Unmanned aerial vehicle low-altitude operation positioning method and system based on multi-modal perception

Abstract

The application relates to the technical field of unmanned aerial vehicle positioning, in particular to a low-altitude operation positioning method and system for unmanned aerial vehicles based on multi-modal sensing. The method first determines a navigation system acceleration and a navigation system speed of the unmanned aerial vehicle; in combination with the navigation system acceleration, the navigation system speed of the unmanned aerial vehicle and the change of three-dimensional data at the current time, initial speed conversion coefficients of the three-dimensional data are determined respectively; point cloud data is classified to obtain point cloud clusters; the volume and type of the point cloud clusters collected at different times are matched to obtain autonomous movement features of each point cloud cluster; in combination with the autonomous movement features and ideal displacement of the unmanned aerial vehicle, an environment correction coefficient at the current time is determined; according to the environment correction coefficient, the initial speed conversion coefficients of the three-dimensional data are adjusted to obtain corrected speed conversion coefficients; according to the corrected speed conversion coefficients, an estimated coordinate position at the next time is determined. The application improves the positioning accuracy of the unmanned aerial vehicle in a poor signal condition.

Description

Technical Field

[0001] This invention relates to the field of UAV positioning technology, specifically to a UAV low-altitude operation positioning method and system based on multimodal perception. Background Technology

[0002] With the advancement of low-altitude airspace opening policies and the development of drone technology, unmanned aerial vehicles (UAVs) are becoming increasingly popular in low-altitude operation scenarios such as agricultural inspection, power line inspection, security patrol, emergency rescue, and disaster assessment. However, low-altitude operation environments often experience satellite signal blockage and dynamic environmental interference, making it difficult to meet the positioning requirements during operations using a single sensor. Therefore, achieving high-altitude, real-time, and interference-resistant UAV positioning has become a key aspect of ensuring operational safety and efficiency.

[0003] For example, in complex low-altitude operating environments such as cities, forests, and mountainous areas, UAV positioning mainly relies on GNSS modules to obtain real-time location information. However, this signal is susceptible to obstruction, multipath effects, and environmental interference, leading to decreased positioning accuracy or signal interruption, seriously threatening UAV operational safety. To ensure high-precision and continuous positioning of UAVs even under abnormal GNSS signal conditions, it is necessary to analyze position information using multimodal data. For example, analysis can be performed by combining IMU and LiDAR data. However, in actual operations, due to the complexity of the low-altitude environment, when adjusting UAV positioning information using LiDAR data to assist IMU data, it may be affected by moving obstacles such as birds, leading to inaccurate adjustments and affecting positioning accuracy under poor satellite signal conditions. Summary of the Invention

[0004] To address the technical problem that adjusting UAV positioning information is affected by moving obstacles, leading to poor positioning accuracy in situations with weak satellite signals, this invention aims to provide a multimodal perception-based UAV low-altitude operation positioning method and system. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of the present invention provide a method for low-altitude operation positioning of unmanned aerial vehicles (UAVs) based on multimodal perception, the method comprising:

[0006] Acquire positioning data of the UAV during low-altitude operations; analyze the periods when the positioning data may exhibit signal anomalies to determine the motion reference period;

[0007] Obtain the navigation system acceleration and UAV navigation system velocity during the motion reference period; combine the navigation system acceleration, the UAV navigation system velocity, and the changes in the current three-dimensional data to determine the initial velocity conversion coefficient of the three-dimensional data respectively;

[0008] Point cloud data is acquired during low-altitude operations and classified to obtain point cloud clusters; the volume and type of point cloud clusters collected at different times are matched to obtain the autonomous movement characteristics of each point cloud cluster; combined with the autonomous movement characteristics and the ideal displacement of the UAV, the environmental correction coefficient of the point cloud data at the current time is determined.

[0009] Based on the environmental correction coefficient, the initial velocity conversion coefficient of the three-dimensional data is adjusted to obtain the corrected velocity conversion coefficient; based on the corrected velocity conversion coefficient, the estimated coordinate position of the UAV at the next moment is determined.

[0010] Furthermore, the analysis of the positioning data during periods when signal anomalies may occur, and the determination of motion reference periods, includes:

[0011] The positioning data includes: the PDOP value of the UAV when performing low-altitude operations, the positioning distance between adjacent time points, and the UAV acceleration data;

[0012] By combining the PDOP value, the positioning distance between adjacent times, and the UAV acceleration data, the satellite signal loss degree of the UAV is determined; when the satellite signal loss degree of the UAV is abnormal, the abnormal time is recorded; the time period between the current time and the previous abnormal time is used as the motion reference time period.

[0013] Furthermore, determining the satellite signal loss rate of the UAV by combining the PDOP value, the positioning distance between adjacent time points, and the UAV acceleration data includes:

[0014] The product of the PDOP value and the positioning distance between adjacent times is used as the numerator, and the UAV acceleration data is used as the denominator. The ratio of the numerator and denominator is used as the satellite signal loss degree of the UAV.

[0015] Furthermore, the method for obtaining the UAV navigation system velocity is as follows: using the point attitude angle at each moment in the motion reference period to construct a DCM matrix, converting the three-axis acceleration of the machine system to the navigation system to obtain the navigation system acceleration; and using the integration method to combine the navigation system acceleration to obtain the UAV navigation system velocity at the current moment.

[0016] Furthermore, the step of determining the initial velocity conversion coefficient of the three-dimensional data by combining the acceleration of the navigation system, the velocity of the UAV navigation system, and the changes in the three-dimensional data at the current moment includes:

[0017] Using data of any dimension in three-dimensional data as the target dimension data;

[0018] Obtain the change value of the target dimension data between the current time and the previous time, and use it as the dimension change value;

[0019] The weights of the navigation system acceleration and the UAV navigation system velocity are determined based on the time length between the current moment and the previous moment. The navigation system acceleration and the UAV navigation system velocity are then weighted and summed to obtain the velocity parameter value.

[0020] Based on the dimensional change value and the velocity parameter value, the initial velocity conversion coefficient of the target dimensional data is determined; wherein, the dimensional change value is positively correlated with the initial velocity conversion coefficient, and the velocity parameter value is negatively correlated with the initial velocity conversion coefficient.

[0021] Furthermore, the matching of the volume and type of point cloud clusters collected at different times to obtain the autonomous movement characteristics of each point cloud cluster includes:

[0022] The matching object for each point cloud cluster is obtained by matching the volume and type of different types of point cloud data collected at different times;

[0023] Obtain the actual displacement distance and direction of motion between the point cloud cluster and the matched object at the current moment;

[0024] Calculate the directional similarity between the direction of motion and the direction of the drone's acceleration at the current moment;

[0025] Based on the current speed of the drone and the ideal displacement distance of the drone at the next moment under the current acceleration conditions;

[0026] The initial displacement features are obtained by comparing the actual displacement distance with the ideal displacement distance; combined with the initial displacement features and direction similarity, the autonomous movement features of the object corresponding to each point cloud cluster are obtained.

[0027] Furthermore, the determination of the environmental correction coefficient for the point cloud data at the current moment, combining the autonomous movement characteristics and the ideal displacement of the UAV, includes:

[0028] The first correction parameter is determined based on the autonomous movement characteristics of the point cloud cluster at each time step;

[0029] The second correction parameter is determined based on the actual displacement distance and the ideal displacement distance of the point cloud cluster;

[0030] By combining the first correction parameter and the second correction parameter, the environmental correction parameter is obtained.

[0031] Further, adjusting the initial velocity conversion coefficient of the three-dimensional data according to the environmental correction coefficient to obtain the corrected velocity conversion coefficient includes:

[0032] Use any dimension of the three-dimensional data as the target dimension data; use any moment within the motion reference time period as the target moment.

[0033] For the target time, the initial velocity conversion coefficient of the target dimension data is weighted by the environmental correction coefficient to obtain the target conversion coefficient at the target time.

[0034] By performing a negative correlation mapping on the satellite signal loss rate at the target time, the conversion adjustment coefficient at the target time is obtained;

[0035] The target conversion coefficient is weighted by using the conversion adjustment coefficient as the weight to obtain the time conversion coefficient; the sum of the time conversion coefficients of all times within the motion reference period is used as the corrected velocity conversion coefficient of the target dimension data.

[0036] Further, determining the estimated coordinate position of the UAV at the next moment based on the corrected velocity conversion coefficient includes:

[0037] Use any dimension of the three-dimensional data as the target dimension data; use any moment within the motion reference time period as the target moment.

[0038] The normalized value of the conversion coefficient of the target dimension data at the current moment is weighted by the UAV navigation system velocity corresponding to the target dimension data to obtain the navigation system correction velocity.

[0039] Based on the navigation system correction velocity, navigation system acceleration, and the time difference between the current moment and the next moment, determine the coordinates of the target dimension data at the next moment;

[0040] The estimated coordinates of the UAV at the next moment are constructed from the coordinates of the 3D data at the next moment.

[0041] Secondly, a low-altitude operation positioning system for unmanned aerial vehicles (UAVs) based on multimodal perception is provided, the system comprising the following modules:

[0042] The time period determination module is used to acquire positioning data of the UAV during low-altitude operations; analyze the time periods when the positioning data may have signal anomalies, and determine the motion reference time period;

[0043] An initial determination module is used to acquire the navigation system acceleration and the UAV navigation system velocity during the motion reference time period; and to determine the initial velocity conversion coefficient of the three-dimensional data by combining the navigation system acceleration, the UAV navigation system velocity, and the changes in the three-dimensional data at the current moment.

[0044] The environmental correction module is used to acquire point cloud data during low-altitude operations, classify the point cloud data to obtain point cloud clusters, match the volume and type of point cloud clusters collected at different times to obtain the autonomous movement characteristics of each point cloud cluster, and combine the autonomous movement characteristics with the ideal displacement of the UAV to determine the environmental correction coefficient of the point cloud data at the current time.

[0045] The positioning and determination module is used to adjust the initial velocity conversion coefficient of the three-dimensional data according to the environmental correction coefficient to obtain the corrected velocity conversion coefficient; and to determine the estimated coordinate position of the UAV at the next moment according to the corrected velocity conversion coefficient.

[0046] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, it implements the various possible implementations of the first aspect.

[0047] Fourthly, embodiments of the present invention provide a computer program product comprising: computer program code, which, when executed on a computer, causes the computer to perform the method described in the first aspect or any possible implementation thereof.

[0048] Fifthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the various possible implementations of the first aspect.

[0049] The embodiments of the present invention have at least the following beneficial effects:

[0050] This invention proposes a multimodal perception-based method and system for UAV low-altitude operation positioning. First, the received satellite signals are analyzed to assess signal anomalies in the current low-altitude operation area, determining a motion reference time period. Based on coordinate change characteristics within this reference time period and monitoring data from the IMU module, an initial velocity conversion coefficient for the UAV is calculated. A point cloud of the current environment is constructed using LiDAR data, and an environmental correction coefficient is calculated based on the point cloud data change characteristics at different times. The initial velocity conversion coefficient is adjusted according to the environmental correction coefficient, and the coordinate movement relationship of the current position is analyzed to obtain a corrected velocity conversion coefficient. Based on the corrected velocity conversion coefficient, the estimated position coordinates for the next moment are calculated, achieving relatively accurate UAV positioning. This method reduces interference from accumulated errors in the IMU unit and minimizes the impact of autonomously moving objects on the analysis process in low-altitude environments. It can obtain relatively accurate UAV position estimates even with weak GNSS module signals, providing a foundation for subsequent UAV flight control, route planning, obstacle avoidance, operation point positioning, and target tracking operations. Attached Figure Description

[0051] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of a method for low-altitude UAV positioning based on multimodal perception, provided in an embodiment of the present invention.

[0053] Figure 2 This is a schematic diagram of UAV attitude angle data provided in one embodiment of the present invention;

[0054] Figure 3 This is a system block diagram of a UAV low-altitude operation positioning system based on multimodal perception, provided in one embodiment of the present invention.

[0055] Figure 4 This is a schematic diagram of the structure of a computer device provided in one embodiment of the present invention. Detailed Implementation

[0056] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the UAV low-altitude operation positioning method and system based on multimodal perception proposed in this invention.

[0057] In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0058] In the description of the embodiments of the present invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of the present invention, "multiple" means two or more.

[0059] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0061] The embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided by the embodiments of the present invention are also applicable to similar technical problems.

[0062] The following description, in conjunction with the accompanying drawings, details the specific scheme of the UAV low-altitude operation positioning method and system based on multimodal perception provided by this invention.

[0063] Example 1:

[0064] Please see Figure 1 The diagram illustrates a flowchart of a multimodal perception-based UAV low-altitude operation positioning method according to an embodiment of the present invention. The method includes the following steps:

[0065] Step S100: Acquire positioning data of the UAV during low-altitude operations; analyze the periods when the positioning data may show signal anomalies, and determine the motion reference period.

[0066] To achieve flight positioning for UAVs operating at low altitudes, using a single sensor is susceptible to interference from specific environmental factors, which can lead to failure or a sharp drop in accuracy. These factors include satellite signal obstruction, loss of visual texture, and LiDAR encountering specific materials. Therefore, it is necessary to install sensors on the UAV to collect multimodal data during low-altitude operations.

[0067] The current low-altitude target location area is obtained, and the A* algorithm is used for path planning to generate the initial UAV operational flight path. It should be noted that the process of obtaining the UAV path using the A* algorithm is an existing step and will not be elaborated upon here.

[0068] The system is equipped with a variety of sensors, including GNSS (Global Navigation Satellite System) modules (such as U-blox F9P), IMU (Inertial Measurement Unit, such as ADIS16470), and LiDAR (Livox Mid-360), to collect attitude data of the UAV as it flies along the initial operational flight path.

[0069] Specifically: the GNSS antenna is mounted on the top of the drone to ensure unobstructed view; the IMU is mounted in the core flight control unit with a fixed orientation; and the LiDAR is mounted in front of the drone to ensure unobstructed view.

[0070] The relative positions of the GNSS antenna phase center, IMU centroid, and LiDAR coordinate system were determined using a checkerboard calibration board, with the error controlled within ±1cm.

[0071] Microsecond-level hardware synchronization is achieved using the PTP protocol, and the IMU (200Hz), LiDAR (20Hz), and GNSS (10Hz) are aligned via FPGA trigger signals.

[0072] After the drone starts up, its sensors collect data in real time. The main data collected by each sensor includes:

[0073] GNSS: Receives satellite signals through an antenna, analyzes and outputs WGS-84 coordinates (specifically including latitude, longitude, altitude, and PDOP);

[0074] IMU: Acquires acceleration and angular velocity;

[0075] LiDAR: Scans the surrounding environment point cloud and outputs point cloud frames with synchronized timestamps.

[0076] The acquired pose data is transmitted to the data processing module of the flight control main board for data cleaning. Then, the angular velocity data is used to obtain the UAV's attitude angle data. The specific calculation process uses existing integration methods, including yaw, roll, and pitch angles. Please refer to [link to relevant documentation]. Figure 2 This is a schematic diagram of the collected drone attitude angle data.

[0077] The drone uses its onboard camera to capture images in real time, which are then transmitted to the data processing module on the flight control board. Non-local mean filtering is then used to reduce noise in the images.

[0078] When operating at low altitudes, the GNSS module signal of a UAV is easily interfered with by the complex environment at low altitudes, resulting in poor signal accuracy and affecting the judgment of the UAV's current attitude. However, the UAV's flight control, real-time route planning, obstacle avoidance, work point positioning, target tracking and other operations all depend on the real-time calculation of its own attitude. Therefore, in the low-altitude operating environment, it is necessary to combine the monitoring data of other sensing devices to analyze the current attitude.

[0079] During actual low-altitude operations, the drone's position data is obtained through GNSS module analysis, and then IMU and LiDAR sensor data are used to assist in the drone's positioning. This precise positioning relies on the satellite positioning data possessing a certain level of accuracy.

[0080] In addition, the LiDAR or camera carried by the drone is usually designed to scan downward or forward, and the environmental data collected is mainly for the terrain below the drone or obstacles in front of the drone. However, in the low-altitude operation environment, there may also be some environmental obstacles above the drone. When there are obstacles above, the accuracy of GNSS signal is significantly affected by obstruction, multipath and attitude interference. Therefore, the obstruction of environmental obstacles above the drone can be analyzed by the data change characteristics of the GNSS module.

[0081] The Position Dilution of Precision (PDOP) value is a numerical indicator that measures the spatial geometric distribution of satellites currently used for positioning calculations, and is one of the output data of the satellite positioning process. It is only related to the position of the satellites relative to the receiver. The larger the PDOP value, the more concentrated the satellite distribution or the closer it is to a straight line / plane, the worse the geometric configuration, the more the measurement error will be amplified, and the lower the potential positioning accuracy.

[0082] In addition, GNSS signals are also affected by obstruction and multipath effects. In low-altitude environments, this is mainly manifested as a reduction in the number of satellite signals received by the GNSS module and abrupt changes in positioning data. However, during low-altitude operations, UAVs usually do not travel at a constant speed. Therefore, the positioning residual of the UAV at the current position should also be analyzed in conjunction with the acceleration data measured by the IMU unit.

[0083] The PDOP value calculated by the GNSS module is obtained when the UAV is performing low-altitude operations; the Euclidean distance between the UAV's position data obtained by the GNSS module at the current moment and the position data at the previous moment is calculated as the positioning distance between adjacent moments. The position data includes longitude, latitude, and altitude.

[0084] To obtain the drone's acceleration data from the previous moment, it should be noted that the drone's acceleration data is a normalized value.

[0085] By combining the PDOP value, the positioning distance between adjacent time points, and the UAV acceleration data, the satellite signal loss rate of the UAV is determined.

[0086] In this embodiment of the invention, the product of the PDOP value and the positioning distance between adjacent times is used as the numerator, and the UAV acceleration data is used as the denominator. The ratio formed by the numerator and the denominator is used as the satellite signal loss degree of the UAV.

[0087] The formula for calculating the satellite signal loss rate is as follows: Where Z is the satellite signal loss rate; P is the PDOP value; D is the positioning distance between adjacent time points; a is the UAV acceleration data; and tanh is the hyperbolic tangent function.

[0088] The greater the satellite signal loss degree Z at the current location of the drone, the stronger the GNSS positioning signal is affected by the obstruction and interference, and the lower the accuracy of the signal positioning. It also indicates that the degree of environmental obstruction above the drone in the low-altitude operation environment is higher and the current low-altitude operation environment is more complex.

[0089] When the satellite signal loss of the drone exceeds the preset loss threshold If the satellite signal loss rate of the UAV is abnormal at a certain time, this moment is designated as an abnormal moment. This indicates that the positioning signal received by the GNSS module at this time requires analysis of the data from the IMU and LiDAR sensors to determine the UAV's positioning information. In this embodiment of the invention, the preset loss threshold value is 0.45. In other embodiments, this value can be adjusted by the implementer according to the actual situation.

[0090] The signal loss rate from the current time point to the previous satellite is greater than The time period between the current moment and the previous abnormal moment is used as the motion reference period. It should be noted that the current moment is not included in the motion reference period.

[0091] Step S200: Construct a DCM matrix using the point attitude angles at each moment during the motion reference period, convert the three-axis acceleration of the machine system to the navigation system, and obtain the navigation system acceleration and the UAV navigation system velocity; combine the navigation system acceleration, the UAV navigation system velocity, and the changes in the three-dimensional data at the current moment to determine the initial velocity conversion coefficient of the three-dimensional data respectively.

[0092] During low-altitude operations, when GNSS signals are lost, it is necessary to jointly analyze the UAV's position using IMU and LiDAR data to achieve continuous positioning. The coordinate change characteristics during periods of good GNSS signal are selected, and the initial velocity conversion coefficient of the UAV is calculated by combining this with monitoring data from the IMU module.

[0093] When the GNSS module signal is poor, the current positioning situation can be analyzed by using the current acceleration data and direction of the UAV. First, it is necessary to evaluate the speed of the IMU sensor when the GNSS signal is relatively ideal: analyzing the relationship between the change of the UAV's positioning coordinates in adjacent moments and the flight speed can provide a reference for the UAV's navigation and positioning process reflected by the IMU data sensor when the GNSS signal is poor.

[0094] When analyzing the correlation between UAV coordinate changes and velocity conditions, it is also necessary to analyze the current three-axis angular velocities of the UAV: ​​the velocity and acceleration directions measured by the IMU are relative to the UAV system. The UAV attitude changes in real time, causing the directional correspondence between the UAV system and the geographic coordinate system to change continuously. If the IMU velocity data is not converted to the geographic coordinate system in real time using attitude angles (Roll, Pitch, Yaw), the actual relationship between the velocity vector and GNSS coordinate changes cannot be accurately reflected.

[0095] During the motion reference period, a Direct Motion Control (DCM) matrix is ​​constructed based on the attitude angles (Yaw, Pitch, Roll) at each time point. The DCM matrix is ​​then used to convert the three-axis accelerations of the machine system to the navigation system (ENU), yielding the navigation system accelerations. The process of converting the machine system's three-axis accelerations to the navigation system follows existing methods, specifically multiplying the column vector of the three-axis accelerations by the DCM matrix, which corresponds to the accelerations in the east, north, and up directions, respectively. It should be noted that the correspondence between the east, north, and up directions and latitude / longitude / altitude is as follows: east corresponds to longitude, north corresponds to latitude, and up corresponds to altitude.

[0096] Subtracting the gravitational acceleration constant G from the upward acceleration yields the navigation system acceleration after deducting the gravitational component, denoted as A. E A N A U The current UAV navigation system velocity is calculated using the integral method combined with the navigation system acceleration, and denoted as V. E V N V U In this embodiment of the invention, the gravitational acceleration constant G is taken as 9.81 m / s². 2 .

[0097] Obtain the differences in longitude, latitude, and altitude between each time point and the previous time point, denoted as ΔE, ΔN, and ΔU, respectively.

[0098] By combining the acceleration of the navigation system, the velocity of the UAV navigation system, and the changes in the current 3D data, the initial velocity conversion coefficients for the 3D data are determined. It should be noted that the 3D data here refers to longitude, latitude, and altitude. The initial velocity conversion coefficients for the 3D data include: the initial velocity conversion coefficient for longitude, the initial velocity conversion coefficient for latitude, and the initial velocity conversion coefficient for altitude.

[0099] Using data of any dimension in three-dimensional data as the target dimension data;

[0100] Obtain the change value of the target dimension data between the current time and the previous time, and use it as the dimension change value. Specifically, use the difference between the target dimension data between the current time and the previous time as the dimension change value.

[0101] The weights of the navigation system acceleration and the UAV navigation system velocity are determined based on the time length between the current moment and the previous moment. The navigation system acceleration and the UAV navigation system velocity are then weighted and summed to obtain the velocity parameter value.

[0102] Based on the dimensional change value and the velocity parameter value, the initial velocity conversion coefficient of the target dimensional data is determined; wherein, the dimensional change value is positively correlated with the initial velocity conversion coefficient, and the velocity parameter value is negatively correlated with the initial velocity conversion coefficient.

[0103] In this embodiment of the invention, the ratio of the dimensional change value to the velocity parameter value can be used as the initial velocity conversion coefficient.

[0104] In some embodiments of the present invention, the initial velocity conversion coefficient I corresponding to the longitude data E The calculation formula is: Where ΔE is the dimensional change value of the longitude data; t is the time length between the current moment and the previous moment; t 2 ×A E +t×V E The speed parameter value is for longitude data; A E The navigation system acceleration for longitude data; V E The speed of the drone navigation system based on longitude data.

[0105] In some embodiments of the present invention, the initial velocity conversion coefficient I corresponding to the latitude data N The calculation formula is: Where ΔN is the dimensionality change value of the latitudinal data; t is the time length between the current time and the previous time; t 2 ×A N +t×V N The velocity parameter value for latitude data; A N The navigation system acceleration for latitude data; V N The speed of the drone navigation system based on latitude data.

[0106] In some embodiments of the present invention, the initial velocity conversion coefficient I corresponding to the altitude data is... N The calculation formula is: Where ΔU is the dimensionality change value of the altitude data; t is the time length between the current moment and the previous moment; t 2 ×A U +t×V U The velocity parameter value for altitude data; A U Acceleration of the navigation system for altitude data; V U The speed of the drone navigation system based on altitude data.

[0107] Step S300: Acquire point cloud data and classify the point cloud data; match the volume and type of different types of point cloud data collected at different times to obtain the autonomous movement characteristics of each type of point cloud data; combine the autonomous movement characteristics with the ideal displacement of the UAV to perform environmental correction on the point cloud data at the current time to obtain the environmental correction coefficient; adjust the initial velocity conversion coefficient of the three-dimensional data according to the environmental correction coefficient to obtain the corrected velocity conversion coefficient; determine the estimated coordinate position at the next time step according to the corrected velocity conversion coefficient.

[0108] Because of the error accumulation problem in the IMU measurement process, the speed of IMU-based calculations will gradually deviate from the true value over a long period of time. Accurate position estimation cannot be obtained solely from IMU sensor data. Therefore, it is necessary to optimize the initial coordinate displacement coefficient using the environmental point cloud change characteristics constructed from LiDAR data. The current environmental point cloud is constructed using LiDAR data, and environmental correction coefficients are calculated based on the point cloud data change characteristics at different times. Specifically, in the LiDAR scan data processing, the reflection intensity is first normalized, and a preset reflection intensity threshold of 0.2 is used. In other embodiments, the implementer can adjust the threshold according to the actual situation. The preset reflection intensity threshold ranges from 0.1 to 0.3. The number of echoes is set to less than or equal to 2, typically 1 or 2.

[0109] However, in low-altitude operating environments, LiDAR may scan point cloud data of moving objects such as birds. When adjusting the initial velocity conversion coefficient by the distance changes between the UAV and the same part of the point cloud data at different time points, it is necessary to analyze the data change characteristics of different objects in the point cloud data to reduce the impact of object displacement on the UAV's velocity correction assessment.

[0110] First, a point cloud space is established within the motion reference time period. For the scan data within the motion reference time period: the raw distance data of the LiDAR scan is converted into three-dimensional spatial coordinates (X,Y,Z), and attributes such as reflection intensity and echo count of each point are recorded to generate raw point cloud data in LAS format.

[0111] Noise points are removed using point cloud processing tools, such as LiDAR360. These tools employ a Statistical Outlier Removal (SOR) algorithm to reduce data redundancy and improve point cloud density uniformity and computational efficiency. In this embodiment, the value of k in the SOR algorithm ranges from 10 to 30. Specifically, k is set to 30 when the scene density is high and to 10 when the scene density is low. The value of k is determined by the implementer based on the actual situation within this range. The outlier determination criterion ranges from 1.0 to 2.0, with a commonly used default value of 1.0.

[0112] By combining the DCM matrix obtained from GNSS positioning and IMU data, coordinate transformation is performed. Centered on the LiDAR, the point cloud data corresponding to the current moment is transformed from the sensor coordinate system to the global coordinate system, obtaining the longitude, latitude, and altitude data of each point in the point cloud. In this embodiment of the invention, the translation error in the pose data accuracy during coordinate transformation is less than 0.05m, and the attitude error is less than 0.5°.

[0113] The RangeNet++ algorithm is used to perform semantic segmentation on the point cloud data corresponding to the current time step, outputting the semantic category label for each voxel point to obtain a point cloud with category labels. In this embodiment of the invention, the semantic categories are defined as: ground, vehicle, pedestrian, vegetation, building, pole, and others, which can be referred to in the SemanticKITTI standard for details; the RangeNet++ network structure is as follows: encoder: Darknet-53, output dimension: 64, batch size: 8-16, learning rate: 0.001, input resolution: 64×1024 (Vertical×Horizontal).

[0114] The voxels in the current point cloud data with the same semantic category label are clustered to obtain several clusters. Each cluster is denoted as a point cloud cluster, and each cluster is considered as the same object. In this embodiment of the invention, K-means clustering can be used, with the clustering relationship being the Euclidean distance between voxels. The clustering parameter K is obtained by using the elbow method to obtain several clusters.

[0115] Based on the number of voxels of point cloud clusters with the same semantic category label at the current time and the next time, the volume similarity of the objects corresponding to the point cloud clusters with the same semantic category label is determined.

[0116] For an object corresponding to a point cloud cluster in the point cloud data at the current time, in the point cloud cluster at the next time, calculate the volume similarity V between the object corresponding to the current point cloud cluster and the object corresponding to the point cloud cluster with the same semantic category label. kl , in, This represents the number of voxels in the k-th point cloud cluster at time t. This represents the number of voxels in the l-th point cloud cluster that shares the same semantic category label as the k-th object at time t+1.

[0117] And calculate the chamfer distance between two point cloud clusters with the same semantic category label, denoted as C. kl .

[0118] By combining the volume similarity and the chamfer distance, the matching relationship between objects corresponding to two point cloud clusters with the same semantic category label at the current time and the next time is determined.

[0119] In this embodiment of the invention, the reciprocal of the product of volume similarity and chamfer distance is used as the matching relationship between objects corresponding to two point cloud clusters with the same semantic category label.

[0120] Based on the matching relationship, determine the matching objects corresponding to all point cloud clusters at the current time. Specifically:

[0121] The point cloud cluster with the largest matching relationship with the same semantic category label at time t+1 is selected as the matching object of the k-th point cloud cluster at time t, and the matching relationship is marked to avoid many-to-many matching.

[0122] After obtaining the matching object, the autonomous movement characteristics of the object corresponding to each point cloud cluster are evaluated by combining the attitude and flight direction of the UAV between the two time points.

[0123] First, obtain the actual displacement distance and direction of motion between the point cloud cluster and the matching object at the current moment. Specifically, taking the k-th point cloud cluster at time t as an example, the actual displacement distance between the k-th point cloud cluster at time t and the matching object at time t+1 is the Euclidean distance between the centroids of the two point cloud clusters; and take the direction from the centroid of the k-th point cloud cluster at time t to the centroid of the matching object at time t+1 as the direction of motion.

[0124] Calculate the directional similarity between the motion direction and the current acceleration direction of the UAV; use the cosine similarity between the motion direction and the current acceleration direction of the UAV as the directional similarity.

[0125] The ideal displacement distance at the next moment is determined based on the drone's current speed and acceleration conditions. It should be noted that determining the object's position at the next moment based on its current position, speed, and acceleration is existing technology and will not be elaborated upon here.

[0126] The initial displacement features are obtained by comparing the actual displacement distance with the ideal displacement distance; combined with the initial displacement features and direction similarity, the autonomous movement features of the object corresponding to each point cloud cluster are obtained.

[0127] In some embodiments, the initial displacement feature is used as the numerator, the orientation similarity is used as the denominator, and the normalized value of the corresponding ratio is used as the autonomous movement feature of the object corresponding to the point cloud cluster.

[0128] The formula for calculating the autonomous movement feature M is as follows: Where norm is the normalization function; D is the actual displacement distance; α represents the ideal displacement distance; α represents the directional similarity. This represents the initial displacement characteristic.

[0129] A larger initial displacement characteristic indicates a higher probability that the k-th point cloud cluster at time t possesses autonomous movement capability. When selecting points in the point cloud for UAV velocity correction, the computational weight of objects with larger autonomous movement characteristics should be reduced to improve correction accuracy and reduce the interference of LiDAR data uncertainty of autonomously moving objects on the correction.

[0130] Based on the autonomous movement characteristics of the point cloud cluster at each time step, a first correction parameter is determined. Specifically, for any point cloud cluster at the current time step, the result value obtained by negatively correlated mapping of the autonomous movement characteristics of the point cloud cluster is used as the first correction parameter. The computational weight of objects that do not have autonomous movement characteristics is amplified through negative correlation mapping.

[0131] The second correction parameter is determined based on the actual displacement distance and the ideal displacement distance of the point cloud cluster; specifically, the ratio of the actual displacement distance to the ideal displacement distance of the point cloud cluster is used as the second correction parameter.

[0132] The second correction parameter represents the difference between the actual displacement distance and the ideal displacement distance. As a correction coefficient for radar data to IMU velocity data, the greater the difference, the further the second correction parameter is from 1, and the stronger the subsequent correction of the initial velocity conversion coefficient.

[0133] Combining the first and second correction parameters, the environmental correction parameters are obtained. For any point cloud cluster at the current time, the product of the first and second correction parameters of the point cloud cluster is used as the single correction factor of the point cloud cluster; the sum of the single correction factors of all point cloud clusters at the current time is used as the environmental correction coefficient at the current time.

[0134] In some embodiments, the environmental correction factor E is calculated using the following formula: Where m is the number of point cloud clusters at the current moment; M j D represents the autonomous movement feature of the j-th point cloud cluster; j The actual displacement distance of the j-th point cloud cluster; For ideal displacement distance; 1-M j This is the first correction parameter; This is the second correction parameter; is a single correction factor for the j-th point cloud cluster.

[0135] Step S400: Adjust the initial velocity conversion coefficient of the three-dimensional data according to the environmental correction coefficient to obtain the corrected velocity conversion coefficient; determine the estimated coordinate position at the next moment according to the corrected velocity conversion coefficient.

[0136] After obtaining the environmental correction coefficient, the initial velocity conversion coefficient is adjusted based on the environmental correction coefficient, and the coordinate movement relationship of the current position is analyzed to obtain the corrected velocity conversion coefficient.

[0137] After obtaining the environmental correction coefficients of the point cloud for each frame, the initial velocity conversion coefficient is corrected using the environmental correction coefficients to evaluate the actual relationship between IMU data and GNSS coordinate changes under the current low-altitude operating environment.

[0138] During the motion reference period, the initial velocity conversion coefficient is adjusted according to the environmental correction coefficient E corresponding to each moment to obtain the corrected velocity conversion coefficient.

[0139] Use any dimension of the three-dimensional data as the target dimension data; use any moment within the motion reference time period as the target moment.

[0140] For the target time, the initial velocity conversion coefficient of the target dimension data is weighted by the environmental correction coefficient to obtain the target conversion coefficient at the target time.

[0141] A negative correlation mapping is performed on the satellite signal loss at the target time to obtain the conversion adjustment coefficient at the target time. The satellite signal loss after negative correlation mapping is used to amplify the calculation results with higher GNSS coordinate reliability when the satellite signal loss is small, and to calculate the actual relationship between IMU data and GNSS coordinate changes in the current operating environment during the past motion reference period.

[0142] The weighted time conversion coefficient is obtained by using the conversion adjustment coefficient as the weight of the target conversion coefficient. The sum of the time conversion coefficients of all moments within the motion reference period is used as the corrected velocity conversion coefficient of the target dimension data. The larger the distance between the corrected velocity conversion coefficient and 1, the greater the degree of cumulative error in the IMU velocity of the UAV's 3D data at the current moment. That is, the greater the degree of cumulative error in the IMU velocity in the longitude, latitude, and altitude directions, the greater the degree of adjustment required when calculating the UAV's position at the next moment using the IMU.

[0143] Taking time i as the current time as an example, the specific adjustment process is as follows:

[0144]

[0145] Among them, I · E Z is the correction speed conversion factor for the longitude data at the current moment;i E represents the current satellite signal loss rate. i I is the environmental correction factor for the current moment; Ei The initial velocity conversion factor for the longitude data at the current moment; n is the number of moments within the motion reference period; I · N The correction velocity conversion coefficient for the latitude data at the current moment; I Ni I is the initial velocity conversion coefficient for the latitude data at the current moment. · U The correction speed conversion coefficient for the current altitude data; I Ui The initial velocity conversion coefficient for the current altitude data; 1-Z i E is the conversion adjustment coefficient at the current moment. i ×I Ei The target conversion coefficient for the longitude data at the current moment; (1-Z i )×(E i ×I Ei E represents the time conversion factor for the current longitude data; i ×I Ni The target transformation coefficient for the latitude data at the current moment; (1-Z i )×(E i ×I Ni E represents the time conversion factor for the latitude data at the current moment. i ×I Ui The target conversion coefficient for the current altitude data; (1-Z i )×(E i ×I Ui ) represents the time conversion coefficient of the current altitude data.

[0146] Finally, based on the corrected velocity conversion coefficient, the estimated coordinate position at the next moment is determined.

[0147] Using the current location's IMU-obtained UAV navigation system velocity and acceleration, the longitude, latitude, and altitude coordinates for the next moment are estimated. Taking longitude coordinates as an example: Where X represents the longitude coordinate at the next moment, and X0 is the longitude coordinate at the current moment. V is the normalized value of the corrected velocity conversion coefficient for the longitude velocity at the current moment, V and A are the UAV navigation system velocity and navigation system acceleration for the latitude data at the current moment, respectively, and t is the time difference between the current moment and the next moment.

[0148] This allows us to obtain the estimated coordinates of each dimension at the next moment, i.e., the coordinates under each dimension, and thus the estimated coordinate position of the UAV at the next moment. For example, if the longitude coordinate of the next moment is c1, the latitude coordinate of the next moment is c2, and the altitude coordinate of the next moment is c3, then the estimated coordinate position of the UAV at the next moment (c1, c2, c3) is obtained from c1, c2, and c3, thus realizing the UAV positioning.

[0149] This method estimates the coordinate position, reducing interference from the accumulated error of the IMU unit and minimizing the impact of autonomously moving objects on the analysis process in low-altitude environments. It enables relatively accurate UAV position estimation even with weak GNSS module signals. This provides a foundation for subsequent UAV operations such as flight control, route planning, obstacle avoidance, work site location, and target tracking. Accurate knowledge of the UAV's own position and attitude is a prerequisite and foundation for safe flight, stable hovering, path planning, and precise operations. Without accurate self-positioning, the UAV cannot reliably navigate or interact with its environment / targets.

[0150] Example 2:

[0151] Please see Figure 3 The diagram illustrates a system block diagram of a UAV low-altitude operation positioning system based on multimodal perception, according to an embodiment of the present invention. The system includes the following modules:

[0152] The time period determination module is used to acquire positioning data of the UAV during low-altitude operations; analyze the time periods when the positioning data may have signal anomalies, and determine the motion reference time period;

[0153] An initial determination module is used to acquire the navigation system acceleration and the UAV navigation system velocity during the motion reference time period; and to determine the initial velocity conversion coefficient of the three-dimensional data by combining the navigation system acceleration, the UAV navigation system velocity, and the changes in the three-dimensional data at the current moment.

[0154] The environmental correction module is used to acquire point cloud data during low-altitude operations, classify the point cloud data to obtain point cloud clusters, match the volume and type of point cloud clusters collected at different times to obtain the autonomous movement characteristics of each point cloud cluster, and combine the autonomous movement characteristics with the ideal displacement of the UAV to determine the environmental correction coefficient of the point cloud data at the current time.

[0155] The positioning and determination module is used to adjust the initial velocity conversion coefficient of the three-dimensional data according to the environmental correction coefficient to obtain the corrected velocity conversion coefficient; and to determine the estimated coordinate position of the UAV at the next moment according to the corrected velocity conversion coefficient.

[0156] Optionally, the transmission medium can be a wired link, such as, but not limited to, coaxial cable, fiber optic cable and digital subscriber line, or a wireless link, such as, but not limited to, wireless Fidelity (WIFI), Bluetooth and mobile device networks.

[0157] It should be noted that the device provided in the above embodiments is only an example of the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above.

[0158] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. For example, as shown... Figure 4 As shown, the computer device 500 includes: a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and running on the processor 520, wherein when the processor 520 executes the computer program 530, the computer device can execute any of the aforementioned multimodal perception-based UAV low-altitude operation positioning methods.

[0159] Furthermore, embodiments of the present invention also protect an apparatus that may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to perform the UAV low-altitude operation positioning method based on multimodal perception provided in embodiments of the present invention.

[0160] In this embodiment of the invention, the device can be divided into functional modules according to the above method example. For example, each module can correspond to a separate function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and is only a logical functional division. In actual implementation, there may be other division methods.

[0161] When each module is divided according to its function, the device may also include a signal uploading module, a determination module, and an adjustment module. It should be noted that all relevant content of each step involved in the above method embodiments can be referenced from the functional descriptions of the corresponding functional modules, and will not be repeated here.

[0162] It should be understood that the apparatus provided in this embodiment of the invention is used to execute the above-described method for low-altitude UAV operation positioning based on multimodal perception, and therefore can achieve the same effect as the above-described implementation method.

[0163] When using integrated units, the device may include a processing module and a storage module. When applied to a device, the processing module can be used to control and manage the device's operations. The storage module can be used to support the device in executing program code, etc. The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits as described in this disclosure. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of Digital Signal Processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0164] In addition, the device provided in the embodiments of the present invention may specifically be a chip, component or module. The chip may include a connected processor and a memory. The memory is used to store instructions. When the processor calls and executes the instructions, the chip can execute the UAV low-altitude operation positioning method based on multimodal perception provided in the above embodiments.

[0165] This invention also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, the computer executes the aforementioned method steps to implement the UAV low-altitude operation positioning method based on multimodal perception provided in the above embodiments.

[0166] This invention also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement the UAV low-altitude operation positioning method based on multimodal perception provided in the above embodiments.

[0167] In this invention, the apparatus, computer-readable storage medium, computer program product, or chip provided in the embodiments are all used to execute the corresponding methods described above. Therefore, the beneficial effects they achieve can be referred to the beneficial effects in the corresponding methods described above, and will not be repeated here. Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways.

[0168] The device embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0169] It should also be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0170] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0171] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0172] The above content is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for positioning of low-altitude operation of a UAV based on multi-modal perception, characterized in that, The method includes the following steps: Acquire positioning data of the UAV during low-altitude operations; analyze the periods when the positioning data may exhibit signal anomalies to determine the motion reference period; Obtain the navigation system acceleration and UAV navigation system velocity during the motion reference period; combine the navigation system acceleration, the UAV navigation system velocity, and the changes in the three-dimensional data at the current moment to determine the initial velocity conversion coefficient of the three-dimensional data, including: using arbitrary dimension data in the three-dimensional data as the target dimension data; Obtain the change value of the target dimension data between the current time and the previous time, and use it as the dimension change value; The weights of the navigation system acceleration and the UAV navigation system velocity are determined based on the time length between the current moment and the previous moment. The navigation system acceleration and the UAV navigation system velocity are then weighted and summed to obtain the velocity parameter value. Based on the dimensional change value and the velocity parameter value, the initial velocity conversion coefficient of the target dimensional data is determined; wherein, the dimensional change value is positively correlated with the initial velocity conversion coefficient, and the velocity parameter value is negatively correlated with the initial velocity conversion coefficient; Acquire point cloud data during low-altitude operations and classify the point cloud data to obtain point cloud clusters; match the volume and type of point cloud clusters collected at different times to obtain the autonomous movement characteristics of each point cloud cluster, including: matching the volume and type of different types of point cloud data collected at different times to obtain the matching object for each point cloud cluster. Obtain the actual displacement distance and direction of motion between the point cloud cluster and the matched object at the current moment; Calculate the directional similarity between the direction of motion and the direction of the drone's acceleration at the current moment; Calculate the current velocity of the drone and the ideal displacement distance of the drone at the next moment under the current acceleration conditions; The initial displacement features are obtained by comparing the actual displacement distance with the ideal displacement distance; the autonomous movement features of the object corresponding to each point cloud cluster are obtained by combining the initial displacement features and the direction similarity; and the environmental correction coefficient of the point cloud data at the current moment is determined by combining the autonomous movement features and the ideal displacement of the UAV. Based on the environmental correction coefficient, the initial velocity conversion coefficient of the three-dimensional data is adjusted to obtain the corrected velocity conversion coefficient, including: using any dimension of the three-dimensional data as the target dimension data; and using any moment within the motion reference time period as the target moment. For the target time, the initial velocity conversion coefficient of the target dimension data is weighted by the environmental correction coefficient to obtain the target conversion coefficient at the target time. By performing a negative correlation mapping on the satellite signal loss rate at the target time, the conversion adjustment coefficient at the target time is obtained; The target conversion coefficient is weighted by the conversion adjustment coefficient to obtain the time conversion coefficient; the sum of the time conversion coefficients of all times within the motion reference period is used as the corrected velocity conversion coefficient of the target dimension data; the estimated coordinate position of the UAV at the next time moment is determined based on the corrected velocity conversion coefficient. 2.The multi-modal perception based UAV low altitude operation positioning method of claim 1, wherein, The analysis of the positioning data may reveal periods of signal anomalies, determining motion reference periods, including: The positioning data includes: the PDOP value of the UAV when performing low-altitude operations, the positioning distance between adjacent time points, and the UAV acceleration data; By combining the PDOP value, the positioning distance between adjacent times, and the UAV acceleration data, the satellite signal loss degree of the UAV is determined; when the satellite signal loss degree of the UAV is abnormal, the abnormal time is recorded; the time period between the current time and the previous abnormal time is used as the motion reference time period. 3.The multi-modal perception based UAV low altitude operation positioning method of claim 2, wherein, The determination of the UAV's satellite signal loss rate by combining the PDOP value, the positioning distance between adjacent time points, and the UAV's acceleration data includes: The product of the PDOP value and the positioning distance between adjacent times is used as the numerator, and the UAV acceleration data is used as the denominator. The ratio of the numerator and denominator is used as the satellite signal loss degree of the UAV. 4.The multi-modal perception based UAV low altitude operation positioning method of claim 1, wherein, The method for obtaining the speed of the UAV navigation system is as follows: using the point attitude angle at each moment in the motion reference period to construct a DCM matrix, the three-axis acceleration of the machine system is converted to the navigation system to obtain the navigation system acceleration; The current UAV navigation system velocity is obtained by combining the integral method with the navigation system acceleration. 5.The multi-modal perception based UAV low altitude operation positioning method of claim 1, wherein, The determination of the environmental correction coefficient for the point cloud data at the current moment, combining the autonomous movement characteristics and the ideal displacement of the UAV, includes: The first correction coefficient is determined based on the autonomous movement characteristics of the point cloud cluster at each time step; The second correction factor is determined based on the actual displacement distance and the ideal displacement distance of the point cloud cluster; Combining the first correction factor and the second correction factor, the environmental correction factor is obtained. 6.The multi-modal perception based UAV low altitude operation positioning method of claim 1, wherein, Determining the estimated coordinate position of the UAV at the next moment based on the corrected velocity conversion coefficient includes: Use any dimension of the three-dimensional data as the target dimension data; use any moment within the motion reference time period as the target moment. The normalized value of the conversion coefficient of the target dimension data at the current moment is weighted by the UAV navigation system velocity corresponding to the target dimension data to obtain the navigation system correction velocity. Based on the navigation system correction velocity, navigation system acceleration, and the time difference between the current moment and the next moment, determine the coordinates of the target dimension data at the next moment; The estimated coordinates of the UAV at the next moment are constructed from the coordinates of the 3D data at the next moment.

7. A low-altitude operation positioning system for unmanned aerial vehicles (UAVs) based on multimodal perception, characterized in that, For executing the UAV low-altitude operation positioning method based on multimodal perception as described in any one of claims 1 to 6, the system includes the following modules: The time period determination module is used to acquire positioning data when the drone is performing low-altitude operations; Analyze the time periods when the positioning data may show signal anomalies to determine the motion reference time period; The initial determination module is used to obtain the navigation system acceleration and the UAV navigation system velocity during the motion reference time period; By combining the acceleration of the navigation system, the velocity of the UAV navigation system, and the changes in the three-dimensional data at the current moment, the initial velocity conversion coefficients of the three-dimensional data are determined respectively; The environment correction module is used to acquire point cloud data during low-altitude operations, classify the point cloud data to obtain point cloud clusters, and match the volume and type of point cloud clusters collected at different times to obtain the autonomous movement characteristics of each point cloud cluster. Determine the environment correction coefficient of the point cloud data at the current time in combination with the autonomous movement feature and the ideal displacement of the unmanned aerial vehicle; The positioning determination module is configured to adjust an initial speed conversion coefficient of the three-dimensional data according to the environment correction coefficient to obtain a corrected speed conversion coefficient, and determine an estimated coordinate position of the unmanned aerial vehicle at the next time according to the corrected speed conversion coefficient.

Images