Glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence

Abstract

The invention discloses a glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence, particularly relates to the field of unmanned aerial vehicle inspection, and is used for solving the problem that curvature optimization and dynamic airspace management lack collaboration in airport glass curtain wall inspection. Track generation immediately inherits real-time airspace information, safety monitoring in a subsequent flight stage triggers a pause signal by distance-attitude double features, track correction only performs normal translation on an unexecuted section and synchronizes to a control tower, global recalculation is avoided, experience iteration is continuously written into a learning library, and safety bandwidth is adaptively converged along with task accumulation; multiple units are connected end to end to form a wall-attached inspection-airspace scheduling collaborative closed chain, the visual axis of a camera is kept to be attached to the curtain wall, a dynamic corridor is avoided in real time, inspection is not interrupted, scheduling is zero-interference, the algorithm load is stabilized at the millisecond level, the flight path pause time of one-time inspection of the airport curtain wall is integrally and remarkably shortened, and the manual intervention frequency is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of unmanned aerial vehicle inspection, more particularly, the present application relates to a glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence. BACKGROUND

[0002] In the glass curtain wall outer facade of the airport terminal, the inspection operation relies on the unmanned aerial vehicle to fly along the double curvature wall, and the heading is output in real time by the curvature smoothing algorithm, so that the camera visual axis is always perpendicular to the glass surface. At the same time, the airport operation command center will repeatedly adjust the clearance corridor around the curtain wall according to the flight in and out and the ground vehicle walking condition, and the corridor boundary will shrink or expand outward according to the dispatch rhythm, especially near the corner and concave-convex nodes, which is extremely close to the inspection route, forming a dynamic interleaved airspace situation.

[0003] The current automatic inspection method only pays attention to curvature smoothing and endurance consumption when planning the route, and does not take the real-time change of the clearance corridor into account. Once the wall-attached trajectory generated by the algorithm is fixed, it will not be updated, and when the clearance boundary approaches the curtain wall, the unmanned aerial vehicle still advances according to the original trajectory, finally crosses into the controlled airspace, causing the tower to interrupt the inspection and return, and the flight scheduling is also forced to be rearranged. The root cause of the problem is that the curvature optimization and airspace management lack coordination, simply expanding the safety distance cannot fundamentally solve the conflict, and the dynamic clearance model must be integrated into the route generation stage to enable the curtain wall inspection and airport operation to coexist for a long time.

[0004] In order to solve the above problems, a technical scheme is provided. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence, which solves the problem of lack of coordination between curvature optimization and dynamic airspace management in airport glass curtain wall inspection. This is achieved by first integrating the curvature grid and the clearance profile in a unified coordinate frame, so that the flight path generation immediately inherits real-time airspace information, the safety monitoring in the subsequent flight stage triggers a pause signal with distance-attitude dual features, the trajectory correction only implements normal translation for the unexecuted segment and synchronously gives the tower, avoiding global recalculation, and the experience iteration continuously writes into the learning library, so that the safety bandwidth adaptsively converges with task accumulation; the multi-unit head-to-tail connection forms a coordinated closed chain of wall-attached inspection-airspace scheduling, which not only keeps the camera visual axis attached to the curtain wall but also avoids the dynamic corridor in real time, thereby solving the problems raised in the background art.

[0006] To achieve the above purpose, the present application provides the following technical scheme: The coordinate fusion unit: before taking off, the curtain wall curvature grid and the clearance corridor profile are superimposed under the unified coordinate reference, a shared coordinate table is generated and loaded into the flight terminal; Trajectory generation unit: read shared coordinate table, generate mixed straight line and arc trajectory according to minimum curvature principle, and write minimum safety zone at each node; Safety monitoring unit: during flight, output residual distance curve and angular velocity signal synchronously with radar and inertial navigation, extract key features to perform safety judgment, if the result exceeds the standard, pause and request the latest corridor profile; Trajectory correction unit: after the ground station pushes the latest corridor profile, the path planner translates the unexecuted trajectory along the normal direction and broadcasts the correction result for tower monitoring, and the flight control then resumes execution of the updated trajectory; Experience iteration unit: after landing, write flight records into the learning library, preset safety bandwidth for the next task and reduce online downtime.

[0007] In a preferred embodiment, the coordinate fusion unit includes the following: Before takeoff, extract the curtain wall curvature grid containing position coordinates and local curvature values from the airport building database, obtain the clearance corridor profile composed of polygon boundary line segments from the airport operation command center, verify the consistency of point cloud density, select the reference point set to calculate the transformation matrix to minimize the sum of Euclidean distances to establish a unified coordinate reference, calculate the projection distance of each curtain wall curvature grid point to the clearance corridor profile to identify the overlapping area and generate a fusion grid, extract the shared coordinate table organized by fusion grid points, and verify and load it into the flight terminal storage module through a wireless transmission protocol.

[0008] In a preferred embodiment, the trajectory generation unit includes the following: The trajectory generation unit extracts the shared coordinate table from the storage module, verifies the integrity of the matrix row number, filters points with curvature values exceeding the preset threshold as curvature key points, calculates the minimum curvature path between adjacent curvature key points to generate a mixed straight line and arc trajectory, calculates the shortest normal distance of each path point to the clearance corridor profile to quantify the clearance and obtain the minimum safety bandwidth value, and appends the minimum safety bandwidth value to the path point sequence as a floating point number format attribute field.

[0009] In a preferred embodiment, the shared coordinate table is a matrix, each row corresponding to a point, including unified coordinates, curvature values, and profile flags; The unified coordinates refer to the three-dimensional position coordinate values of each point in the fusion grid under the unified coordinate reference, including x component, y component, and z component; The curvature value refers to the local curvature value derived from the curtain wall curvature grid, which directly inherits the original local curvature value for points derived from the curtain wall curvature grid, and fills in zero for points derived from the clearance corridor profile to represent no curvature information; The contour mark refers to a binary identification value, which is assigned a value of one for a point belonging to the boundary of the clearance corridor contour and a value of zero for a point not belonging to the boundary of the clearance corridor contour.

[0010] In a preferred embodiment, the safety monitoring unit comprises the following: During flight, the residual distance readings are collected by the distance radar to form a residual distance curve, the three-axis angular velocity signals are recorded by the inertial navigation sensor, the sequence data are aligned in time clock, the key features are extracted, including the margin contraction speed ratio and the attitude amplitude density, the pre-trained lightweight random forest inputs the margin contraction speed ratio and the attitude amplitude density to output the attitude stability adjustment coefficient and determine whether it is over-standard, if yes, the latest clearance corridor contour is requested and the high-precision coordinates are locked.

[0011] In a preferred embodiment, the margin contraction speed ratio calculates the slope difference of adjacent time points based on the residual distance curve, and selects the maximum negative slope value to quantify the clearance margin reduction rate.

[0012] In a preferred embodiment, the attitude amplitude density applies a short window fast Fourier transform to the three-axis angular velocity signals, integrates the part exceeding the medium frequency threshold after obtaining the frequency spectrum amplitude, to evaluate the attitude oscillation intensity.

[0013] In a preferred embodiment, the trajectory correction unit comprises the following: After the ground station pushes the latest clearance corridor contour, the boundary line segment integrity is verified, the unexecuted flight path sub-sequence is extracted based on the current UAV position binary search matching index, the normal distance of each point to the latest clearance corridor contour is calculated to translate the point position to generate the corrected unexecuted flight path sub-sequence, the point position of the corrected unexecuted flight path sub-sequence and the minimum safety bandwidth value are sent, and the tower station confirms to replace the original unexecuted flight path sub-sequence and release the pause to start the advance after synchronizing the timing.

[0014] In a preferred embodiment, the experience iteration unit comprises the following: After landing, the flight record is extracted and the sequence integrity is verified, the flight record is appended to the learning library and the structured field is defined, the multi-task flight record is queried to calculate the preset safety bandwidth intention index decay, the minimum safety bandwidth value of the shared coordinate table is updated and version marking control is performed, and the buffer threshold limit is expanded during flight.

[0015] The technical effects and advantages of the glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence are as follows: The application fuses the curvature grid and the clearance profile in a unified coordinate frame in advance, the flight path generation immediately inherits real-time airspace information, the subsequent flight phase safety monitoring triggers a pause signal with distance-attitude double features, the trajectory correction only implements normal translation for the unexecuted segment and synchronously gives the tower, avoiding global recalculation, the experience iteration continuously writes into the learning library, making the safety bandwidth adaptively converge with task accumulation; the multi-unit head-tail connection forms a cooperative closed chain of wall-hugging inspection-airspace scheduling, which not only keeps the camera visual axis adhering to the curtain wall but also avoids dynamic corridors in real time, the inspection is uninterrupted, the scheduling is zero-interference, and the algorithm load is stable at the millisecond level, which significantly shortens the flight path pause time of airport curtain wall once inspection and reduces the frequency of manual intervention. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a structural schematic diagram of the glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0018] Embodiment 1: Figure 1 The application provides a glass curtain wall unmanned aerial vehicle automatic airport inspection system based on artificial intelligence, which comprises: A coordinate fusion unit: before taking off, the curtain wall curvature grid and the clearance corridor profile are superimposed under a unified coordinate reference to generate a shared coordinate table and load it to a flight terminal.

[0019] A flight path generation unit: the shared coordinate table is read, a wall-hugging straight line and circular arc hybrid flight path is generated according to the minimum curvature principle, and the minimum safety bandwidth is written at each node.

[0020] A safety monitoring unit: during flight, the residual distance curve and angular velocity signal are synchronously output from the distance radar and the inertial navigation system, key features are extracted to perform wall-hugging safety judgment, and if the judgment result is out of standard, the progress is paused and the latest corridor profile is requested.

[0021] A trajectory correction unit: after the ground station pushes the latest corridor profile, the path planner corrects the unexecuted flight path along the normal direction, broadcasts the correction result for the tower to monitor, and the flight control subsequently releases the pause to execute the updated flight path.

[0022] An experience iteration unit: after landing, the flight record is written into the learning library, the safety bandwidth is preset for the next task, and the online pause is reduced.

[0023] In the inspection operation of the glass curtain wall exterior of the airport terminal building, the UAV needs to fly along the double curvature surface to ensure that the camera's visual axis is always perpendicular to the glass surface. At the same time, the airport operation command center adjusts the clearance corridor boundary around the curtain wall in real time according to the arrival and departure of flights and the movement of ground vehicles. These boundaries are extremely close to the inspection route at the corner and concave-convex nodes, forming a dynamic and interlaced airspace environment. The current inspection method only focuses on smooth curvature and endurance consumption, and does not take into account the real-time changes of the clearance corridor in the route planning constraints, resulting in the generated wall-clinging trajectory being fixed and unable to adapt to the boundary approximation, causing the UAV to enter the controlled airspace and interrupt the inspection.

[0024] To solve the problem of lack of coordination between curvature optimization and airspace management, the coordinate fusion unit integrates the curtain wall curvature grid and the clearance corridor profile under the unified coordinate reference before takeoff, generates a shared coordinate table and loads it to the flight terminal, thereby integrating the dynamic clearance model at the route generation stage and ensuring that the subsequent inspection route inherits real-time airspace information from the source, achieving long-term coexistence of inspection and airport operation.

[0025] The processing method of this scheme is to integrate the curtain wall curvature grid and the clearance corridor profile in the same coordinate frame, avoid the deviation caused by independent coordinates, and generate a shared coordinate table that can be directly called by downstream units, thereby providing a unified basis for curvature and airspace constraints for route generation.

[0026] The processing of the coordinate fusion unit is divided into the following steps: 1-1. The process of collecting the curtain wall curvature grid and the clearance corridor profile is to obtain the surface geometry information and airspace boundary data required for inspection from the source, thereby laying a accurate input basis before fusion.

[0027] Before takeoff, the curtain wall curvature grid is extracted from the airport building database, which is represented by a three-dimensional point cloud, and each point contains position coordinates and local curvature value. At the same time, the clearance corridor profile is obtained from the airport operation command center in real time, which is composed of polygon boundary line segments, and each line segment is labeled with start and end point coordinates. When extracting, verify the consistency of point cloud density by comparing the average distance between adjacent points to ensure that the grid resolution matches the length of the contour line segment, avoiding scale deviation during fusion.

[0028] This step outputs the original data set of the curtain wall curvature grid and the clearance corridor profile, which facilitates the direct reference of these point information by subsequent mapping processing, and realizes the reliable preparation of input data.

[0029] 1-2. Establish a unified coordinate reference to eliminate coordinate deviations between data from different sources and ensure that the integration operation is performed in the same framework, thereby improving the geometric consistency of the fusion.

[0030] The curtain wall curvature grid and the clearance corridor profile are mapped into the airport global coordinate frame, and the reference points are aligned by the least square method. Specifically, a set of reference points in the curtain wall curvature grid and a corresponding set of reference points in the clearance corridor profile are selected, and a transformation matrix is calculated to minimize the sum of the Euclidean distances between the two sets of points. In the calculation, the transformation matrix is first initialized as an identity matrix, and then the matrix parameters are iteratively adjusted to minimize the sum of the squared distances between each reference point after transformation and the corresponding point. This summing process involves accumulating the distance calculations for all pairs of reference points, and gradually approaching the optimal solution through gradient descent, where the distance calculation is the square root of the sum of the squares of the transformed coordinates minus the target coordinates.

[0031] This step outputs the curtain wall curvature grid and the clearance corridor profile under the unified coordinate reference, ensuring that the downstream overlay directly uses these aligned point positions, avoiding airspace misjudgment caused by deviation accumulation.

[0032] 1-3. Overlay the curtain wall curvature grid and the clearance corridor profile to identify the geometric overlapping area, thereby labeling potential conflict points in the shared table and achieving preliminary integration of curvature and airspace.

[0033] Based on the curtain wall curvature grid and the clearance corridor profile under the unified coordinate reference, geometric overlay is performed in three-dimensional space to identify overlapping areas. Specifically, the projection distance of each curtain wall curvature grid point within the clearance corridor profile is calculated. This distance is calculated by first projecting the point onto each profile segment, with the projection point being the perpendicular foot coordinate from the grid point to the segment. Then, the straight-line distance from the grid point to the projection point is calculated, and the minimum distance value among all segments is taken. If the minimum distance value is negative or zero, it is determined that the point is located within the clearance corridor profile.

[0034] This step outputs the fused grid after overlaying, and labels the overlapping state of each point, facilitating the inheritance of these labeling information by subsequent table generation and improving the response accuracy of the inspection path to dynamic boundaries.

[0035] 1-4. Generate a shared coordinate table to organize the fused information into a structured format, thereby facilitating efficient reading of curvature and profile constraints by flight terminals and achieving standardization of data sharing.

[0036] Based on the fused grid after overlaying, the coordinates, curvature values, and profile boundary information of all points are extracted and organized into a table structure. Specifically, the shared coordinate table is a matrix, with each row corresponding to a point, including the unified coordinates, the curvature value (if derived from the curtain wall curvature grid, the original value is directly inherited, otherwise, it is filled with zero to represent no curvature information), and the profile flag (assigned a value of one to indicate that it belongs to the clearance corridor profile boundary, and a value of zero to indicate a non-boundary point). The filling rule ensures that the curvature value is only valid at the grid source point, and the flag bit is binary based on the overlapping state.

[0037] This step ensures that the shared coordinate table contains complete fusion information, making it easy for downstream units to directly parse the matrix row and achieve seamless information transmission.

[0038] 1-5. Load the shared coordinate table to the flight terminal and implant the fusion data into the unmanned aerial vehicle control module in real time to support immediate flight path adjustment after takeoff and achieve synchronization of ground and air data.

[0039] Based on the generated shared coordinate table, upload it to the storage module of the unmanned aerial vehicle flight terminal through the wireless transmission protocol. Add a check code during transmission for verification, i.e., calculate the hash value of the table data and attach it for sending. The receiving end recalculates the hash to confirm the integrity.

[0040] The coordinate fusion unit realizes the unified superposition of the curtain wall curvature grid and the clearance corridor profile before takeoff, ensuring that the generated shared coordinate table inherits the curvature constraints and dynamic airspace boundary information under the airport glass curtain wall inspection scenario, thereby providing a fusion basis for subsequent flight path generation and avoiding airspace conflicts caused by independent coordinates.

[0041] The coordinate fusion unit realizes the unified superposition of the curtain wall curvature grid and the clearance corridor profile before takeoff, and generates a shared coordinate table loaded to the flight terminal, thereby providing a unified data basis for the fusion of curvature constraints and dynamic airspace boundaries for inspection route planning. However, the current method ignores clearance changes during the planning phase, leading to frequent conflicts after trajectory fixation. Therefore, the flight path generation unit generates a wall-hugging straight line and circular arc hybrid trajectory by reading the shared coordinate table and following the minimum curvature principle, while writing the minimum safety zone at the nodes to ensure that the generated trajectory inherits real-time airspace information from the source, achieving preliminary coordination of camera visual axis vertical adhesion and dynamic clearance corridor avoidance in glass curtain wall inspection.

[0042] The processing of the flight path generation unit is divided into the following steps: 2-1. Read the shared coordinate table to obtain the fused curvature and profile information, thereby directly incorporating the curtain wall geometry and airspace constraints in the flight path planning.

[0043] The flight terminal extracts the shared coordinate table from the storage module. The shared coordinate table is a matrix structure, each row containing unified coordinates, curvature values, and profile flags. The unified coordinates are three-dimensional position coordinate values, the curvature values are local curvature numerical values or zero padding, and the profile flags are binary identification values. When extracting, the number of matrix rows is checked to match the expected total number, ensuring data integrity.

[0044] 2-2. Extract the curvature key points to identify high curvature areas, thereby guiding the selection of straight line and circular arc hybrid trajectory segmentation points.

[0045] Based on the complete matrix of the shared coordinate table, traverse each row of curvature values, and filter out the points with curvature values exceeding the preset threshold as curvature key points. This facilitates the downstream generation of using these points as the turning positions of the flight path to avoid the deviation of the visual axis caused by insufficient smoothing.

[0046] 2-3. Generate a mixed flight path of straight lines and circular arcs that adhere to the glass surface, and minimize the cumulative curvature to ensure that the visual axis of the camera is always perpendicular to the glass surface during the flight of the UAV.

[0047] Based on the set of curvature key points, connect adjacent points to form a mixed trajectory based on the principle of minimum curvature. Specifically, calculate the minimum curvature path between adjacent curvature key points, and the total length is obtained by adding the length of the straight line segment and the length of the circular arc segment. The length of the straight line segment is approximated by dividing the path into small segments along one coordinate axis, calculating the square of the slope between two points for each small segment, and then taking the square root of the sum of all small segments. The slope is obtained by dividing the difference between the adjacent points on the other two coordinate axes by the difference on that axis. The length of the circular arc segment is obtained by multiplying the radius of each circular arc by the angular radian. The radius and angular radian are iteratively adjusted to minimize the total curvature integral, i.e., repeatedly calculating the curvature of the path under different parameters and selecting the minimum one.

[0048] This calculation ensures that the straight line segment dominates in the low curvature area, and the circular arc segment smoothly transitions in the high curvature area, with the unit being length. This step outputs the path point sequence of the mixed flight path of straight lines and circular arcs that adhere to the glass surface, which facilitates the direct insertion of safety information along these points in subsequent node processing, and improves the adhesion of the trajectory to the hyperbolic surface of the curtain wall.

[0049] 2-4. Calculate the minimum safety zone to quantify the clearance at the nodes of the flight path, thereby preventing boundary intrusion when the dynamic corridor shrinks.

[0050] Based on the path point sequence of the mixed flight path of straight lines and circular arcs that adhere to the glass surface, calculate the shortest distance from each path point to the clearance corridor contour. Specifically, use the contour markers in the shared coordinate table to identify boundary points, traverse all points with contour markers equal to one, calculate the Euclidean distance between the uniform coordinate components of each path point and the uniform coordinate components of the boundary points, select the minimum Euclidean distance, and then multiply it by the cosine value of the angle between the normal vector of the path point and the distance vector. The cosine value is obtained by multiplying the corresponding components of the two vectors and dividing the product of the respective modules, ensuring that it is projected along the normal.

[0051] This calculation ensures that the safety zone is projected along the normal, with the unit being length. Output the minimum safety zone width value for each path point to facilitate the direct addition of these values in the write operation, and implement buffer protection for the airspace intersection near the corner concave-convex nodes.

[0052] The minimum safety band refers to the shortest normal distance from each path point of the mixed straight-line and circular arc flight path to the clearance corridor contour, which is used to quantify the buffer margin between the UAV and the dynamic airspace boundary. Specifically, the minimum safety band is determined by calculating the vertical distance from the uniform coordinates of the path point to the nearest clearance corridor boundary line, the distance is projected along the curtain surface normal, the unit is length, which ensures that enough space is provided during the inspection to avoid trajectory conflicts caused by the shrinkage or expansion of the clearance corridor, thereby maintaining the vertical fit of the camera visual axis and flight safety.

[0053] 2-5. Write the minimum safety band to the node, embed the safety information into the flight path structure, so as to provide real-time judgment of the margin for the downstream monitoring unit. Based on the minimum safety band width value of each path point, these values are attached to the path point sequence of the mixed straight-line and circular arc flight path to form extended path data. Specifically, the minimum safety band width value is appended as an attribute field after each node of the path point sequence, and the field is defined as a floating point number format for unified storage and reading. This step ensures the completeness of the extended path data, which facilitates the safety monitoring during flight to directly read these fields and realize the dynamic adaptability of the inspection flight path.

[0054] The flight path generation unit realizes the generation of the mixed straight-line and circular arc flight path based on the shared coordinate table, and writes the minimum safety band at the node, ensuring the comprehensive integration of flight path planning from curvature minimization to airspace margin buffering in the airport glass curtain wall inspection scenario, thereby avoiding trajectory conflicts caused by changes in the clearance corridor.

[0055] The flight path generation unit reads the shared coordinate table to generate the mixed straight-line and circular arc flight path and writes the minimum safety band, thereby providing an initial path that inherits the curvature and airspace constraints for the flight phase. However, the current inspection ignores real-time clearance changes during flight, leading to frequent trajectory conflicts. Therefore, the safety monitoring unit outputs the residual distance curve and angular velocity signal synchronously and extracts key features to perform judgment, ensuring the stability of the UAV attitude and the immediate response of the dynamic corridor avoidance in the glass curtain wall inspection.

[0056] The key features include the margin shrinkage speed ratio and the attitude amplitude density.

[0057] The processing of the safety monitoring unit is divided into the following steps: 3-1. Synchronously output the residual distance curve and angular velocity signal to capture the instantaneous distance change of the UAV to the clearance corridor boundary and the attitude dynamics during flight, thereby integrating the distance radar and inertial navigation data in monitoring.

[0058] During flight, the residual distance readings of the clearance corridor boundary are continuously collected by the distance radar, forming a residual distance curve composed of time series distance values; at the same time, the inertial navigation sensor records the three-axis angular velocity signals in real time, including the angular velocity values of the roll, pitch and yaw axes. During collection, the clock is synchronized to align the residual distance readings and the three-axis angular velocity signals to the millisecond level, avoiding the influence of timing deviation on the accuracy of feature extraction. The output residual distance curve and three-axis angular velocity signal facilitate real-time tracking of dynamic airspace interlacing.

[0059] 3-2. Calculate the residual shrinkage speed ratio to quantify the rate of clearance reduction, so as to identify the risk of boundary approaching. Based on the residual distance curve, the slope difference of adjacent time points is calculated, and the maximum negative slope is selected as the residual shrinkage speed ratio. The specific calculation is to first take the derivative of the residual distance curve point by point, that is, the slope is obtained by dividing the difference between the next distance value and the current distance value by the time interval at each time point, then the difference value of adjacent slopes is calculated, only the negative difference value sequence is retained, and the maximum value is selected from it as the residual shrinkage speed ratio, wherein the negative difference value represents acceleration of shrinkage, ensuring that the accelerated shrinkage is captured. Output the residual shrinkage speed ratio, which facilitates the use of the residual shrinkage speed ratio value by downstream features to improve the sensitivity to corridor shrinkage near the corner node.

[0060] 3-3. Calculate the attitude wave amplitude density to evaluate the strength of attitude oscillation, so as to detect the stability deviation in the wall-attached flight.

[0061] Based on the three-axis angular velocity signal, a short window fast Fourier transform is applied to each axis signal to obtain the frequency spectrum amplitude, and then the amplitude values exceeding the medium frequency threshold are integrated as the attitude wave amplitude density. First, the three-axis angular velocity signal is divided into short time windows, and the fast Fourier transform is performed in each window to obtain the frequency domain amplitude spectrum, then the part of the amplitude spectrum with a frequency higher than the medium frequency threshold is selected, and the integral sum of these amplitudes from the medium frequency threshold to infinity is calculated, and the average of the three-axis integral results is taken as the attitude wave amplitude density, ensuring the quantification of high-frequency disturbance.

[0062] Output the attitude wave amplitude density to facilitate the judgment of the directly input attitude wave amplitude density value, and realize the accurate monitoring of the flight attitude of the hyperbolic curvature surface.

[0063] 3-4. Perform wall-attached safety judgment, evaluate the attitude stability state comprehensively, and decide whether to trigger the pause.

[0064] Based on the residual shrinkage speed ratio and the attitude wave amplitude density, as input to the lightweight random forest model, the attitude stability adjustment coefficient is output. If the attitude stability adjustment coefficient exceeds the preset limit value, it is determined to be out of limit.

[0065] The construction process of the lightweight random forest is based on ensemble learning, combining multiple decision trees to form an efficient model to meet real-time response requirements, such as millisecond-level computation. In the initial stage of construction, a bootstrap sample is generated from the training dataset, i.e., a subset of the same size is extracted from the original dataset by sampling with replacement, so that some data points may be repeated, and about one-third of the data is left as an out-of-bag sample for subsequent validation. Each decision tree is built independently, first randomly selecting a subset of features from the bootstrap sample to reduce tree correlation and improve generalization. The construction of the decision tree starts from the root node, based on the input features, i.e., the residual contraction rate and the attitude amplitude density, using a pre-set branch threshold for recursive splitting: specifically, for each node, evaluate whether the feature value exceeds the pre-set threshold, if yes, continue splitting along the "yes" branch, otherwise along the "no" branch; the splitting criterion uses mean squared error as a measure to ensure that each split reduces the variance of the data within the node. The recursive process continues until the leaf node conditions are met, such as the number of node samples being below the minimum sample splitting threshold or reaching the maximum depth limit, forming a shallow tree structure to keep the model lightweight. After all decision trees are constructed, they are integrated into a random forest through an aggregation mechanism, with the number of trees controlled within a small range to optimize computational efficiency and ensure the overall model outputs values quickly in attitude stability evaluation.

[0066] The training algorithm of the lightweight random forest uses a bootstrap aggregation method combined with feature randomness to achieve efficient learning and minimize the risk of overfitting. Training starts with dataset division, where the original dataset is used to generate multiple bootstrap samples, each created by sampling with replacement to introduce diversity. For each decision tree, the classification and regression tree algorithm is used for training: starting from the root node, calculate the reduction in mean squared error of the node for each possible branch threshold, select the branch threshold that results in the maximum reduction for splitting; for the two input features, residual contraction rate and attitude amplitude density, the pre-set branch threshold is used as the splitting point, recursively assigning data subsets to child nodes until the leaf nodes, which store the average value as the output. During training, each tree only considers a randomly selected subset of features, usually the square root of the total number of features, to reduce computational complexity and tree correlation. The out-of-bag sample is used for internal cross-validation: for each tree, input the out-of-bag sample that did not participate in training into the model to calculate the prediction error to evaluate performance. After all trees are trained, the prediction results are aggregated by averaging all tree outputs to form the attitude stability adjustment coefficient as the final output. This training mechanism ensures that the model converges quickly under limited resources, suitable for real-time processing of distance and attitude features in flight monitoring scenarios.

[0067] The optimization of the lightweight random forest focuses on improving the prediction accuracy and computational efficiency, reducing overfitting and highlighting key features through various techniques. First, variance is reduced by aggregating the independence between trees: each tree is trained on independent bootstrap samples, and the average output smooths the noise of individual trees, ensuring the robustness of the overall model. Second, feature importance is evaluated to reduce input: using the mean reduction impurity method, the model accuracy decreases after excluding the residual contraction speed or attitude amplitude density; or using permutation importance, the contribution is quantified by randomly permuting feature values in the out-of-bag samples and observing the average decrease in accuracy, so as to prioritize high importance features. The optimization also includes handling missing values: for inputs such as attitude amplitude density that may be missing, use feature randomness to estimate the missing values and maintain the integrity of the model. In addition, hyperparameter tuning techniques are used, such as grid search systematically traverses the combination of tree number and maximum depth, and evaluates performance metrics such as mean squared error on the validation set to select the parameter set that minimizes the error; or random search samples in the parameter space to efficiently explore a wide range of settings. This optimization process ensures that the model maintains high accuracy in millisecond-level response while adapting to the real-time needs of dynamic airspace monitoring.

[0068] The parameter settings of the lightweight random forest aim to balance model complexity and computational overhead, and all parameters are fixed before training to support efficient deployment. The main parameters include node size, which is the minimum number of samples in a leaf node, used to control tree depth and prevent overfitting, usually set to a small value to allow shallow tree structure; the number of trees, which is the total number of decision trees in the ensemble, is set to a limited number such as tens of trees to achieve lightweight and ensure millisecond-level inference time; the number of sampled features, which is the size of the feature subset considered when splitting each node, is usually set to the square root or all of the total number of features for models with only two input features to maintain diversity. In addition, the maximum depth parameter limits the number of splitting layers of the tree, and is set to a shallow value to reduce computational load; the minimum sample split threshold defines the minimum number of samples required for node splitting, and is preset to be higher than the default to speed up training; the default branch threshold is specifically set for residual contraction speed and attitude amplitude density, such as a fixed threshold range based on domain knowledge, used for recursive splitting rather than data-driven learning to simplify the model. Other parameters such as the bootstrap sample ratio are set to the size of the dataset by default, and the out-of-bag samples are used for validation and not involved in the setting. These parameters are finally determined through validation set tuning to ensure that the dimensionless output of the attitude stability adjustment coefficient accurately triggers the judgment when it exceeds the limit value, optimizing the response efficiency in the inspection scenario.

[0069] 3-5. In response to the judgment of exceeding the standard, interrupt the progress in the risk time and update the information, so as to maintain the continuity of the inspection. Based on the judgment result, if the standard is exceeded, the unmanned aerial vehicle is suspended, and the ground station is requested to request the latest clear corridor profile. Specifically, the suspension command locks the current position, uses high-precision coordinate recording to ensure that the request contains the accurate position, and the wireless link sends the current coordinates to obtain the updated profile. This step ensures flight safety and facilitates subsequent processing by the trajectory correction unit, thereby reducing the intervention of the tower station.

[0070] The safety monitoring unit realizes the extraction and judgment of the distance-attitude feature during flight, ensures the real-time response of the unmanned aerial vehicle to the change of the clear corridor in the scene of the airport glass curtain wall inspection, and minimizes the interruption caused by trajectory conflict.

[0071] The safety monitoring unit extracts key features to judge the exceeding standard and suspend the progress to request the latest clear corridor profile, thereby providing dynamic updated airspace boundary information for the correction stage. However, the current inspection needs to be globally re-planned when the boundary changes, resulting in a sharp increase in computational load. Therefore, the trajectory correction unit corrects the unexecuted flight path along the normal translation and broadcasts the results, ensuring local efficiency of flight path adjustment in glass curtain wall inspection and synchronization of tower monitoring, and realizing rapid recovery after interruption.

[0072] The processing of the trajectory correction unit is divided into the following steps: 4-1. Receive the latest clear corridor profile, obtain the updated boundary data pushed by the ground station, and integrate the real-time airspace changes in the correction. The ground station pushes the latest clear corridor profile in response to the request, and the latest clear corridor profile is composed of polygon boundary line segments, each of which is marked with start point and end point coordinates. When pushing, check the integrity of the boundary line segment, that is, check whether the start point and end point coordinates are continuous, to avoid transmission errors affecting the correction accuracy.

[0073] 4-2. Identify the unexecuted flight path, isolate the flown part, and only adjust the remaining path.

[0074] Based on the boundary data set of the latest clear corridor profile and the current unmanned aerial vehicle position, the path point sequence of the mixed flight path of the wall-attached straight line and circular arc is queried, and the remaining points starting from the current position are extracted as the unexecuted flight path sub-sequence. Specifically, the index of the current position to the path point sequence is matched by the binary search algorithm, that is, the sequence is repeatedly divided and compared with the coordinate value until the position is located, and all points are extracted from the located index to form the unexecuted flight path sub-sequence. This facilitates direct operation of the unexecuted flight path sub-sequence by the downstream correction, avoiding global recalculation to improve computational efficiency.

[0075] 4-3. Perform normal translation correction to offset the unexecuted flight path to avoid the new boundary and maintain the fitting of the camera visual axis to the glass surface.

[0076] Based on the unexecuted flight path sub-sequence, the normal distance to the latest clear corridor profile is calculated for each point, and the point is translated along the normal. The translation vector is calculated by first finding the normal distance of the point to the nearest boundary line, i.e. the minimum perpendicular distance of the projected point to the polygon boundary line segment, and assigning a positive or negative sign according to the position inside or outside, and then multiplying the normal distance by the unit normal vector of the curtain surface at the point to get the vector component, ensuring that the offset is perpendicular to the surface. The corrected unexecuted flight path sub-sequence is output, which is convenient for the broadcast to directly use these offset points to achieve accurate avoidance of the staggered space near the corner concave and convex nodes.

[0077] 4-4. Broadcast the correction result to inform the tower to monitor and update the path, and coordinate the airport dispatch to avoid interference.

[0078] Based on the corrected unexecuted flight path sub-sequence, the point positions and minimum safety bandwidth values of the unexecuted flight path sub-sequence are sent to the tower through the wireless link. This ensures that the tower receives complete correction data, making it easy to monitor these points later, achieving zero interference synchronization between patrol and flight operation.

[0079] 4-5. Remove the pause execution update flight path and restore the flight and apply the corrected path to complete the continuous patrol after interruption. Based on the broadcast confirmation, the flight control module replaces the original unexecuted flight path sub-sequence with the corrected unexecuted flight path sub-sequence, and removes the pause command to start advancing. The tower confirms the synchronization timing, i.e. waits for a response signal within a fixed time window to align the update, ensuring no delay integration. This ensures seamless integration of the updated flight path, making it easy for the experience iteration unit to record later, achieving a significant reduction in flight path pause time.

[0080] The trajectory correction unit realizes the normal translation correction of the unexecuted flight path and broadcasts it for the tower to monitor, ensuring local adjustment of dynamic clearances in the airport glass curtain wall patrol scenario, thereby minimizing the computational load and quickly resuming flight.

[0081] The trajectory correction unit corrects the unexecuted flight path by normal translation and broadcasts it for the tower to monitor, thereby achieving rapid resumption of flight after interruption. However, the current patrol lacks task experience accumulation, resulting in fixed safety parameters. Therefore, the experience iteration unit writes flight records to the learning library and presets the safety bandwidth, ensuring adaptive optimization of safety parameters in glass curtain wall patrol, and achieving reduced pauses under multi-task accumulation.

[0082] 5-1. Collect flight records to capture all data during this patrol, and accumulate distance, attitude and correction event information in iterations.

[0083] After landing, flight records are extracted from the flight terminal, including residual distance curve, three-axis angular velocity signal, attitude stabilization adjustment coefficient, modified non-executed flight path sub-sequence, and pause count. When extracting, the integrity of the record sequence is checked, i.e. the number of residual distance curve points, the number of three-axis angular velocity signal axes, the value of attitude stabilization adjustment coefficient, the length of modified non-executed flight path sub-sequence, and the count of pause count are checked item by item to match the expected value, to avoid incomplete data affecting the accuracy of iteration. Comprehensive archiving of dynamic airspace interleaving events is achieved.

[0084] 5-2. Write learning library, persist flight records for cross-task analysis.

[0085] Based on the complete data set of flight records, the data set is appended to the learning library, which is a time series database structure, indexed by task number. Create a new entry to insert residual distance curve, three-axis angular velocity signal, attitude stabilization adjustment coefficient, modified non-executed flight path sub-sequence, and pause count, and define the entry as a structured field in a unified format to ensure query consistency. Output the updated learning library to facilitate downstream analysis to directly query these historical entries and improve the continuity of experience accumulation.

[0086] 5-3. Analyze historical records to extract patterns to guide parameter adjustment.

[0087] Based on the updated learning library, query multi-task flight records, calculate average pause duration and margin contraction frequency. Calculate the preset safety bandwidth, first divide the total number of historical pauses by the total number of tasks to get the ratio, then take the negative ratio as the index, take the natural base as the power to get the decay factor, and finally multiply the initial safety bandwidth by the decay factor to get the preset safety bandwidth, where the initial safety bandwidth is a fixed reference length, the total number of historical pauses is the cumulative number of all tasks, and the total number of tasks is the number of executed tasks. This calculation ensures that the bandwidth decays exponentially with experience. Output the preset safety bandwidth to facilitate downstream applications to directly use the preset safety bandwidth value to achieve adaptive buffering of the clear corridor changes.

[0088] 5-4. Pre-set safety bandwidth, initialize parameters for next task, reduce online adjustment requirements.

[0089] Based on the preset safety bandwidth, update the minimum safety bandwidth value in the shared coordinate table of the flight path generation unit to the preset safety bandwidth. Cover the minimum safety bandwidth value of each node and use version tag control update, i.e. append task number label to prevent multiple version conflicts. Ensure that the next task inherits the optimized parameters to facilitate the safety monitoring unit to reduce the judgment of exceeding the standard frequency and achieve the cumulative compression of the patrol stop.

[0090] 5-5. Reduce online pause, apply iteration results to optimize flight logic, improve overall efficiency.

[0091] Based on the updated minimum safe bandwidth value, the buffer threshold limit is expanded during flight, and the preset threshold value of the attitude stabilization adjustment coefficient is adjusted by the up-regulation ratio, which is calculated inversely proportional to the total number of historical pauses, to ensure that the threshold limit is dynamically increased to reduce the triggering frequency. Ensure that the online pause is reduced, facilitate closed loop formation, and realize significant reduction of the flight path pause time.

[0092] The above formulas are dimensionless values calculated, and the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0093] It should be noted that the system of the present application can be deployed in the device itself to realize embedded application, or can be run on PC or other terminal with user interface, so as to meet various hardware environments and use requirements.

[0094] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

[0095] It should be noted that in this paper, if there are relationship terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "includes" "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0096] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An artificial intelligence-based unmanned aerial vehicle automatic airport inspection system for glass curtain walls, characterized in that, include: Coordinate fusion unit: Before takeoff, the curtain wall curvature grid and the airspace corridor outline are superimposed under a unified coordinate reference to generate a shared coordinate table and load it into the flight terminal; Track generation unit: Reads the shared coordinate table, generates a hybrid straight-circular-arc track that adheres to the wall according to the principle of minimum curvature, and writes the minimum safety band at each node; Safety monitoring unit: During flight, the distance radar and inertial navigation synchronously output residual distance curves and angular velocity signals, extract key features to perform wall-hugging safety judgments, and if the judgment results exceed the limits, pause the advance and request the latest corridor outline; Track correction unit: After the ground station pushes the latest corridor outline, the path planner corrects the unexecuted trajectory along the normal direction and broadcasts the correction results for tower monitoring. The flight control then releases the pause on updating the trajectory. Experience Iteration Unit: After landing, the flight record is written to the learning library to preset a safe bandwidth for the next mission and reduce online downtime. 2.The artificial intelligence-based unmanned aerial vehicle automatic airport inspection system for glass curtain walls according to claim 1, characterized in that, The coordinate fusion unit includes the following: Before takeoff, the curtain wall curvature mesh containing location coordinates and local curvature values ​​is extracted from the airport building database. The airspace corridor outline composed of polygonal boundary segments is obtained from the airport operations command center. The consistency of point cloud density is verified. A set of reference points is selected to calculate the transformation matrix to minimize the sum of Euclidean distances and thus establish a unified coordinate reference. The projection distance from each curtain wall curvature mesh point to the airspace corridor outline is calculated to identify overlapping areas and generate a fused mesh. A shared coordinate table organized by the points of the fused mesh is extracted. A check code is added for verification via wireless transmission protocol and loaded into the flight terminal storage module. 3.The artificial intelligence-based glass curtain wall unmanned aerial vehicle automatic airport inspection system according to claim 2, characterized in that, The trajectory generation unit includes the following: The flight terminal of the trajectory generation unit extracts the shared coordinate table from the storage module, verifies the integrity of the matrix rows, traverses the curvature values ​​to filter points that exceed the preset threshold as curvature key points, calculates the minimum curvature path between adjacent curvature key points to minimize the overall curvature accumulation and generate a wall-hugging straight-circular-arc hybrid trajectory, calculates the shortest normal distance from each path point to the clearance corridor outline to quantify the clearance margin and obtain the minimum safety belt width value, and appends the minimum safety belt width value to the path point sequence as a floating-point format attribute field.

4. The artificial intelligence-based glass curtain wall unmanned aerial vehicle automatic airport inspection system according to claim 2, Its features are: The shared coordinate table is a matrix, with each row corresponding to a point, including uniform coordinates, curvature values, and contour markers; Unified coordinates refer to the three-dimensional position coordinates of each point in the fused mesh under a unified coordinate reference, including the x, y and z components; The curvature value refers to the local curvature value derived from the curtain wall curvature grid. Points derived from the curtain wall curvature grid directly inherit the original local curvature value, while points derived from the clear corridor outline are filled with zero to indicate no curvature information. The outline marker refers to a binary identifier value, which is assigned a value of one for points belonging to the boundary of the clearance corridor outline and a value of zero for points not belonging to the boundary of the clearance corridor outline. 5.The artificial intelligence-based unmanned aerial vehicle automatic airport inspection system for glass curtain walls of claim 3, characterized in that, The security monitoring unit includes the following components: Residual distance readings are collected by the radar during the flight to form a residual distance curve, and three-axis angular velocity signals are recorded by the inertial navigation sensor. The sequence data are aligned in time and synchronized, and key features are extracted, including the margin contraction speed ratio and the attitude amplitude density. A pre-trained lightweight random forest is used to input the margin contraction speed ratio and the attitude amplitude density to output the attitude stabilization adjustment coefficient and determine whether it exceeds the standard. If it does, the forward request is suspended, the latest clearance corridor profile is requested, and the high-precision coordinates are locked.

6. The artificial intelligence-based unmanned aerial vehicle automatic airport inspection system for glass curtain walls according to claim 5, characterized in that: The margin contraction speed ratio is calculated based on the residual distance curve to obtain the slope difference at adjacent time points, and the maximum negative slope value is selected to quantify the clearance margin reduction rate.

7. The artificial intelligence-based unmanned aerial vehicle automatic airport inspection system for glass curtain walls according to claim 5, characterized in that: The attitude amplitude density applies a short window fast Fourier transform to the three-axis angular velocity signals, integrates the part exceeding the intermediate frequency threshold after obtaining the frequency spectrum amplitude, and evaluates the attitude oscillation strength. 8.The artificial intelligence-based glass curtain wall unmanned aerial vehicle automatic airport inspection system according to claim 5, characterized in that, The trajectory correction unit includes the following: After the ground station pushes the latest clearance corridor profile, the boundary line segment integrity is verified, the unexecuted flight path sub-sequence is extracted based on the current unmanned aerial vehicle position binary search matching index, the normal distance from each point to the latest clearance corridor profile is calculated to translate the point position to generate the corrected unexecuted flight path sub-sequence, and the corrected unexecuted flight path sub-sequence point position and the minimum safety bandwidth value are sent. After the tower station confirms, the original unexecuted flight path sub-sequence is replaced and the suspension is started. 9.The artificial intelligence-based glass curtain wall unmanned aerial vehicle automatic airport inspection system according to claim 8, characterized in that, The experience iteration unit includes the following: After landing, the flight record is extracted and the sequence integrity is verified, the flight record is appended to the learning library and the structured field is defined, the multi-task flight record is queried to calculate the preset safety bandwidth intention index decay, the minimum safety bandwidth value of the shared coordinate table is updated and version marking control is performed, and the buffer threshold limit is expanded during the flight.

Images