Low-altitude UAV intelligent early warning system and method based on multi-source data fusion

Abstract

The present invention relates to the field of drone early warning technology, and in particular to a low-altitude drone intelligent early warning system and method based on multi-source data fusion. The system comprises: an acquisition module, an extraction module, a determination module, a type determination module, a level determination module, an adjustment module, and an early warning module. The present invention achieves accurate identification and intelligent early warning of ice adhesion disturbances during low-altitude drone flight in foggy weather by fusing multi-source data. By collecting parameters such as water vapor concentration, image information, and wake velocity, the flight environment and aircraft state can be comprehensively reflected, enhancing the system's sensitivity to interference conditions. Using characteristic parameters such as particle radius, distribution density, and ice area and thickness extracted from images, a multi-dimensional judgment basis for interference identification and graded judgment is established, effectively solving the problem of low early warning accuracy when dealing with complex flight interference caused by a single data source or static prediction model.

Description

Technical Field

[0001] The present invention relates to the field of UAV early warning technology, and in particular to a low-altitude UAV intelligent early warning system and method based on multi-source data fusion. Background Art

[0002] With the continuous development of drone technology, the use of drones in low-altitude flight is gradually increasing, especially in areas such as surveillance, transportation, agriculture, and exploration. However, drones face numerous safety risks during low-altitude flight, especially in adverse weather conditions and complex environments, where flight stability and safety are seriously threatened. Natural factors such as fog, frost, and ice, as well as complex geographical and meteorological conditions, can cause drones to malfunction or deviate from their intended trajectory, creating unpredictable safety hazards. Therefore, intelligent early warning and monitoring of low-altitude drones is particularly important. To effectively ensure drone flight safety, how to accurately monitor the drone's flight status, environmental factors, and potential interference in real time, and issue timely warnings, has become a key issue in the current development of drone technology.

[0003] Patent document with publication number CN113867391A discloses a low-altitude safety warning and monitoring method and system for drones based on digital twins. The method includes: step 1, real-time collection of drone operation-related data, including drone operation trajectory and operation performance data, drone operation geographical environment data, drone operation meteorological environment actual and spatial position information data, and drone operation restricted area data; step 2, preprocessing the real-time collected drone operation-related data; step 3, determining the drone operation risk type based on the preprocessed drone operation-related data; step 4, selecting a pre-trained neural network model corresponding to the drone operation risk type determined in step 3, inputting the preprocessed drone operation-related data into the neural network model, predicting the probability of risk occurrence, and obtaining a risk prediction result; step 5, classifying the drone operation low-altitude safety risk according to the risk prediction result, and completing the risk level assessment; step 6, issuing a risk warning based on the risk level assessment result.

[0004] It can be seen that the digital twin-based drone low-altitude safety warning and monitoring method has the following problems: this method mainly focuses on the digital twin model and cannot dynamically adapt to the complex flight environment. When encountering specific interference conditions, it will cause the warning system to respond lag or lack accuracy; this method relies on predicted static data and cannot adjust parameters in real time according to environmental changes, which will lead to large errors in the prediction results and affect the safety monitoring and warning of drones; this method relies too much on a single neural network model, resulting in ignoring the combination of physical models and environmental data, and will not be able to effectively capture various changes during the flight process. Summary of the Invention

[0005] To this end, the present invention provides a low-altitude UAV intelligent early warning system and method based on multi-source data fusion, which is used to overcome the problem of low early warning accuracy in the existing technology when dealing with complex flight interference due to a single data source or static prediction model through real-time data collection and dynamic threshold adjustment.

[0006] To achieve the above objectives, the present invention provides, on the one hand, a low-altitude UAV intelligent early warning system based on multi-source data fusion, comprising:

[0007] The acquisition module is used to collect real-time water vapor concentration, images at the wings, and wake velocity at the fuselage within a range of a preset acquisition length centered on the fuselage during low-altitude UAV flight in foggy weather;

[0008] an extraction module connected to the acquisition module, for extracting in real time the particle radius, distribution density, ice area, and ice thickness of the water vapor particles in the image;

[0009] a determination module connected to the acquisition module, configured to determine whether the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and form an interference determination result;

[0010] a type determination module, connected to the determination module, the acquisition module, and the extraction module, respectively, for determining that the interference type of the disturbed state is an ice layer attachment disturbance type based on the interference determination result, the particle radius, the ice layer area, and the distribution density;

[0011] a level determination module, connected to the type determination module, the acquisition module, and the extraction module, respectively, for determining whether the interference level of the disturbed state is a severe level according to the ice attachment disturbance type, the ice thickness, and the wake velocity;

[0012] an adjustment module connected to the level determination module, configured to adjust the preset concentration fluctuation threshold or the preset acquisition length according to all the severity levels within a preset adjustment time period;

[0013] An early warning module is connected to the level determination module and is used to issue an alarm for the severity level that is re-determined based on the adjusted preset concentration fluctuation threshold or the preset collection length.

[0014] Furthermore, the type determination module includes:

[0015] a radius distribution calculation unit, for calculating the standard deviation of all the particle radii to obtain a radius distribution;

[0016] a recording unit connected to the radius distribution calculation unit, configured to record a timestamp when the radius distribution degree is greater than a preset standard distribution degree within a preset first determined time period, to form a recorded time period;

[0017] A type determination unit is connected to the recording unit and is used to determine that the interference type of the disturbed state is the ice layer attachment disturbance type according to the ice layer area and the distribution density when the ratio of the recording time to the preset first determination time is greater than the preset standard type ratio.

[0018] Furthermore, the type determination unit includes:

[0019] an area fluctuation calculation subunit, configured to calculate a standard deviation of all ice layer areas from an initial moment to each moment within a preset second determined time period, to obtain a plurality of ice layer area fluctuation values;

[0020] a density fluctuation calculation subunit, configured to calculate a standard deviation of all the distribution densities from an initial moment to each moment within the preset second determined time period, to obtain a plurality of distribution density fluctuation values;

[0021] a drawing subunit, connected to the area fluctuation calculation subunit and the density fluctuation calculation subunit, respectively, for drawing a change curve of the ice layer area fluctuation value within the preset second determined time period to obtain an area fluctuation curve, and drawing a change curve of the distribution density fluctuation value within the preset second determined time period to obtain a density fluctuation curve;

[0022] a consistency calculation subunit, connected to the drawing subunit, for calculating the cosine similarity of the area fluctuation curve and the density fluctuation curve to obtain a change consistency;

[0023] A type determination subunit is connected to the consistency calculation subunit, and is used to determine that the interference type of the disturbed state is the ice layer adhesion disturbance type when the change consistency is greater than a preset standard consistency.

[0024] Furthermore, the level determination module includes:

[0025] a thickness change calculation unit, configured to calculate the difference between the ice layer thickness at each moment within a preset thickness determination time period and the ice layer thickness at a previous moment when the ice layer adhesion disturbance type is determined, to obtain a plurality of thickness change rates;

[0026] a thickness change fluctuation calculation unit, configured to calculate a thickness change fluctuation value based on all of the thickness change rates;

[0027] A level determination unit is connected to the change fluctuation calculation unit and is used to determine that the interference level of the disturbed state is the severe level according to the thickness change fluctuation value and the wake velocity.

[0028] Furthermore, the thickness change fluctuation calculation unit includes:

[0029] A quantity acquisition subunit is used to acquire the number of thickness change rates that are positive numbers within the preset thickness determination time period, and obtain a number of growth quantities;

[0030] The change fluctuation calculation subunit is connected to the quantity acquisition subunit and is used to calculate the standard deviation of the absolute values ​​of all the thickness change rates when the growth quantity is greater than a preset standard quantity to obtain the thickness change fluctuation value.

[0031] Furthermore, the level determination unit includes:

[0032] a wake fluctuation calculation subunit, configured to calculate a standard deviation of the wake velocity within a preset level determination time period to obtain a wake fluctuation value when the thickness variation fluctuation value is greater than a preset thickness variation fluctuation threshold;

[0033] The level determination subunit is connected to the wake fluctuation calculation subunit and is used to determine that the interference level of the disturbed state is the severe level when the wake fluctuation value is greater than a preset wake fluctuation threshold.

[0034] Furthermore, the adjustment module includes:

[0035] a marking unit, configured to mark once the severity level is determined within the preset adjustment time period, and obtain a plurality of severity marks with timestamps;

[0036] a severity distribution calculation unit connected to the marking unit and configured to calculate a standard deviation of the interval between two severity marks with the most recent timestamps to obtain a severity distribution degree;

[0037] an adjustment unit connected to the severity distribution calculation unit, and configured to reduce the preset concentration fluctuation threshold value according to the relative deviation between the severity distribution degree and the maximum value of the preset severity distribution degree range and a preset adjustment coefficient when the severity distribution degree is greater than the maximum value of the preset severity distribution degree range, or to increase the preset acquisition length according to the relative deviation between the minimum value of the preset severity distribution degree range and the severity distribution degree and the preset adjustment coefficient when the severity distribution degree is less than the minimum value of the preset severity distribution degree range.

[0038] Furthermore, the determination module includes:

[0039] a duration acquisition unit, configured to acquire a duration during which the water vapor concentration is continuously greater than a preset standard concentration within a preset determination duration, and obtain a plurality of durations;

[0040] a proportion calculation unit connected to the duration acquisition unit, for calculating the ratio of the sum of all the durations to the preset determination duration to obtain a duration proportion;

[0041] A determination unit is connected to the proportion calculation unit and is used to determine whether the low-altitude UAV is in the interfered state according to the duration proportion and the water vapor concentration, thereby forming the interference determination result.

[0042] Furthermore, the determination unit includes:

[0043] a concentration fluctuation calculation subunit, configured to calculate the standard deviation of all the water vapor concentrations within the preset determination time period when the duration ratio is greater than a preset standard ratio, to obtain a concentration fluctuation value;

[0044] The determination subunit is connected to the concentration fluctuation calculation subunit and is used to determine that the low-altitude UAV is in the interfered state when the concentration fluctuation value is greater than a preset concentration fluctuation threshold, thereby forming the interference determination result.

[0045] On the other hand, the present invention also provides a low-altitude UAV intelligent early warning method based on multi-source data fusion, comprising:

[0046] Real-time collection of water vapor concentration, wing images, and fuselage wake velocity within a range of a preset collection length centered on the fuselage during low-altitude UAV flight in foggy weather;

[0047] extracting in real time the particle radius, distribution density, ice layer area, and ice layer thickness of the water vapor particles in the image;

[0048] Determining that the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and forming an interference determination result;

[0049] Determining, according to the interference determination result, the particle radius, the ice layer area, and the distribution density, that the interference type of the disturbed state is an ice layer adhesion disturbance type;

[0050] Determining, according to the ice attachment disturbance type, the ice thickness, and the wake velocity, that the interference level of the disturbed state is a severe level;

[0051] Adjusting the preset concentration fluctuation threshold or the preset collection length according to all the severity levels within a preset adjustment time;

[0052] An alarm is issued for the severity level that is re-determined based on the adjusted preset concentration fluctuation threshold or the preset sampling length.

[0053] Compared with the existing technology, the beneficial effect of the present invention lies in that it can achieve accurate identification and intelligent warning of ice adhesion disturbances during low-altitude UAV flights in foggy weather by fusing multi-source data, with the advantages of strong real-time performance, accurate identification, and flexible response. By collecting parameters such as water vapor concentration, image information, and wake velocity, the flight environment and aircraft status can be fully reflected, and the system's sensitivity to interference conditions can be enhanced. By using characteristic parameters such as particle radius, distribution density, and ice layer area and thickness extracted from images, the formation mechanism of fog particles can be reflected, and the impact of ice on flight can be quantified, establishing a multi-dimensional judgment basis for interference identification and classification. By coupling the analysis of ice layer thickness and wake velocity, the system can more accurately determine the disturbance level, improving the physical logic rationality and scientific nature of interference judgment. In addition, through dynamic statistics of severe events, the system can adaptively optimize the collection radius and concentration threshold, realize dynamic correction and precise matching of threshold parameters, improve the robustness and environmental adaptability of the early warning strategy, and thus more effectively ensure the flight safety of low-altitude UAVs in complex weather conditions, and effectively solve the problem of low early warning accuracy when dealing with complex flight interference due to a single data source or static prediction model.

[0054] Furthermore, by taking the standard deviation of the particle radius of water vapor particles as an important indicator reflecting the stability of the water vapor particle distribution in a foggy environment, combined with the preset first determination time and standard distribution degree, the preliminary identification of abnormal water vapor state is achieved; then by recording the duration of the anomaly and comparing it with the preset first determination time, it is ensured that the identified disturbance is continuous and representative. The ice layer area reflects the scale of the ice area, and the distribution density reflects the intensity of particle aggregation. The three together construct a logical closed loop for interference type judgment, which has strong stability and interpretability, can effectively distinguish ice layer attachment disturbances from other types of interference, and significantly improve the accuracy and practicality of system warnings.

[0055] Furthermore, the fluctuations in ice area and distribution density reflect the aggregation and growth of particles on the aircraft surface during ice adhesion, and their time-dependent fluctuations directly impact flight safety. Plotting the two curves separately and matching them using cosine similarity effectively reveals the consistency of their changes. A high degree of consistency indicates synchronization between particle deposition and ice formation, making an ice adhesion disturbance more likely. Therefore, variation consistency, a core parameter for determining ice disturbance types, depends on the coupling strength of area and density fluctuations, serving as a dynamic disturbance identification indicator. By jointly analyzing the dynamic fluctuation patterns of ice area and particle density and introducing cosine similarity to quantitatively compare their changing trends, the method upgrades disturbance type determination from a single static threshold to multi-dimensional dynamic trend matching, effectively improving the accuracy and timeliness of ice adhesion disturbance identification.

[0056] Furthermore, the thickness change rate reflects the instantaneous trend of ice attachment growth, the thickness change fluctuation value further reveals its stability and continuity, and the wake velocity reflects the response characteristics of the flight state to disturbances. The three work together to form the core judgment basis for the determination of disturbance level, ensuring that the judgment result not only focuses on the change of structural load, but also takes into account the response of the flight environment. It is a key parameter system for achieving high-reliability graded warnings. By analyzing the rate of change of ice thickness over time and its fluctuations, combined with flight dynamic parameters such as wake velocity, it is possible to more accurately assess the level of impact of ice attachment disturbances on drones, effectively avoiding the situation where the risk is underestimated due to a small increase in thickness but severe disturbance, thereby improving the scientific nature and actual protection capabilities of the warning level judgment and enhancing the system's intelligent response effect in complex environments.

[0057] Furthermore, the growth rate reflects the persistence of ice growth; only after the ice has grown to a certain magnitude does it become meaningful to analyze its fluctuation characteristics. The standard deviation of the absolute value of the thickness change rate characterizes the fluctuation of growth. The two are linked and progressively judged, jointly improving the accuracy of interference level identification. By initiating the fluctuation value calculation logic only after the continuous growth rate meets certain conditions, it can avoid misjudgments caused by occasional changes, thereby improving the accuracy and stability of the judgment. The introduction of the standard deviation of the thickness change rate quantifies the fluctuation amplitude of the ice growth rate, and combined with the wake velocity, it further determines the disturbance level, effectively improving the drone's ability to respond to icing risks in complex foggy environments, and enhancing the practicality and robustness of the system.

[0058] Furthermore, the preset thickness variation fluctuation threshold is closely correlated with the standard deviation of the wake fluctuation. The former indicates changes in ice adhesion, while the latter reflects the changing trend of the wake during flight. If the ice layer changes too much, the stability of the wake will be affected, and the wake fluctuation value will need to be used to determine whether it has reached a serious interference level to ensure flight safety. By combining the calculation of the wake fluctuation, a more comprehensive assessment of the wake interference factors of the drone during flight can be achieved, especially in the case of wake instability caused by ice adhesion. This method not only accurately captures the details of the interference, but also can promptly identify the level of danger, improving the response speed to potential threats in the flight environment. By using the dual judgment of wake fluctuation and thickness variation fluctuation, the false positive rate can be effectively reduced, and the accuracy and reliability of the early warning system can be improved.

[0059] Furthermore, the severity distribution is used to characterize the time interval fluctuations of severity levels. After comparing it with the maximum and minimum values ​​of the preset severity distribution range, combined with the relative deviation and the preset adjustment coefficient, it drives the adaptive adjustment of the threshold and acquisition length. This automatically optimizes the judgment sensitivity and monitoring range when the frequency of interference events is too high or too low, ensuring that the system is neither too sensitive nor too slow, and achieving a balance between response efficiency and accuracy. By introducing the time distribution characteristics of severity levels and their fluctuations, it is possible to dynamically adjust the system's acquisition and judgment parameters, improve the system's response adaptability when interference patterns change frequently or the interference level tends to be concentrated, avoid misjudgments and missed judgments, and improve the stability, flexibility, and accuracy of the overall early warning system.

[0060] Furthermore, the preset judgment duration determines the total length of the judgment cycle, affecting the time base for the overall interference trend analysis; the preset standard concentration is the concentration threshold for identifying the interference state and is the key condition for triggering recording; and the duration ratio is a combined indicator of the two, which comprehensively evaluates the duration of the interference. The three work together to ensure that the system can not only quickly respond to sudden concentration increases, but also accurately identify persistent interference, thereby improving the stability and accuracy of the overall judgment. By introducing the parameter "duration ratio", it can effectively avoid misjudgments caused by short-term, instantaneous concentration mutations, and more stably reflect the persistence and severity of interference in the environment. This solution combines concentration values ​​and persistence trends to improve the accuracy of judgment and noise resistance. It is especially suitable for complex and changeable low-altitude flight environments, helping the system to quickly identify potential risks and carry out subsequent processing.

[0061] Furthermore, there is a close logical relationship between the preset standard ratio, concentration fluctuation threshold, and standard deviation of water vapor concentration. The standard ratio determines the extent to which sustained changes in water vapor concentration over a given period of time influence the determination of interference status. The concentration fluctuation threshold ensures that interference is only determined when water vapor concentration fluctuates significantly, thereby avoiding misjudgments of interference due to small fluctuations. The standard deviation measures the amplitude of water vapor concentration fluctuations, ensuring accurate determination of interference status and enabling real-time and precise monitoring of environmental interference experienced by low-altitude drones during flight, particularly changes in water vapor concentration. Effective analysis of water vapor concentration fluctuations enables timely detection of anomalies in the flight environment and prevents potential flight risks. Calculating the ratio of concentration fluctuation value to duration helps ensure the actual occurrence of interference, thereby improving the accuracy and reliability of interference determination.

[0062] Furthermore, through the fusion of multi-source data, precise real-time monitoring and dynamic adjustments are achieved, effectively improving the drone's ability to cope with complex flight environments. By adjusting parameters (such as the preset concentration fluctuation threshold and acquisition radius), the early warning system can ensure efficient response in different environments, avoid excessive or insufficient interference judgments, and ensure accurate interference level and timely response. The logical correlation between these parameters lies in the interaction between parameters such as the concentration fluctuation threshold, acquisition radius, and wake velocity. Adjusting these parameters ensures the system's adaptability and reliability in different flight environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the low-altitude UAV intelligent early warning system based on multi-source data fusion in this embodiment;

[0064] Figure 2 A logic decision diagram for determining the type of ice adhesion disturbance by the type determination subunit of this embodiment;

[0065] Figure 3 As shown, it is a decision logic diagram of the severity level determined by the level determination subunit of this embodiment;

[0066] Figure 4 This is a flow chart of the low-altitude UAV intelligent early warning method based on multi-source data fusion in this embodiment. DETAILED DESCRIPTION

[0067] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0068] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0069] On the one hand, see Figure 1 , which is a schematic diagram of the low-altitude UAV intelligent early warning system based on multi-source data fusion in this embodiment;

[0070] This embodiment provides a low-altitude UAV intelligent early warning system based on multi-source data fusion, including:

[0071] The acquisition module is used to collect real-time water vapor concentration, images at the wings, and wake velocity at the fuselage within a range of a preset acquisition length centered on the fuselage during low-altitude UAV flight in foggy weather;

[0072] an extraction module connected to the acquisition module, for extracting in real time the particle radius, distribution density, ice area, and ice thickness of the water vapor particles in the image;

[0073] a determination module connected to the acquisition module, configured to determine whether the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and form an interference determination result;

[0074] a type determination module, connected to the determination module, the acquisition module, and the extraction module, respectively, for determining that the interference type of the disturbed state is an ice layer attachment disturbance type based on the interference determination result, the particle radius, the ice layer area, and the distribution density;

[0075] a level determination module, connected to the type determination module, the acquisition module, and the extraction module, respectively, for determining whether the interference level of the disturbed state is a severe level according to the ice attachment disturbance type, the ice thickness, and the wake velocity;

[0076] an adjustment module connected to the level determination module, configured to adjust the preset concentration fluctuation threshold or the preset acquisition length according to all the severity levels within a preset adjustment time period;

[0077] An early warning module is connected to the level determination module and is used to issue an alarm for the severity level that is re-determined based on the adjusted preset concentration fluctuation threshold or the preset collection length.

[0078] When the judgment result is a serious level, the early warning module will simultaneously issue a multi-modal alarm by emitting a high-frequency sound through the buzzer, flashing the LED light, and a real-time pop-up alarm prompt on the image transmission interface, ensuring that the operator obtains risk information and responds in the first time.

[0079] During system operation, the acquisition module uses infrared water vapor sensors to detect water vapor concentration within a radius of a preset acquisition length centered on the low-altitude drone's flight path, thereby determining the local humidity distribution. A multispectral imaging camera mounted on the wing captures real-time flight images in foggy conditions, offering high fog-penetration capabilities and accurately capturing the distribution of water vapor particles in low-light and low-contrast conditions. A micro-turbulence sensor or laser Doppler velocimeter is deployed at the rear of the fuselage for real-time measurement of wake airflow velocity. The extraction module, based on the acquired images, first extracts water vapor particle regions using image enhancement and segmentation algorithms. It then uses image recognition techniques based on light scattering properties (such as the Mie scattering model) to calculate the particle radius and distribution density. Furthermore, a deep learning semantic segmentation network (such as U-Net) is used to annotate ice regions in the image. The surface area and thickness of the ice layer are then estimated by combining image grayscale information with a depth estimation algorithm, enabling the precise extraction and structured output of key disturbance parameters.

[0080] The preset collection length refers to the spatial radius range when parameter collection is performed with the drone body as the center. It depends on the influence radius of water vapor disturbance in the foggy environment and the effective detection distance of the sensor. It is usually set between 1 meter and 3 meters. In this embodiment, it is set to 2 meters, which can take into account the comprehensiveness of water vapor collection and real-time response capabilities.

[0081] The preset concentration fluctuation threshold refers to the critical value at which the water vapor concentration change rate reaches the triggering interference judgment. It depends on the natural fluctuation range of water vapor concentration during flight and the significant distinction between abnormal disturbances. It is usually set between ±5% and ±15%. In this embodiment, it is set to ±10%, which can effectively identify potential disturbances and reduce the misjudgment rate.

[0082] The preset adjustment time refers to the time window used by the system to cumulatively analyze severe interference events and trigger dynamic parameter adjustment. It depends on the typical distribution characteristics of the UAV flight stability requirements and the duration of interference. It is usually set between 30 seconds and 180 seconds. In this embodiment, it is set to 60 seconds, which can balance the response speed and the stability of interference trend judgment.

[0083] The acquisition module obtains key environmental and flight data of low-altitude drones flying in fog in real time, including water vapor concentration, image information and wake velocity; the extraction module extracts water vapor particle characteristics and ice layer information from the image; the judgment module determines whether there is interference based on changes in water vapor concentration; the type determination module further combines the image and extracted features to determine whether the interference type is ice attachment disturbance; the level determination module determines the interference level based on ice thickness and wake velocity; the adjustment module dynamically optimizes the concentration threshold or acquisition range based on multiple severe levels of interference; and finally, the warning module issues an intelligent warning based on the adjusted parameters and level results.

[0084] By integrating multi-source data, the system accurately identifies and intelligently warns low-altitude UAVs of ice-attached disturbances during foggy flights, offering the advantages of real-time performance, accurate identification, and flexible response. By collecting parameters such as water vapor concentration, image information, and wake velocity, it comprehensively reflects the flight environment and aircraft status, enhancing the system's sensitivity to disturbances. Image-extracted characteristic parameters such as particle radius, distribution density, and ice area and thickness reveal the formation mechanism of fog particles and quantify the impact of ice on flight, establishing a multi-dimensional basis for interference identification and classification. By coupling ice thickness with wake velocity, the system more accurately determines disturbance levels, enhancing the physical logic and scientific rationality of interference assessment. Furthermore, by dynamically analyzing severity events, the system adaptively optimizes preset collection radius and concentration thresholds, enabling dynamic adjustment and precise matching of threshold parameters. This improves the robustness and environmental adaptability of the warning strategy, effectively ensuring the safety of low-altitude UAVs in complex weather conditions and addressing the issue of low warning accuracy for complex flight disturbances caused by single data sources or static prediction models.

[0085] Specifically, the type determination module includes:

[0086] a radius distribution calculation unit, for calculating the standard deviation of all the particle radii to obtain a radius distribution;

[0087] a recording unit connected to the radius distribution calculation unit, configured to record a timestamp when the radius distribution degree is greater than a preset standard distribution degree within a preset first determined time period, to form a recorded time period;

[0088] A type determination unit is connected to the recording unit and is used to determine that the interference type of the disturbed state is the ice layer attachment disturbance type according to the ice layer area and the distribution density when the ratio of the recording time to the preset first determination time is greater than the preset standard type ratio.

[0089] The preset first determination time length is the duration threshold for recording the particle radius distribution, which depends on the minimum time required for the low-altitude UAV to cross different areas in a typical foggy environment. It is usually set between 5 seconds and 30 seconds. In this embodiment, it is set to 15 seconds, which can ensure that the recorded abnormal distribution has time continuity and eliminates short-term fluctuation interference.

[0090] The preset standard distribution degree is the judgment threshold of the standard deviation of the water vapor particle radius, which depends on the normal fluctuation range of the water vapor particle radius in a common foggy environment. It is usually set between 0.5μm and 2μm. In this embodiment, it is set to 1μm. It can determine whether the particle radius distribution is in an abnormal diffusion state and provide a basis for type identification.

[0091] The preset standard type ratio is a threshold value of the ratio of the recording duration to the type determination duration, which depends on the sensitivity of the disturbance's impact on flight safety. It is usually set between 0.4 and 0.8. In this embodiment, it is set to 0.6. This ensures that an interference type is determined only when an abnormal state exists for a larger proportion of the time, thereby improving the rigor of the judgment.

[0092] First, the radius distribution calculation unit calculates the standard deviation of the particle radius of all water vapor particles in the collected image to obtain the radius distribution degree of the particles; when the distribution degree is continuously greater than the set standard distribution degree within a preset first determination time, the recording unit records the time point when the threshold is exceeded and accumulates the records to form a recording time; then, the type determination unit calculates the ratio of the recording time to the preset first determination time. If the ratio exceeds the set standard type ratio, it is further combined with the ice layer area and distribution density parameters extracted from the image to comprehensively determine that the current interference type is an ice layer attachment disturbance type.

[0093] By using the standard deviation of the particle radius of water vapor particles as an important indicator to reflect the stability of the water vapor particle distribution in a foggy environment, combined with the preset first determination time and standard distribution degree, the preliminary identification of abnormal water vapor status is achieved; then by recording the duration of the anomaly and comparing it with the preset first determination time, it is ensured that the identified disturbance is continuous and representative. Further combined with the ice layer area to reflect the scale of the icing area and the distribution density to reflect the intensity of particle aggregation, the three together construct a logical closed loop for interference type judgment, which has strong stability and interpretability, can effectively distinguish ice layer attachment disturbances from other types of interference, and significantly improve the accuracy and practicality of the system warning.

[0094] Please continue reading Figure 2 As shown, it is a logic determination diagram of the type determination subunit of this embodiment for determining the type of ice adhesion disturbance;

[0095] The type determination unit includes:

[0096] an area fluctuation calculation subunit, configured to calculate a standard deviation of all ice layer areas from an initial moment to each moment within a preset second determined time period, to obtain a plurality of ice layer area fluctuation values;

[0097] a density fluctuation calculation subunit, configured to calculate a standard deviation of all the distribution densities from an initial moment to each moment within the preset second determined time period, to obtain a plurality of distribution density fluctuation values;

[0098] a drawing subunit, connected to the area fluctuation calculation subunit and the density fluctuation calculation subunit, respectively, for drawing a change curve of the ice layer area fluctuation value within the preset second determined time period to obtain an area fluctuation curve, and drawing a change curve of the distribution density fluctuation value within the preset second determined time period to obtain a density fluctuation curve;

[0099] a consistency calculation subunit, connected to the drawing subunit, for calculating the cosine similarity of the area fluctuation curve and the density fluctuation curve to obtain a change consistency;

[0100] A type determination subunit is connected to the consistency calculation subunit, and is used to determine that the interference type of the disturbed state is the ice layer adhesion disturbance type when the change consistency is greater than a preset standard consistency.

[0101] To calculate the cosine similarity of the area fluctuation curve and the density fluctuation curve, it is necessary to first vectorize the two curves and then calculate the cosine similarity. The entire calculation process is prior art and will not be described in detail here.

[0102] The preset second determination time length is the time window used by the system to analyze the fluctuation trend of ice layer area and distribution density, which depends on the typical time range of ice layer attachment disturbance formation and is usually set between 5 and 30 seconds. In this embodiment, it is set to 20 seconds, which can ensure that key fluctuation information in the process of disturbance type change is captured.

[0103] The preset standard consistency is the threshold value for measuring the similarity between the changing trends of the ice layer area fluctuation curve and the distribution density fluctuation curve. It depends on the similarity range of the two types of parameter curves when ice layer disturbance occurs in experimental statistics. It is usually set between 0.85 and 0.98. In this embodiment, it is set to 0.9, which can accurately identify ice layer attachment disturbance and avoid confusion with other types of interference.

[0104] The area fluctuation calculation subunit and the density fluctuation calculation subunit calculate the standard deviations of the ice layer area and particle distribution density within the preset second determination time period, respectively, to generate multiple fluctuation values; the drawing subunit then draws these two fluctuation values ​​into an area fluctuation curve and a density fluctuation curve; the consistency calculation subunit then performs cosine similarity calculation on the two curves to obtain the change consistency; finally, the type determination subunit compares the change consistency with the preset standard consistency. If the change consistency is high, the interference type is determined to be an ice layer attachment disturbance type.

[0105] The fluctuations in ice area and distribution density reflect the aggregation and growth of particles on the aircraft surface during ice adhesion, and their time-dependent fluctuations directly impact flight safety. Plotting the two curves separately and matching them using cosine similarity effectively reveals the consistency of their changes. A high degree of consistency indicates synchronization between particle deposition and ice formation, making an ice adhesion disturbance more likely. Therefore, variation consistency, a core parameter for determining ice disturbance types, depends on the coupling strength between area and density fluctuations, serving as a dynamic disturbance identification indicator. By jointly analyzing the dynamic fluctuation patterns of ice area and particle density and introducing cosine similarity to quantitatively compare their changing trends, the identification of disturbance types has been upgraded from a single static threshold to a multi-dimensional dynamic trend matching approach, effectively improving the accuracy and timeliness of ice adhesion disturbance identification.

[0106] Specifically, the level determination module includes:

[0107] a thickness change calculation unit, configured to calculate the difference between the ice layer thickness at each moment within a preset thickness determination time period and the ice layer thickness at a previous moment when the ice layer adhesion disturbance type is determined, to obtain a plurality of thickness change rates;

[0108] a thickness change fluctuation calculation unit, configured to calculate a thickness change fluctuation value based on all of the thickness change rates;

[0109] A level determination unit is connected to the change fluctuation calculation unit and is used to determine that the interference level of the disturbed state is the severe level according to the thickness change fluctuation value and the wake velocity.

[0110] On the premise that the type of ice attachment disturbance has been identified, the thickness change calculation unit first calculates the rate of change of the ice thickness hourly within a preset thickness determination time, obtaining multiple thickness change rate values; then, the thickness change fluctuation calculation unit performs statistical analysis on these rate values ​​and calculates their fluctuation values ​​to reflect the severity of the ice thickness change; finally, the level determination unit combines the thickness change fluctuation values ​​with the wake velocity to comprehensively judge the impact of the disturbance on flight stability and safety, and outputs the severity level of the disturbance state.

[0111] The thickness change rate reflects the instantaneous trend of ice attachment growth, the thickness change fluctuation value further reveals its stability and continuity, and the wake velocity reflects the response characteristics of the flight state to disturbances. The three work together to form the core judgment basis for determining the disturbance level, ensuring that the judgment result not only focuses on the change in structural load, but also takes into account the response of the flight environment. It is a key parameter system for achieving high-reliability graded warnings. By analyzing the rate of change of ice thickness over time and its fluctuations, combined with flight dynamic parameters such as wake velocity, it is possible to more accurately assess the level of impact of ice attachment disturbances on drones, effectively avoiding the situation where the risk is underestimated due to a small increase in thickness but severe disturbance, thereby improving the scientific nature of the warning level judgment and the actual protection capability, and enhancing the system's intelligent response effect in complex environments.

[0112] Specifically, the thickness change fluctuation calculation unit includes:

[0113] A quantity acquisition subunit is used to acquire the number of thickness change rates that are positive numbers within the preset thickness determination time period, and obtain a number of growth quantities;

[0114] The change fluctuation calculation subunit is connected to the quantity acquisition subunit and is used to calculate the standard deviation of the absolute values ​​of all the thickness change rates when the growth quantity is greater than a preset standard quantity to obtain the thickness change fluctuation value.

[0115] The preset standard number is the minimum counting threshold for the thickness change rate to be a positive number, which depends on the need to judge the continuous growth behavior of the ice layer. It is usually set between 5 and 15. In this embodiment, it is set to 10. It can ensure that the fluctuation value analysis is only performed when the ice layer growth has a certain degree of continuity, thereby improving the stability and accuracy of the disturbance level judgment.

[0116] Through the quantity acquisition subunit, the number of moments when the ice layer thickness change rate is positive within the preset thickness determination time is first counted, that is, the number of times the ice layer continues to grow, as the growth quantity; if the growth quantity exceeds the preset standard quantity, the change fluctuation calculation subunit calculates the standard deviation of the absolute value of the ice layer thickness change rate within the period, thereby obtaining the thickness change fluctuation value, which is used to reflect the fluctuation characteristics of the ice layer growth rate and provide a basis for the subsequent refined determination of the interference level.

[0117] The growth rate reflects the persistence of ice growth. Only after the ice has grown to a certain magnitude does it become meaningful to analyze its fluctuation characteristics. The standard deviation of the absolute value of the thickness change rate characterizes the fluctuation of growth. The two are linked and judged progressively, jointly improving the accuracy of interference level identification. By initiating the fluctuation value calculation logic only after the continuous growth rate meets certain conditions, it can avoid misjudgments caused by occasional changes, thereby improving the accuracy and stability of the judgment. The introduction of the standard deviation of the thickness change rate quantifies the fluctuation amplitude of the ice growth rate, and combined with the wake velocity, it further determines the disturbance level, effectively improving the drone's ability to respond to icing risks in complex foggy environments, and enhancing the practicality and robustness of the system.

[0118] Please continue reading Figure 3 As shown, it is a decision logic diagram of the severity level determined by the level determination subunit of this embodiment;

[0119] The level determination unit includes:

[0120] a wake fluctuation calculation subunit, configured to calculate a standard deviation of the wake velocity within a preset level determination time period to obtain a wake fluctuation value when the thickness variation fluctuation value is greater than a preset thickness variation fluctuation threshold;

[0121] The level determination subunit is connected to the wake fluctuation calculation subunit and is used to determine that the interference level of the disturbed state is the severe level when the wake fluctuation value is greater than a preset wake fluctuation threshold.

[0122] In the level determination unit, when the thickness fluctuation value of the ice attachment disturbance is detected to exceed a preset threshold, the wake fluctuation calculation subunit first calculates the standard deviation of the wake velocity within a preset level determination period to obtain a wake fluctuation value. The level determination subunit then compares the calculated wake fluctuation value with a preset wake fluctuation threshold. If the wake fluctuation value exceeds the threshold, the low-altitude UAV is determined to be in a severe interference state and its interference level is determined to be severe.

[0123] The preset thickness change fluctuation threshold is used to determine the range of variation of ice adhesion disturbances. It depends on the sensitivity of ice adhesion disturbances and changes in the flight environment. It is usually set between 0.1 mm and 1.0 mm. In this embodiment, it is set to 0.5 mm, which can effectively distinguish between normal flight and severe disturbances caused by ice adhesion.

[0124] The preset wake fluctuation threshold is used to determine the fluctuation range of the wake velocity. It depends on the aircraft's wake interference sensitivity and the influence of the external environment. It is usually set between 0.2 m / s and 2.0 m / s. In this embodiment, it is set to 1.0 m / s. This can accurately identify the severity of the wake disturbance and promptly determine the interference level.

[0125] The preset thickness variation fluctuation threshold is closely correlated with the standard deviation of the wake fluctuation. The former indicates changes in ice adhesion, while the latter reflects the changing trend of the wake during flight. Excessive ice variation can affect the stability of the wake, necessitating the use of the wake fluctuation value to determine whether severe interference has occurred to ensure flight safety. By combining the calculation of wake fluctuations, a more comprehensive assessment of the factors affecting the UAV's wake interference during flight can be achieved, particularly in the case of wake instability caused by ice adhesion. This method not only accurately captures the details of the interference but also enables timely identification of the level of danger, improving the response speed to potential threats in the flight environment. By using both wake fluctuations and thickness variation as dual assessments, the false positive rate can be effectively reduced, improving the accuracy and reliability of the early warning system.

[0126] Specifically, the adjustment module includes:

[0127] a marking unit, configured to mark once the severity level is determined within the preset adjustment time period, and obtain a plurality of severity marks with timestamps;

[0128] a severity distribution calculation unit connected to the marking unit and configured to calculate a standard deviation of the interval between two severity marks with the most recent timestamps to obtain a severity distribution degree;

[0129] an adjustment unit connected to the severity distribution calculation unit, and configured to reduce the preset concentration fluctuation threshold value according to the relative deviation between the severity distribution degree and the maximum value of the preset severity distribution degree range and a preset adjustment coefficient when the severity distribution degree is greater than the maximum value of the preset severity distribution degree range, or to increase the preset acquisition length according to the relative deviation between the minimum value of the preset severity distribution degree range and the severity distribution degree and the preset adjustment coefficient when the severity distribution degree is less than the minimum value of the preset severity distribution degree range.

[0130] The preset severity distribution range is the range of fluctuation values ​​of the severity level time interval allowed by the system, which depends on the tolerance of the interference response frequency in the actual flight mission and the statistical results of historical flight data. It is usually set between 2 seconds and 8 seconds. In this embodiment, it is set to [3 seconds, 7 seconds]. It can ensure the timeliness of the warning while avoiding misjudgment or overreaction due to occasional abnormal interference frequency.

[0131] The preset adjustment coefficient is a proportional factor used to adjust the acquisition parameters according to the deviation amplitude of the severity distribution. It depends on the adjustment sensitivity and adjustment amplitude control strategy required by the system. It is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.25. It can achieve progressive adaptive adjustment of the concentration fluctuation threshold or acquisition length, ensuring the stability and reliability of the system in a changing environment.

[0132] First, the marking unit performs a timestamp-based severity mark each time a severity level is identified within a preset adjustment time period, forming a time series record of the severity level; then, the severity distribution calculation unit calculates the standard deviation of the interval time between each two most recent severity marks to obtain the severity distribution degree, which reflects the temporal distribution pattern of the severity level; finally, the adjustment unit dynamically adjusts the preset concentration fluctuation threshold or acquisition length based on the deviation between the severity distribution degree and the maximum or minimum value of the preset severity distribution degree range, as well as the preset adjustment coefficient, thereby optimizing the system's perception sensitivity or perception range to disturbances.

[0133] The severity distribution is used to characterize the time interval fluctuations of severity levels. After comparing it with the maximum and minimum values ​​of the preset severity distribution range, combined with the relative deviation and the preset adjustment coefficient, it drives the adaptive adjustment of the threshold and acquisition length. This automatically optimizes the judgment sensitivity and monitoring range when the frequency of interference events is too high or too low, ensuring that the system is neither too sensitive nor too slow, and achieving a balance between response efficiency and accuracy. By introducing the temporal distribution characteristics of severity levels and their fluctuations, it is possible to dynamically adjust the system's acquisition and judgment parameters, improving the system's response adaptability when interference patterns change frequently or the interference level tends to be concentrated, avoiding misjudgments and missed judgments, and improving the stability, flexibility, and accuracy of the overall early warning system.

[0134] Specifically, the determination module includes:

[0135] a duration acquisition unit, configured to acquire a duration during which the water vapor concentration is continuously greater than a preset standard concentration within a preset determination duration, and obtain a plurality of durations;

[0136] a proportion calculation unit connected to the duration acquisition unit, for calculating the ratio of the sum of all the durations to the preset determination duration to obtain a duration proportion;

[0137] A determination unit is connected to the proportion calculation unit and is used to determine whether the low-altitude UAV is in the interfered state according to the duration proportion and the water vapor concentration, thereby forming the interference determination result.

[0138] The preset judgment time is the time range used by the low-altitude UAV to continuously monitor the water vapor concentration during flight. It depends on the duration of the target flight mission and the average duration of the interference event. It is usually set between 10 seconds and 60 seconds. In this embodiment, it is set to 30 seconds, which can ensure that the persistence of water vapor interference is accurately assessed within a reasonable time window.

[0139] The preset standard concentration is a concentration threshold used to determine whether the water vapor concentration has reached an interference level. It depends on the tolerance limit of the UAV's waterproof and anti-coagulation structure and flight safety requirements. It is usually set between 5g / m³ and 15g / m³. In this embodiment, it is set to 10g / m³, which can effectively identify high-concentration moisture areas that pose a risk to flight performance.

[0140] First, the duration acquisition unit continuously monitors the water vapor concentration of low-altitude drones over a preset determination duration, recording periods of time when the concentration is continuously above the preset standard concentration and calculating all qualifying durations. Subsequently, the ratio calculation unit accumulates these durations and calculates the ratio with the entire preset determination duration to determine the duration ratio. Finally, the determination unit uses this duration ratio and the current water vapor concentration value to determine whether the drone is in an interference state and generates an interference determination result.

[0141] The preset judgment duration determines the total length of the judgment cycle, affecting the time base for the overall interference trend analysis; the preset standard concentration is the concentration threshold for identifying the interference state and is the key condition for triggering recording; and the duration ratio is a combined indicator of the two, which comprehensively evaluates the duration of the interference. The three work together to ensure that the system can not only quickly respond to sudden concentration increases, but also accurately identify persistent interference, thereby improving the stability and accuracy of the overall judgment. By introducing the parameter "duration ratio", it can effectively avoid misjudgments caused by short-term, instantaneous concentration mutations, and more stably reflect the persistence and severity of interference in the environment. This solution combines concentration values ​​with persistence trends to improve the accuracy of judgment and noise resistance. It is especially suitable for complex and changeable low-altitude flight environments, helping the system to quickly identify potential risks and carry out subsequent processing.

[0142] Specifically, the determination unit includes:

[0143] a concentration fluctuation calculation subunit, configured to calculate the standard deviation of all the water vapor concentrations within the preset determination time period when the duration ratio is greater than a preset standard ratio, to obtain a concentration fluctuation value;

[0144] The determination subunit is connected to the concentration fluctuation calculation subunit and is used to determine that the low-altitude UAV is in the interfered state when the concentration fluctuation value is greater than a preset concentration fluctuation threshold, thereby forming the interference determination result.

[0145] First, the concentration fluctuation calculation subunit determines the extent of water vapor concentration fluctuations within a preset judgment duration. If the duration exceeds the preset standard, the system calculates the standard deviation of the water vapor concentration to determine the concentration fluctuation value. Next, the judgment subunit compares the concentration fluctuation value with the preset concentration fluctuation threshold to determine whether the low-altitude drone is in an interference state. If the concentration fluctuation value exceeds the preset threshold, the drone is deemed to be in an interference state, resulting in an interference judgment.

[0146] There is a close logical relationship between the preset standard ratio, concentration fluctuation threshold, and standard deviation of water vapor concentration. The standard ratio determines the extent to which sustained changes in water vapor concentration over a given period of time influence the determination of interference status. The concentration fluctuation threshold ensures that interference is only determined when water vapor concentration fluctuates significantly, thereby avoiding misjudgments of interference due to small fluctuations. The standard deviation measures the amplitude of water vapor concentration fluctuations, ensuring accurate determination of interference status. This allows for real-time and precise monitoring of environmental interference experienced by low-altitude drones during flight, particularly changes in water vapor concentration. Effective analysis of water vapor concentration fluctuations enables timely detection of anomalies in the flight environment and prevents potential flight risks. Calculating the ratio of concentration fluctuation value to duration helps ensure that interference actually occurs, thereby improving the accuracy and reliability of interference determination.

[0147] On the other hand, please continue to see Figure 4 As shown, it is a flow chart of the low-altitude UAV intelligent early warning method based on multi-source data fusion in this embodiment;

[0148] This embodiment also provides a low-altitude UAV intelligent early warning method based on multi-source data fusion, including:

[0149] Real-time collection of water vapor concentration, wing images, and fuselage wake velocity within a range of a preset collection length centered on the fuselage during low-altitude UAV flight in foggy weather;

[0150] extracting in real time the particle radius, distribution density, ice layer area, and ice layer thickness of the water vapor particles in the image;

[0151] Determining that the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and forming an interference determination result;

[0152] Determining, according to the interference determination result, the particle radius, the ice layer area, and the distribution density, that the interference type of the disturbed state is an ice layer adhesion disturbance type;

[0153] Determining, according to the ice attachment disturbance type, the ice thickness, and the wake velocity, that the interference level of the disturbed state is a severe level;

[0154] Adjusting the preset concentration fluctuation threshold or the preset collection length according to all the severity levels within a preset adjustment time;

[0155] An alarm is issued for the severity level that is re-determined based on the adjusted preset concentration fluctuation threshold or the preset sampling length.

[0156] By combining multiple real-time data points for comprehensive monitoring and intelligent analysis, the system ensures drone flight safety in adverse weather conditions. This system collects comprehensive data on the drone's flight environment through real-time collection of water vapor concentration, wing images, and wake velocity. Combined with this image data, it extracts parameters such as the radius, distribution density, ice area, and thickness of water vapor particles. Further analysis of water vapor concentration fluctuations determines whether the drone is experiencing interference. The interference level is assessed based on the type of interference (such as ice attachment disturbance type) and the degree of interference (such as ice thickness and wake velocity). If necessary, the preset concentration fluctuation threshold or collection radius is adjusted. Finally, based on the adjusted parameters, the severity level is reassessed and an alarm is issued, providing dynamic early warning of drone flight safety.

[0157] By integrating multi-source data, precise real-time monitoring and dynamic adjustments are achieved, effectively improving the drone's ability to cope with complex flight environments. By adjusting parameters (such as the preset concentration fluctuation threshold and acquisition radius), the early warning system can ensure efficient response in different environments, avoid excessive or insufficient interference judgments, and ensure accurate interference levels and timely responses. The logical correlation between these parameters lies in the interaction between parameters such as the concentration fluctuation threshold, acquisition radius, and wake velocity. Adjusting these parameters ensures the system's adaptability and reliability in different flight environments.

[0158] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A low-altitude UAV intelligent early warning system based on multi-source data fusion, characterized by: include: The acquisition module is used to collect real-time water vapor concentration, images at the wings, and wake velocity at the fuselage within a range of a preset acquisition length centered on the fuselage during low-altitude UAV flight in foggy weather; an extraction module connected to the acquisition module, for extracting in real time the particle radius, distribution density, ice area, and ice thickness of the water vapor particles in the image; a determination module connected to the acquisition module, configured to determine whether the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and form an interference determination result; a type determination module, connected to the determination module, the acquisition module, and the extraction module, respectively, for determining that the interference type of the disturbed state is an ice layer attachment disturbance type based on the interference determination result, the particle radius, the ice layer area, and the distribution density; a level determination module, connected to the type determination module, the acquisition module, and the extraction module, respectively, for determining whether the interference level of the disturbed state is a severe level according to the ice attachment disturbance type, the ice thickness, and the wake velocity; an adjustment module connected to the level determination module, configured to adjust the preset concentration fluctuation threshold or the preset acquisition length according to all the severity levels within a preset adjustment time period; an early warning module connected to the level determination module, configured to issue an alarm for the severity level re-determined based on the adjusted preset concentration fluctuation threshold or the preset collection length; The type determination module includes: a radius distribution calculation unit, for calculating the standard deviation of all the particle radii to obtain a radius distribution; a recording unit connected to the radius distribution calculation unit, configured to record a timestamp when the radius distribution degree is greater than a preset standard distribution degree within a preset first determined time period, to form a recorded time period; A type determination unit is connected to the recording unit and is used to determine that the interference type of the disturbed state is the ice layer attachment disturbance type according to the ice layer area and the distribution density when the ratio of the recording time to the preset first determination time is greater than the preset standard type ratio.

2. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 1 is characterized in that: The type determination unit includes: an area fluctuation calculation subunit, configured to calculate a standard deviation of all ice layer areas from an initial moment to each moment within a preset second determined time period, to obtain a plurality of ice layer area fluctuation values; a density fluctuation calculation subunit, configured to calculate a standard deviation of all the distribution densities from an initial moment to each moment within the preset second determined time period, to obtain a plurality of distribution density fluctuation values; a drawing subunit, connected to the area fluctuation calculation subunit and the density fluctuation calculation subunit, respectively, for drawing a change curve of the ice layer area fluctuation value within the preset second determined time period to obtain an area fluctuation curve, and drawing a change curve of the distribution density fluctuation value within the preset second determined time period to obtain a density fluctuation curve; a consistency calculation subunit, connected to the drawing subunit, for calculating the cosine similarity of the area fluctuation curve and the density fluctuation curve to obtain a change consistency; A type determination subunit is connected to the consistency calculation subunit, and is used to determine that the interference type of the disturbed state is the ice layer adhesion disturbance type when the change consistency is greater than a preset standard consistency.

3. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 2 is characterized in that: The level determination module includes: a thickness change calculation unit, configured to calculate the difference between the ice layer thickness at each moment within a preset thickness determination time period and the ice layer thickness at a previous moment when the ice layer adhesion disturbance type is determined, to obtain a plurality of thickness change rates; a thickness change fluctuation calculation unit, configured to calculate a thickness change fluctuation value based on all of the thickness change rates; A level determination unit is connected to the change fluctuation calculation unit and is used to determine that the interference level of the disturbed state is the severe level according to the thickness change fluctuation value and the wake velocity.

4. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 3 is characterized in that: The thickness change fluctuation calculation unit includes: A quantity acquisition subunit is used to acquire the number of thickness change rates that are positive numbers within the preset thickness determination time period, and obtain a number of growth quantities; The change fluctuation calculation subunit is connected to the quantity acquisition subunit and is used to calculate the standard deviation of the absolute values ​​of all the thickness change rates when the growth quantity is greater than a preset standard quantity to obtain the thickness change fluctuation value.

5. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 4 is characterized in that: The level determination unit includes: a wake fluctuation calculation subunit, configured to calculate a standard deviation of the wake velocity within a preset level determination time period to obtain a wake fluctuation value when the thickness variation fluctuation value is greater than a preset thickness variation fluctuation threshold; The level determination subunit is connected to the wake fluctuation calculation subunit and is used to determine that the interference level of the disturbed state is the severe level when the wake fluctuation value is greater than a preset wake fluctuation threshold.

6. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 5 is characterized in that: The adjustment module includes: a marking unit, configured to mark once the severity level is determined within the preset adjustment time period, and obtain a plurality of severity marks with timestamps; a severity distribution calculation unit connected to the marking unit and configured to calculate a standard deviation of the interval between two severity marks with the most recent timestamps to obtain a severity distribution degree; an adjustment unit connected to the severity distribution calculation unit, and configured to reduce the preset concentration fluctuation threshold value according to the relative deviation between the severity distribution degree and the maximum value of the preset severity distribution degree range and a preset adjustment coefficient when the severity distribution degree is greater than the maximum value of the preset severity distribution degree range, or to increase the preset acquisition length according to the relative deviation between the minimum value of the preset severity distribution degree range and the severity distribution degree and the preset adjustment coefficient when the severity distribution degree is less than the minimum value of the preset severity distribution degree range.

7. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 6 is characterized in that: The determination module includes: a duration acquisition unit, configured to acquire a duration during which the water vapor concentration is continuously greater than a preset standard concentration within a preset determination duration, and obtain a plurality of durations; a proportion calculation unit connected to the duration acquisition unit, for calculating the ratio of the sum of all the durations to the preset determination duration to obtain a duration proportion; A determination unit is connected to the proportion calculation unit and is used to determine whether the low-altitude UAV is in the interfered state according to the duration proportion and the water vapor concentration, thereby forming the interference determination result.

8. The low-altitude UAV intelligent early warning system based on multi-source data fusion according to claim 7 is characterized in that: The determination unit includes: a concentration fluctuation calculation subunit, configured to calculate the standard deviation of all the water vapor concentrations within the preset determination time period when the duration ratio is greater than a preset standard ratio, to obtain a concentration fluctuation value; The determination subunit is connected to the concentration fluctuation calculation subunit and is used to determine that the low-altitude UAV is in the interfered state when the concentration fluctuation value is greater than a preset concentration fluctuation threshold, thereby forming the interference determination result.

9. A low-altitude UAV intelligent early warning method based on multi-source data fusion, applied to the low-altitude UAV intelligent early warning system based on multi-source data fusion according to any one of claims 1 to 8, characterized in that: include: Real-time collection of water vapor concentration, wing images, and fuselage wake velocity within a range of a preset collection length centered on the fuselage during low-altitude UAV flight in foggy weather; extracting in real time the particle radius, distribution density, ice layer area, and ice layer thickness of the water vapor particles in the image; Determining that the low-altitude UAV is in an interfered state based on the water vapor concentration and a preset concentration fluctuation threshold, and forming an interference determination result; Determining, according to the interference determination result, the particle radius, the ice layer area, and the distribution density, that the interference type of the disturbed state is an ice layer adhesion disturbance type; Determining, according to the ice attachment disturbance type, the ice thickness, and the wake velocity, that the interference level of the disturbed state is a severe level; Adjusting the preset concentration fluctuation threshold or the preset collection length according to all the severity levels within a preset adjustment time; An alarm is issued for the severity level that is re-determined based on the adjusted preset concentration fluctuation threshold or the preset sampling length.

Images