median of all days within the lifecycle; Climatic Geometric Contour: Take the union of the contours of all days within the channel's lifecycle as the representative contour for visualization and spatial querying. If a single representative contour is needed, take the middle day of the lifecycle (D). count When the number is odd, take the (D)th number. count +1) / 2 days, for even numbers, take the connected component contour of the Dcount / 2th day; LCS strength: It reflects the Lagrange stability of the channel.

[0067] The airship channel identification method provided in this application transforms purely hydrodynamic results into engineering-usable parameters by extracting climatological average wind speed, area, and geometric contour. The extraction of multi-dimensional attributes enables the planning algorithm to perform multi-target matching and optimal path selection based on mission payload, target area, and flight duration, greatly improving the precision of mission planning. The number of consecutive maintenance days directly reflects the channel's lifespan, and the LCS intensity reflects the channel boundary's anti-interference capability. By extracting the start date, the number of consecutive maintenance days, and the LCS intensity, the risk of mid-flight failure in long-duration missions is reduced.

[0068] In an optional embodiment, the airship path identification method further includes: Based on the number of consecutive days of each climatic channel, the persistence of each climatic channel is classified. Based on the climatological average wind speed of each of the aforementioned climatological channels, the channel speed is classified for each of the aforementioned climatological channels; Based on the persistence classification results and channel speed classification results of each of the aforementioned climatic channels, channel labels are generated for each of the aforementioned channels.

[0069] In this embodiment, the number of consecutive days of each climatic channel is obtained, and it is mapped to the corresponding persistence level, such as short-term, regular, and long-term, according to a preset persistence classification rule, such as a number of days threshold range. The climatic average wind speed of each climatic channel is obtained, and it is mapped to the corresponding speed level, such as low speed, medium speed, and high speed, according to a preset speed classification rule, such as a wind speed threshold range. The above classification results are combined, and a unique channel label is generated for each climatic channel according to a predefined label generation rule. This label is written into the channel database and serves as the core index for the channel's external interactions.

[0070] In the specific implementation process, the persistence classification is shown in Table 1 below, and the channel speed classification is shown in Table 2 below. The wind speed boundaries for each level are adjusted according to the characteristics of the stratospheric jet stream and the mission requirements. The combination format is [Persistence]-[Speed] stable channel. For example, a long-term-high-speed stable channel indicates that the channel has been maintained continuously for more than 15 days with an average wind speed exceeding 30 m / s, and a medium-term-medium-speed stable channel indicates that it has been maintained continuously for 10 to 15 days with an average wind speed of 20 to 30 m / s.

[0071] Table 1 Persistence Classification Channel sustainability level definition Judgment criteria Long-term shipping lanes Maintaining stability for more than 15 consecutive days <![CDATA[D count ≥ 15 days Mid-term Channel Maintain stable for 10-15 days <![CDATA[10 ≤ D count <15 days] Short-term shipping lanes Maintain stable for 5-10 consecutive days <![CDATA[5 ≤ D count <10 days]]> Table 2 Channel Speed ​​Classification

[0072] The airship channel identification method provided in this application discretizes continuous physical quantities into levels with business semantics through continuous grading and channel speed grading, reducing the decision complexity of upper-layer algorithms. By generating channel labels, massive amounts of data are pre-calculated and dimensionality reduced. When an airship receives a mission instruction requiring a high-speed and long-life channel, the system can quickly filter out a pool of candidate channels simply by matching labels, significantly improving the response speed and computational efficiency of long-endurance mission planning. Furthermore, it enables the airship channel identification system to have standardized and productized output capabilities, solving the problem that existing technologies lack route base maps that can be directly accessed by GIS and cannot support long-endurance mission planning.

[0073] In an optional embodiment, the retention candidate region is extracted based on the following steps to obtain the retention region: For the connected regions composed of the candidate regions whose backward FTLE is higher than the third preset quantile of the global FTLE of the day, connected regions with a geometric closure degree greater than the closure degree threshold are selected. For the closed ridge contours of each of the connected domains, a polygon filling algorithm is used to generate their internal regions; The intersection of the candidate regions where the forward FTLE is lower than the fourth preset quantile of the global FTLE for that day, and the internal regions, is taken as the candidate connected region; If the area of ​​the candidate connected region is greater than the area threshold, and the geometric closure of the candidate connected region is greater than the closure threshold, then the candidate connected region is determined to be a stagnant region.

[0074] In this embodiment, the candidate stagnation areas for the day are traversed, and grids with a backward FTLE higher than the third preset quantile (e.g., 70%) are selected. These grids are then marked as connected components, and the geometric closure of these connected components is calculated. Only connected components with a closure greater than a closure threshold are retained. These connected components represent highly closed vortex boundaries in the stratosphere, i.e., closed ridge profiles. For each retained closed ridge profile, a polygon filling algorithm is used to mark all the discrete spatial grids inside the closed curve. This step expands the one-dimensional boundary line into a two-dimensional vortex interior. On the same day, globally, candidate regions with forward FTLE below the fourth preset quantile (e.g., 30%) are extracted. These regions are then spatially intersected with the interior surfaces of the vortex to obtain candidate connected regions. This means that only regions that are both inside the vortex and possess forward stability will be retained. The candidate connected regions are then finally verified: their area is calculated to see if it is greater than the area threshold, and their geometric closure is verified to see if it is greater than the closure threshold, ensuring that they are still closed structures. Candidate connected regions that meet both of these conditions are finally confirmed as true hidden retention regions.

[0075] In practical implementation, the stagnation zone is defined as the area where the aerostat may hover or circulate at low speed for an extended period. This application uses the LCS bidirectional joint criterion to identify the stagnation zone. It needs to be clarified that: the high-value area of ​​the backward FTLE represents the attraction ridge, and the inner region it surrounds (the low-value area of ​​the backward FTLE) is the vortex core where the aerostat is converged and bound; the low-value area of ​​the forward FTLE confirms that particles do not disperse. The two criteria are used together to locate the Lagrange-type stagnation zone. Therefore, the extraction logic for the candidate stagnation zone is as follows: (1) Extract the backward FTLE closed ridge contour right Grid points above the 70th percentile of the global total for that day are labeled as 8-connected components, and the geometric closure of each connected component is calculated as c = 4πA / P. 2 The connected regions with c>0.6 are retained as the attracting ridge profile with closed circulation characteristics.

[0076] (2) Generate the internal region enclosed by the ridge line. For each closed ridge contour, a polygon filling algorithm (such as scanline filling) is used to generate its internal region R. attract This region is the backward FTLE low value region (vortex core).

[0077] (3) Intersection with the low-value region of the forward FTLE Take R attract With forward FTLE low value region ( The intersection of regions (below the 30th percentile of the global average on that day) satisfies the dual conditions of the aerostat not dispersing (low forward FTLE) and being bound by vortex attraction (within the backward FTLE attraction ridge), thus becoming a candidate region for the LCS-type loitering zone. For all formed candidate connected domains, the following filtering condition must be further met: A > 5000 km 2 It can travel at speeds between 4,000 and 10,000 km. 2 Interval fine-tuning avoids identifying a large number of small-scale, negligible local stagnation zones.

[0078] For all candidate connected components, recalculate their geometric closure c. If c > 0.6 (adjustable range 0.55~0.70), the connected component is determined to be a closed structure with vortex circulation characteristics; otherwise, it is determined to be an open structure.

[0079] The airship channel identification method provided in this application identifies the invisible boundary of the vortex from a dynamic perspective by extracting connected regions with high backward FTLE and high closure. The boundary is transformed into an internal region through a polygon filling algorithm and intersected with the region with low forward FTLE to extract the core vortex region, which has both a closed boundary and a relatively stable interior, ensuring that the identified stagnation area is physically valid. The area threshold eliminates tiny patches that do not pose a real threat and verifies the closure again, ensuring that the final output stagnation area is still a complete closed region in terms of shape.

[0080] In an optional embodiment, the retention area is verified based on the following steps: n test particles are uniformly released using the Monte Carlo method. The forward Lagrange trajectory of the test particles is integrated over D days, and the number of escaped particles that successfully cross the boundary of the retention area after D days is counted. The escape rate is determined based on the ratio of the number of escaping particles to the number of test particles; Based on the escape rate, the risk level of the retention area is determined.

[0081] In this embodiment of the application, the Monte Carlo method is used to uniformly release the material within the identified stagnation region's connected domain. n A test particle, using high-precision wind field data, to this n Each particle undergoes a period of time D The forward Lagrange trajectory integral is used to simulate the actual motion of the airship after it accidentally enters the region. During or at the end of the integration, the spatial coordinates of each test particle are monitored in real time to see if they cross the boundary of the confinement zone, and statistics are compiled. D The number of particles that successfully escaped by the Queen crossing the boundary, divided by the total number of released particles. n The escape rate of the loitering zone is calculated; the calculated escape rate is compared with a preset risk threshold range to determine the risk level of the loitering zone, such as extremely high risk, high risk, medium risk, etc., providing a direct avoidance basis for airship mission planning.

[0082] In the specific implementation process, n test particles (n=100~500) are uniformly released within the vortex boundary. After forward integration for T=7 days, the number of particles n leaving the boundary is counted. out Escape rate η=n out / n.

[0083] Optionally, the boundary is defined as follows: the vortex boundary is The boundary grid points of the closed high-value connected domain are used to sample test particles uniformly within the connected domain using the Monte Carlo method, with FTLE ridges ( >P 70As a more precise LCS boundary, only particles that cross the ridge are counted as leaving the boundary.

[0084] Optionally, the escape rate classification is as follows: high restraint corresponds to η<0.2 (strong vortex restraint, highly dangerous); medium restraint corresponds to 0.2≤η≤0.5 (medium restraint, moderately dangerous); low restraint corresponds to η>0.5 (weak restraint, slight impact).

[0085] In this embodiment, the escape rate is used as a Lagrangian verification method for LCS binding strength. When the escape rate and the backward FTLE value are inconsistent, the escape rate is used as the standard because it directly reflects particle behavior and is a more direct physical verification. Compared with the traditional wind speed threshold method, the LCS bidirectional joint criterion of this application can identify vortex-bound stagnation zones with high wind speeds but low escape rates, which traditional methods would miss. In the final product label, the LCS binding strength field is taken from the escape rate classification result.

[0086] The airship channel identification method provided in this application releases test particles using the Monte Carlo method and performs forward trajectory integration. If the area is a true physical confinement zone, the particles will be bounced back or bound by the boundary within D days, regardless of where they originate from. Conversely, if any particles escape, it indicates that there is a loophole in the boundary. By calculating the escape rate and determining the risk level, the abstract fluid dynamics characteristics are transformed into intuitive level indicators.

[0087] In an optional embodiment, the airship path identification method further includes: Based on the number of consecutive days each candidate connected component is maintained, the candidate connected components are classified according to their persistence. Based on the climatological average wind speed of each candidate connected domain, the channel speed is classified for each candidate connected domain. Based on the persistence classification results and channel speed classification results of each candidate connected domain, a stagnation zone label is generated for each stagnation zone.

[0088] In this embodiment of the application, the same cross-day continuous tracking method as the above-mentioned channel candidate area is adopted: calculating the adjacent day mask IoU and centroid drift distance, making joint judgment based on dual criteria, and the cross-day matching algorithm is the same as the above-mentioned greedy matching strategy.

[0089] From the set of candidate connected components for 365 days of the year, all continuous and stable stagnation regions are traced, and for each stagnation region, the following is extracted: Start date; Maximum consecutive maintenance days D count Climatic average wind speed Unified with the channel calculation method; climatological area; climatological centroid location; climatological geometric profile: take the union of all daily profiles within the life cycle, or take the intermediate daily profile; LCS binding strength: take the escape rate verification classification results (high binding / medium binding / low binding).

[0090] In this embodiment, the number of consecutive days of each candidate connected domain is obtained, and it is mapped to the corresponding persistence level, such as transient, normal, and long-term, according to a preset persistence classification rule, such as a number of days threshold range. The climatological average wind speed during the life cycle of each candidate connected domain is obtained, and it is mapped to the corresponding speed level, such as low speed, medium speed, and high speed, according to a preset speed classification rule, such as a wind speed threshold range. The above classification results are combined, and a unique detention area label is generated for each candidate connected domain according to a predefined label generation rule. This label is written into the waterway database as the core hazard index for the external interaction of the detention area.

[0091] In the specific implementation process, the persistence classification is shown in Table 3 below, and the velocity classification of the stagnation zone is shown in Table 4 below. The combination format is [Persistence]-[Velocity]Bound Stagnation Zone. For example, the long-term-weak wind stagnation zone means that it is maintained for more than 10 days with an average wind speed of 5 to 10 m / s, and the medium-term-very-low speed stagnation zone means that it is maintained for 5 to 10 days with an average wind speed of less than 3 m / s.

[0092] Table 3 Persistence Classification Persistence level of the detention area definition Judgment criteria Long-term stay area Maintain for more than 10 consecutive days <![CDATA[D count ≥ 10 days Mid-term retention area Maintain for 5 to 10 consecutive days. <![CDATA[5 ≤ D count <10 days]]> Short-term stay area Maintain for 3-5 consecutive days <![CDATA[3 ≤ D count <5 days]]> Table 4 Velocity Classification in the Retention Zone

[0093] The airship route identification method provided in this application discretizes continuous physical quantities into levels with business semantics through continuous grading and route speed grading, reducing the decision complexity of upper-level algorithms. By generating detention area labels, this solution constructs an efficient retrieval index for a massive climatological detention area database. It can be deduced that by pre-computing and dimensionality reduction of massive data, when an airship receives a mission instruction for safe long-endurance flight, the system can filter out high-risk detention areas or absolute no-go zones that need to be absolutely avoided within milliseconds simply by matching labels. This significantly improves the response speed and computational efficiency of long-endurance mission planning and enables the airship route identification system to have standardized and productized output capabilities, solving the problems of existing technologies lacking route base maps that can be directly accessed by GIS and being unable to support long-endurance mission planning.

[0094] Furthermore, in order to address the issues of inconsistent output formats and inability to directly interface with GIS systems in existing technologies, this application establishes a standardized GeoJSON product specification.

[0095] The product employs a two-tier architecture. The index file is used for global retrieval and does not store geometric data. It only records a list of data for all identified channels / retention zones under the isobaric surface, including a globally unique ID, start calendar day, duration in days, persistence level, velocity level, location in the corresponding shard file, and its sequence number within the shard file. The index file is small in size, supports resident memory or fast loading, and enables global retrieval within seconds.

[0096] (1) Index file Globally unique ID generation rule: {product type}{isobaric surface}{starting calendar day}_{serial number}, such as route_250hPa_180_001, which represents the first waterway starting on the 180th day of the 250hPa isobaric surface.

[0097] The file partitioning strategy is as follows: partitioning is based on isobaric surfaces and product types. Each isobaric surface and each product type has one index file and multiple single-channel files. The single-channel files are arranged in ID order, and the sequence number in the index file strictly matches the Feature order in the single-channel files.

[0098] File naming convention: index_{product type}_{isobaric surface}.geojson; where product type is divided into stagnation (stagnation area) and route (route).

[0099] (2) Single-channel file A single-channel, independent GeoJSON file stores complete geometric and attribute data, while the properties file stores complete attributes consistent with the index file, allowing index retrieval results to directly return complete information without concatenating with the single-channel file; geometric data is loaded only as needed for visualization or spatial analysis.

[0100] File naming convention: {Product Type}{Isobaric Surface}{id}.geojson; All channel and detention area features follow the GeoJSONRFC7946 standard, with built-in geometry objects. Channels and detention areas use Polygon / MultiPolygon, storing spatial boundary latitude and longitude point sets, supporting direct GIS visualization and spatial retrieval.

[0101] Furthermore, the channel feature attribute fields are as follows: feature_type is a fixed value "route"; The persistence_level is "long" / "medium" / "short", corresponding to the long-term / medium-term / short course; The speed_level is "high" / "medium" / "low", corresponding to high speed / medium speed / low speed; The label is a combination of tags, such as "long-term - high-speed"; calendar_start_day is the starting calendar day, such as 1 corresponding to January 1st; duration_days is the maximum number of consecutive stable days; mean_wind_speed is the average wind speed in the channel, in m / s. area is the channel area, in km². 2 ; The data_quality_flag is a data quality flag. "complete" indicates that the data is complete, while "incomplete" indicates that more than 5% of the grid points are permanently missing on that day.

[0102] The Feature attribute fields for the retention area are as follows: feature_type is a fixed value "stagnation"; persistence_level is "long" / "medium" / "short"; The speed_level is "extreme_low" / "low" / "weak", corresponding to extremely low speed / low speed / weak wind. The label is a combination of tags, such as "long-term - extremely low speed"; calendar_start_day is the starting calendar day; duration_days is the maximum number of consecutive days that can be maintained. mean_wind_speed is the average wind speed in the area, in m / s; area refers to the area of ​​the retention zone, in km². 2 ; The closure is the geometric closure degree, ranging from 0 to 1; escape_rate is the escape rate η; data_quality_flag is a data quality flag, similar to the one used in the navigation channel.

[0103] Furthermore, all products are pre-generated offline and stored in a database or file system. Users can directly retrieve the corresponding product by specifying the isobaric surface altitude (e.g., 250 hPa) and the takeoff calendar date (e.g., day 180, corresponding to late June). The system loads the index file corresponding to that isobaric surface (e.g., index_route_250hPa.geojson), requiring no online calculations and providing query responses in seconds. Users can retrieve information on airways and detention areas corresponding to a specified calendar date from the index. Each record in the index has calendar_start_day (starting calendar day) and duration_days (duration days), supporting searches and matches for any date; upon a match, the corresponding single-channel file is loaded as needed to obtain the geometric data.

[0104] Query logic: Retrieve channels / detention areas that satisfy the condition calendar_start_day ≤ query day ≤ calendar_start_day + duration_days - 1. duration_days includes the start date (e.g., start date 170, duration 10 days, covering days 170-179). For channels whose start date equals the query day, the condition (calendar_start_day ≤ query day) is also satisfied. For example, for "global route charts departing on day 180 at 250 hPa", the system loads the index_route_250hPa.geojson index, retrieves a list of channel IDs where calendar_start_day ≤ 180 ≤ calendar_start_day + duration_days - 1, and loads the corresponding single-channel file to obtain geometric data as needed.

[0105] Furthermore, independent historical airship flight data (such as the trans-Pacific flight trajectory of the Google Loon project from 2017 to 2020) were selected as a validation set. The actual flight trajectories were overlaid on the annual continuous climatological route base map of the corresponding dates to evaluate the spatial consistency between the trajectory and the identification channel and the coverage of loitering events.

[0106] Using daily wind field data from years not included in climatological statistical calculations (e.g., 2021–2023), Lagrange particle tracking was performed. The proportion of particle trajectories falling within the corresponding calendar day's channel and stagnation zone was statistically analyzed to quantify the product's prediction accuracy. The total data period was 18 years (2006–2023), with 15 years (2006–2020) used for climatological construction and 3 years (2021–2023) used for independent validation.

[0107] like Figure 2 As shown above, this application includes: Data input and preprocessing: Read multi-year ERA5 reanalysis data, complete wind speed synthesis and missing value interpolation, and provide a standardized input field for subsequent climatological statistics; The climatological FTLE field was constructed by aggregating hourly data from 15 years. First, a multi-year average climatological wind field sequence was constructed, and then Lagrange integration was performed with a fixed integration duration of |T|=7 days. The forward / backward FTLE fields were calculated, and the continuous climatological LCS field for 365 days of the year was obtained by integrating the multi-year average wind field. LCS Stable Channel Identification and Classification: Based on forward FTLE low-value open areas, candidate transport channels are extracted. After multiple screenings including wind speed, geometry, and exclusion of vortex areas, continuity is evaluated through cross-day IoU tracking, and standardized channel products are output. LCS stagnation zone identification and classification: Based on LCS bidirectional joint discrimination (backward FTLE closed high value zone + forward FTLE low value zone confirmation) to extract stagnation risk candidate zones, with escape rate as auxiliary verification, and continuousness is evaluated through cross-day IoU tracking, and standardized stagnation zone products are output. The standardized GeoJSON product output adopts a two-level architecture of index + single channel, organizes continuous annual climatological data by isobaric surface, has complete metadata and standardized attribute fields, and supports users to retrieve and query by any calendar date with a response time of seconds. The product query and supplementary verification provides a product retrieval interface with a second-level response based on isobaric surface altitude and takeoff calendar date, supporting multi-dimensional attribute filtering and on-demand loading of geometric data; at the same time, it quantitatively verifies the product spatial fit and prediction accuracy through historical trajectory back-substitution and Lagrange particle tracking simulation.

[0108] The airship channel identification device provided in the embodiments of this application is described below. The airship channel identification device described below and the airship channel identification method described above can be referred to in correspondence.

[0109] Figure 3 This is a schematic diagram of the airship navigation identification device provided in this application, as shown below. Figure 3 As shown, the airship navigation identification device may include, but is not limited to: Data preprocessing module 310 is used to: aggregate and average the wind field data of the same period over many years on an hourly basis to obtain the climatological wind field sequence for each hour of the year; The FTLE field construction module 320 is used to: calculate the finite-time Lyapunov exponent FTLE for each discrete spatial grid on a daily basis, based on the climatological wind field sequence, wherein the FTLE includes forward FTLE and backward FTLE; The region division module 330 is used to: for any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. The channel determination module 340 is used to: identify a channel based on the channel candidate area; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0110] It should be noted that the airship channel identification device provided in this application embodiment can execute the airship channel identification method described in any of the above embodiments during specific operation, and this embodiment will not elaborate on this.

[0111] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an airship path identification method, which includes: By aggregating and averaging the wind field data from the same period over many years on an hourly basis, we can obtain the climatological wind field sequence for each hour of the year. Based on the climatological wind field sequence, the finite-time Lyapunov exponent FTLE is calculated for each discrete spatial grid on a daily basis. The FTLE includes forward FTLE and backward FTLE. For any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. Based on the candidate waterway areas, waterways are identified; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0112] Furthermore, the logical instructions in the aforementioned memory 430 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 this application, in essence, 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 this application. 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.

[0113] On the other hand, this application 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 is able to execute the airship path identification method provided by the above methods, the method including: By aggregating and averaging the wind field data from the same period over many years on an hourly basis, we can obtain the climatological wind field sequence for each hour of the year. Based on the climatological wind field sequence, the finite-time Lyapunov exponent FTLE is calculated for each discrete spatial grid on a daily basis. The FTLE includes forward FTLE and backward FTLE. For any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. Based on the candidate waterway areas, waterways are identified; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0114] In another aspect, this application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the airship path identification method provided by the above methods, the method comprising: By aggregating and averaging the wind field data from the same period over many years on an hourly basis, we can obtain the climatological wind field sequence for each hour of the year. Based on the climatological wind field sequence, the finite-time Lyapunov exponent FTLE is calculated for each discrete spatial grid on a daily basis. The FTLE includes forward FTLE and backward FTLE. For any discrete spatial grid on any given day, if both the forward FTLE and backward FTLE are low, then the discrete spatial grid is determined to be a channel candidate area on that day; within a closed area surrounded by a high backward FTLE area, if the forward FTLE is low, then the closed area is determined to be a stagnation candidate area on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a danger zone on that day; if both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. Based on the candidate waterway areas, waterways are identified; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

[0115] 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.

[0116] 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.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such 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 this application.

Claims

1. A method for identifying the flight path of an airship, characterized in that, include: By aggregating and averaging the wind field data from the same period over many years on an hourly basis, we can obtain the climatological wind field sequence for each hour of the year. Based on the climatological wind field sequence, the finite-time Lyapunov exponent FTLE is calculated for each discrete spatial grid on a daily basis. The FTLE includes forward FTLE and backward FTLE. For any discrete spatial grid on any given day, if the forward FTLE is low and the backward FTLE is low, then the discrete spatial grid is determined to be a channel candidate area on that day. Within a closed area surrounded by a high backward FTLE value region, if the forward FTLE is low, the closed area is determined to be a candidate area for stagnation on that day; if the forward FTLE is high and the backward FTLE is low, the discrete spatial grid is determined to be a danger zone on that day. If both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. Based on the candidate waterway areas, waterways are identified; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

2. The airship channel identification method according to claim 1, characterized in that, For each day, the candidate waterway areas are selected based on the following criteria: First condition: The forward FTLE of the candidate channel area is lower than the first preset quantile of the global FTLE on that day; Second condition: The average wind speed of the day reaches the wind speed threshold. Third condition: For connected components that satisfy the first and second conditions, calculate the aspect ratio of the minimum area bounding rectangle, where the aspect ratio is greater than the aspect ratio threshold. Fourth condition: For a connected region composed of channel candidate areas whose backward FTLE is higher than the second preset quantile of the global FTLE of the day, if the geometric closure of the connected region is greater than the closure threshold, the connected region is removed.

3. The airship channel identification method according to claim 2, characterized in that, After screening the candidate waterway areas, the process further includes: The candidate waterway areas that meet the conditions are marked with 8-connected components, and the connected components with an area greater than the area threshold are retained to obtain the daily connected component set of the climate state for the whole year.

4. The airship path identification method according to claim 3, characterized in that, The process of identifying a waterway based on the candidate waterway area includes: For the i-th connected component on day d and the j-th connected component on day d+1, calculate the area intersection-union ratio and centroid drift distance of the two connected components; If the area intersection-union ratio is greater than the spatial overlap threshold and the centroid drift distance is less than the drift distance threshold, then the i-th connected component on day d and the j-th connected component on day d+1 are determined to be two adjacent days of the same climatological channel, and a cross-day correspondence is established. After obtaining the cross-day correspondence throughout the year, a time-limited waterway is determined based on the life cycle of each climatological channel.

5. The airship path identification method according to claim 4, characterized in that, The airship channel identification method also includes: The following attributes are extracted for each climatological channel: start date, number of consecutive days, climatological average wind speed, climatological area, climatological centroid location, climatological geometric profile, and Lagrange coherence structure intensity.

6. The airship path identification method according to claim 5, characterized in that, The airship channel identification method also includes: Based on the number of consecutive days of each climatic channel, the persistence of each climatic channel is classified. Based on the climatological average wind speed of each of the aforementioned climatological channels, the channel speed is classified for each of the aforementioned climatological channels; Based on the persistence classification results and channel speed classification results of each of the aforementioned climatic channels, channel labels are generated for each of the aforementioned channels.

7. The airship channel identification method according to claim 1, characterized in that, The retention candidate region is extracted based on the following steps to obtain the retention region: For the connected regions composed of the candidate regions whose backward FTLE is higher than the third preset quantile of the global FTLE of the day, connected regions with a geometric closure degree greater than the closure degree threshold are selected. For the closed ridge contours of each of the connected domains, a polygon filling algorithm is used to generate their internal regions; The intersection of the candidate regions where the forward FTLE is lower than the fourth preset quantile of the global FTLE for that day, and the internal regions, is taken as the candidate connected region; If the area of ​​the candidate connected region is greater than the area threshold, and the geometric closure of the candidate connected region is greater than the closure threshold, then the candidate connected region is determined to be a stagnant region.

8. The airship path identification method according to claim 7, characterized in that, The retention area is verified based on the following steps: n test particles are uniformly released using the Monte Carlo method. The forward Lagrange trajectory of the test particles is integrated over D days, and the number of escaped particles that successfully cross the boundary of the retention area after D days is counted. The escape rate is determined based on the ratio of the number of escaping particles to the number of test particles; Based on the escape rate, the risk level of the retention area is determined.

9. The airship path identification method according to claim 7, characterized in that, The airship channel identification method also includes: Based on the number of consecutive days each candidate connected component is maintained, the candidate connected components are classified according to their persistence. Based on the climatological average wind speed of each candidate connected domain, the channel speed is classified for each candidate connected domain. Based on the persistence classification results and channel speed classification results of each candidate connected domain, a stagnation zone label is generated for each stagnation zone.

10. An airship navigation path identification device, characterized in that, include: The data preprocessing module is used to: aggregate and average the wind field data from the same period over many years on an hourly basis to obtain the climatological wind field sequence for each hour of the year; The FTLE field construction module is used to: calculate the finite-time Lyapunov exponent FTLE for each discrete spatial grid on a daily basis, based on the climatological wind field sequence, wherein the FTLE includes forward FTLE and backward FTLE; The region division module is used to: for any discrete spatial grid on any day, if the forward FTLE is low and the backward FTLE is low, then determine that the discrete spatial grid is a channel candidate area on that day; Within a closed area surrounded by a high backward FTLE value region, if the forward FTLE is low, the closed area is determined to be a candidate area for stagnation on that day; if the forward FTLE is high and the backward FTLE is low, the discrete spatial grid is determined to be a danger zone on that day. If both the forward FTLE and backward FTLE are high, then the discrete spatial grid is determined to be a restricted area on that day. The channel determination module is used to: identify a channel based on the channel candidate area; Wherein, "FTLE low" means that FTLE is lower than a first threshold, and "FTLE high" means that FTLE is higher than a second threshold.

Images