A high-precision electronic fence control method and system for low-altitude flight

Abstract

The embodiment of the present application provides a high-precision electronic fence control method and system for low-altitude flight. By acquiring low-altitude flight area data and extracting its spatial features, accurate assessment of the risk level of the flight area and effective annotation of flight knowledge are achieved. For areas with higher risk levels, when non-no-fly zones are confirmed and contain multiple grid areas, high-precision electronic fences that are highly matched with the grid areas can be automatically generated. By further calculating and verifying the deviations between these high-precision electronic fences, the accuracy and consistency of the fence settings are ensured. Finally, qualified high-precision electronic fences are established as the flight restriction boundaries of each grid area, which significantly improves the level of refinement of low-altitude flight management, enhances flight safety monitoring capabilities, effectively avoids potential risks during flight, and provides technical support for low-altitude flight.

Description

Technical Field

[0001] The present invention relates to the field of low-altitude flight technology, and in particular to a high-precision electronic fence control method and system for low-altitude flight. Background Art

[0002] With the rapid development and widespread application of drone technology, low-altitude flight activities are becoming increasingly frequent, which has put forward higher requirements for flight safety supervision. Traditional airspace management methods often rely on manually set no-fly zones or simple geometric electronic fences, which are difficult to adapt to the complex and changeable low-altitude flight environment. Especially in areas with complex geographical features and diverse flight needs such as cities, mountainous areas, and waters, traditional electronic fences often lack sufficient accuracy and flexibility, cannot effectively ensure flight safety, and also restrict the orderly development of low-altitude flight activities. Summary of the invention

[0003] In order to at least overcome the above-mentioned deficiencies in the prior art, an object of the embodiments of the present application is to provide a high-precision electronic fence control method and system for low-altitude flight.

[0004] According to one aspect of the present application, a high-precision electronic fence control method for low-altitude flight is provided, the method comprising:

[0005] Obtain low-altitude flight area data;

[0006] Acquire spatial feature data of the low-altitude flight area data, and convert the spatial feature data into a risk level and a flight knowledge label of the low-altitude flight area data;

[0007] If the risk level is greater than the threshold level, the flight knowledge tag indicates that the low-altitude flight area data does not include a no-fly zone, and the low-altitude flight area data includes multiple grid areas, obtaining high-precision electronic fences corresponding to each of the multiple grid areas, and the matching degree between the grid area and the corresponding high-precision electronic fence is greater than a set matching degree;

[0008] Determine the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas;

[0009] If the fence deviation is less than the set deviation, the high-precision electronic fences corresponding to each of the multiple grid areas are used as flight restriction boundaries of the corresponding grid areas.

[0010] In the technical solutions provided in some embodiments of the present application, the embodiments of the present application achieve accurate assessment of the risk level of the flight area and effective annotation of flight knowledge by acquiring low-altitude flight area data and extracting its spatial features. For areas with higher risk levels, when non-no-fly zones are confirmed and contain multiple grid areas, a high-precision electronic fence that is highly matched with the grid area can be automatically generated. By further calculating and verifying the deviation between these high-precision electronic fences, the accuracy and consistency of the fence setting are ensured. Finally, the qualified high-precision electronic fences are established as the flight restriction boundaries of each grid area, which significantly improves the level of refinement of low-altitude flight management, enhances flight safety monitoring capabilities, effectively avoids potential risks during flight, and provides technical support for low-altitude flight. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be extracted in combination with these drawings without creative work.

[0012] Figure 1 A schematic diagram of a process flow of a high-precision electronic fence control method for low-altitude flight provided in an embodiment of the present application;

[0013] Figure 2 A schematic block diagram of the architecture of a high-precision electronic fence control system for low-altitude flight provided in an embodiment of the present application. DETAILED DESCRIPTION

[0014] The following description is intended to enable one of ordinary skill in the art to implement and combine the present application, and the description is provided in the context of a specific application scenario and its requirements. It will be apparent to one of ordinary skill in the art that various changes may be made to the disclosed embodiments, and that the general principles defined in the present application may be applied to other embodiments and application scenarios without departing from the principles and scope of the present application. Therefore, the present application is not limited to the described embodiments, but should be given the broadest scope consistent with the claims.

[0015] Figure 1 It is a flow chart of a high-precision electronic fence control method for low-altitude flight provided by an embodiment of the present application. The high-precision electronic fence control method for low-altitude flight is introduced in detail below.

[0016] Step S110, obtaining low-altitude flight area data.

[0017] In this embodiment, in an aviation management system, the server manages low-altitude flight-related affairs in a specific airspace. The server is connected to multiple data sources, including aviation survey equipment, geographic information systems (GIS), and previously stored historical data of flight areas. The aviation survey equipment regularly sends geographic information, meteorological information, and air traffic flow data of each area in the airspace, which constitute part of the low-altitude flight area data. For example, the aviation survey equipment can detect the topography of a certain area through high-precision radars and sensors, such as the height of mountains, the depth of valleys, and the location of rivers and lakes, and can also obtain real-time meteorological data of the area, such as wind direction, wind speed, air pressure, etc. The GIS system provides more comprehensive geospatial information, including land use types (urban areas, agricultural areas, nature reserves, etc.) and detailed map coordinate information. The data obtained by the server from these data sources is the low-altitude flight area data. Suppose in the airspace near a mountainous area, the data obtained by the server shows that there are multiple peaks in the area with altitudes between 1,000 and 2,000 meters. At the same time, meteorological data shows that there are often strong winds in the area with changeable wind directions. In addition, from the GIS system, it is known that part of the area is a nature reserve and part is a rural residential area. These are the specific contents of the low-altitude flight area data obtained by the server.

[0018] Step S120, obtaining the spatial feature data of the low-altitude flight area data, and converting the spatial feature data into the risk level and flight knowledge label of the low-altitude flight area data.

[0019] In this embodiment, after receiving the low-altitude flight area data, the server starts to process it. First, the server extracts the spatial element coordinate distribution of the low-altitude flight area data. For example, for the airspace data near the mountain area previously acquired, the server determines the coordinates of the vertices of each mountain peak, the coordinates of the bottom of the valley, the coordinates of the edges of rivers and lakes, etc. These coordinate information constitute the spatial element coordinate distribution. Then, the server extracts features through the spatial feature coding network based on these spatial element coordinate distributions. This spatial feature coding network is pre-trained and can recognize the patterns of various spatial features. During the feature extraction process, the server inputs the spatial element coordinate distribution into the spatial feature coding network, and the spatial feature coding network generates spatial feature data of the low-altitude flight area data through algorithms and neuron calculations. For example, the spatial feature data may contain information such as the undulation of the terrain and the spatial distance relationship between different areas.

[0020] Next, the server converts the spatial feature data into the risk level and flight knowledge label of the low-altitude flight area data. The server performs attention processing on the spatial feature data through the attention unit in the first airspace feature prediction model. Taking the mountainous airspace as an example, the attention unit will focus on those features related to flight risks, such as the height of the peak, the strong wind area, etc., to generate the first attention feature data. Then, the first attention feature data is converted into the first confidence of multiple reference risk levels through the fully connected mapping unit in the first airspace feature prediction model. Assuming that the reference risk levels are divided into low risk, medium risk and high risk, the fully connected mapping unit calculates that the confidence of low risk is 0.3, the confidence of medium risk is 0.5, and the confidence of high risk is 0.2. Since the confidence of medium risk is the highest and meets the preset condition (here the preset condition can be the level with the highest confidence), the server uses medium risk as the risk level of the low-altitude flight area data. This risk level is used to indicate the score for generating the flight restriction boundary based on the low-altitude flight area data.

[0021] At the same time, the server performs attention processing on the spatial feature data through the attention unit in the second airspace feature prediction model to generate second attention feature data. For example, the attention unit will pay attention to whether there are no-fly-related signs or special areas in the area. Then, the second attention feature data is converted into a second confidence level that maps multiple flight knowledge labels through the fully connected mapping unit in the first airspace feature prediction model. Assuming that the flight knowledge labels include "including no-fly zones" and "excluding no-fly zones", the server calculates that the confidence level of "including no-fly zones" is 0.1 and the confidence level of "excluding no-fly zones" is 0.9. Because the confidence level of "excluding no-fly zones" meets the preset conditions (such as the confidence level is greater than 0.5), the server uses "excluding no-fly zones" as the flight knowledge label for low-altitude flight area data.

[0022] Step S130, if the risk level is greater than the threshold level, the flight knowledge label indicates that the low-altitude flight area data does not include a no-fly zone, and the low-altitude flight area data includes multiple grid areas, obtain the high-precision electronic fences corresponding to each of the multiple grid areas, and the matching degree between the grid area and the corresponding high-precision electronic fence is greater than the set matching degree.

[0023] In this embodiment, it is assumed that the threshold level set by the server is medium risk, the risk level calculated previously is high risk, and the flight knowledge label is "excluding no-fly zones", and the low-altitude flight area data is divided into multiple grid areas under the processing of the server, for example, the airspace near the mountainous area is divided into 10 grid areas according to certain coordinate rules. Each grid area has its own unique geographical features and flight-related information.

[0024] The server starts to obtain the high-precision electronic fence corresponding to each grid area. A high-precision electronic fence is an accurate virtual boundary used to restrict flight areas. The data of these electronic fences may be stored in a special aviation safety database. The server will search for a matching high-precision electronic fence from the database based on the geographic coordinate range, terrain features and other information of each grid area. For example, for one of the grid areas close to a mountain, the server will search for an electronic fence that can accurately surround the area and take into account the terrain of the mountain. During the search process, the server will calculate the matching degree between the grid area and the corresponding high-precision electronic fence. The calculation of the matching degree involves multiple factors, such as the degree of overlap of the boundaries, the inclusion relationship of geographical features, etc. Assuming that the matching degree is set to 0.8, the server calculates the matching degree between a grid area and a high-precision electronic fence through a complex algorithm to be 0.9. This matching degree is greater than the set matching degree, so the server will use this high-precision electronic fence as a candidate high-precision electronic fence for the grid area. For all grid areas, the server will perform such operations to obtain high-precision electronic fences that meet the conditions.

[0025] Step S140: determining the fence deviation between the high-precision electronic fences corresponding to the plurality of grid areas.

[0026] In this embodiment, after obtaining the high-precision electronic fences corresponding to each of the multiple grid areas, the server begins to determine the fence deviation between them. For each high-precision electronic fence, the server first determines the mean value of the spatial element coordinates of each boundary segment in the high-precision electronic fence. For example, a high-precision electronic fence is composed of four boundaries, and the server calculates the mean value of each coordinate point on these four boundaries respectively. Assuming that there are 10 coordinate points on the first boundary segment, the server adds the horizontal coordinates of these 10 coordinate points and divides them by 10, and adds the vertical coordinates and divides them by 10 to obtain the mean value of the spatial element coordinates of the first boundary segment. In the same way, the server obtains the mean value of the spatial element coordinates of the other three boundaries, and then combines the mean value of the spatial element coordinates of each boundary segment into the electronic fence feature vector of the high-precision electronic fence.

[0027] Next, the server determines the discreteness of the electronic fence feature vectors of the high-precision electronic fences corresponding to the multiple grid areas, and uses the discreteness as the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas. The calculation of the discreteness may involve some statistical methods, such as calculating the variance or standard deviation. Assume that the server calculates that the discreteness of the electronic fence feature vectors of the high-precision electronic fences corresponding to the three grid areas is 0.1, and this value represents the fence deviation between the three high-precision electronic fences.

[0028] Step S150: If the fence deviation is less than the set deviation, the high-precision electronic fences corresponding to each of the multiple grid areas are used as flight restriction boundaries of the corresponding grid areas.

[0029] In this embodiment, it is assumed that the set deviation is 0.2, and the fence deviation calculated previously is 0.1, and this fence deviation is less than the set deviation. The server will use the high-precision electronic fences corresponding to each of the multiple grid areas as the flight restriction boundaries of the corresponding grid areas. For example, for each of the 10 grid areas mentioned previously, if its corresponding high-precision electronic fence meets this condition, then this high-precision electronic fence becomes the flight restriction boundary of the grid area. This means that in actual low-altitude flight management, when the aircraft is flying within the grid area, it cannot cross the range defined by the high-precision electronic fence. The server will store this flight restriction boundary information and use it in the flight monitoring and management process to ensure the safety and compliance of low-altitude flight.

[0030] Based on the above steps, the embodiment of the present application achieves accurate assessment of the risk level of the flight area and effective annotation of flight knowledge by acquiring low-altitude flight area data and extracting its spatial features. For areas with higher risk levels, when non-no-fly zones are confirmed and contain multiple grid areas, a high-precision electronic fence that matches the grid area highly can be automatically generated. By further calculating and verifying the deviation between these high-precision electronic fences, the accuracy and consistency of the fence setting are ensured. Ultimately, the qualified high-precision electronic fence is established as the flight restriction boundary of each grid area, which significantly improves the level of refinement of low-altitude flight management, enhances flight safety monitoring capabilities, effectively avoids potential risks during flight, and provides technical support for low-altitude flight.

[0031] In a possible implementation, the method further includes: if the fence deviation is not less than the set deviation, further subdividing the multiple grid areas into multiple new low-altitude flight area data.

[0032] If the fence deviation is not less than the set deviation when determining the fence deviation, the server will perform the operation of further subdividing multiple grid areas into multiple new low-altitude flight area data. For example, the airspace near the mountain mentioned above is divided into 10 grid areas. When calculating the fence deviation, it is found that it is not less than the set deviation. At this time, the server will re-divide these grid areas according to more detailed standards. The server may subdivide based on factors such as topography, meteorological conditions, or air traffic flow. For example, for one of the grid areas, it contains a mountain peak and a valley. Due to the complex terrain, the previous high-precision electronic fence deviation is large. The server will use the ridgeline of the mountain as the boundary to subdivide this grid area into two new low-altitude flight area data, one is the area around the mountain peak, and the other is the valley area. For other grid areas, similar subdivision operations will be performed according to their respective situations. It may be to divide a larger grid area into two new low-altitude flight area data under the influence of different wind directions according to wind direction, or to subdivide according to the different density of air traffic flow. The new low-altitude flight area data after subdivision will be more conducive to the precise management of low-altitude flight activities.

[0033] In a possible implementation, the method further includes:

[0034] If the risk level is greater than the threshold level and the flight knowledge tag indicates that the low-altitude flight area data includes a no-fly zone, the low-altitude flight area data is removed.

[0035] If the risk level is not greater than the threshold level, the flight restriction boundary of the low-altitude flight area data is archived into the boundary sequence of the airspace according to the constructed time order, and different low-altitude flight area data are obtained from the airspace in sequence.

[0036] In another case, if the risk level is greater than the threshold level, and the flight knowledge label indicates that the low-altitude flight area data includes a no-fly zone, the server will remove the low-altitude flight area data. Assume that the threshold level is medium risk, the risk level of a certain area is calculated to be high risk, and the flight knowledge label indicates that the area contains a no-fly zone, such as a military base or an important scientific research experimental area. After detecting this situation, the server will remove the low-altitude flight area data from the current processing flow. This means that in subsequent flight management operations, this area will not be treated as an ordinary low-altitude flight area, but will be managed in accordance with the strict management regulations of the no-fly zone. The server will update its internal flight area management list and mark this area as a no-fly zone. The related data has been removed to avoid confusion in subsequent operations, while also ensuring flight safety and the confidentiality and security of related areas.

[0037] When the risk level is not greater than the threshold level, the server will file the flight restriction boundaries of the low-altitude flight area data into the boundary sequence of the airspace according to the constructed time order. Different low-altitude flight area data are obtained from the airspace in sequence. For example, when processing multiple low-altitude flight area data, the server obtains data from different areas in sequence and calculates their flight restriction boundaries. For areas with risk levels lower than the threshold level, such as a low-altitude flight area in a plain area, the risk level is low risk. After the server calculates the flight restriction boundary of the area, it will file this flight restriction boundary into the boundary sequence of the airspace in the order in which the data was obtained. Assuming that the low-altitude flight area data of the plain area is obtained first, after calculating the flight restriction boundary, it will be filed as the first element in the boundary sequence. If a low-altitude flight area data near the city with a risk level not greater than the threshold level is obtained later, after calculating the flight restriction boundary, it will be filed in the next position of the boundary sequence in sequence. Doing so will make it easier for the server to query and use the flight restriction boundary information in sequence when managing low-altitude flight activities in the entire airspace, thereby more orderly planning and managing low-altitude flight routes and flight ranges, ensuring that low-altitude flight activities in different areas are carried out within their respective appropriate boundaries.

[0038] In a possible implementation, step S120 includes:

[0039] Step S121, extracting the spatial element coordinate distribution of the low-altitude flight area data, and based on the spatial element coordinate distribution, performing feature extraction through a spatial feature coding network to generate spatial feature data of the low-altitude flight area data.

[0040] Step S122: Perform attention processing on the spatial feature data through the attention unit in the first spatial feature prediction model to generate first attention feature data; convert the first attention feature data into first confidence levels of multiple reference risk levels through the fully connected mapping unit in the first spatial feature prediction model; and use the reference risk level corresponding to the first confidence level that meets the preset conditions as the risk level of the low-altitude flight area data; the risk level is used to indicate the score for generating a flight restriction boundary based on the low-altitude flight area data.

[0041] Step S123, performing attention processing on the spatial feature data through the attention unit in the second spatial feature prediction model to generate second attention feature data, converting the second attention feature data into a second confidence level that maps multiple flight knowledge labels through the fully connected mapping unit in the first spatial feature prediction model, and using the flight knowledge label corresponding to the second confidence level that meets the preset conditions as the flight knowledge label of the low-altitude flight area data, the flight knowledge label includes: including no-fly zones and excluding no-fly zones.

[0042] In this embodiment, for the link of obtaining the spatial feature data of the low-altitude flight area data, the low-altitude flight area data is first deeply analyzed to extract the spatial element coordinate distribution. Taking the airspace near the mountain area as an example, the low-altitude flight area data obtained by the server from a variety of data sources contains a lot of geographical information and other contents. The server will accurately extract the spatial element coordinate distribution from it. For example, for each peak in the mountain area, the server will determine the coordinates of its vertex and the coordinates of the key turning points on the hillside. For the valley, the coordinates of the valley bottom are determined. For water bodies such as rivers and lakes, the coordinates of their edges are determined. These coordinates are obtained accurately to a certain accuracy standard, such as several digits after the decimal point of longitude and latitude to ensure accuracy. After obtaining the distribution of these spatial element coordinates, the server extracts features based on these data through a pre-constructed spatial feature coding network. This spatial feature coding network is trained based on a large amount of historical flight area data and related geographic spatial data. When the server inputs the spatial element coordinate distribution data into the spatial feature coding network, each neuron in the network processes the data according to a predetermined algorithm. For example, some neurons in the network may be responsible for identifying the distance relationship between coordinates, while others may be responsible for analyzing the clustering pattern of coordinate points. Through the collaborative operation of these neurons, spatial feature data of low-altitude flight area data is finally generated. These spatial feature data may contain information such as the complexity of the terrain and the spatial correlation between different geographical elements. This information will provide an important basis for the subsequent determination of risk levels and flight knowledge labels.

[0043] When converting the spatial feature data into the risk level and flight knowledge label of the low-altitude flight area data, the server operates using two airspace feature prediction models. First, the server performs attention processing on the spatial feature data through the attention unit in the first airspace feature prediction model to generate the first attention feature data. Still taking the airspace near the mountain area as an example, the spatial feature data contains a variety of information such as the complexity of the mountain terrain and meteorological conditions. The attention unit in the first airspace feature prediction model will focus on those features that are highly relevant to risk assessment. For example, for mountainous areas, the attention unit will focus on factors such as the height and slope of the peak and the strong wind area in the meteorological data. By focusing on these key factors, the first attention feature data is generated. Compared with the original spatial feature data, this first attention feature data highlights the feature information that has an important impact on risk assessment. Then, the server converts the first attention feature data into the first confidence of multiple reference risk levels through the fully connected mapping unit in the first airspace feature prediction model. Assume that the reference risk level is divided into three levels: low risk, medium risk and high risk. The fully connected mapping unit calculates the first attention feature data according to its internal mapping relationship and algorithm. For example, after calculation, the first confidence of low risk is 0.3, the first confidence of medium risk is 0.5, and the first confidence of high risk is 0.2. Since the first confidence of medium risk is the highest, according to the preset conditions (here the preset conditions are set as the reference risk level with the highest first confidence as the final risk level), the server uses medium risk as the risk level of low-altitude flight area data. This risk level will serve as an important indicator to indicate the score for generating flight restriction boundaries based on low-altitude flight area data. The higher the risk level, the more restrictive factors may need to be considered when generating flight restriction boundaries.

[0044] At the same time, the server also performs attention processing on the spatial feature data through the attention unit in the second airspace feature prediction model to generate second attention feature data. Various information in the spatial feature data is analyzed again, and at this time the attention unit in the second airspace feature prediction model will focus on the features related to the no-fly zone determination. For example, in the airspace data near the mountainous area, if there is identification information of specific military facilities or special protection areas, the attention unit will focus on processing this information. After generating the second attention feature data in this way, the server converts the second attention feature data into a second confidence mapping multiple flight knowledge labels through the fully connected mapping unit in the first airspace feature prediction model. The flight knowledge labels here include "including no-fly zones" and "excluding no-fly zones". Assume that after calculation, the second confidence of "including no-fly zones" is 0.1, and the second confidence of "excluding no-fly zones" is 0.9. According to the preset condition (for example, the preset condition is that the flight knowledge label with a second confidence greater than 0.5 is the final label), the server uses "excluding no-fly zones" as the flight knowledge label of the low-altitude flight area data. In this way, through in-depth processing of spatial feature data, the server successfully determined the risk level and flight knowledge label of low-altitude flight area data, providing key decision-making basis for subsequent low-altitude flight management operations, such as determining flight restriction boundaries and handling different types of flight areas.

[0045] In a possible implementation, step S140 includes:

[0046] Step S141, for each of the high-precision electronic fences, determine the mean value of the spatial element coordinates of each boundary segment in the high-precision electronic fence, and combine the mean value of the spatial element coordinates of each boundary segment into an electronic fence feature vector of the high-precision electronic fence.

[0047] Step S142, determining the discreteness of the electronic fence feature vectors of the high-precision electronic fences corresponding to the multiple grid areas, and using the discreteness as the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas.

[0048] In a possible implementation, the method further includes:

[0049] Step A110, obtaining a target combination template sequence, wherein the target combination template sequence includes multiple target combinations, and the target combination includes template low-altitude flight area data and template high-precision electronic fence.

[0050] Step A120, according to the ascending order of the multiple defined boundary points, determine the matching parameters of the target combination template sequence at the boundary points of this round, the matching parameters are the statistics of the matching target combination, the presidential metric of the multiple target combinations. The matching degree between the template low-altitude flight area data and the template high-precision electronic fence in the matching target combination is not less than the boundary points of this round.

[0051] Step A130, if the matching parameter at the boundary value point of this round is not less than the set matching parameter, the boundary value point of this round is set to the set matching degree.

[0052] Step A140, determining the discreteness of the electronic fence feature vector of the template high-precision electronic fence in the matching target combination, and using the discreteness of the electronic fence feature vector of the template high-precision electronic fence in the matching target combination as the set deviation.

[0053] In this embodiment, when determining the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas, the airspace near the mountain area mentioned above is divided into multiple grid areas as an example, and each grid area has a corresponding high-precision electronic fence. For each high-precision electronic fence, the server first determines the mean value of the spatial element coordinates of each segment of its boundary. Assuming that a certain high-precision electronic fence surrounds a grid area containing valleys and hills, this electronic fence may be composed of four segments of boundaries. The server will obtain the information of all spatial element coordinate points on the first segment of the boundary, including the longitude and latitude values ​​of each coordinate point. Then, these longitude values ​​are added and divided by the number of coordinate points to obtain the mean value of the spatial element coordinates of the segment of the boundary in the longitude direction; similarly, the same operation is performed on the latitude value to obtain the mean value of the spatial element coordinates in the latitude direction. According to this method, the server also calculates the mean value of the spatial element coordinates for the other three segments of the high-precision electronic fence, and then combines the mean value of the spatial element coordinates of each segment of the boundary to form the electronic fence feature vector of the high-precision electronic fence. This electronic fence feature vector can accurately describe the spatial characteristics of the high-precision electronic fence and is an important basis for subsequent calculations.

[0054] Next, the server determines the discreteness of the electronic fence feature vectors of the high-precision electronic fences corresponding to the multiple grid areas, and uses this discreteness as the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas. The server collects the electronic fence feature vectors of the high-precision electronic fences corresponding to each grid area. For example, there are three grid areas, each corresponding to three high-precision electronic fences. After the server obtains the three electronic fence feature vectors, it will use a specific algorithm to calculate the discreteness between them. This algorithm may involve calculating statistics such as the distance or variance between vectors. For example, the server calculates the differences between these electronic fence feature vectors in various dimensions (such as the dimensions corresponding to the longitude and latitude directions) and comprehensively obtains a value that can reflect the degree of discreteness between them. This value is the discreteness of the electronic fence feature vector, that is, the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas. This fence deviation can intuitively reflect the degree of difference between different high-precision electronic fences, which is of great significance in subsequent operations such as determining whether the high-precision electronic fence can be used as a flight restriction boundary.

[0055] In a possible implementation, the server also needs to perform a series of operations to determine the set matching degree and the set deviation degree. First, the server needs to obtain a target combination template sequence, which contains multiple target combinations, each of which consists of template low-altitude flight area data and template high-precision electronic fences. These template data are standard data samples pre-stored in the server to assist in determining various parameters. For example, these template data may be typical data combinations summarized based on the actual situation of a large number of low-altitude flight areas and high-precision electronic fences in the past.

[0056] Then, the server determines the matching parameters of the target combination template sequence at the current round of boundary points according to the ascending order of the multiple defined boundary points. This matching parameter is the relationship value between the statistic of the matching target combination and the presidential metric of multiple target combinations. Specifically, for each boundary point, the server checks each target combination in the target combination template sequence to determine whether the matching degree between the template low-altitude flight area data and the template high-precision electronic fence is not less than the current round of boundary points. For example, the server starts checking from the smallest boundary point. For a boundary point of 0.5, the server calculates the matching degree between the template low-altitude flight area data and the template high-precision electronic fence in each target combination. The calculation of this matching degree may involve comparing the spatial coverage of the two, the overlap of geographical features, and other factors. If the matching degree of a certain target combination is not less than 0.5, then this target combination is regarded as a matching target combination. Then, the server counts the number of matching target combinations and divides this number by the total number of target combinations to obtain the matching parameters at the current round of boundary points.

[0057] If the matching parameter at the boundary value point of this round is not less than the set matching parameter, the server will set the boundary value point of this round as the set matching degree. Assuming the set matching parameter is 0.8, when the matching parameter calculated by the server at a certain boundary value point (such as 0.85) is 0.9, since 0.9 is not less than 0.8, the server will set the boundary value point of 0.85 as the set matching degree. This set matching degree will be used as an important judgment criterion when obtaining the high-precision electronic fence corresponding to the grid area in the future. Only when the matching degree between the grid area and the high-precision electronic fence is greater than this set matching degree, the high-precision electronic fence will be used as a candidate object.

[0058] Finally, the server determines the discreteness of the electronic fence feature vector of the template high-precision electronic fence in the matching target combination, and uses this discreteness as the set deviation. After the matching target combination is determined in the previous step, the server processes the template high-precision electronic fence in these matching target combinations. For each template high-precision electronic fence in the matching target combination, the server obtains its electronic fence feature vector according to the method of calculating the electronic fence feature vector previously. Then, the server calculates the discreteness of the electronic fence feature vectors of these template high-precision electronic fences. For example, the server calculates the degree of discreteness between these electronic fence feature vectors by analyzing the differences in various dimensions of these electronic fence feature vectors using a specific mathematical formula. This discreteness is set as the set deviation. This set deviation will be used as a standard when judging whether the fence deviation between the high-precision electronic fences corresponding to each of the multiple grid areas meets the requirements. Only when the fence deviation is less than this set deviation will the high-precision electronic fences corresponding to each of the multiple grid areas be used as the flight restriction boundary of the corresponding grid area. Through these steps, the server can accurately determine various parameters, thereby more effectively managing the low-altitude flight area and ensuring flight safety and order.

[0059] In a possible implementation, if the low-altitude flight area data is compliant flight area data, multiple construction boundaries of the low-altitude flight area data are obtained, the flight restriction boundaries of the low-altitude flight area data are determined from the multiple construction boundaries, and the flight restriction boundaries of the low-altitude flight area data are stored, and the compliant flight area data is flight area data that does not meet the violation data requirements, and the violation data requirements include: the risk level is greater than the threshold level, the flight knowledge label indicates that the low-altitude flight area data does not include a no-fly zone, and the low-altitude flight area data includes multiple grid areas. The fence deviation is not less than the set deviation. The risk level is greater than the threshold level, and the flight knowledge label indicates that the low-altitude flight area data includes a no-fly zone.

[0060] In this embodiment, when the low-altitude flight area data is determined to be compliant flight area data, that is, the low-altitude flight area data does not meet specific violation data requirements, these violation data requirements include that the risk level is greater than the threshold level, the flight knowledge label indicates that the low-altitude flight area data does not include a no-fly zone and the low-altitude flight area data contains multiple grid areas, or the fence deviation is not less than the set deviation, or the risk level is greater than the threshold level and the flight knowledge label indicates that the low-altitude flight area data includes a no-fly zone, etc.

[0061] For compliant low-altitude flight area data, the server must first obtain multiple construction boundaries of the low-altitude flight area data. Taking the airspace near the mountainous area as an example, the server has detailed low-altitude flight area data for the area. Since it is compliant data, the server begins to build boundary-related operations. The server will build boundaries based on various information in the existing low-altitude flight area data. This includes geographic information, such as the coordinates of mountains, the direction of valleys, the location of rivers and lakes, etc. It also includes meteorological-related data, such as wind direction, wind speed and other factors that may affect flight, as well as special signs in the airspace and other information.

[0062] In the process of building boundaries, the server will comprehensively consider the relationship between multiple data elements. For example, for terrain data in mountainous areas, the server will determine some key boundary points based on the height and slope of the peaks. If a mountain is high and has a steep slope, the ridgeline of the peak or a safe height position on the mountainside will be set as part of the boundary when building the boundary. For valleys, the boundary may be determined based on the width and direction of the valley and the meteorological conditions in the valley (such as the influence range of valley wind). At the same time, considering water bodies such as rivers and lakes, the boundary will be constructed based on their edge positions and the scope of influence on flight safety (for example, avoiding aircraft flying close to the water at low altitudes to prevent the risk of accidental falling into the water). In addition, special signs in the airspace, such as the location of specific navigation points or communication base stations, will also affect the construction of boundaries to ensure that aircraft can effectively use these facilities and maintain a safe distance during flight.

[0063] After obtaining multiple construction boundaries, the server needs to determine the flight restriction boundaries of the low-altitude flight area data from these construction boundaries. Suppose in the airspace near the mountainous area, the server obtains three construction boundaries. The first construction boundary is mainly constructed based on the direction and height of the mountain range, covering most of the mountain peak area; the second construction boundary considers more the situation in the valley and the influence range of the valley wind; the third construction boundary combines the influence of rivers, lakes and some special signs nearby. The server will analyze and evaluate these three construction boundaries.

[0064] The server will determine the final flight restriction boundary based on factors such as the type of flying objects in the low-altitude flight area (such as small private aircraft, commercial helicopters, etc.), the airspace environment (such as the complexity of the terrain, the stability of meteorological conditions, etc.), and auxiliary facilities (such as navigation facilities, meteorological monitoring stations, etc.). For each constructed boundary, the server will analyze its adaptability to these factors. For example, for small private aircraft, if a certain constructed boundary is located in an area where there are often strong air currents (unstable airspace environment), then this constructed boundary may not be suitable as a flight restriction boundary, because small private aircraft have a higher risk of flying in strong airflow areas. On the contrary, if there are more navigation facilities (abundant auxiliary facilities) around a constructed boundary and the terrain is relatively flat (better airspace environment), this constructed boundary is more likely to be determined as a flight restriction boundary.

[0065] After comprehensively considering various factors, the server determines the most suitable flight restriction boundary. Finally, the server stores the flight restriction boundary of the low-altitude flight area data. This storage operation is crucial because in actual aviation management, the server needs to quickly and accurately obtain this information in order to effectively manage low-altitude flights. When an aircraft requests to fly in this low-altitude flight area, the server can quickly query the stored flight restriction boundary information to ensure that the aircraft flies within the prescribed boundaries, ensure flight safety, avoid flight conflicts, and make rational use of airspace resources. Through this process, the server completes a series of operations from constructing boundaries to determining flight restriction boundaries and storing them for compliant low-altitude flight area data, thereby providing an accurate and reliable basis for aviation management.

[0066] In a possible implementation, the step of obtaining a plurality of construction boundaries of the low-altitude flight area data includes:

[0067] Step B110, performing feature extraction on the low-altitude flight area data to generate spatial feature data of the low-altitude flight area data and an electronic fence feature vector corresponding to the spatial feature data of the low-altitude flight area data.

[0068] Step B120, performing interference processing on the electronic fence feature vector to generate an interference electronic fence feature vector, fusing the spatial feature data and the interference electronic fence feature vector to generate a fused feature vector, performing de-interference processing on the fused feature vector to generate a target electronic fence feature vector, performing boundary construction on the target electronic fence feature vector to generate multiple construction boundaries.

[0069] In this embodiment, first, the server performs feature extraction on the low-altitude flight area data. Taking the low-altitude flight area data near the mountainous area as an example, after the server obtains the detailed information of the area from many data sources, for the extraction of spatial feature data, the server will analyze the coordinate information of various geographical elements in the area, such as the coordinates of the peak, the coordinates of the lowest point of the valley, and the coordinates of the edge of the river and lake, etc. These coordinate information accurately depicts the topographic and geomorphic features of the area. At the same time, weather-related data features will also be extracted, such as wind speed, wind direction, air pressure and other information in different regions, because these meteorological factors have an important impact on low-altitude flight. In addition, special geographical identifications in the airspace, such as the location of specific navigation facilities, will also be extracted as part of the spatial feature data. While extracting the spatial feature data, the server will also generate an electronic fence feature vector corresponding to the spatial feature data of the low-altitude flight area data. This electronic fence feature vector is constructed based on the spatial feature data, and it contains various feature information that can define the boundary of the low-altitude flight area, such as the approximate shape of the boundary, the terrain slope near the boundary, and other information, which are represented in the form of vectors for subsequent processing.

[0070] Next, the server performs interference processing on the electronic fence feature vector to generate an interference electronic fence feature vector. In actual aviation management scenarios, in order to simulate some special situations or test the stability and robustness of the system, interference factors need to be introduced. For example, the positioning error caused by some equipment failure may be simulated. The server will add some random small deviations to the coordinate information of the electronic fence feature vector, or change the weight value of some boundary features to generate an interference electronic fence feature vector. This is like introducing some uncertain factors in a normal flight environment to test the system's ability to cope with complex situations.

[0071] Then, the server fuses the previously generated spatial feature data and the interference electronic fence feature vector to generate a fused feature vector. This fusion operation combines the topographic, meteorological and other information in the spatial feature data with the electronic fence feature vector after interference processing. For example, the terrain height information of the mountainous area is fused with the boundary shape information with interference, so that the fused feature vector contains both the original geographical and meteorological information and the boundary related information with interference. The purpose of this is to take into account the combined influence of multiple factors in subsequent processing and accurately construct the boundary in the presence of interference.

[0072] Afterwards, the server performs interference removal processing on the fused feature vector to generate the target electronic fence feature vector. Since interference factors were introduced before, these interferences need to be removed in actual applications to obtain accurate boundary information. The server will identify and remove the deviations and error information added during the interference processing process according to pre-set algorithms and rules. For example, for the previously added coordinate deviation, by comparing the coordinate reference in the original spatial feature data and the deviation information in the interfering electronic fence feature vector, the coordinate value is restored to a reasonable range, and the weight value of the changed boundary feature is adjusted to restore it to a value that conforms to the actual situation, thereby obtaining the target electronic fence feature vector.

[0073] Finally, the server constructs the boundary of the target electronic fence feature vector and generates multiple constructed boundaries. In this process, the server will perform detailed boundary construction operations based on various information in the target electronic fence feature vector. For example, the boundary of the mountain area is determined based on the terrain height change information in the vector, the safety buffer boundary that needs to be reserved in certain areas is determined based on the wind direction and wind speed information, and the special boundaries around these facilities are determined based on the location of the navigation facilities. The server will comprehensively consider all the information in the vector and construct multiple construction boundaries that meet the requirements in accordance with the rules and safety standards of aviation management. These construction boundaries will provide the basis for the subsequent determination of the final flight restriction boundaries.

[0074] In a possible implementation manner, when the low-altitude flight area data is non-initial low-altitude flight area data acquired from an airspace, determining the flight restriction boundary of the low-altitude flight area data from the multiple constructed boundaries includes:

[0075] Step C110, obtaining the electronic fence feature vectors corresponding to each of the multiple constructed boundaries, extracting the airspace feature members in the low-altitude flight area data, performing feature extraction on the airspace feature members, generating a regional feature member vector, and for each of the constructed boundaries, determining the matching degree between the electronic fence feature vector of the constructed boundary and the regional feature member vector, using the matching degree as the matching degree between the constructed boundary and the low-altitude flight area data, and using the constructed boundary corresponding to a matching degree greater than the set matching degree as a candidate constructed boundary.

[0076] In this embodiment, taking the airspace near the mountain area as an example, assuming that there are three construction boundaries, the server will operate on these three construction boundaries respectively. For each construction boundary, its electronic fence feature vector contains various feature information about the boundary, such as the shape of the boundary, the terrain features near the boundary (such as slope, altitude, etc.), the relative position relationship between the boundary and special geographical markers (such as navigation facilities, communication base stations, etc.), etc. These electronic fence feature vectors are obtained by processing the low-altitude flight area data before, and they are an important basis for describing the characteristics of the construction boundary.

[0077] Step C120: If the candidate constructed boundary includes an airspace feature member in the low-altitude flight area data, and at least one prior feature member in the airspace feature library includes the airspace feature member, extract the prior feature member vector of the airspace feature member from the airspace feature library.

[0078] Next, the server extracts airspace feature members from the low-altitude flight area data. In mountainous airspace scenarios, airspace feature members may include the types of flying objects that frequently appear in the area (such as small private aircraft, commercial helicopters, etc.), airspace environment-related information (such as complex mountainous terrain, changeable weather conditions, etc.), and auxiliary facilities (such as navigation facilities, meteorological monitoring stations, etc.). The server then extracts features from these airspace feature members to generate a regional feature member vector. For example, for the type of flying object, if the area is mainly small private aircraft flying, the server will extract the flight performance characteristics of these aircraft (such as flight speed range, flight altitude range, etc.), represent these features in the form of vectors, and combine them with features related to the airspace environment and auxiliary facilities to form a regional feature member vector.

[0079] Step C130, determining the first feature distances between the candidate construction boundary and the prior feature member vectors of different feature member dimensions, wherein the feature member dimensions of the prior feature member include: flying objects, airspace environment, and auxiliary facilities.

[0080] Step C140, performing fusion calculation on the first feature distances between the candidate constructed boundary and the prior feature member vectors of different feature member dimensions, generating the first feature distance between the candidate constructed boundary and the prior feature member vector, and extracting the electronic fence feature vector of the prior flight restriction boundary of the prior low-altitude flight area data from the airspace feature library, and determining the second feature distance between the candidate constructed boundary and the prior flight restriction boundary based on the electronic fence feature vector of the prior flight restriction boundary and the electronic fence feature vector of the candidate constructed boundary.

[0081] Step C150: performing a fusion calculation on the first feature distance and the second feature distance to generate a target feature distance of the candidate construction boundary.

[0082] Step C160, setting the candidate construction boundary corresponding to the target feature distance that meets the preset conditions as the flight restriction boundary of the low-altitude flight area data.

[0083] For each construction boundary, the server determines the matching degree between the geo-fence feature vector of the construction boundary and the regional feature member vector, and uses this matching degree as the matching degree between the construction boundary and the low-altitude flight area data. Taking the first construction boundary as an example, the server compares its geo-fence feature vector with the regional feature member vector. If the shape and position of the boundary in the geo-fence feature vector match the flight path of the flying object in the regional feature member vector, the terrain in the airspace environment, and the distribution of auxiliary facilities, then the matching degree will be high. For example, the shape of the construction boundary can reasonably avoid complex terrain and be close to navigation facilities, which matches the information in the regional feature member vector, and the calculated matching degree is 0.8. In the same way, the server also calculates other construction boundaries, and uses the construction boundaries corresponding to the matching degree greater than the set matching degree (assuming the set matching degree is 0.7) as candidate construction boundaries. Assuming that the matching degree of the second construction boundary is 0.75 and the matching degree of the third construction boundary is 0.9, then the second and third construction boundaries become candidate construction boundaries.

[0084] If the candidate construction boundary includes an airspace feature member in the low-altitude flight area data, and at least one prior feature member in the airspace feature library includes the airspace feature member, the server will extract the prior feature member vector of the airspace feature member from the airspace feature library. For example, for airspace environment feature members (such as mountainous terrain) in the candidate construction boundary, if there is a prior feature member about mountainous terrain in the airspace feature library, the server will extract the prior feature member vector. This prior feature member vector contains the results of previous analysis of similar airspace environments, such as the best flight path for flying objects under such terrain, areas prone to danger, and other information.

[0085] Then, the server determines the first characteristic distance between the candidate construction boundary and the prior characteristic member vectors of different characteristic member dimensions. The calculation is performed separately for the three characteristic member dimensions of the prior characteristic members, namely, flying objects, airspace environment, and auxiliary facilities. Taking the second candidate construction boundary as an example, in the flying object dimension, if the flying habits of the flying objects recorded in the prior characteristic member vector (such as flight altitude preference, etc.) are different from the flying conditions of the flying objects adapted by the candidate construction boundary, the server will calculate the first characteristic distance in this dimension according to a certain algorithm. In the airspace environment dimension, the terrain features (such as the direction and slope of the mountains, etc.) in the prior characteristic member vector are compared with the airspace environment conditions corresponding to the candidate construction boundary to calculate the corresponding first characteristic distance. For the auxiliary facility dimension, the first characteristic distance is calculated based on the auxiliary facility layout in the prior characteristic member vector and the distribution of auxiliary facilities around the candidate construction boundary.

[0086] Next, the server performs a fusion calculation on the first feature distances between the candidate construction boundary and the prior feature member vectors of different feature member dimensions, and generates the first feature distance between the candidate construction boundary and the prior feature member vector. This fusion calculation may involve operations such as weighted summation of feature distances of different dimensions, assigning different weights according to the importance of each dimension. For example, in mountainous airspace, the airspace environment dimension may be assigned a higher weight because the terrain has a greater impact on flight safety. The first feature distance between the second candidate construction boundary and the prior feature member vector is calculated to be a specific value.

[0087] At the same time, the server extracts the electronic fence feature vector of the prior flight restriction boundary of the prior low-altitude flight area data from the airspace feature library, and determines the second feature distance between the candidate construction boundary and the prior flight restriction boundary based on the electronic fence feature vector of the prior flight restriction boundary and the electronic fence feature vector of the candidate construction boundary. For the second candidate construction boundary, the server compares its electronic fence feature vector with the electronic fence feature vector of the prior flight restriction boundary. If the two vectors differ in terms of boundary shape, position, etc., the second feature distance is calculated according to the corresponding algorithm.

[0088] Finally, the server performs a fusion calculation on the first feature distance and the second feature distance to generate the target feature distance of the candidate construction boundary. This fusion calculation may also involve weighting and other operations to comprehensively consider the influence of the first feature distance and the second feature distance. For example, the first feature distance is multiplied by a weight of 0.4, and the second feature distance is multiplied by a weight of 0.6, and then added to obtain the target feature distance of the candidate construction boundary. Then, the candidate construction boundary corresponding to the target feature distance that meets the preset conditions (such as the target feature distance is less than a certain threshold) is set as the flight restriction boundary of the low-altitude flight area data. Assuming that the target feature distance of the second candidate construction boundary meets the preset conditions, then this candidate construction boundary is determined as the flight restriction boundary of the low-altitude flight area data, and the server will store this result so as to accurately restrict and guide low-altitude flights in aviation management.

[0089] For example, in a possible implementation, step B120 includes:

[0090] Step B121, initialize various construction parameters and data structures required in the boundary construction process, wherein the construction parameters include the accuracy threshold of the boundary construction and the direction identification parameters, and the data structure is used to store the intermediate results and the boundary information of the final construction, wherein a list data structure for storing the coordinates of the boundary points is specifically created and initialized to empty, and a flag variable for recording the status of the construction process is initialized to a default state.

[0091] Step B122, parsing the target electronic fence feature vector according to a predefined rule, decomposing the target electronic fence feature vector into a plurality of feature vector components, each of which corresponds to a feature representation of the electronic fence in different dimensions.

[0092] Step B123, preliminarily determining the boundary construction direction according to the relevant information in the parsed feature vector component, wherein the boundary construction direction is determined based on a preset spatial direction identifier or an association relationship with other relevant features.

[0093] Step B124, determining the starting point of construction according to the initially determined boundary construction direction and the relevant information in the parsed feature vector components, wherein the determination of the starting point involves numerical analysis of the feature vector components or matching with predefined starting point rules.

[0094] Step B125, starting from the starting point of the construction, according to the preliminarily determined boundary construction direction, the boundary points are gradually constructed according to the parsed feature vector components to generate a preliminarily constructed boundary point sequence, wherein during the construction process, the offset or coordinate change of each boundary point relative to the previous boundary point is determined according to the numerical information in the feature vector components, and the accuracy threshold in the construction parameter set is referenced to ensure that the construction of each boundary point meets the accuracy requirements.

[0095] Step B126, performing an integrity check on the initially constructed boundary point sequence. When the initially constructed boundary point sequence has an integrity problem, adjustments are made based on the relevant information in the parsed feature vector component, wherein the integrity check includes whether the boundary is closed and whether the spacing between boundary points meets preset requirements.

[0096] Step B127, based on the set identifier or numerical range in the parsed feature vector component, determine whether it is necessary to construct multiple boundaries. If the judgment result is yes, segment the target electronic fence feature vector according to the set identifier or numerical range in the feature vector component, determine the feature vector component corresponding to each boundary construction, and repeat the above process for each segmented feature vector component to construct a boundary point sequence of multiple boundaries.

[0097] Step B128, optimizing the constructed multiple boundary point sequences, the optimization operation includes removing redundant boundary points and smoothing the coordinates of the boundary points, wherein the redundant boundary points are removed by comparing the boundary point distance between adjacent boundary points with the redundant judgment threshold in the construction parameter set, if the boundary point distance is less than the set distance threshold, the corresponding boundary point is determined to be a redundant point and is removed, and the process of smoothing the coordinates of the boundary points specifically calculates new coordinate values ​​based on the coordinate values ​​of adjacent boundary points and predefined weight coefficients, thereby obtaining an optimized constructed boundary.

[0098] In this embodiment, firstly, various construction parameters and data structures required in the boundary construction process are initialized. Construction parameters play a key role in constructing the boundary of the low-altitude flight area. Among them, the precision threshold of boundary construction is an important construction parameter, which determines the accuracy of the constructed boundary points. For example, in the airspace near the mountainous area, the precision threshold may be set to a smaller value in meters. This is because the mountainous terrain is complex, and a smaller precision threshold can more accurately depict the boundary and prevent the aircraft from entering the dangerous area due to inaccurate boundaries. The direction identification parameter is used to determine the direction reference of the boundary construction, which helps to accurately define the scope of the flight area in three-dimensional space.

[0099] At the same time, the server creates a specific data structure to store the intermediate results and the boundary information of the final build. Specifically, the server creates a list data structure for storing the coordinates of the boundary points, which is initialized to empty. This list will be gradually filled with the coordinate information of the boundary points during the subsequent build process. In addition, a flag variable is created to record the status of the build process, which is initialized to the default state. This flag variable can be used to mark various states in the build process, such as whether it is being built, whether special circumstances are encountered, etc., so that the server can perform corresponding processing according to different states during the build process.

[0100] Next, the server parses the target electronic fence feature vector according to predefined rules and decomposes it into multiple feature vector components. The target electronic fence feature vector contains rich information about the electronic fence in different dimensions. Taking mountainous airspace as an example, there may be terrain height dimensions, distance dimensions from special geographical markers (such as navigation facilities, communication base stations, etc.), and meteorological conditions. Through parsing, the server decomposes the target electronic fence feature vector into multiple feature vector components according to these different dimensions. Each feature vector component corresponds to the feature representation of the electronic fence in different dimensions. For example, the feature vector component of the terrain height dimension may contain mountain height data, valley depth data, etc. These data are represented in the form of vectors and can accurately reflect the characteristics of the electronic fence under this dimension.

[0101] Based on the relevant information in the parsed eigenvector components, the server preliminarily determines the boundary construction direction. The determination of this boundary construction direction is based on a preset spatial direction identifier or an association with other relevant features. In mountainous airspace, if the preset spatial direction identifier is to build the boundary according to the direction of the mountain range, the server will determine the direction of the mountain range based on the eigenvector component of the terrain height dimension in the target electronic fence feature vector, thereby determining the boundary construction direction. Alternatively, if there is an association with other relevant features, such as the impact of wind direction on flight, if the wind direction is from west to east, in order to avoid the risk of the aircraft flying against the wind, the boundary construction direction may consider relationships such as perpendicular or parallel to the wind direction, thereby determining a suitable boundary construction direction.

[0102] After determining the boundary construction direction, the server determines the construction starting point based on the preliminary boundary construction direction and the relevant information in the parsed feature vector component. The determination of the starting point involves numerical analysis of the feature vector component or matching with the predefined starting point rule. For example, when determining the starting point based on the feature vector component of the terrain height dimension, if it is specified that the starting point should be located at a special location such as a specific height position in the mountain range or the lowest point in the valley, the server will analyze the numerical value in the feature vector component to find a coordinate point that meets the requirements as the starting point. Alternatively, if the starting point is determined according to the feature vector component of the distance dimension from a special geographic marker, and the predefined starting point rule is the point closest to the navigation facility, the server will look for the coordinate point closest to the navigation facility in the feature vector component as the starting point.

[0103] Starting from the determined starting point, the server gradually constructs boundary points according to the initially determined boundary construction direction and the resolved feature vector components to generate a preliminarily constructed boundary point sequence. In this process, the offset or coordinate change of each boundary point relative to the previous boundary point is determined based on the numerical information in the feature vector component, and the accuracy threshold in the construction parameter set is referenced to ensure that the construction of each boundary point meets the accuracy requirements. Taking the feature vector component of the terrain height dimension as an example, if the slope data of a certain section of the mountain range is represented as a certain value in the feature vector component, the server will calculate the vertical coordinate change of each boundary point relative to the previous boundary point based on this value, and calculate the corresponding coordinate change in the horizontal direction based on the relationship with other features (such as the distance of special geographical signs, etc.). When constructing each boundary point, the accuracy threshold will be referenced. If the calculated boundary point coordinate change causes the distance between the boundary points to be less than the accuracy threshold, the server will adjust the coordinate change to meet the accuracy requirements, thereby ensuring that the constructed boundary point sequence has sufficient accuracy.

[0104] After generating the preliminary constructed boundary point sequence, the server performs an integrity check on the preliminary constructed boundary point sequence. The contents of the integrity check include whether the boundary is closed, whether the spacing between boundary points meets the preset requirements, etc. In mountainous airspace, if the constructed boundary is a flight area around a mountainous area, then the boundary should be closed, that is, the starting point and the end point should coincide. The server will check whether the coordinates of the first boundary point and the last boundary point in the preliminary constructed boundary point sequence are the same. If they are not the same, it means that the boundary is not closed and there is an integrity problem. At the same time, for the spacing between boundary points, the server will check whether the distance between adjacent boundary points meets the preset requirements. For example, the preset requirement is that the distance between adjacent boundary points cannot exceed a certain value. If the distance between adjacent boundary points is too large or too small (less than the accuracy threshold), it means that there is an integrity problem. When there is an integrity problem in the preliminary constructed boundary point sequence, the server will make adjustments based on the relevant information in the parsed feature vector component. If the boundary is not closed, the server may adjust the coordinates of the last boundary point based on the information in the feature vector components, such as the terrain height dimension or the distance dimension to the special geographic landmark, so that it coincides with the first boundary point; if the spacing between the boundary points does not meet the requirements, the server will recalculate the coordinate changes between the boundary points and adjust the positions of the boundary points based on the numerical information in the feature vector components to meet the preset spacing requirements.

[0105] According to the set identifier or value range in the parsed feature vector component, the server determines whether it is necessary to construct multiple boundaries. In mountainous airspace, there may be different flight area division requirements or different terrain, meteorological and other conditions that require the construction of multiple boundaries. For example, if the set identifier in the parsed feature vector component indicates that there are different flight rules in a certain area (such as a strictly restricted area close to the core area of ​​the mountain and a relatively loose area on the periphery), or the numerical range shows that the terrain or meteorological conditions of a certain area are significantly different from other areas (such as a strong wind area on one side and a relatively calm area on the other side), then multiple boundaries need to be constructed. If the judgment result is yes, the server will segment the target electronic fence feature vector according to the set identifier or value range in the feature vector component, and determine the feature vector component corresponding to each boundary construction. For example, the feature vector component corresponding to the strong wind area is segmented to construct the boundary of the strong wind area, and the feature vector component corresponding to the core area of ​​the mountain area is segmented to construct the boundary of the core area, etc. Then for each segmented feature vector component, the server repeats the above process to construct a boundary point sequence of multiple boundaries.

[0106] Finally, the server optimizes the constructed multiple boundary point sequences. The optimization operation includes removing redundant boundary points and smoothing the coordinates of boundary points. When removing redundant boundary points, it is achieved by comparing the boundary point distance between adjacent boundary points and the redundancy judgment threshold in the construction parameter set. For example, in mountainous airspace, if the distance between adjacent boundary points in the constructed boundary point sequence is very small (less than the set distance threshold), it means that these boundary points may be redundant. The server will determine that the corresponding boundary point is a redundant point and remove it to simplify the boundary representation. The process of smoothing the coordinates of the boundary points is to calculate the new coordinate values ​​according to the coordinate values ​​of the adjacent boundary points and the pre-defined weight coefficients, so as to obtain the optimized constructed boundary. For example, for a certain section of the boundary point sequence, the server will calculate the new coordinate values ​​according to the longitude and latitude coordinate values ​​of the adjacent boundary points according to the pre-defined weight coefficients (such as the weight of the previous boundary point is 0.4, and the weight of the next boundary point is 0.6), so that the coordinate transition of the boundary points can be smoother, and sharp corners of the boundary can be avoided, so as to obtain an optimized constructed boundary that better meets the actual flight needs. Through the above series of steps, the server successfully constructs multiple construction boundaries based on the target electronic fence feature vector. These construction boundaries will provide an important basis for subsequent low-altitude flight management.

[0107] In a possible implementation, the method further includes:

[0108] When the low-altitude flight area data is initial low-altitude flight area data obtained from an airspace, the step of optimizing the airspace feature library comprises:

[0109] The airspace feature members in the low-altitude flight area data are used as the priori feature members, and the priori feature members and the corresponding priori feature member vectors are archived in the airspace feature library.

[0110] The flight restriction boundary of the low-altitude flight area data is used as a priori boundary, and the priori boundary is archived in the airspace feature library.

[0111] And, when the low-altitude flight area data is non-initial low-altitude flight area data obtained from the airspace, the step of optimizing the airspace feature library includes:

[0112] The feature vectors of the spatial feature members appearing in the candidate construction boundary and the prior feature member vectors of the prior feature members of the same index attribute in the spatial feature library are fused and calculated, and the optimized prior feature member vectors overwrite the prior feature member vectors before optimization.

[0113] The electronic fence feature vector of the candidate construction boundary and the electronic fence feature vector of the prior flight restriction boundary in the airspace feature library are fused and calculated, and the optimized electronic fence feature vector covers the electronic fence feature vector before optimization.

[0114] In this embodiment, when the low-altitude flight area data is the initial low-altitude flight area data obtained from the airspace, the server will perform specific optimization steps on the airspace feature library. Taking the airspace near the mountainous area as an example, when the server first obtains the low-altitude flight area data of the area, it will perform a series of operations to enrich the airspace feature library. First, the server will use the airspace feature members in the low-altitude flight area data as prior feature members. In this mountainous airspace scenario, the airspace feature members contain multiple elements, such as the type of flying objects unique to the area. Assuming that there are often small private aircraft and some drones used for terrain surveys flying here, the types of these flying objects, flight performance (such as flight speed range, flight altitude range, endurance, etc.) and other information become part of the airspace feature members. In addition, airspace environmental characteristics are also an important component, such as the complex terrain of the mountainous area, including the height, slope, depth and direction of the valley, as well as the changing meteorological conditions, such as local strong wind areas, unstable airflow areas and other information. There are also auxiliary facilities, such as the existing navigation facilities, the location of meteorological monitoring stations, coverage and other information.

[0115] The server archives these airspace feature members and the corresponding prior feature member vectors into the airspace feature library. The prior feature member vector is the result of quantifying and encoding the airspace feature members. For example, for the type of flying object, the relevant performance data of small private aircraft and drones may be converted into vector form according to a certain encoding rule. For terrain information, a vector that can represent the terrain characteristics is constructed based on the coordinates, height and other data of the terrain. These prior feature member vectors are stored in the airspace feature library together with the corresponding prior feature members, so that they can be used as a reference when processing other low-altitude flight area data later.

[0116] At the same time, the server will use the flight restriction boundary of the low-altitude flight area data as a priori boundary and archive it in the airspace feature library. This flight restriction boundary is based on the analysis and processing of the initial low-altitude flight area data. For example, the boundary is determined based on factors such as the terrain, meteorological conditions, and flight habits of flying objects in the mountainous area. This boundary may be drawn along the ridgeline of the mountain, the edge of the valley, or outside the safe area that takes into account the influence of meteorological conditions. Archiving this flight restriction boundary in the airspace feature library can provide an initial boundary reference for subsequent processing of low-altitude flight area data in similar areas, which helps to improve the efficiency and accuracy of subsequent boundary determination.

[0117] When the low-altitude flight area data is non-initial low-altitude flight area data obtained from the airspace, the server's optimization steps for the airspace feature library are different. Suppose the server is processing new low-altitude flight area data in an airspace near a mountainous area that already has a certain airspace feature library foundation. The server will first fuse the feature vectors of the airspace feature members that appear in the candidate construction boundary and the prior feature member vectors of the prior feature members with the same index attribute in the airspace feature library. For example, in the candidate construction boundary, if a new feature vector about the impact of meteorological conditions on flight is found, such as the presence of an airflow pattern that has not been fully recorded at a certain altitude, and the airspace feature library already has prior feature member vectors about meteorological conditions, they may all contain information on relevant dimensions such as wind speed, wind direction, and airflow stability.

[0118] The server will fuse these two vectors according to a specific fusion calculation algorithm. This algorithm may take into account the weight of new data. For example, newly discovered airflow pattern data may be given a higher weight because it is a new observation result, and the prior data in the airspace feature library will also be given corresponding weights based on factors such as its historical reliability. Through this fusion calculation, the optimized prior feature member vector is obtained, and then the optimized prior feature member vector is overwritten with the prior feature member vector before optimization, thereby updating the meteorological condition-related prior feature member vector in the airspace feature library to include the latest airspace feature information.

[0119] In addition, the server will also fuse the electronic fence feature vector of the candidate construction boundary and the electronic fence feature vector of the prior flight restriction boundary in the airspace feature library. Taking the mountainous airspace as an example, the electronic fence feature vector of the candidate construction boundary contains information such as the shape, location, and relationship with the surrounding geographical features and auxiliary facilities of the newly constructed boundary, while the electronic fence feature vector of the prior flight restriction boundary is the boundary-related information previously stored in the airspace feature library. The server will operate according to the pre-set fusion calculation rules. For example, if the new candidate construction boundary takes into account the impact of new navigation facilities, and the electronic fence feature vector of the prior flight restriction boundary does not fully contain this information, then during the fusion calculation, the new navigation facility-related information will be integrated into the electronic fence feature vector of the prior flight restriction boundary.

[0120] Similarly, this fusion calculation will take into account the weights of different information. The new candidate construction boundary electronic fence feature vector may be given a higher weight because it is based on the latest low-altitude flight area data. The optimized electronic fence feature vector is obtained through this fusion calculation, and then the optimized electronic fence feature vector is overwritten with the electronic fence feature vector before optimization, thereby updating the electronic fence feature vector of the prior flight restriction boundary in the airspace feature library. In this way, the airspace feature library can continuously absorb useful information from new low-altitude flight area data, maintain the accuracy and completeness of its description of airspace features, and provide better decision-making basis for subsequent low-altitude flight management.

[0121] Figure 2 The low-altitude high-precision electronic fence control system 100 shown includes: a processor 1001 and a memory 1003. The processor 1001 and the memory 1003 are connected, such as through a bus 1002. Optionally, the low-altitude high-precision electronic fence control system 100 may also include a transceiver 1004, which can be used for data interaction between the server and other servers, such as data transmission and / or data reception. It should be noted that the transceiver 1004 is not limited to one in actual scheduling, and the structure of the low-altitude high-precision electronic fence control system 100 does not constitute a limitation on the embodiments of the present application.

[0122] Processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0123] The bus 1002 may include a path to transmit information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0124] The memory 1003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium, other magnetic storage devices, or any other medium that can be used to carry or store program code and can be read by a computer, without limitation herein.

[0125] The memory 1003 is used to store program codes for executing the embodiments of the present application, and the execution is controlled by the processor 1001. The processor 1001 is used to execute the program codes stored in the memory 1003 to implement the steps shown in the above method embodiments.

[0126] In addition, an embodiment of the present application also provides a readable storage medium, in which computer executable instructions are pre-installed. When the processor executes the computer executable instructions, the high-precision electronic fence control method for low-altitude flight as above is implemented.

[0127] Similarly, it should be noted that in order to simplify the description disclosed in this application, thereby helping to understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, multiple features are sometimes combined into one embodiment, drawings, or descriptions thereof. Similarly, it should be noted that in order to simplify the description disclosed in this application, thereby helping to understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, multiple features are sometimes combined into one embodiment, drawings, or descriptions thereof.

Claims

1. A high-precision electronic fence control method for low-altitude flight, characterized in that: The method comprises: Obtain low-altitude flight area data; Acquire spatial feature data of the low-altitude flight area data, and convert the spatial feature data into a risk level and a flight knowledge label of the low-altitude flight area data; If the risk level is greater than the threshold level, the flight knowledge tag indicates that the low-altitude flight area data does not include a no-fly zone, and the low-altitude flight area data includes multiple grid areas, obtaining high-precision electronic fences corresponding to each of the multiple grid areas, and the matching degree between the grid area and the corresponding high-precision electronic fence is greater than a set matching degree; Determine the fence deviation between the high-precision electronic fences corresponding to the multiple grid areas; If the fence deviation is less than the set deviation, the high-precision electronic fences corresponding to each of the plurality of grid areas are used as the flight restriction boundaries of the corresponding grid areas; The determining of the fence deviation between the high-precision electronic fences corresponding to the plurality of grid areas includes: For each of the high-precision electronic fences, determine the mean value of the spatial element coordinates of each boundary segment in the high-precision electronic fence, and combine the mean value of the spatial element coordinates of each boundary segment into an electronic fence feature vector of the high-precision electronic fence; Determine the discreteness of the electronic fence feature vectors of the high-precision electronic fences corresponding to each of the multiple grid areas, and use the discreteness as the fence deviation between the high-precision electronic fences corresponding to each of the multiple grid areas.

2. The high-precision electronic fence control method for low-altitude flight according to claim 1 is characterized in that: The method further comprises: If the fence deviation is not less than the set deviation, the multiple grid areas are further subdivided into multiple new low-altitude flight area data.

3. The high-precision electronic fence control method for low-altitude flight according to claim 1 is characterized in that: The step of acquiring the spatial feature data of the low-altitude flight area data comprises: Extracting the spatial element coordinate distribution of the low-altitude flight area data, and performing feature extraction through a spatial feature coding network based on the spatial element coordinate distribution to generate spatial feature data of the low-altitude flight area data; The step of converting the spatial feature data into the risk level and flight knowledge label of the low-altitude flight area data comprises: Performing attention processing on the spatial feature data through an attention unit in a first airspace feature prediction model to generate first attention feature data, converting the first attention feature data into a first confidence level of a plurality of reference risk levels through a fully connected mapping unit in the first airspace feature prediction model, and using the reference risk level corresponding to the first confidence level that meets a preset condition as the risk level of the low-altitude flight area data, wherein the risk level is used to indicate a score for generating a flight restriction boundary based on the low-altitude flight area data; The spatial feature data is subjected to attention processing by the attention unit in the second spatial feature prediction model to generate second attention feature data, and the second attention feature data is converted into a second confidence level that maps multiple flight knowledge labels by the fully connected mapping unit in the first spatial feature prediction model, and the flight knowledge label corresponding to the second confidence level that meets the preset conditions is used as the flight knowledge label of the low-altitude flight area data, and the flight knowledge label includes: including no-fly zones and excluding no-fly zones.

4. The high-precision electronic fence control method for low-altitude flight according to claim 1 is characterized in that: The method further comprises: If the risk level is greater than the threshold level, and the flight knowledge tag indicates that the low-altitude flight area data includes a no-fly zone, removing the low-altitude flight area data; If the risk level is not greater than the threshold level, the flight restriction boundary of the low-altitude flight area data is archived into the boundary sequence of the airspace according to the constructed time order, and different low-altitude flight area data are obtained from the airspace in sequence.

5. The high-precision electronic fence control method for low-altitude flight according to claim 1 is characterized in that: The method further comprises: Acquire a target combination template sequence, wherein the target combination template sequence includes a plurality of target combinations, wherein the target combination includes template low-altitude flight area data and template high-precision electronic fence; According to the ascending order of the multiple defined boundary points, the matching parameters of the target combination template sequence at the boundary points of this round are determined, and the matching parameters are the statistics of the matching target combination divided by the presidential statistic of the multiple target combinations; the matching degree between the template low-altitude flight area data and the template high-precision electronic fence in the matching target combination is not less than the boundary points of this round; If the matching parameter at the current round boundary value point is not less than the set matching parameter, the current round boundary value point is set to the set matching degree; The discreteness of the electronic fence feature vector of the template high-precision electronic fence in the matching target combination is determined, and the discreteness of the electronic fence feature vector of the template high-precision electronic fence in the matching target combination is used as the set deviation.

6. The high-precision electronic fence control method for low-altitude flight according to any one of claims 1 to 5, characterized in that: The method further comprises: If the low-altitude flight area data is compliant flight area data, multiple construction boundaries of the low-altitude flight area data are obtained, the flight restriction boundaries of the low-altitude flight area data are determined from the multiple construction boundaries, and the flight restriction boundaries of the low-altitude flight area data are stored, the compliant flight area data is flight area data that does not meet the violation data requirements, and the violation data requirements include: the risk level is greater than the threshold level, the flight knowledge label indicates that the low-altitude flight area data does not include a no-fly zone, and the low-altitude flight area data includes multiple grid areas; the fence deviation is not less than the set deviation; the risk level is greater than the threshold level, and the flight knowledge label indicates that the low-altitude flight area data includes a no-fly zone.

7. The high-precision electronic fence control method for low-altitude flight according to claim 6 is characterized in that: The step of obtaining a plurality of construction boundaries of the low-altitude flight area data comprises: Performing feature extraction on the low-altitude flight area data to generate spatial feature data of the low-altitude flight area data and an electronic fence feature vector corresponding to the spatial feature data of the low-altitude flight area data; Performing interference processing on the electronic fence feature vector to generate an interference electronic fence feature vector, fusing the spatial feature data with the interference electronic fence feature vector to generate a fused feature vector, performing de-interference processing on the fused feature vector to generate a target electronic fence feature vector, and performing boundary construction on the target electronic fence feature vector to generate a plurality of construction boundaries; And, when the low-altitude flight area data is non-initial low-altitude flight area data obtained from an airspace, determining the flight restriction boundary of the low-altitude flight area data from the multiple constructed boundaries includes: Acquire the electronic fence feature vectors corresponding to each of the multiple construction boundaries, extract the airspace feature members in the low-altitude flight area data, perform feature extraction on the airspace feature members, generate a regional feature member vector, and for each of the construction boundaries, determine the matching degree between the electronic fence feature vector of the construction boundary and the regional feature member vector, use the matching degree as the matching degree between the construction boundary and the low-altitude flight area data, and use the construction boundary corresponding to the matching degree greater than the set matching degree as a candidate construction boundary; If the candidate constructed boundary includes an airspace feature member in the low-altitude flight area data, and at least one prior feature member in the airspace feature library includes the airspace feature member, extracting a prior feature member vector of the airspace feature member from the airspace feature library; Determine the first feature distances between the candidate construction boundary and the prior feature member vectors of different feature member dimensions, wherein the feature member dimensions of the prior feature member include: flying objects, airspace environment, and auxiliary facilities; Performing fusion calculation on the first feature distances between the candidate constructed boundary and the prior feature member vectors of different feature member dimensions, respectively, to generate the first feature distance between the candidate constructed boundary and the prior feature member vector, and extracting the electronic fence feature vector of the prior flight restriction boundary of the prior low-altitude flight area data from the airspace feature library, and determining the second feature distance between the candidate constructed boundary and the prior flight restriction boundary based on the electronic fence feature vector of the prior flight restriction boundary and the electronic fence feature vector of the candidate constructed boundary; Performing a fusion calculation on the first feature distance and the second feature distance to generate a target feature distance of the candidate construction boundary; The candidate construction boundary corresponding to the target feature distance that meets the preset conditions is set as the flight restriction boundary of the low-altitude flight area data.

8. The high-precision electronic fence control method for low-altitude flight according to claim 7 is characterized in that: The method further comprises: When the low-altitude flight area data is initial low-altitude flight area data obtained from an airspace, the step of optimizing the airspace feature library comprises: Using the airspace feature members in the low-altitude flight area data as the prior feature members, and filing the prior feature members and the corresponding prior feature member vectors into the airspace feature library; Using the flight restriction boundary of the low-altitude flight area data as a priori boundary, and filing the priori boundary into the airspace feature library; And, when the low-altitude flight area data is non-initial low-altitude flight area data obtained from the airspace, the step of optimizing the airspace feature library includes: The feature vectors of the spatial feature members appearing in the candidate construction boundary and the prior feature member vectors of the prior feature members of the same index attribute in the spatial feature library are fused and calculated, and the optimized prior feature member vectors are used to cover the prior feature member vectors before optimization; The electronic fence feature vector of the candidate construction boundary and the electronic fence feature vector of the prior flight restriction boundary in the airspace feature library are fused and calculated, and the optimized electronic fence feature vector covers the electronic fence feature vector before optimization.

9. A high-precision electronic fence control system for low-altitude flight, characterized in that: The high-precision electronic fence control system for low-altitude flight includes a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions, and the machine-executable instructions are loaded and executed by the processor to implement the high-precision electronic fence control method for low-altitude flight described in any one of claims 1-8.

Images