Electromagnetic environment abnormity identification system and method based on unmanned aerial vehicle

Abstract

The invention discloses an electromagnetic environment abnormity identification system and method based on an unmanned aerial vehicle, and the method comprises the steps: obtaining a to-be-monitored area, dividing the to-be-monitored area into space grid units, planning an identification path based on the space grid units, setting a monitoring task frequency band comprising a plurality of sub-frequency bands, and distributing the sub-frequency bands to different space grid units; the unmanned aerial vehicle flies along an identification path, and collects low-resolution spectrum data and high-resolution spectrum data of corresponding sub-bands in each space grid unit; the method comprises the following steps: aligning, grouping and splicing high-resolution frequency spectrum data, and performing frequency spectrum interpolation estimation by combining low-resolution frequency spectrum data to generate a frequency spectrum distribution model; according to the method, a spectrum body model is constructed according to a spectrum distribution model, and the electromagnetic environment abnormity is identified based on the intensity difference of a main unit and an adjacent subunit at the same frequency position. According to the invention, high-efficiency and accurate electromagnetic environment abnormity identification can be realized under limited flight conditions.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic environment detection technology, specifically to an electromagnetic environment anomaly identification system and method based on an unmanned aerial vehicle (UAV). Background Technology

[0002] Unmanned aerial vehicles (UAVs) have advantages such as high mobility, flexible deployment, wide coverage, and the ability to operate in complex terrain and dangerous areas. They can quickly enter areas that are difficult to reach by traditional ground monitoring methods to carry out electromagnetic environment inspection and monitoring. Therefore, they have broad application prospects in fields such as electromagnetic environment anomaly identification, illegal wireless power source investigation, major event security, disaster emergency communication assessment, and safety monitoring around important facilities.

[0003] However, in practical applications, UAVs have limited payload and endurance, making it difficult to perform high-resolution continuous scanning across a wide frequency band for extended periods. Furthermore, the complex propagation characteristics of electromagnetic signals in space mean that data acquired from a single location within a short timeframe cannot fully reflect the true electromagnetic environment. Therefore, how to acquire effective electromagnetic environment data with both high spectral resolution and wide spatial coverage within a limited flight time has become a critical problem that urgently needs to be solved when using UAVs for electromagnetic anomaly identification. Summary of the Invention

[0004] The purpose of this invention is to provide an electromagnetic environment anomaly identification system and method based on unmanned aerial vehicles (UAVs) to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs), the method comprising the following steps: The monitoring area is acquired, and the UAV performs electromagnetic environment anomaly identification within the monitoring area; the monitoring area is divided into several spatial grid units, and an identification path is planned based on the spatial grid units; a monitoring task frequency band is set, which includes several sub-frequency bands, and the sub-frequency bands are assigned to different spatial grid units; Furthermore, the process of planning the identification path and assigning the sub-frequency band to different spatial grid cells includes: Based on the requirements of scanning accuracy, a three-dimensional size is preset, and the area to be monitored is divided according to the preset three-dimensional size to obtain the spatial grid unit; the geometric center of each spatial grid unit is used as the path node, and the arrangement order is set. The path nodes are arranged according to the arrangement order to form the recognition path. Specifically, the arrangement order is set according to a preset traversal rule so that the drone can reach each path node in sequence; the preset traversal rule can be set according to task requirements, such as arranging by row and column, arranging by column and row, serpentine back and forth, or arranging by height layer and then layer by layer, etc. Based on the requirements of scanning accuracy, the monitoring task frequency band is divided into several consecutive sub-frequency bands according to frequency order, and the number of sub-frequency bands is less than or equal to the number of spatial grid cells; The sub-frequency bands are sequentially assigned to the corresponding spatial grid cells according to their arrangement order; when the number of sub-frequency bands is less than the number of spatial grid cells, the sub-frequency bands are cyclically assigned to the remaining spatial grid cells until all spatial grid cells have their corresponding sub-frequency bands. Specifically, the starting and ending frequencies of the monitoring task frequency band are obtained, the frequency direction is determined according to the starting and ending frequencies, and the monitoring task frequency band is divided into several continuous sub-frequency bands along the frequency direction; each sub-frequency band is connected end to end in frequency, and the ending frequency of the previous sub-frequency band is adjacent to the starting frequency of the next sub-frequency band. The sub-bands are numbered in ascending order of frequency to obtain a sub-band sequence. The sub-band with the lowest frequency is designated as the first sub-band, the sub-band with the second lowest frequency is designated as the second sub-band, and so on, until the nth sub-band is obtained. The spatial grid unit sequence is obtained based on the identification path. The spatial grid unit located at the beginning of the identification path is taken as the first spatial grid unit, and the remaining spatial grid units passed along the identification path are taken as the second spatial grid unit, the third spatial grid unit, and so on until the last spatial grid unit. The sub-frequency band sequence is allocated sequentially according to the order of the spatial grid cell sequence, that is, the first sub-frequency band is allocated to the first spatial grid cell, the second sub-frequency band is allocated to the second spatial grid cell, and so on; when the nth sub-frequency band is allocated, the nth sub-frequency band is allocated to the nth spatial grid cell corresponding to the path sequence, so as to establish the correspondence between the spatial grid cell and the sub-frequency band. When the number of sub-bands is less than the number of spatial grid cells, after completing the first round of sequential allocation, the remaining spatial grid cells that have not yet been allocated sub-bands continue to be repeatedly allocated; after completing one round of allocation from the first sub-band to the nth sub-band, the first sub-band is allocated to the next unallocated spatial grid cell, and the second sub-band is allocated to the next unallocated spatial grid cell, until all spatial grid cells have obtained the corresponding sub-bands; The UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band and sub-frequency band. The spectrum data includes low-resolution spectrum data and high-resolution spectrum data. Furthermore, the process by which the UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band includes: During the flight of the UAV, the spatial grid cell in which the UAV is currently located is acquired in real time as the current position cell; the sub-frequency band corresponding to the current position cell is acquired as the current frequency band. A broadband omnidirectional antenna is deployed on the UAV to perform a spectrum scan based on the current frequency band and the monitoring task frequency band, and to obtain spectrum data through the spectrum scan. The spectrum data is used to characterize the distribution of electromagnetic signals in the frequency domain, and is the signal strength value corresponding to each frequency position. Furthermore, the spectral scanning process includes: The scanning mode is set, which includes a first spectrum scanning mode and a second spectrum scanning mode. The first spectrum scanning mode and the second spectrum scanning mode respectively include the corresponding frequency resolution bandwidth, step frequency and observation dwell time. Specifically, frequency resolution bandwidth refers to the parameter used to determine the ability of a spectrum scan to distinguish adjacent frequency components on the frequency axis. It characterizes the frequency resolution scale corresponding to a single spectrum scan. The smaller the frequency resolution bandwidth, the stronger the resolution of the spectrum on the frequency axis; the larger the frequency resolution bandwidth, the coarser the expression of the spectrum on the frequency axis. The step frequency refers to the frequency interval between adjacent frequency positions during spectrum scanning, which represents the sampling step size on the frequency axis. The smaller the step frequency, the denser the sampling points on the frequency axis, the more complete and smooth the spectrum profile will be, and the more fully the narrowband structure will be captured. The larger the step frequency, the sparser the sampling points will be, and it will be easy to skip local details. Observation dwell time refers to the length of time during which an observation is performed at a single frequency position during a spectrum scan. It represents the duration of observation at a single frequency position. The longer the observation dwell time, the more complete the samples collected at that frequency position, and the more stable the spectrum estimation is usually. The shorter the observation dwell time, the faster the measurement speed, but the less information is available from a single point observation. Among them, the frequency resolution bandwidth and step frequency of the first spectrum scanning method are both smaller than those of the second spectrum scanning method, and the observation dwell time of the first spectrum scanning method is greater than that of the second spectrum scanning method. A first spectrum scanning method is performed on the current frequency band of the current location unit, and the obtained spectrum data is used as the high-resolution spectrum data; a second spectrum scanning method is performed on the monitoring task frequency band of the current location unit, and the obtained spectrum data is used as the low-resolution spectrum data. The high-resolution spectrum data obtained by the first spectrum scanning method is more detailed within the local frequency range; although the frequency expression of the second spectrum scanning method is coarser, the coverage of the low-resolution spectrum data is more complete.

[0006] The high-resolution spectrum data is stitched together to construct a spectrum data set; the low-resolution spectrum data is then subjected to spectrum interpolation estimation based on the spectrum data set to generate a spectrum distribution model. Furthermore, the process of stitching together the high-resolution spectrum data to construct a spectrum data set includes: The high-resolution spectrum data is grouped according to sub-frequency bands to obtain high-resolution spectrum data groups divided by sub-frequency bands, and each high-resolution spectrum data group corresponds to a non-repeating sub-frequency band. Specifically, the corresponding sub-bands are extracted from the high-resolution spectrum data, and then the data is further divided according to the sub-bands, so that the high-resolution spectrum data belonging to the same sub-band are organized into the same high-resolution spectrum data group. The high-resolution spectrum data is aligned according to a uniform frequency grid to obtain aligned high-resolution spectrum data. The aligned high-resolution spectrum data groups are stitched together according to frequency order to obtain stitched spectrum data. Establish a spectrum data set to associate and store each high-resolution spectrum data group in the stitched spectrum data with its corresponding spatial grid cell.

[0007] The frequency order specifically refers to the order in which the start and end frequencies of the sub-bands corresponding to each high-resolution spectrum data group are arranged in the monitoring task frequency band, generally from low frequency to high frequency. Specifically, a uniform frequency grid refers to a set of continuously arranged frequency positions that are preset based on accuracy requirements throughout the entire spectrum data processing process, covering the entire monitoring task frequency band, where the interval between each adjacent frequency position is consistent; the uniform frequency grid is used to ensure that high-resolution spectrum data, low-resolution spectrum data, spectrum estimation data, and intensity values ​​in the spectrum model are processed at the same frequency position. The alignment process includes: Read the original frequency positions in the high-resolution spectrum data, calculate the frequency difference between each original frequency position and each frequency position in the unified frequency grid, and map each original frequency position to the frequency position with the smallest frequency difference to obtain the aligned high-resolution spectrum data. The frequency order of each aligned high-resolution spectrum data is obtained, and the aligned high-resolution spectrum data groups are spliced ​​together according to the frequency order to obtain spliced ​​spectrum data. A spectrum data set is then established to store the spliced ​​spectrum data.

[0008] Each aligned high-resolution spectrum data corresponds to a sub-band identifier. Therefore, each aligned high-resolution spectrum data corresponds to a specific frequency range in the monitoring task frequency band. The frequency order is the arrangement order of each aligned high-resolution spectrum data on the corresponding frequency axis of the monitoring task frequency band. Specifically, the sorted and aligned high-resolution spectrum data are read sequentially, and each aligned high-resolution spectrum data is sorted by frequency position and written into the same continuous data sequence, thereby forming a continuous spliced ​​spectrum data. Furthermore, the process of generating a spectral distribution model by performing spectral interpolation estimation on the low-resolution spectral data according to the aforementioned spectral data set includes: The low-resolution spectrum data is aligned according to a uniform frequency grid to obtain aligned low-resolution spectrum data as data to be estimated, and the spatial grid cell corresponding to the data to be estimated is used as the spatial grid cell to be estimated. Extract the high-resolution spectrum data set corresponding to the spatial grid cell adjacent to the spatial grid cell to be estimated from the spectrum data set, and use it as a reference set; Based on the reference set, interpolation estimation is performed on the data to be estimated corresponding to the spatial grid cell to be estimated to obtain the spectrum estimation data corresponding to the spatial grid cell to be estimated; The spectral estimation data corresponding to each spatial grid cell to be estimated is organized according to the spatial grid cell to obtain the spectral distribution model; Furthermore, the interpolation estimation process includes: The low-resolution intensity values ​​of the data to be estimated are combined with the high-resolution intensity values ​​of adjacent spatial grid cells in the reference set at the same frequency position to obtain the spectrum estimation data: ; in Represents the spatial grid cell to be estimated In the uniform frequency grid Frequency position The estimated intensity value at that location is used to represent the spectrum estimation data corresponding to the data to be estimated. Represents the spatial grid cell to be estimated Intensity value at frequency position Low-resolution intensity values ​​at that location; Representative and spatial grid cell to be estimated The adjacent first Each spatial grid cell at frequency position High-resolution intensity values ​​at that location; Representative and spatial grid cell to be estimated The number of adjacent spatial grid cells; Furthermore, the process of organizing the spectral estimation data corresponding to each spatial grid cell to be estimated to obtain the spectral distribution model includes: Using spatial grid cells as row indices and frequency positions corresponding to uniform frequency grids as column indices; constructing a two-dimensional matrix based on the row and column indices; and writing the spectrum estimation data corresponding to the spatial grid cells to be estimated into the two-dimensional matrix to obtain the spectrum distribution model. The essence of constructing a spectrum distribution model is to organize the discrete spectrum estimation data obtained in the previous steps for different spatial grid cells into a structured data model that has both spatial location attributes and frequency location attributes. This allows each spectrum estimation data to be determined by a unique spatial grid cell and a unique frequency location, thereby transforming the originally scattered spectrum estimation data into a joint spatial frequency distribution expression for the entire monitored area.

[0009] A spectral model of the area to be monitored is generated based on the spectral distribution model, and electromagnetic environment anomalies are identified based on the spectral model. Furthermore, the process of generating a spectral model of the region to be monitored based on the aforementioned spectral distribution model includes: The aligned high-resolution spectrum data is input into the spectrum distribution model, and the spectrum estimation data with the same spatial grid cells and the same sub-frequency bands as the aligned high-resolution spectrum data is replaced to obtain the judgment data. The judgment data includes the high-resolution spectrum data in the spectrum distribution model and the remaining spectrum estimation data. The aligned low-resolution spectral data is input into the spectral distribution model, and the spectral distribution model is expanded into a spectral volume model by using a scanning method as a new dimension. The aligned low-resolution spectrum data is used to trace the spectrum data; the first spectrum scanning method in the scanning method is used to point to the judgment data, and the second spectrum scanning method is used to point to the aligned low-resolution spectrum data. The spectral model is a three-dimensional model, with the three dimensions being spatial grid cells, frequency positions corresponding to uniform frequency grids, and scanning methods; the frequency positions corresponding to the same uniform frequency grid contain judgment data and aligned low-resolution spectral data. Furthermore, the process of identifying electromagnetic environment anomalies based on the aforementioned spectral model includes: Electromagnetic environment anomalies are identified sequentially for each spatial grid cell in the order they are arranged. In the spectral model, the spatial grid cell to be identified is taken as the main cell, and the spatial grid cell adjacent to the main cell in the spectral model is taken as the sub-cell; the judgment data corresponding to the main cell is read as the main data, and the judgment data corresponding to the sub-cell is read as the sub-data; Using the frequency position corresponding to a uniform frequency grid as a comparison index, the difference between the intensity value of the main data and the intensity value of the sub-data is calculated under the same comparison index, which is taken as the spectral intensity difference; when the spectral intensity difference is greater than a preset intensity difference threshold, it is considered an electromagnetic environment anomaly, and the spatial grid cell to be identified corresponding to the main data is taken as an abnormal spatial grid cell. The intensity difference threshold is pre-calibrated based on historically collected normal electromagnetic environment data, and represents the maximum normal intensity difference that adjacent spatial grid cells are allowed to occur at the same frequency location.

[0010] An electromagnetic environment anomaly identification system based on UAV, the system includes a grid allocation module, a path acquisition module, a splicing interpolation module and a spectrum recognition module; The grid allocation module is used to allocate sub-frequency bands to different spatial grid cells; The path acquisition module is used to plan and identify paths and acquire the spectrum data of spatial grid cells according to the frequency band of the monitoring task. The splicing interpolation module is used to construct a spectrum dataset and generate a spectrum distribution model; The spectral identification module is used to generate a spectral model of the area to be monitored and to identify electromagnetic environment anomalies based on the spectral model. The output of the grid allocation module is connected to the input of the path acquisition module; the output of the path acquisition module is connected to the input of the splicing interpolation module; and the output of the splicing interpolation module is connected to the input of the spectral recognition module.

[0011] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention divides the area to be monitored into spatial grid units, plans identification paths, and divides the monitoring task frequency band into several sub-frequency bands and assigns them to different spatial grid units. This enables UAVs to simultaneously complete local high-resolution scanning and full-band low-resolution scanning under limited payload capacity and endurance conditions, thereby balancing the ability to acquire spectrum details and regional coverage efficiency, and improving the feasibility and efficiency of electromagnetic environment anomaly identification.

[0012] 2. This invention constructs a spectrum data set by grouping, aligning and stitching the high-resolution spectrum data collected from each spatial grid cell according to sub-frequency bands. Then, it combines the high-resolution spectrum data of adjacent spatial grid cells to interpolate and estimate the low-resolution spectrum data. This can supplement the spectrum information of the spatial grid cell to be estimated without increasing the overall scanning burden, thereby improving the completeness and accuracy of the spectrum distribution model.

[0013] 3. This invention extends the spectrum distribution model into a spectrum volume model that includes spatial grid cells, frequency positions, and scanning methods. It determines anomalies based on the intensity differences between the main cell and adjacent sub-cells at the same frequency position. This not only enhances the ability to identify abnormal signals by utilizing spatial neighborhood relationships, but also enables result tracing and verification by using low-resolution spectrum data, thereby improving the reliability and accuracy of electromagnetic environment anomaly identification. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating an electromagnetic environment anomaly identification method based on unmanned aerial vehicle (UAV) according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example: Figure 1 As shown, the present invention provides a technical solution, a method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs), the method comprising the following steps: The monitoring area is acquired, and the UAV performs electromagnetic environment anomaly identification within the monitoring area; the monitoring area is divided into several spatial grid units, and an identification path is planned based on the spatial grid units; a monitoring task frequency band is set, which includes several sub-frequency bands, and the sub-frequency bands are assigned to different spatial grid units; The process of planning identification paths and assigning the sub-frequency bands to different spatial grid cells includes: The area to be monitored is divided into spatial grid units according to a preset three-dimensional size; the geometric center of each spatial grid unit is used as a path node, and the arrangement order is set. The path nodes are arranged according to the arrangement order to form an identification path. The monitoring task frequency band is divided into several consecutive sub-frequency bands according to frequency order, and the number of sub-frequency bands is less than or equal to the number of spatial grid cells; The sub-frequency bands are sequentially assigned to the corresponding spatial grid cells according to their arrangement order; when the number of sub-frequency bands is less than the number of spatial grid cells, the sub-frequency bands are cyclically assigned to the remaining spatial grid cells until all spatial grid cells have their corresponding sub-frequency bands.

[0017] In this embodiment, the area to be monitored is 300m×200m×60m, and the area is divided into 12 spatial grid units. The three-dimensional dimensions of each spatial grid unit are 50m×50m×30m. There are 6 columns along the length direction, 4 columns along the width direction, and 12 representative grid units in 2 layers along the height direction. This size keeps the flight dwell time and spectrum scanning time within a single grid cell on the same order of magnitude, eliminating the need for frequent turning by the UAV, while also preserving sufficient spatial differentiation between adjacent cells to facilitate subsequent comparison of the intensity differences between the main cell and the sub-cell; In practical applications, the grid side length can be selected by combining the target area scale, the drone's cruising speed, and the time taken to scan a single point. In this embodiment, the identification path adopts a serpentine back-and-forth arrangement, passing through G1, G2, G3, G4, G5, G6, G12, G11, G10, G9, G8, and G7 in sequence, in order to reduce the large-angle back-and-forth at the end of each row, which is more conducive to completing the full grid coverage under the same battery life. In this embodiment, the monitoring task frequency band is set to 88.0MHz to 108.0MHz, with a total bandwidth of 20.0MHz. This band is divided into four consecutive sub-bands from low to high frequency, as follows: B1: 88.0MHz to 93.0MHz; B2: 93.0MHz to 98.0MHz; B3: 98.0MHz to 103.0MHz; B4: 103.0MHz to 108.0MHz; If the total bandwidth to be monitored increases, it can be divided into equal or approximately equal widths according to the upper limit of the bandwidth that a single high-resolution scan can withstand; The sub-bands are assigned cyclically according to the identification path order: G1 corresponds to B1, G2 corresponds to B2, G3 corresponds to B3, G4 corresponds to B4, G5 corresponds to B1 again, G6 corresponds to B2, G12 corresponds to B3, G11 corresponds to B4, G10 corresponds to B1, G9 corresponds to B2, G8 corresponds to B3, and G7 corresponds to B4. This ensures that each sub-band appears repeatedly in space, which facilitates the extraction of reference sets from adjacent spatial grid cells.

[0018] The UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band and sub-frequency band. The spectrum data includes low-resolution spectrum data and high-resolution spectrum data. The process by which the UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band includes: During the flight of the UAV, the spatial grid cell in which the UAV is currently located is acquired in real time as the current position cell; the sub-frequency band corresponding to the current position cell is acquired as the current frequency band. A broadband omnidirectional antenna is deployed on the UAV to perform spectrum scanning based on the current frequency band and the monitoring task frequency band, and to obtain spectrum data through the spectrum scanning. The process of spectrum scanning includes: The scanning mode is set, which includes a first spectrum scanning mode and a second spectrum scanning mode. The scanning mode includes frequency resolution bandwidth, step frequency and observation dwell time. Among them, the frequency resolution bandwidth and step frequency of the first spectrum scanning method are both smaller than those of the second spectrum scanning method, and the observation dwell time of the first spectrum scanning method is greater than that of the second spectrum scanning method. A first spectrum scanning method is performed on the current frequency band of the current location unit, and the obtained spectrum data is used as the high-resolution spectrum data; a second spectrum scanning method is performed on the monitoring task frequency band of the current location unit, and the obtained spectrum data is used as the low-resolution spectrum data.

[0019] In this embodiment, the drone's cruising speed is set to 8 m / s. After reaching each path node, it hovers near the center of the corresponding spatial grid cell for 6 seconds. During the hovering period, it completes the first spectrum scan of the current frequency band and the second spectrum scan of the entire mission frequency band. The frequency resolution bandwidth of the first spectrum scanning method is set to 25kHz, the step frequency is set to 25kHz, and the observation dwell time is set to 20ms, which is used to acquire high-resolution spectrum data. The frequency resolution bandwidth of the second spectrum scanning method is set to 200kHz, the step frequency is set to 200kHz, and the observation dwell time is set to 5ms, which is used to collect low-resolution spectrum data. The first spectrum scanning method has a smaller frequency resolution bandwidth and step frequency, and at the same time uses a longer dwell time, which can more fully present narrowband details; The second spectrum scanning method focuses on rapid full-bandwidth coverage, and therefore uses coarser frequency sampling. If the task emphasizes local details, the step frequency of the first spectrum scanning method can be further reduced to 10kHz to 20kHz; if the task emphasizes large-scale rapid survey, the step frequency of the second spectrum scanning method can be increased to 250kHz to 500kHz.

[0020] The high-resolution spectrum data is stitched together to construct a spectrum data set; the low-resolution spectrum data is then subjected to spectrum interpolation estimation based on the spectrum data set to generate a spectrum distribution model. The process of stitching together the high-resolution spectrum data to construct a spectrum data set includes: The high-resolution spectrum data is grouped according to sub-frequency bands to obtain high-resolution spectrum data groups divided by sub-frequency bands, and each high-resolution spectrum data group corresponds to a non-repeating sub-frequency band. The high-resolution spectrum data group is aligned according to a uniform frequency grid to obtain an aligned high-resolution spectrum data group. The aligned high-resolution spectrum data groups are stitched together according to frequency order to obtain stitched spectrum data. Establish a spectrum data set to associate and store each high-resolution spectrum data group in the stitched spectrum data with its corresponding spatial grid cell.

[0021] In this embodiment, the uniform frequency grid is set to a 25kHz interval, covering 88.0MHz to 108.0MHz; Taking G2 as an example, G2 corresponds to sub-band B2, which is from 93.0MHz to 98.0MHz; After G2 performed the first spectrum scan, the high-resolution intensity values ​​measured at 93.000MHz, 93.025MHz, and 93.050MHz were -81dBm, -79dBm, and -80dBm, respectively. After G2 performed the second spectrum scan, the low-resolution intensity values ​​measured at 93.0MHz, 93.2MHz, and 93.4MHz were -82dBm, -78dBm, and -79dBm, respectively. In this embodiment, the uniform frequency grid is set to be consistent with the high-resolution step frequency to avoid excessive interpolation after stitching. If the high-resolution scan step frequency is adjusted, it is generally safer to keep the uniform frequency grid consistent with the smaller step frequency. During alignment, the original frequency position is mapped to the uniform frequency grid position with the smallest frequency difference. Thus, 93.000MHz, 93.025MHz, and 93.050MHz are directly aligned to the grid points of the same name, and 93.2MHz corresponds to the 93.200MHz grid point.

[0022] The process of generating a spectral distribution model by performing spectral interpolation estimation on low-resolution spectral data based on the aforementioned spectral data set includes: The low-resolution spectrum data is aligned according to a uniform frequency grid to obtain aligned low-resolution spectrum data as data to be estimated, and the spatial grid cell corresponding to the data to be estimated is used as the spatial grid cell to be estimated. Extract the high-resolution spectrum data set corresponding to the spatial grid cell adjacent to the spatial grid cell to be estimated from the spectrum data set, and use it as a reference set; Based on the reference set, interpolation estimation is performed on the data to be estimated corresponding to the spatial grid cell to be estimated to obtain the spectrum estimation data corresponding to the spatial grid cell to be estimated; In this embodiment, G6 is taken as the spatial grid cell to be estimated. G6 corresponds to sub-frequency band B2, and its adjacent spatial grid cells are G2, G5, and G11. The number of adjacent cells is... Choose 3; if the mesh is dense, prioritize adjacent cells that share edges or faces; Suppose that at 93.200MHz, the low-resolution intensity value L6 of G6 is -77dBm, and the high-resolution intensity values ​​of G2, G5, and G11 at the same frequency are -76dBm, -78dBm, and -77dBm, respectively. Then... =6, =3 Substitute into the combination calculation formula: ; get =-77dBm; The spectral estimation data corresponding to each spatial grid cell to be estimated are organized to obtain the spectral distribution model; The process of organizing the spectral estimation data corresponding to each spatial grid cell to be estimated to obtain the spectral distribution model includes: Using spatial grid cells as row indices and frequency positions corresponding to uniform frequency grids as column indices, a two-dimensional matrix is ​​constructed based on the row and column indices. The spectrum estimation data corresponding to the spatial grid cells to be estimated is written into the two-dimensional matrix to obtain the spectrum distribution model.

[0023] In this embodiment, each spatial grid cell is used as a row index, and the uniform frequency grid position from 88.0MHz to 108.0MHz is used as a column index to construct a spectrum distribution model; A spectral model of the area to be monitored is generated based on the spectral distribution model, and electromagnetic environment anomalies are identified based on the spectral model. The process of generating a spectral model of the region to be monitored based on the aforementioned spectral distribution model includes: The aligned high-resolution spectrum data is input into the spectrum distribution model, and the spectrum estimation data with the same spatial grid cells and the same sub-frequency bands as the aligned high-resolution spectrum data is replaced to obtain the judgment data. The judgment data includes the high-resolution spectrum data in the spectrum distribution model and the remaining spectrum estimation data. The aligned low-resolution spectral data is input into the spectral distribution model, and the spectral distribution model is expanded into a spectral volume model by using a scanning method as a new dimension. In this embodiment, the obtained aligned high-resolution spectrum data is backfilled into the corresponding spatial grid cell and the corresponding sub-band position to replace the original spectrum estimation data, thereby forming judgment data; At the same time, the aligned low-resolution spectral data is written as another scanning dimension to obtain the spectral body model; By employing three dimensions—spatial grid cells, frequency location, and scanning method—it can save both fine-grained data for judgment and low-resolution full-bandwidth data for traceability and verification. The process of identifying electromagnetic environment anomalies based on the aforementioned spectral model includes: Electromagnetic environment anomalies are identified sequentially for each spatial grid cell in the order they are arranged. In the spectral model, the spatial grid cell to be identified is taken as the main cell, and the spatial grid cell adjacent to the main cell in the spectral model is taken as the sub-cell; the judgment data corresponding to the main cell is read as the main data, and the judgment data corresponding to the sub-cell is read as the sub-data; Using the frequency position corresponding to a uniform frequency grid as a comparison index, the difference between the intensity value of the main data and the intensity value of the sub-data is calculated under the same comparison index, which is taken as the spectral intensity difference. When the spectral intensity difference is greater than a preset intensity difference threshold, it is considered an electromagnetic environment anomaly, and the spatial grid cell to be identified corresponding to the main data is taken as an abnormal spatial grid cell.

[0024] In this embodiment, G6 is used as the main unit to be identified, and G2, G5, and G11 are used as sub-units. At 95.400MHz, the judgment data intensity value of G6 is -61dBm, and the judgment data intensity values ​​of G2, G5, and G11 are -74dBm, -72dBm, and -73dBm, respectively. Therefore, the spectral intensity differences between the main unit and each subunit are 13dB, 11dB, and 12dB, respectively. In engineering, the distribution of differences between adjacent units can usually be obtained from historical normal electromagnetic environment data. The amplitude corresponding to the higher quantile is then taken as the threshold starting point and corrected by actual measurement. In this embodiment, the intensity difference threshold is set to 8dB. This value can separate normal propagation fluctuations from significant abnormal changes, and will not miss local abnormal emissions due to an excessively high threshold. In this embodiment, if there is at least one adjacent cell with a difference greater than 8dB at 95.400MHz, then G6 is an abnormal spatial grid cell, and if multiple consecutive frequency positions exceed the threshold, then the anomaly determination is more reliable.

[0025] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs), characterized in that: The method includes the following steps: The monitoring area is acquired, and the UAV performs electromagnetic environment anomaly identification within the monitoring area; the monitoring area is divided into several spatial grid units, and an identification path is planned based on the spatial grid units; a monitoring task frequency band is set, which includes several sub-frequency bands, and the sub-frequency bands are assigned to different spatial grid units; The UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band and sub-frequency band. The spectrum data includes low-resolution spectrum data and high-resolution spectrum data. The high-resolution spectrum data is stitched together to construct a spectrum data set; the low-resolution spectrum data is then subjected to spectrum interpolation estimation based on the spectrum data set to generate a spectrum distribution model. A spectral model of the area to be monitored is generated based on the spectral distribution model, and electromagnetic environment anomalies are identified based on the spectral model.

2. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The process of planning identification paths and assigning the sub-frequency bands to different spatial grid cells includes: The area to be monitored is divided into spatial grid units according to a preset three-dimensional size; the geometric center of each spatial grid unit is used as a path node, and the arrangement order is set. The path nodes are arranged according to the arrangement order to form an identification path. The monitoring task frequency band is divided into several consecutive sub-frequency bands according to frequency order, and the number of sub-frequency bands is less than or equal to the number of spatial grid cells; The sub-frequency bands are sequentially assigned to the corresponding spatial grid cells according to their arrangement order; when the number of sub-frequency bands is less than the number of spatial grid cells, the sub-frequency bands are cyclically assigned to the remaining spatial grid cells until all spatial grid cells have their corresponding sub-frequency bands.

3. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 2, characterized in that, The process by which the UAV flies along the identified path and acquires the spectrum data of each spatial grid cell according to the monitoring task frequency band includes: During the flight of the UAV, the spatial grid cell in which the UAV is currently located is acquired in real time as the current position cell; the sub-frequency band corresponding to the current position cell is acquired as the current frequency band. A broadband omnidirectional antenna is deployed on the UAV to perform spectrum scanning based on the current frequency band and the monitoring task frequency band, and to obtain spectrum data through the spectrum scanning.

4. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, The process of spectrum scanning includes: The scanning mode is set, which includes a first spectrum scanning mode and a second spectrum scanning mode. The first spectrum scanning mode and the second spectrum scanning mode respectively include the corresponding frequency resolution bandwidth, step frequency and observation dwell time. Among them, the frequency resolution bandwidth and step frequency of the first spectrum scanning method are both smaller than those of the second spectrum scanning method, and the observation dwell time of the first spectrum scanning method is greater than that of the second spectrum scanning method. A first spectrum scanning method is performed on the current frequency band of the current location unit, and the obtained spectrum data is used as the high-resolution spectrum data; a second spectrum scanning method is performed on the monitoring task frequency band of the current location unit, and the obtained spectrum data is used as the low-resolution spectrum data.

5. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The process of stitching together the high-resolution spectrum data to construct a spectrum data set includes: The high-resolution spectrum data is grouped according to sub-frequency bands to obtain high-resolution spectrum data groups divided by sub-frequency bands, and each high-resolution spectrum data group corresponds to a non-repeating sub-frequency band. The high-resolution spectrum data group is aligned according to a uniform frequency grid to obtain an aligned high-resolution spectrum data group. The aligned high-resolution spectrum data groups are stitched together according to frequency order to obtain stitched spectrum data. Establish a spectrum data set to associate and store each high-resolution spectrum data group in the stitched spectrum data with its corresponding spatial grid cell.

6. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 5, characterized in that, The process of generating a spectral distribution model by performing spectral interpolation estimation on low-resolution spectral data based on the aforementioned spectral data set includes: The low-resolution spectrum data is aligned according to a uniform frequency grid to obtain aligned low-resolution spectrum data as data to be estimated, and the spatial grid cell corresponding to the data to be estimated is used as the spatial grid cell to be estimated. Extract the high-resolution spectrum data set corresponding to the spatial grid cell adjacent to the spatial grid cell to be estimated from the spectrum data set, and use it as a reference set; Based on the reference set, interpolation estimation is performed on the data to be estimated corresponding to the spatial grid cell to be estimated to obtain the spectrum estimation data corresponding to the spatial grid cell to be estimated; The spectral estimation data corresponding to each spatial grid cell to be estimated are organized to obtain the spectral distribution model.

7. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 6, characterized in that, The process of organizing the spectrum estimation data corresponding to each spatial grid cell to be estimated to obtain the spectrum distribution model includes: Using spatial grid cells as row indices and frequency positions corresponding to uniform frequency grids as column indices, a two-dimensional matrix is ​​constructed based on the row and column indices. The spectrum estimation data corresponding to the spatial grid cells to be estimated is written into the two-dimensional matrix to obtain the spectrum distribution model.

8. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 6, characterized in that, The process of generating a spectral model of the region to be monitored based on the aforementioned spectral distribution model includes: The aligned high-resolution spectrum data is input into the spectrum distribution model, and the spectrum estimation data with the same spatial grid cells and the same sub-frequency bands as the aligned high-resolution spectrum data is replaced to obtain the judgment data. The judgment data includes the high-resolution spectrum data in the spectrum distribution model and the remaining spectrum estimation data. The aligned low-resolution spectral data is input into the spectral distribution model, and the spectral distribution model is expanded into a spectral volume model by using a scanning method as a new dimension.

9. The method for identifying electromagnetic environment anomalies based on unmanned aerial vehicles (UAVs) according to claim 8, characterized in that, The process of identifying electromagnetic environment anomalies based on the aforementioned spectral model includes: Electromagnetic environment anomalies are identified sequentially for each spatial grid cell in the order they are arranged. In the spectral model, the spatial grid cell to be identified is taken as the main cell, and the spatial grid cell adjacent to the main cell in the spectral model is taken as the sub-cell; the judgment data corresponding to the main cell is read as the main data, and the judgment data corresponding to the sub-cell is read as the sub-data; Using the frequency position corresponding to a uniform frequency grid as a comparison index, the difference between the intensity value of the main data and the intensity value of the sub-data is calculated under the same comparison index, which is taken as the spectral intensity difference. When the spectral intensity difference is greater than a preset intensity difference threshold, it is considered an electromagnetic environment anomaly, and the spatial grid cell to be identified corresponding to the main data is taken as an abnormal spatial grid cell.

10. An unmanned aerial vehicle (UAV)-based electromagnetic environment anomaly identification system, applied to the UAV-based electromagnetic environment anomaly identification method according to any one of claims 1-9, characterized in that, The system includes a grid allocation module, a path acquisition module, a splicing interpolation module, and a spectral recognition module; The grid allocation module is used to allocate sub-frequency bands to different spatial grid cells; The path acquisition module is used to plan and identify paths and acquire the spectrum data of spatial grid cells according to the frequency band of the monitoring task. The splicing interpolation module is used to construct a spectrum dataset and generate a spectrum distribution model; The spectral identification module is used to generate a spectral model of the area to be monitored and to identify electromagnetic environment anomalies based on the spectral model. The output of the grid allocation module is connected to the input of the path acquisition module; the output of the path acquisition module is connected to the input of the splicing interpolation module; and the output of the splicing interpolation module is connected to the input of the spectral recognition module.

Images