Unmanned aerial vehicle remote sensing data processing system and data processing method

Abstract

The present application relates to the technical field of remote sensing data processing, and particularly relates to a UAV remote sensing data processing system and a data processing method. The method comprises the following steps: collecting remote sensing data of an original survey area by a UAV, eliminating relative errors caused by ground height differences by using height normalization, constructing a survey area coordinate system, obtaining a trajectory sequence and attitude disturbance information of the UAV, projecting the trajectory sequence and the attitude disturbance information into a unified survey area coordinate system, building a three-dimensional disturbance trajectory network, identifying point cloud spatial offset deviation through the network, dynamically compensating each point cloud unit in combination with a time stamp, correcting survey area remote sensing data, matching the corrected remote sensing data with corresponding point cloud unit coordinates, rebuilding a three-dimensional survey area framework, and realizing high-precision topographic feature delineation. The present application combines the trajectory sequence and the attitude disturbance information of the UAV, builds a three-dimensional disturbance trajectory network, and effectively deals with complex surveying and mapping tasks and environmental changes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing data processing, and in particular to a UAV remote sensing data processing system and a data processing method. BACKGROUND

[0002] The existing UAV remote sensing data processing technology has significant deficiencies in data acquisition and processing accuracy. Many methods fail to fully consider the influence of terrain changes on remote sensing data, resulting in significant relative errors in remote sensing data in areas with large elevation changes, affecting the reliability of subsequent analysis. In particular, in complex terrain, traditional elevation normalization techniques often fail to effectively eliminate errors, resulting in a significant reduction in data accuracy, which is particularly evident in practical applications. The existing technology lacks systematic analysis of UAV trajectory and attitude disturbance information and fails to effectively integrate dynamic changes during flight into data processing, resulting in an imprecise three-dimensional trajectory network and an inability to accurately identify point cloud spatial offset deviations, especially during high-speed flight, where attitude disturbance is often ignored, ultimately leading to poor point cloud data correction and affecting subsequent measurement area framework reconstruction. SUMMARY

[0003] Therefore, it is necessary to provide a UAV remote sensing data processing system and a data processing method to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, the UAV remote sensing data processing method comprises the following steps:

[0005] Step S1: collecting original measurement area remote sensing data by a UAV; eliminating relative errors caused by ground height differences in the original measurement area remote sensing data through elevation normalization, and constructing a measurement area coordinate system;

[0006] Step S2: obtaining a UAV trajectory sequence and attitude disturbance information generated during UAV flight; projecting the attitude disturbance information and the UAV trajectory sequence into the same measurement area coordinate system to build a three-dimensional disturbance trajectory network;

[0007] Step S3: identifying point cloud spatial offset deviations in the original measurement area remote sensing data through the three-dimensional disturbance trajectory network, and dynamically compensating each point cloud unit in combination with the time stamp to correct the measurement area remote sensing data;

[0008] Step S4: matching the corrected measurement area remote sensing data with the corresponding point cloud unit coordinates in the measurement area coordinate system; reconstructing a three-dimensional measurement area framework based on the point cloud unit coordinates, the measurement area coordinate system, and the corrected measurement area remote sensing data.

[0009] The present application also provides a UAV remote sensing data processing system for executing the UAV remote sensing data processing method as described above, which comprises:

[0010] a data acquisition module, configured to acquire original survey area remote sensing data by the unmanned aerial vehicle, eliminate relative errors caused by ground height differences in the original survey area remote sensing data through elevation normalization, and construct a survey area coordinate system;

[0011] a disturbance analysis module, configured to acquire a sequence of trajectories of the unmanned aerial vehicle and attitude disturbance information generated in the flight process of the unmanned aerial vehicle, project the attitude disturbance information and the sequence of trajectories of the unmanned aerial vehicle into the same survey area coordinate system, and build a three-dimensional disturbance trajectory network;

[0012] an offset correction module, configured to identify point cloud spatial offset deviations in the original survey area remote sensing data through the three-dimensional disturbance trajectory network, and dynamically compensate each point cloud unit in combination with a time stamp, so as to correct the survey area remote sensing data;

[0013] a remote sensing data reconstruction module, configured to match the corrected survey area remote sensing data and corresponding point cloud unit coordinates in the survey area coordinate system, and reconstruct a three-dimensional survey area framework according to the point cloud unit coordinates in combination with the survey area coordinate system and the corrected survey area remote sensing data.

[0014] The present application has the following advantages: on the one hand, the original remote sensing data is efficiently acquired by the unmanned aerial vehicle, which significantly improves the speed and coverage of data acquisition, adapts to the needs of complex terrain, and can complete the surveying and mapping of a large area in a short time, providing more abundant and accurate geographic information, meeting the dual requirements of efficiency and accuracy of modern surveying and mapping, eliminating errors caused by ground height differences through elevation normalization technology, ensuring the accuracy of the survey area coordinate system, enhancing the reliability of the data, laying a solid foundation for subsequent data processing and analysis, and improving the overall data quality.

[0015] On the other hand, in combination with the sequence of trajectories of the unmanned aerial vehicle and the attitude disturbance information, a three-dimensional disturbance trajectory network is built, providing comprehensive monitoring of flight dynamic changes, enhancing the systematicness of data processing, accurately capturing various disturbances generated in the flight process, providing an important basis for data correction, in addition, the point cloud spatial offset deviations are identified through the three-dimensional disturbance trajectory network, and dynamic compensation is performed in combination with the time stamp, ensuring the correction accuracy of each point cloud unit, reducing the data deviation caused by attitude disturbance, and improving the overall quality of remote sensing data.

[0016] On the other hand, matching the corrected remote sensing data with the point cloud unit coordinates in the survey area coordinate system enhances the correlation between the data and improves the consistency of data processing. Based on the point cloud unit coordinates and the corrected remote sensing data, a three-dimensional survey area framework is reconstructed, the topographic features are accurately described, detailed geographic information support is provided, and the application of unmanned aerial remote sensing technology in environmental monitoring, resource investigation and urban planning is promoted. The overall process improves the efficiency and accuracy of data processing, effectively deals with complex surveying and mapping tasks and environmental changes, and promotes the further development of intelligent surveying and mapping technology. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is a step flowchart of a kind of unmanned aerial vehicle remote sensing data processing method;

[0018] Figure 2 It is a schematic diagram of offset variation under different flight altitudes;

[0019] Figure 3 It is a schematic diagram of the influence of different pitch angle variation;

[0020] Figure 4 It is a schematic diagram of remote sensing data acquisition process.

[0021] The implementation of the object of the application, functional characteristics and advantages will be further described with reference to the drawings. DETAILED DESCRIPTION

[0022] The technical method of the application will be described clearly and completely below in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the application.

[0023] In addition, the drawings are only schematic illustrations of the application, not necessarily drawn to scale. The same reference signs in the drawings represent the same or similar parts, so repeated description of them will be omitted. Some block diagrams shown in the drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0024] It should be understood that, although the terms "first", "second" or the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of the example embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] To achieve the above object, please refer to Figures 1 to 4 The unmanned aerial vehicle remote sensing data processing method comprises the following steps:

[0026] Step S1: collecting original survey area remote sensing data by an unmanned aerial vehicle; eliminating relative errors caused by ground height differences in the original survey area remote sensing data by height normalization, and constructing a survey area coordinate system;

[0027] Please refer to Figure 4 In an implementation manner of the embodiment of the present application, the unmanned aerial vehicle flies over the survey area according to a preset flight route to collect original remote sensing data covering the survey area range, the flight route is set in a parallel flight strip mode to ensure about 30% overlap between adjacent flight strips, the flight height is set to 120 meters to ensure that a large range area can be covered and sufficient resolution is retained, the collected images are imported into a data processing software, height normalization processing is performed, a unified reference plane (for example, sea level or the lowest point of the survey area) is taken as a reference, relative errors caused by ground height fluctuations in the images are eliminated, and a survey area coordinate system is constructed based on the unified spatial reference to obtain standardized survey area original remote sensing data.

[0028] In an implementation manner of the embodiment of the present application, a block region mean comparison method is used to adjust the reference of different terrain heights, and the specific method is to divide the survey area into a plurality of small square regions, calculate the average height value of each region, and uniformly adjust these average values to the same reference plane, thereby effectively avoiding spatial distortion introduced by some local height differences (such as hills or depressions), and unifying the height reference of the entire region.

[0029] Step S2: obtaining an unmanned aerial vehicle trajectory sequence and attitude disturbance information generated in the unmanned aerial vehicle flight process; projecting the attitude disturbance information and the unmanned aerial vehicle trajectory sequence into the same survey area coordinate system to build a three-dimensional disturbance trajectory network;

[0030] In one implementation manner of the embodiment of the present application, the trajectory sequence and the attitude disturbance information of the unmanned aerial vehicle are simultaneously collected by the GPS and the IMU (inertial measurement unit) carried by the unmanned aerial vehicle, wherein the trajectory data includes longitude and latitude and height coordinates sampled once per second, and the attitude information includes roll angle, pitch angle and yaw angle; the data is exported and then imported into the same survey area coordinate system; the trajectory points and the attitude points are spatially superimposed by using a data processing software, so as to build a three-dimensional disturbance trajectory network capable of dynamically describing the flight path and the attitude change.

[0031] In one implementation manner of the embodiment of the present application, if the trajectory points are sampled once per second and the attitude points are sampled once per 0.5 second, the trajectory points need to be interpolated into 0.5-second data by using a linear interpolation method to ensure the timestamp consistency, and then the trajectory data and the attitude data are correspondingly superimposed in the coordinate system, so that the three-dimensional disturbance trajectory can completely and continuously describe the motion state of the unmanned aerial vehicle, and the disconnection between the trajectory and the attitude is avoided.

[0032] Step S3: identifying the point cloud spatial offset deviation in the original survey area remote sensing data by using the three-dimensional disturbance trajectory network, and dynamically compensating each point cloud unit in combination with the timestamp to correct the survey area remote sensing data;

[0033] In one implementation manner of the embodiment of the present application, the acquired remote sensing point cloud data is compared with the three-dimensional disturbance trajectory network; for each point cloud unit, the trajectory and the attitude state of the unmanned aerial vehicle at the corresponding timestamp are found, and then the offset of the point cloud in space is calculated, and the point cloud unit is adjusted back to the correct spatial position by using a compensation algorithm, so as to obtain the corrected point cloud data.

[0034] In one implementation manner of the embodiment of the present application, when the timestamp dynamic compensation is performed, a point-by-point interpolation weighting method is adopted, the nearest two trajectory points and attitude points before and after the timestamp of the point cloud unit are found, the offset difference values of the two points are calculated, the offset of the point cloud unit at the middle time point is calculated by using a weighted average method, and finally the compensation amount is added back to the point cloud unit coordinate, so as to realize smooth transition and avoid sudden jump errors.

[0035] Step S4: matching the corrected survey area remote sensing data with the corresponding point cloud unit coordinates in the survey area coordinate system; and reconstructing a three-dimensional survey area framework according to the point cloud unit coordinates, the survey area coordinate system and the corrected survey area remote sensing data.

[0036] In one implementation manner of the embodiment of the present application, the corrected point cloud data is compared with the point cloud reference unit in the survey area coordinate system point by point, the corrected data is gradually aligned to the survey area coordinate system by point-by-point coordinate matching, and after the completion, the three-dimensional framework of the survey area is reconstructed based on the three-dimensional modeling tool by using the matching results, so as to form a high-precision spatial model.

[0037] In one implementation of the embodiment of the present application, in the point-by-point matching process, a nearest neighbor iteration method (one implementation of the ICP algorithm) is adopted, for each point cloud unit, the nearest reference point is found as the matching object, and the difference between the two is calculated, and after each round of iteration, the position and direction of the point cloud data are adjusted as a whole, so that it gradually converges to the state of complete alignment with the reference coordinate system, and finally a three-dimensional survey area framework with an error within centimeters is obtained.

[0038] Preferably, the step S3 of identifying the point cloud spatial correlation deviation in the original survey area remote sensing data through the three-dimensional disturbance recording network comprises:

[0039] Matching the time sequence of each point cloud unit in the original survey area remote sensing data with the position of the unmanned aerial vehicle at the moment in the three-dimensional disturbance reporting network to obtain position matching data;

[0040] Calculating the spatial deviation between the theoretical flight position of the unmanned aerial vehicle in the position matching data and the actual recorded position in the unmanned aerial vehicle trajectory sequence to identify first deviation spatial distribution data;

[0041] On the basis of the first deviation spatial distribution data, the second deviation spatial distribution data under the combined action of the attitude disturbance information is calculated;

[0042] Integrating the first deviation spatial distribution data and the second deviation spatial distribution data to calculate the point cloud spatial deviation.

[0043] In one implementation of the embodiment of the present application, the time stamp of each point cloud unit in the original survey area remote sensing data is extracted one by one, and compared with the position of the unmanned aerial vehicle at the corresponding moment in the three-dimensional disturbance recording network to generate the corresponding unmanned aerial vehicle position matching data of each point cloud unit.

[0044] In another embodiment of the present application, for each point cloud unit in the original survey area remote sensing data, the collection time stamp is first extracted, and then the closest unmanned aerial vehicle position point to the time stamp is found in the three-dimensional disturbance recording network, and the two are corresponded to obtain the position matching result of the unmanned aerial vehicle at the collection moment, which can establish a direct time and space relationship between the point cloud data and the unmanned aerial vehicle flight path.

[0045] In another embodiment of the present application, in the time stamp comparison process, in order to avoid the time misalignment problem caused by different data sampling intervals, a linear interpolation method is introduced to fill in the unmanned aerial vehicle position sequence, so that the time stamp of the point cloud unit can strictly correspond to the unmanned aerial vehicle position, and the accuracy of the position matching data is ensured.

[0046] In another embodiment of the present application, the ideal position in the theoretical flight path of the unmanned aerial vehicle is compared with the actual position recorded in the trajectory sequence point by point, the offset in the three-dimensional coordinate space is calculated, and the result is taken as the first deviation spatial distribution data.

[0047] In another embodiment of the present application, when calculating the offset, the difference in three-dimensional coordinates is decomposed into three components in X, Y and Z directions, and the spatial deviation in the three directions is recorded respectively, so that the first deviation spatial distribution data can reflect the distribution characteristics of the offset in each direction.

[0048] In another embodiment of the present application, the change amount of pitch angle, yaw angle and roll angle recorded in the attitude disturbance information is superimposed with the first deviation spatial distribution data to obtain the second deviation spatial distribution data considering the influence of attitude.

[0049] In another embodiment of the present application, when superimposing, the angle change is converted into the increment in the direction, and is superimposed on the first deviation spatial distribution data point by point, so that the second deviation spatial distribution data can truly reflect the comprehensive effect of flight attitude and trajectory deviation.

[0050] In another embodiment of the present application, the first deviation spatial distribution data and the second deviation spatial distribution data are merged to generate the final point cloud spatial offset deviation, which is used to correct the position of the point cloud unit in space.

[0051] In another embodiment of the present application, a weighting method is adopted in the merging process, different weights are set for the second deviation spatial distribution data according to the strength of the attitude disturbance in different flight stages, so that the point cloud spatial offset deviation more in line with the actual situation is obtained.

[0052] In another embodiment of the present application, taking a certain urban building group survey area as an example, the unmanned aerial vehicle flies according to the predetermined route to collect point cloud data, extracts the timestamp in the point cloud data, and matches it with the position of the unmanned aerial vehicle in the three-dimensional disturbance recording network to obtain position matching data, calculates the difference between the ideal route position and the actual trajectory position, generates the first deviation spatial distribution data, combines the changes of pitch angle, yaw angle and roll angle recorded in the flight process, adds the influence of attitude to the first deviation data to obtain the second deviation spatial distribution data, integrates the two kinds of deviation data, and calculates the overall spatial offset deviation of the building group point cloud, so as to obtain the corrected urban three-dimensional point cloud data.

[0053] Preferably, the spatial offset between the theoretical flight position of the unmanned aerial vehicle in the position matching data and the actual recorded position in the trajectory sequence of the unmanned aerial vehicle is calculated, and the first deviation spatial distribution data comprises:

[0054] extracting the time corresponding to the theoretical flight position of the unmanned aerial vehicle from each three-dimensional disturbance monitoring network, synchronously reading the actual recorded position at the same time;

[0055] calculating the three-dimensional Euclidean distance between the theoretical flight position of the unmanned aerial vehicle and the actual recorded position as a spatial offset;

[0056] projecting the spatial offset into the ground coordinate system to obtain first deviation spatial distribution data.

[0057] In an implementation manner of the embodiment of the present application, the time corresponding to the theoretical flight position of the unmanned aerial vehicle is extracted from each three-dimensional disturbance monitoring network, the actual recorded position at the same time is synchronously read, the three-dimensional Euclidean distance between the theoretical flight position of the unmanned aerial vehicle and the actual recorded position is calculated, the distance is taken as a spatial offset, and the spatial offset is projected into the ground coordinate system, so that the first deviation spatial distribution data is obtained.

[0058] In another embodiment of the present application, the time corresponding to the theoretical flight position of the unmanned aerial vehicle is extracted from each three-dimensional disturbance monitoring network by establishing a unified time stamp system at the data acquisition end, and each theoretical flight position is marked with a time stamp when stored, and the corresponding record is retrieved according to the time stamp when read.

[0059] In another embodiment of the present application, the coordinate point at a certain time is found from the theoretical flight trajectory of the unmanned aerial vehicle calculated by the system, and the coordinate point actually recorded at the time is read from the actual flight record log, the two points are placed in the same three-dimensional coordinate system, the straight line distance between the two points is calculated, the straight line distance is the spatial offset, and the spatial offsets are uniformly converted into the ground coordinate system, so that the error distribution at different positions, that is, the first deviation spatial distribution data, can be obtained, and through this way, the deviation of the theoretical flight trajectory and the actual flight trajectory at different positions can be clearly seen.

[0060] In another embodiment of the present application, the GNSS data output by the onboard positioning module of the unmanned aerial vehicle contains a time stamp, and the data reading program retrieves and calls the data of the corresponding time in the GNSS according to the time stamp of the theoretical flight position, so as to ensure that the two types of data are completely aligned in the time dimension.

[0061] In another embodiment of the present application, in the calculation of the three-dimensional Euclidean distance between the theoretical flight position of the unmanned aerial vehicle and the actual recorded position, the three-dimensional coordinates of the theoretical position and the actual position are respectively substituted into the formula , so that the spatial offset can be obtained, and it is necessary to ensure that the input coordinates are unified to the same coordinate system in the calculation process.

[0062] In another embodiment of the present application, dx and dy in the spatial offset vector (dx, dy, dz) are kept as the ground projection, while the dz component is ignored, and the two-dimensional projection offset thus obtained can represent the horizontal offset of the UAV in the ground coordinate system.

[0063] In another embodiment of the present application, the spatial distribution of the calculated two-dimensional projection offset is statistically analyzed, and the offset is plotted according to the sampling point position in the geographic coordinate frame of the survey area to form a complete set of deviation spatial distribution data, which can be used for subsequent error correction and trajectory optimization.

[0064] In another embodiment of the present application, for example, the theoretical flight position of the UAV is (120.100, 30.200, 100.0), and the actual recorded position is (120.102, 30.201, 99.8). According to the Euclidean distance formula, the spatial offset is about 0.24 meters, and after projecting it into the ground coordinate system, the horizontal offset is about 0.18 meters. This result is used as part of the deviation distribution data for the survey area deviation statistics and correction.

[0065] Preferably, on the basis of the first deviation spatial distribution data, the second deviation spatial distribution data under the combined action is calculated in combination with the attitude disturbance information, which includes:

[0066] calculating the change amount of the pitch angle in the attitude disturbance information;

[0067] quantifying the influence of the change amount of the pitch angle on the laser beam pointing, and analyzing the laser beam pointing deviation in combination with the change amount;

[0068] converting the laser beam pointing deviation into the offset of the point cloud position;

[0069] calculating the actual offset according to the amplification effect of the flight height in the actual recorded position on the offset;

[0070] calculating the second deviation spatial distribution data under the combined action in combination with the actual offset and the first deviation spatial distribution data.

[0071] In one implementation of the embodiment of the present application, the change amount of the pitch angle in the attitude disturbance information is calculated, the influence of the change amount of the pitch angle on the laser beam pointing is quantified, and the laser beam pointing deviation is analyzed in combination with the change amount. The laser beam pointing deviation is converted into the offset of the point cloud position. The actual offset is calculated according to the amplification effect of the flight height in the actual recorded position on the offset. The second deviation spatial distribution data under the combined action is calculated in combination with the actual offset and the first deviation spatial distribution data.

[0072] In another embodiment of the present application, the pitch angle data recorded during flight is arranged in chronological order, the angle change amount at each time relative to the previous time is calculated, and then the angle change amount is substituted into the outgoing direction of the laser beam, which is equivalent to simulating what kind of deflection the laser beam will follow if the unmanned aerial vehicle nose is raised or lowered, and then corresponding to the position of the ground point cloud, the initial offset of the point cloud can be obtained. Then, considering the influence of flight height, for example, when flying at high altitude, even a small angle change will form a larger displacement on the ground, so the offset needs to be amplified according to the flight height to obtain the actual offset. Finally, the actual offset is superimposed with the first deviation spatial distribution data before, and the second deviation spatial distribution data under the combined action is obtained.

[0073] In another embodiment of the present application, by extracting the attitude angle data in the flight log in the unmanned aerial vehicle flight control system, the pitch angle at the current time is differentially operated with the pitch angle at the previous time to obtain the pitch angle change amount. For example, the pitch angle at the previous time is 5 degrees and the current time is 5.3 degrees, then the pitch angle change amount is 0.3 degrees.

[0074] In another embodiment of the present application, by establishing the geometric mapping relationship between the pitch angle and the laser beam direction vector, assuming that the laser is installed at the bottom of the unmanned aerial vehicle and faces downward, when the pitch angle increases, the projection point of the laser beam on the ground will produce a front and back displacement. The offset amount is calculated by the trigonometric function formula , wherein is the pitch angle change amount, is the flight height, is the ground displacement length, and according to this relationship, the laser beam pointing deviation can be quantitatively obtained.

[0075] In another embodiment of the present application, the laser beam pointing deviation is projected as an error component in the point cloud coordinate system, and the ground displacement corresponding to the deviation angle is mapped to the horizontal or vertical coordinate of a certain point in the point cloud, so as to obtain the point cloud position offset amount. For example, the laser beam is offset by 0.2 meters in the forward direction, and then the y coordinate of the corresponding sampling point in the point cloud is increased by 0.2 meters.

[0076] In another embodiment of the present application, as shown in Figure 2 and Figure 3 , the higher the flight height is under the condition of a certain pitch angle, the greater the ground displacement is. The result can be used as an amplified actual offset amount by using the formula , wherein is the pitch angle change amount, is the actual flight height. For example, when = 0.3 degrees, = 100 meters, 0.52 meters.

[0077] In another embodiment of the present application, the actual offset is superimposed on the corresponding point in the first deviation data, the position of the same timestamp and the same coordinate point in the first deviation spatial distribution data is found, and the actual offset vector is added to the deviation value of the point to obtain new deviation distribution data, so as to generate the second deviation spatial distribution data under the combined action.

[0078] In another embodiment of the present application, for example, the horizontal offset of a point in the first deviation spatial distribution data is 0.18 meters, the flight height is 100 meters, and the change amount of the pitch angle is 0.3 degrees. According to the trigonometric function, the projection point offset of the laser beam on the ground is 0.52 meters. The value is superimposed on the original horizontal offset of 0.18 meters to obtain the second deviation spatial distribution data under the combined action of 0.70 meters. This result can be directly used for subsequent point cloud correction and optimization of aerial survey accuracy.

[0079] Especially important is to quantify the influence of the change amount of the pitch angle on the pointing of the laser beam, and to analyze the pointing deviation of the laser beam in combination with the change amount, including:

[0080] Based on the change amount of the pitch angle, time series mapping is performed to obtain the pitch angle change data under the time stamp of each point,

[0081] The direction projection calculation is performed on the point-by-point pitch angle change data in combination with the initial direction vector of the laser beam to obtain the projection offset of the pitch angle change on the direction of the laser beam,

[0082] According to the projection offset, the up and down directions are determined in combination with the sign of the pitch angle change data to obtain the pitch direction correction data of each point,

[0083] The pitch direction correction data is analyzed for continuity in combination with the sampling time interval to obtain the pitch offset time sequence curve of each point,

[0084] The pitch offset time sequence curve is analyzed for amplitude and the maximum effective offset is extracted to obtain the pitch angle quantization influence data of each point,

[0085] The pitch angle quantization influence data is combined with the original pointing of the laser beam to calculate the difference to obtain the pointing deviation of the laser beam at each point.

[0086] In one implementation form of the embodiment of the application, the pitch angle change at each moment during flight is recorded to form a curve changing with time, and the data is used to express the process of lifting or lowering the nose of the unmanned aerial vehicle, the angle change is substituted into the initial direction of the laser beam, and the amount of forward or backward deviation of the laser beam is simulated if the laser beam is originally vertically downward, and the direction deviation at different time points is calculated. In order to distinguish whether the laser beam deviates upward or downward, it is also necessary to judge the direction according to whether the angle change is positive or negative, and a set of correction data changing with time is obtained. The correction data is connected in time sequence to form a continuous deviation curve. Further analysis of the curve is performed to extract the part with larger amplitude as the effective deviation of the laser beam affected by the pitch angle. The effective deviation is compared with the original direction of the laser beam to calculate the final pointing deviation of the laser beam.

[0087] In another embodiment of the application, the pitch angle change is mapped in time sequence. The pitch angle difference value in the attitude disturbance log derived from the flight control system is arranged according to the sampling time point to form a one-dimensional sequence data table corresponding to the time stamp, so that the pitch angle change at each moment can be synchronized and corresponded to the point cloud sampling.

[0088] In another embodiment of the application, the direction projection calculation is performed on the point-by-point pitch angle change data combined with the initial direction vector of the laser beam. The projection relationship of three-dimensional vectors is established to convert the pitch angle change data into the projection change of the laser beam direction vector in the Z-axis direction, so as to obtain the projection deviation changing with time.

[0089] In another embodiment of the application, the up-down direction of the projection deviation combined with the sign of the pitch angle change data is determined. The deviation direction is determined by judging whether the pitch angle change is positive or negative, for example, a positive value indicates that the nose of the unmanned aerial vehicle is tilted upward, which corresponds to the forward movement of the laser beam, and a negative value indicates that the nose of the unmanned aerial vehicle is tilted downward, which corresponds to the backward movement of the laser beam, so as to obtain the up-down direction correction data of each point.

[0090] In another embodiment of the application, the continuity analysis is performed on the pitch direction correction data combined with the sampling time interval. The difference between the correction data of adjacent sampling points is calculated and a time sequence curve is drawn to detect the smoothness and mutation points of the change, so as to obtain the pitch deviation time sequence curve of each point.

[0091] In another embodiment of the application, the amplitude of the pitch deviation time sequence curve is analyzed and the maximum effective deviation is extracted. The amplitude range of the curve data is calculated and the noise points are excluded to obtain the pitch angle quantization influence data of each point. For example, the pitch deviation amplitude is ±0.5 degrees in a certain period of time, and the maximum effective deviation of 0.5 degrees is taken as the quantization result of the data in this period.

[0092] In another embodiment of the application, the pitch angle quantization impact data is combined with the original laser beam pointing to calculate the difference, which is achieved by superimposing the pitch quantization impact data on the initial direction vector of the point cloud, that is, converting the pitch angle offset into the displacement length of the ground coordinates, and subtracting the original pointing of the laser beam, so as to obtain the point-by-point laser beam pointing deviation.

[0093] In another embodiment of the application, for example, the sampling frequency of the unmanned aerial vehicle during flight is 1 second, and the pitch angle change values recorded in the next 5 seconds are 0.1 degrees, 0.2 degrees, 0.3 degrees, 0.25 degrees, and 0.15 degrees. After mapping these values to the time sequence and combining the laser beam direction, the corresponding projection offset values are 0.17 meters, 0.35 meters, 0.52 meters, 0.43 meters, and 0.26 meters. According to the sign of the pitch angle, a time sequence curve is drawn and the maximum effective offset 0.52 meters is extracted. Finally, the point-by-point laser beam pointing deviation range calculated by combining the original laser beam pointing difference is 0.17 meters to 0.52 meters.

[0094] Preferably, the change in the quantized pitch angle affects the pointing of the laser beam, and the laser beam pointing deviation is analyzed in combination with the change, including:

[0095] Arranging the change in the pitch angle according to the time stamp to obtain a longitudinal pitch angle change sequence;

[0096] Calculating the pitch offset of the laser beam pointing caused by the longitudinal pitch angle change sequence to determine the longitudinal offset component data under the action of the longitudinal pitch angle;

[0097] Point-by-point superimposing the longitudinal offset component data according to the change in the pitch angle to obtain the comprehensive offset component under each time stamp;

[0098] Determining the laser beam pointing deviation through the comprehensive offset component.

[0099] In one implementation manner of the embodiment of the application, the pitch angle values recorded during flight are arranged in time sequence, so that it can be clearly seen whether the unmanned aerial vehicle head is lifted or pressed at different time points, forming a longitudinal angle change time sequence. The longitudinal angle change is point-by-point brought into the direction model of the laser beam, which is equivalent to simulating how much the light beam will deviate forward or backward at different angles. The results at each time are recorded to obtain a set of longitudinal offset component data. The longitudinal offset component is point-by-point accumulated to obtain the accumulated light beam deviation in the entire flight process. The accumulated result is compared with the original ideal direction to obtain the final laser beam pointing deviation.

[0100] In another embodiment of the present application, the pitch angle raw data collected by the flight control system is sorted in chronological order, and the pitch angle difference value at each time is arranged in sequence to form a time sequence, so that each data point can be one-to-one corresponding to the time stamp of the point cloud sampling.

[0101] In another embodiment of the present application, the longitudinal angle change value is substituted into the geometric relationship of the laser beam and the ground angle, and the offset length is calculated through simple trigonometric function relationship, for example, the offset amount = flight height x tan (angle change) is used to calculate the longitudinal offset component.

[0102] In another embodiment of the present application, the overall offset trend caused by continuous change is calculated by cumulative addition, and the longitudinal offset amount at each new time point is increased on the basis of the cumulative result of the previous time, so as to obtain the point-by-point comprehensive offset component.

[0103] In another embodiment of the present application, the comprehensive offset component at each time stamp is compared with the original non-offset position of the laser beam, and the difference is the actual laser beam pointing deviation, forming a point-by-point error distribution curve.

[0104] In another embodiment of the present application, for the calculation method of the above longitudinal offset component data, a small angle correction term can be added to the offset formula to improve the accuracy of the calculation, especially in the case of small angle change, the simple tan function may have amplification error, and the correction term can ensure the stability of the data.

[0105] In another embodiment of the present application, for the point-by-point superposition method, the sliding average method is combined to smooth the superposition curve, avoiding the offset mutation caused by individual abnormal points, so that the comprehensive offset component is more consistent with the continuity of the actual flight process.

[0106] In another embodiment of the present application, for the determination of the final laser beam pointing deviation, the flight height can be layered corrected, that is, different weight adjustments are adopted for the same angle offset at different heights, so as to obtain a more actual deviation result.

[0107] In another embodiment of the application, for example, the flight height of the unmanned aerial vehicle is 100 meters, and the pitch angle change amounts in the next 5 seconds are 0.05 degrees, 0.1 degrees, 0.15 degrees, 0.12 degrees, and 0.08 degrees, respectively. After arranging them according to the time stamp, a longitudinal angle change sequence is formed, and the longitudinal offset component is calculated by substituting the formula offset amount = height x tan(angle) to obtain the longitudinal offset component of 0.087 meters, 0.175 meters, 0.262 meters, 0.209 meters, and 0.140 meters, respectively. The integrated offset components are 0.087 meters, 0.262 meters, 0.524 meters, 0.733 meters, and 0.873 meters, respectively, by stacking these data point by point. The deviation range of the laser beam pointing direction is 0.087 meters to 0.873 meters by comparing the integrated offset components with the original pointing direction of the laser beam.

[0108] Preferably, according to the amplification effect of the flight height on the offset amount in the actual recorded position, calculating the actual offset amount comprises:

[0109] Extracting the flight height record in the unmanned aerial vehicle flight data that is consistent with the position node of the actual recorded position, and pairing the offset amount of the point cloud position at the same time node with the flight height record to generate offset amount and height corresponding data;

[0110] Calculating the height difference of each point cloud unit according to the offset amount and height corresponding data;

[0111] Matching the height difference and the offset amount point by point to form point-by-point offset and height matching data;

[0112] Identifying the height amplification coefficient through the point-by-point offset and height matching data;

[0113] Based on the point-by-point multiplication of the height amplification coefficient and the offset amount, and calculating the actual offset amount after amplification correction.

[0114] In one implementation of the embodiment of the application, the height information corresponding to the point cloud data collection position is extracted from the unmanned aerial vehicle flight record, and the offset amount data and the flight height are matched by one-to-one correspondence through the time stamp. Then, the height difference of different point cloud units is calculated according to the matched data, which reflects the flight state difference of different sampling points. Then, the height difference and the offset amount are combined to form a point-by-point height and offset amount comparison relationship, and then the amplification coefficient of the offset with the height change is identified. The amplification coefficient essentially describes the sensitivity of the flight height to the offset amount. Finally, the original offset amount is corrected using the amplification coefficient, so that the corrected offset amount can more truly reflect the spatial error situation at the actual flight height, thereby obtaining the actual offset amount corrected by the height amplification effect.

[0115] In another embodiment of the present application, by parsing the unmanned aerial vehicle flight log file, the latitude and longitude time stamp is screened out, and the spatial position point time stamp in the point cloud file is aligned, to ensure that the height information and the offset at the same time can be one-to-one corresponding.

[0116] In another embodiment of the present application, by selecting the reference point height as the reference in the same flight trajectory, the height values of other points are subtracted from the reference point height to obtain the point-by-point height difference.

[0117] In another embodiment of the present application, an index table is established, the height difference of each point cloud point is taken as the first column, the offset is taken as the second column, and the time sequence or spatial sequence is arranged, thereby forming a corresponding list of height and offset.

[0118] In another embodiment of the present application, a linear fitting method is used, the height difference is taken as the independent variable, and the offset is taken as the dependent variable. The slope obtained by fitting is the height amplification coefficient. If the data fluctuates greatly, a segmented fitting method can also be used to calculate the height amplification coefficient of different height ranges, thereby obtaining a more refined height amplification coefficient.

[0119] In another embodiment of the present application, the height amplification coefficient can be introduced into the point-by-point offset data, the original offset value is multiplied by the corresponding coefficient point by point to obtain the corrected offset, and then all the corrected offsets are statistically averaged or the maximum value is taken to obtain the actual offset result of the entire flight interval.

[0120] In another embodiment of the present application, it is assumed that the unmanned aerial vehicle performs a flight task in a certain area, the point cloud position offset at a flight height of 100 meters is 0.2 meters, the offset at 150 meters is 0.35 meters, and the offset at 200 meters is 0.55 meters. By comparison, it can be concluded that the height difference is proportional to the offset, and the height amplification coefficient obtained by linear fitting is 0.0035. When the original offset recorded by a certain point is 0.3 meters and the height is 160 meters, the actual offset obtained by correction calculation is , thereby realizing the correction of offset amplification.

[0121] Preferably, in step S3, each point cloud unit is dynamically compensated in combination with the time stamp to correct the survey area remote sensing data, including:

[0122] Time sequence the three-dimensional disturbance trajectory network into a trajectory disturbance sequence;

[0123] Query the missing value of the point cloud spatial offset deviation at any time point in the trajectory disturbance sequence based on the point cloud spatial offset deviation, and supplement the point cloud spatial offset deviation by continuous interpolation;

[0124] Connect the offset deviation of the supplementary point cloud space with the timestamp of each point cloud unit to identify the point-by-point deviation annotation data;

[0125] According to the first deviation spatial distribution data of the point-by-point deviation annotation data, the point-by-point translation correction data is generated by applying the direction and amplitude corresponding to the timestamp;

[0126] Based on the second deviation spatial distribution data of the point-by-point translation correction data, the direction difference and scale difference of each point cloud unit at the corresponding timestamp are calculated;

[0127] The point cloud coordinates in the point-by-point translation correction data are adjusted in combination with the direction difference and scale difference to correct the remote sensing data of the survey area.

[0128] In one implementation of the embodiment of the present application, the three-dimensional disturbance trajectory generated by the unmanned aerial vehicle during flight is organized into a trajectory disturbance sequence in time sequence, which can be synchronized with the point cloud acquisition process. On this basis, when there is a missing part in the trajectory disturbance data, the missing part is filled by continuous interpolation, making the entire trajectory sequence continuous and complete. The supplemented disturbance information is corresponded to the timestamp of the point cloud unit one by one, so as to obtain data with point-by-point deviation annotation. The direction and amplitude in these point-by-point deviation annotation data are used to correct the point cloud unit point by point, so that the data is preliminarily corrected in space. Through the second deviation spatial distribution data, the direction difference and scale difference of each point cloud unit at the corresponding timestamp are further calculated. In combination with the direction difference and scale difference, the point cloud coordinates are further adjusted, so as to improve the spatial accuracy of remote sensing data as a whole.

[0129] In another embodiment of the present application, the three-dimensional disturbance trajectory is converted into a trajectory disturbance sequence. The attitude angle and position change data recorded during the flight of the unmanned aerial vehicle are arranged in time sequence to convert them into a one-dimensional time sequence file.

[0130] In another embodiment of the present application, the missing value in the trajectory disturbance sequence can be supplemented by using linear interpolation method, i.e. finding the known values before and after the missing point, calculating the value of the missing point in the middle according to their time interval and value difference, or using cubic spline interpolation method to make the inserted value more smooth in curve change.

[0131] In another embodiment of the present application, a two-column correspondence table is established. The first column stores the timestamp of the point cloud unit, and the second column stores the offset deviation value at the corresponding time. Through this correspondence table, the error situation of each point cloud unit can be clearly seen, and an intuitive input structure for generating point-by-point deviation annotation data is provided.

[0132] In another embodiment of the present application, the direction information and amplitude information in the point-by-point deviation annotation data are applied to the original coordinates of the point cloud one by one, the direction information is used to determine the vector direction of the offset correction, and the amplitude information is used to determine the size of the correction, so that each point cloud point will obtain a separate correction result, and finally form a point-by-point translation correction data file.

[0133] In another embodiment of the present application, a vector difference calculation formula is used to obtain the direction difference by comparing the included angle of the point cloud vectors before and after correction, and the scale difference is calculated by the ratio of the lengths of the point cloud vectors before and after correction.

[0134] In another embodiment of the present application, the direction difference and the scale difference are superimposed on the original translation correction data point by point, the direction difference is used to correct the offset direction, and the scale difference is used to correct the scaling degree of the point cloud, so as to obtain the point cloud coordinate result which is more consistent with the real space structure.

[0135] In another embodiment of the present application, assuming that the time stamp of a certain point cloud unit recorded by the unmanned aerial vehicle during the flight in the survey area is 10 seconds, the height disturbance data is missing, the offset at 9 seconds is 0.15 meters, and the offset at 11 seconds is 0.25 meters, the compensation offset at 10 seconds can be calculated by linear interpolation as 0.20 meters, the value is paired with the time stamp 10 seconds to generate point-by-point deviation annotation data, and the direction and amplitude correction is applied to obtain the translation correction data, and the direction difference before and after correction is 2 degrees, and the scale difference is 1.03, finally the coordinates of the point are adjusted in direction and scale to obtain the corrected coordinate value, so as to realize the repair of the point cloud data.

[0136] Especially important is that the first deviation space allocation data according to the point-by-point deviation annotation data is applied point by point according to the corresponding direction and amplitude of the time stamp, including:

[0137] Analyzing the time stamp in the point-by-point deviation annotation data determines the deviation action time corresponding to each point cloud unit in the point-by-point deviation annotation data;

[0138] Extracting the three-dimensional direction vector and displacement amplitude of the first deviation space allocation data at each deviation action time;

[0139] Matching the three-dimensional direction vector and displacement amplitude with the original coordinates of the point cloud unit point by point, and repeatedly adjusting the original coordinates based on the three-dimensional direction vector and displacement amplitude to generate point-by-point translation correction data.

[0140] In one implementation manner of the embodiment of the present application, by analyzing the time stamp in the point-by-point deviation marking data, the point cloud unit and its corresponding deviation action time are established accurate correspondence, the three-dimensional direction vector and displacement amplitude in the first deviation space distribution data are extracted at each deviation action time, wherein the direction vector is used to indicate the spatial direction of the offset, and the displacement amplitude is used to quantify the size of the offset, the two parameters are combined with the original coordinates of the point cloud unit point by point, and in the operation process, the coordinates are moved according to the spatial direction indicated by the direction vector, and the moving distance is controlled by the displacement amplitude. In order to improve the accuracy, the adjustment of each point cloud unit is repeatedly calculated to ensure that the accumulated offset can fully reflect the actual disturbance effect, and after the point-by-point repeated calculation, the point-by-point translation correction data is generated.

[0141] In another embodiment of the present application, the time value is extracted one by one in the form of reading the time column in the point cloud data table, and the one-to-one correspondence between the point cloud unit and the time stamp is established. If the point cloud data is large, the point-by-point reading can be performed after sorting according to the time stamp.

[0142] In another embodiment of the present application, by reading the vector components X, Y, Z and the corresponding amplitude values in the deviation distribution file, they are converted into a three-dimensional array and stored in the memory, which is convenient for subsequent batch calculation with point cloud coordinates. The direction vector is used to determine the spatial direction of the correction, and the amplitude directly determines the correction distance.

[0143] In another embodiment of the present application, the method of vector addition is used to add the original coordinates of the point cloud unit and the direction vector bit by bit, and then multiply the displacement amplitude to obtain the corrected coordinates. For example, the original coordinates are (x, y, z), the direction vector is (dx, dy, dz), and the amplitude is m. Then the corrected coordinates are , which realizes the point-by-point spatial translation.

[0144] In another embodiment of the present application, the same offset formula is applied to each point cloud unit according to the corresponding time stamp, so that all point clouds can obtain corrected coordinates. If there are multiple deviations acting on the same point cloud unit, the offset can be superimposed multiple times according to the time sequence.

[0145] In another embodiment of the present application, the corrected point cloud coordinates are stored in a new data table, which includes point cloud number, time stamp, corrected X, Y, Z coordinates.

[0146] In another embodiment of the present application, assuming that the timestamp of a certain point cloud unit is 20 seconds, the original coordinates are (100, 200, 50), the three-dimensional direction vector at 20 seconds is (0.2, 0.1, 0.05), and the displacement amplitude is 2 meters, then the corrected coordinates according to the formula are (100+0.2*2, 200+0.1*2, 50+0.05*2), i.e. (100.4, 200.2, 50.1), and after storing the result in the point-by-point translation correction data file, the real position of the point after correction can be obtained.

[0147] Preferably, calculating the direction difference and the scale difference of each point cloud unit at the corresponding timestamp based on the second deviation spatial distribution data of the point-by-point translation correction data comprises:

[0148] Querying the second deviation spatial distribution data of each point cloud unit at the corresponding timestamp in the point-by-point translation correction data as point-by-point deviation data;

[0149] Decomposing the direction component of the point-by-point deviation data to extract point-by-point direction decomposition data around the X-axis, the Y-axis, and the Z-axis;

[0150] Determining the direction difference of each point cloud unit at the corresponding timestamp according to the point-by-point direction decomposition data;

[0151] Analyzing the scale difference value of the point-by-point deviation data in the X-axis, the Y-axis, and the Z-axis directions;

[0152] Determining the scale difference of each point cloud unit at the corresponding timestamp based on the scale difference value.

[0153] In one implementation manner of the embodiment of the present application, the second deviation spatial distribution data of each point cloud unit at the corresponding timestamp in the point-by-point translation correction data is queried and processed as point-by-point deviation data. The direction component of the point-by-point deviation data is decomposed to extract direction decomposition data around the X-axis, the Y-axis, and the Z-axis, respectively. The direction difference of each point cloud unit at the corresponding timestamp is calculated according to the point-by-point direction decomposition data. The scale difference value of the point-by-point deviation data in the X-axis, the Y-axis, and the Z-axis directions is analyzed. The scale difference of each point cloud unit at the corresponding timestamp is determined based on the scale difference value, so as to obtain the point-by-point direction difference and scale difference data.

[0154] In another embodiment of the present application, by reading the number and timestamp fields of each point cloud unit in the correction data table, the corresponding second deviation spatial distribution data is extracted, and a one-to-one mapping relationship between the point cloud unit and the timestamp is established, so as to ensure that the deviation data of each point cloud unit accurately corresponds to the time point.

[0155] In another embodiment of the present application, the point-by-point deviation vector is projected along three-dimensional coordinate axes to obtain directional components of X-axis, Y-axis and Z-axis respectively, for example, the deviation vector (dx, dy, dz) is resolved into three components dx, dy and dz respectively, so as to separately calculate the directional difference of each axis.

[0156] In another embodiment of the present application, the directional difference is determined by comparing the included angle between the point-by-point directional decomposition data and the reference directional vector, and the angle difference value can be calculated by using the vector dot product formula to obtain the deflection angle of each point cloud unit in X, Y and Z axes, which is used to represent the size and direction of the directional deviation.

[0157] In another embodiment of the present application, the amplitude variation of the point-by-point deviation data on the three coordinate axes is directly read as the scale difference value of the point cloud unit, for example, the absolute value or amplitude variation of dx, dy and dz, to form a scale difference matrix.

[0158] In another embodiment of the present application, the scale difference of each point cloud unit is determined based on the scale difference, and the scale differences in X, Y and Z directions can be integrated or weighted to obtain a comprehensive scale difference.

[0159] In another embodiment of the present application, assuming that the second deviation space allocation data of a point cloud unit at a time stamp of 10 seconds is a deviation vector (0.3, 0.1, 0.05), the X-axis directional difference is 0.3, the Y-axis directional difference is 0.1, and the Z-axis directional difference is 0.05, and the corresponding scale difference can also be resolved as 0.3, 0.1 and 0.05 by the same value, and through these directional differences and scale differences, the point cloud unit in the three-dimensional space can be accurately corrected.

[0160] Preferably, determining the directional difference of each point cloud unit at the corresponding time stamp according to the point-by-point directional decomposition data comprises:

[0161] sequentially arranging the point-by-point directional decomposition data into point-by-point directional time mapping data according to the time stamp of each point cloud unit;

[0162] corresponding the point-by-point directional time mapping data to the reference coordinate directions of X-axis, Y-axis and Z-axis respectively, and calculating the variation of the directional component of the point cloud unit under adjacent time stamps to obtain point-by-point directional difference data;

[0163] integrating the point-by-point directional difference data in a vector synthesis manner to identify the directional difference.

[0164] In one implementation form of the embodiment of the application, the point-by-point direction decomposition data is arranged in time sequence according to the time stamp of each point cloud unit to obtain point-by-point direction time mapping data, the point-by-point direction time mapping data is corresponded to the reference coordinate directions of the X axis, the Y axis and the Z axis, the change amount of the direction component under adjacent time stamps is calculated, point-by-point direction differential data is generated, the point-by-point direction differential data is integrated by vector synthesis, and the direction difference of the point cloud unit under the corresponding time stamp is identified, so as to obtain the direction difference information of each point cloud unit under the time sequence.

[0165] In another embodiment of the application, the data table of the point cloud unit is sorted in ascending order according to the time field, so that the point-by-point direction decomposition data strictly corresponds to the time stamp, and the calculation process of the direction change is continuous and stable.

[0166] In another embodiment of the application, the point-by-point direction time mapping data is projected onto the X, Y and Z three-axis directions, for example, a direction vector (vx, vy, vz) is directly decomposed into three coordinate direction components, so as to calculate the change amount of each axis subsequently.

[0167] In another embodiment of the application, the differential method can be used, that is, the direction components under adjacent two time stamps are subtracted, for example , so as to obtain point-by-point direction differential data.

[0168] In another embodiment of the application, the square sum and square root method is used to calculate the synthetic value of the direction difference, that is This method can integrate the differential data in the X, Y and Z three-axis directions into a total direction difference.

[0169] In another embodiment of the application, the axial component can be weighted before vector synthesis, for example, in some cases, the error of the Z-axis direction has a greater impact, and the Z-axis direction component can be amplified by setting a weight coefficient, so that the final direction difference calculation result is more in line with the characteristics of the actual flight environment.

[0170] In another embodiment of the application, it is assumed that the direction vector of a certain point cloud unit is (0.5, 0.2, 0.1) at the time stamp of 5 seconds, and the direction vector is (0.6, 0.25, 0.15) at the time stamp of 6 seconds, the direction component change amount is dx=0.1, dy=0.05, dz=0.05, and the direction difference is obtained by vector synthesis, and such direction difference result can be used for dynamic correction of the point cloud.

[0171] Therefore, the embodiments should be regarded, at any point, as being exemplary and not limiting, the scope of the application being defined by the appended claims and not by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

[0172] The foregoing is considered as illustrative only of the principles of the application. Numerous modifications and changes will readily occur to those skilled in the art, and the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for processing unmanned aerial vehicle (UAV) remote sensing data, characterized in that, Includes the following steps: Step S1: Collect raw remote sensing data of the survey area using a drone; The relative error caused by the ground height difference in the original remote sensing data of the survey area is eliminated by elevation normalization, and a coordinate system for the survey area is constructed. Step S2: Obtain the UAV trajectory sequence and attitude disturbance information generated during UAV flight; project the attitude disturbance information and UAV trajectory sequence into the same survey area coordinate system to build a three-dimensional disturbance trajectory network; Step S3: Identify the spatial offset deviation of point cloud in the original remote sensing data of the survey area through a three-dimensional perturbation trajectory network, and dynamically compensate each point cloud unit in combination with timestamps to correct the remote sensing data of the survey area. The spatial correlation bias of point clouds in the original remote sensing data of the survey area identified through the three-dimensional perturbation recording network includes: The location matching data is obtained by matching the temporal sequence of each point cloud unit in the original remote sensing data of the survey area with the UAV position at a certain time in the three-dimensional disturbance report network. Calculate the spatial offset between the theoretical flight position of the UAV in the position matching data and the actual recorded position in the UAV trajectory sequence, and identify the first deviation spatial allocation data; Based on the first deviation space allocation data, the second deviation space allocation data under the combined action is calculated by combining attitude disturbance information; Integrate the first deviation spatial allocation data and the second deviation spatial allocation data to calculate the point cloud spatial offset deviation; The calculation of the second deviation space allocation data under the combined effect, based on the first deviation space allocation data and combined with attitude disturbance information, includes: Calculate the change in pitch angle in attitude disturbance information; The effect of the change in pitch angle on the laser beam pointing is quantified, and the laser beam pointing deviation is analyzed in combination with the change. Convert the laser beam pointing deviation into an offset of the point cloud position; The actual offset is calculated based on the amplification effect of the flight altitude at the actual recorded location on the offset. Calculate the second deviation space allocation data under the combined action by combining the actual offset and the first deviation space allocation data; The quantification of the effect of the change in pitch angle on the laser beam pointing, and the analysis of laser beam pointing deviation based on the change, include: Arrange the changes in pitch angle according to the timestamps to obtain the longitudinal pitch angle change sequence; Calculate the effect of the longitudinal elevation angle change sequence on the pitch offset of the laser beam direction, and determine the longitudinal offset component data under the action of the longitudinal elevation angle; Based on the change in pitch angle, the longitudinal offset component data is superimposed point by point until the comprehensive offset component under each timestamp is obtained. The laser beam pointing deviation is determined by integrating the offset components; The calculation of the actual offset based on the amplification effect of the flight altitude in the actual recorded location includes: Extract flight altitude records from the UAV flight data that match the actual recorded location nodes, and pair the offset of the point cloud location with the flight altitude records at the same time node to generate offset and altitude correspondence data; Calculate the height difference of each point cloud unit based on the offset and height correspondence data; The height difference and offset are matched point by point to form point-by-point offset and height matching data; The height magnification factor is identified by point-by-point offset and height matching data; The actual offset after magnification correction is calculated by multiplying the height magnification factor by the offset point by point. Step S4: matching the corrected survey area remote sensing data with the corresponding point cloud unit coordinates in the survey area coordinate system; and reconstructing a three-dimensional survey area framework based on the point cloud unit coordinates, the survey area coordinate system, and the corrected survey area remote sensing data. 2.The UAV remote sensing data processing method of claim 1, wherein, The spatial offset between the theoretical flight position of the unmanned aerial vehicle in the position matching data and the actual recorded position in the unmanned aerial vehicle trajectory sequence is calculated, and the first deviation spatial distribution data is identified, including: Extracting the time-corresponding theoretical flight position of the unmanned aerial vehicle from each three-dimensional disturbance monitoring network, and synchronously reading the actual recorded position at the same time; Calculating the three-dimensional Euclidean distance between the theoretical flight position of the unmanned aerial vehicle and the actual recorded position as the spatial offset; Projecting the spatial offset into the ground coordinate system to obtain the first deviation spatial distribution data. 3.The method of claim 2, wherein, The dynamic compensation of each point cloud unit in step S3 in combination with the time stamp to correct the survey area remote sensing data includes: Time-sequencing the three-dimensional disturbance trajectory network into a trajectory disturbance sequence; Querying the missing values of the point cloud spatial offset deviation at any time point in the trajectory disturbance sequence, and supplementing the point cloud spatial offset deviation through continuous interpolation; Connecting the supplemented point cloud spatial offset deviation with the time stamp of each point cloud unit to identify the point-by-point deviation annotation data; According to the first deviation spatial distribution data of the point-by-point deviation annotation data, the point-by-point translation correction data is generated by applying the direction and amplitude corresponding to the time stamp; Based on the second deviation spatial distribution data of the point-by-point translation correction data, the direction difference and the scale difference of each point cloud unit at the corresponding time stamp are calculated. Adjusting the point cloud coordinates in the point-by-point translation correction data in combination with the direction difference and the scale difference to correct the survey area remote sensing data. 4.The method of claim 2, wherein, Based on the second deviation spatial distribution data of the point-by-point translation correction data, the direction difference and the scale difference of each point cloud unit at the corresponding time stamp are calculated, including: Querying the second deviation spatial distribution data of each point cloud unit at the corresponding time stamp in the point-by-point translation correction data as the point-by-point deviation data; Decomposing the direction component of the point-by-point deviation data to extract the point-by-point direction decomposition data around the X-axis, Y-axis, and Z-axis; Determining the direction difference of each point cloud unit at the corresponding time stamp according to the point-by-point direction decomposition data; Analyzing the scale difference value of the point-by-point deviation data in the X-axis, Y-axis, and Z-axis directions; Determining the scale difference of each point cloud unit at the corresponding time stamp based on the scale difference value.

5. The unmanned aerial vehicle remote sensing data processing method according to claim 4, characterized in that, Determining the direction difference of each point cloud unit at the corresponding time stamp according to the point-by-point direction decomposition data includes: Arranging the point-by-point direction decomposition data in sequence into point-by-point direction time mapping data according to the time stamp of each point cloud unit; Corresponding the point-by-point direction time mapping data to the reference coordinate directions of the X-axis, Y-axis, and Z-axis, respectively, and calculating the change amount of the direction component of the point cloud unit at adjacent time stamps to obtain the point-by-point direction difference data; Integrating the point-by-point direction difference data in a vector synthesis manner to identify the direction difference.

6. A UAV remote sensing data processing system, characterized in that, The unmanned aerial vehicle remote sensing data processing system for executing the unmanned aerial vehicle remote sensing data processing method as claimed in claim 1, comprising: A data acquisition module for collecting original survey area remote sensing data through an unmanned aerial vehicle; eliminating the relative error caused by the ground height difference in the original survey area remote sensing data through height normalization, and constructing a survey area coordinate system; The disturbance analysis module is configured to acquire a UAV trajectory sequence and attitude disturbance information generated in a UAV flight process; project the attitude disturbance information and the UAV trajectory sequence into a same survey area coordinate system, and build a three-dimensional disturbance trajectory network; The offset correction module is configured to identify a point cloud spatial offset deviation in original survey area remote sensing data through the three-dimensional disturbance trajectory network, and dynamically compensate each point cloud unit in combination with a time stamp, so as to correct the survey area remote sensing data; The remote sensing data reconstruction module is configured to match the corrected survey area remote sensing data and a corresponding point cloud unit coordinate in the survey area coordinate system; and reconstruct a three-dimensional survey area framework according to the point cloud unit coordinate, the survey area coordinate system and the corrected survey area remote sensing data.

Images