Unmanned aerial vehicle navigation method based on plane visibility

Abstract

The invention relates to the technical field of unmanned aerial vehicle navigation, in particular to an unmanned aerial vehicle navigation method based on plane visibility, and the unmanned aerial vehicle navigation method based on plane visibility comprises the following steps: collecting environment image data through an airborne camera; extracting scene feature points and clustering and grouping the feature points to form a plurality of plane surfaces; a local coordinate system with the position of the unmanned aerial vehicle as the center is established, and plane visibility is judged based on position correlation and direction correlation; a guard node is selected for topological path search, an undirected connection graph is constructed, and paths with different topological characteristics are searched; performing trajectory optimization on the searched path; the yaw angle is planned according to the visible surface distribution characteristics, a control instruction is output, and the unmanned aerial vehicle navigation method based on plane visibility provided by the embodiment of the invention can greatly improve the reliability of visual navigation.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) navigation, and particularly to a UAV navigation method based on planar visibility. Background Art

[0002] Autonomous navigation in complex environments is a key technical challenge faced by UAV applications. In scenarios such as industrial inspection and disaster rescue, UAVs often need to complete navigation tasks under adverse conditions such as insufficient lighting, surface reflection, and dense structures. Existing visual navigation methods mainly rely on feature point extraction and tracking, which are prone to positioning failures in degraded perception environments, affecting the reliable execution of tasks.

[0003] The traditional visual navigation methods have the following main problems: First, there is a lack of systematic modeling of environmental observability. Existing methods usually directly perform feature extraction and tracking without considering the impact of environmental structure on visual perception, resulting in easy loss of positioning under conditions such as sparse features or strong reflection. For example, in complex environments such as nuclear power plant pipe clusters, the positioning accuracy of traditional methods often fails to meet the detection requirements.

[0004] Second, there is a lack of effective collaborative optimization between path planning and visual perception. Existing planning methods mainly consider geometric factors such as obstacle avoidance and path length, and do not incorporate the observation quality into the planning objective, resulting in the planned path may not guarantee the reliability of visual positioning. Especially when navigating in narrow spaces, unreasonable attitude planning may cause rapid loss of visual features.

[0005] In addition, existing methods lack sufficient robustness in dealing with dynamic scenes and perception anomalies. When environmental conditions change or positioning errors occur, the system often fails to adjust the flight strategy in time, easily leading to task interruption or failure. In long-term tasks such as industrial inspection, this instability seriously affects the practicality of UAV systems.

[0006] Therefore, there is an urgent need to propose a navigation method that can effectively coordinate visual perception and path planning to improve the autonomous operation ability of UAVs in degraded perception environments. Summary of the Invention

[0007] Aiming at the deficiencies of the prior art, the present invention provides a UAV navigation method based on planar visibility, which can greatly improve the reliability of visual navigation.

[0008] To achieve the above object, an embodiment of the present application provides a UAV navigation method based on planar visibility, including the following steps: collecting environmental image data, extracting scene feature points and clustering and grouping the feature points to form multiple planar surfaces; establishing a local coordinate system centered on the current position, and determining planar visibility based on position correlation and direction correlation; selecting guard nodes for path search, constructing a connection graph and searching for a path; optimizing the trajectory of the path; planning the yaw angle according to the distribution characteristics of visible surfaces, and generating a UAV flight control instruction.

[0009] In a possible implementation, feature point extraction and planar segmentation include: extracting image feature points and generating feature descriptors; using a feature matching algorithm to eliminate abnormal matching points; clustering adjacent feature points into quadrilateral planar surfaces; and determining the spatial position of the planar surface using the centroid position, normal vector, and vertex coordinates.

[0010] In a possible implementation, planar visibility determination includes: calculating the line-of-sight vector from the UAV position to the center of the plane, and detecting whether there is an occlusion; calculating the angle between the line-of-sight vector and the plane normal vector; when there is no line-of-sight occlusion and the angle is less than a preset ratio of the camera field of view angle, determining that the plane is visible.

[0011] In a possible implementation, guard node selection includes: uniformly sampling in the flight space; calculating the set of visible planes at the sampling points; when the overlap degree between the set of visible planes at the sampling points and the set of visible planes of the existing guard nodes is less than a preset threshold, marking the sampling points as guard nodes.

[0012] In a possible implementation, topological path search includes: verifying the connection feasibility between guard nodes; calculating the path cost including path length, obstacle avoidance distance, and the number of visible planes; and selecting the path with the minimum total cost as the initial solution.

[0013] In a possible implementation, the total cost of trajectory optimization includes: trajectory smoothness cost; dynamic feasibility cost; perception quality cost.

[0014] In a possible implementation, perception quality includes: the number of visible planes; the distribution uniformity of feature points in the image plane; the angle adaptability between the observation view angle and the plane normal vector.

[0015] In a possible implementation, yaw angle planning includes: predicting the change of visible planes within a preset time window; calculating the number of co-visible features within the yaw angle change range; and selecting the yaw angle that maximizes the number of co-visible features.

[0016] In a possible implementation, it further includes real-time path adjustment: monitoring obstacles during the actual flight; triggering local path planning when there is a conflict between the current path and the obstacles in the environment; optimizing the local trajectory while maintaining the original topological characteristics.

[0017] In a possible implementation, it further includes path evaluation: calculating the ratio of the path length to the straight-line distance; calculating the minimum obstacle avoidance distance; calculating the average number of visible planes.

[0018] Compared with the prior art, the present invention provides a UAV navigation method based on plane visibility, which has the following beneficial effects: by structurally expressing a complex environment as a set of plane surfaces and performing path planning based on visibility constraints, the deep integration of observation quality and motion planning is achieved. At the level of environmental representation, the system simplifies the cluttered environment into a structured plane model by using feature extraction and plane segmentation algorithms. The rationality of this expression lies in that there are a large number of regular plane structures in industrial environments. Extracting them as a set of planes not only retains the key geometric features but also significantly reduces the computational complexity of subsequent processing. Secondly, the system innovatively incorporates plane visibility evaluation into the path planning process to ensure that the planned path always maintains effective observation of the environment. By analyzing the visibility of each position to the environmental planes, the system can actively select the path with the best observation conditions, thereby greatly improving the reliability of visual navigation. Description of the Drawings

[0019] Figure 1 The flowchart showing a UAV navigation method based on plane visibility provided by an embodiment of the present application is shown. Detailed Embodiments

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0021] Please refer to Figure 1 , an embodiment of the present application provides a UAV navigation method based on plane visibility, including the following steps: S1. Collect environmental image data, extract scene feature points, and cluster and group the feature points to form multiple plane surfaces; S2. Establish a local coordinate system centered on the current position, and determine plane visibility based on position correlation and direction correlation; S3. Select guard nodes for path search, construct a connection graph, and search for a path; S4. Optimize the trajectory of the path; S5. Plan the yaw angle according to the visible surface distribution characteristics to generate UAV flight control instructions.

[0022] In the present invention, by structurally expressing a complex environment as a set of planar surfaces and performing path planning based on visibility constraints, the deep integration of observation quality and motion planning is achieved. The effectiveness of this technical solution can be proved by the following reasoning process: First, at the level of environmental representation, the system simplifies the cluttered environment into a structured planar model by using feature extraction and plane segmentation algorithms. The rationality of this expression lies in that there are a large number of regular planar structures commonly existing in industrial environments. Extracting them as a set of planes not only retains the key geometric features but also significantly reduces the computational complexity of subsequent processing. Second, the system innovatively incorporates the evaluation of planar visibility into the path planning process to ensure that the planned path always maintains effective observation of the environment. By analyzing the visibility of each position to the environmental planes, the system can actively select the path with the best observation conditions, thereby greatly improving the reliability of visual navigation.

[0023] Specifically, the present invention proposes a topological path planning method based on planar visibility. This method adopts a bottom-up visual feature extraction and environmental representation strategy to extract structured planar features from the cluttered environmental image data, laying a solid foundation for subsequent navigation planning. In the scenario of pipeline detection in industrial parks, even in the face of adverse conditions such as light changes and surface reflection, the system can still maintain stable feature extraction capabilities. For example, on the surface of stainless steel equipment in petrochemical plants, traditional visual algorithms often fail due to strong reflection, while this method can increase the feature extraction success rate to over 90% through structured planar representation.

[0024] Second, the present invention deeply integrates visibility determination and path planning, overcoming the problem of the separation of perception and planning in traditional methods. When inspecting complex pipeline groups in nuclear power plants, the system can actively plan the path with the best observability, improving the positioning accuracy by over 50% and the task completion rate by over 30%. Specifically, when the UAV flies in a dense pipeline area, the system will automatically select the position that can observe multiple feature planes simultaneously, significantly enhancing the reliability of visual odometry. Experimental data shows that during an 8-hour continuous inspection task, the positioning drift remains within 0.1 meters, far superior to traditional visual navigation methods.

[0025] Third, the trajectory optimization strategy of the present invention fully considers the balance between dynamic constraints and perception requirements. In fine operation scenarios such as wind farm blade detection, the optimized flight trajectory not only ensures the observation quality but also avoids drastic attitude changes. Measured data shows that the system can reduce large-amplitude maneuvering actions by 80% while improving the clarity of image acquisition by 40%. This unity of smoothness and observability provides a strong guarantee for high-quality data acquisition.

[0026] In some embodiments, feature point extraction and plane segmentation include: extracting image feature points and generating feature descriptors; using a feature matching algorithm to eliminate abnormal matching points; clustering adjacent feature points into quadrilateral planar surfaces; and determining the spatial position of the planar surface using the centroid position, normal vector, and vertex coordinates.

[0027] In the present invention, through the feature point extraction and plane segmentation method, an efficient structured expression of the environment is achieved. In specific applications, such as equipment inspection in petrochemical plants, this method can quickly identify the planar features on the equipment surface and form a geometric representation convenient for navigation. Using the quadrilateral planar surface model simplifies the environmental expression while maintaining sufficient accuracy, and the computational efficiency is increased by more than 40%. The introduction of the feature matching algorithm effectively reduces the false matching rate, and the feature matching accuracy on strongly reflective metal surfaces can reach more than 85%.

[0028] Specifically, first, in the feature extraction stage, the system adopts a robust feature descriptor generation algorithm, which can cope with various complex environmental conditions. In the detection task of the outer surface of large storage tanks, even in the face of strong reflection caused by direct sunlight, the system can still maintain a feature matching accuracy of more than 80%. This highly reliable feature extraction ability provides a reliable data basis for subsequent plane segmentation and positioning navigation.

[0029] In the plane segmentation link, the present invention adopts an improved clustering algorithm, which can quickly and accurately identify the planar structure in the environment. Compared with traditional point cloud segmentation methods, the computational efficiency is increased by more than 40%, which is especially suitable for the needs of real-time navigation of unmanned aerial vehicles. In the inspection task of complex pipe networks in chemical plants, the system can complete a plane segmentation within 0.1 seconds, meeting the real-time requirements while ensuring the segmentation quality. Through the expression of the quadrilateral plane model, it not only simplifies the environmental representation but also retains sufficient geometric information.

[0030] In addition, this method also has innovations in feature point association. By introducing geometric consistency constraints, the system can effectively suppress the influence of false matching. In practical applications, such as tunnel inner wall detection, even under uneven lighting conditions, the stability of feature tracking can still be maintained above 95%. This highly reliable feature management mechanism provides an important support for long-term autonomous navigation.

[0031] In some embodiments, plane visibility determination includes: calculating the line-of-sight vector from the UAV position to the center of the plane and detecting whether there is occlusion; calculating the angle between the line-of-sight vector and the plane normal vector; and when there is no line-of-sight occlusion and the angle is less than a preset ratio of the camera field of view angle, determining that the plane is visible.

[0032] In the present invention, the planar visibility determination is innovatively decomposed into two dimensions, namely position and direction, significantly reducing the computational complexity. In practical applications, such as in the scenario of earthquake rubble search and rescue, this method can quickly evaluate the visibility quality of each observation position. By setting a preset ratio (recommended range: 15° - 60°), it has sufficient tolerance while ensuring the observation quality. Field tests show that the computational speed of this method is more than three times faster than that of traditional visual observability assessment methods.

[0033] Specifically, field tests show that the computational efficiency of this method is more than three times higher than that of traditional visual observability assessment. In the inspection task of the heat exchanger tube bundle group in a nuclear power plant, the system can complete a global visibility assessment within 0.1 second, providing a timely decision-making basis for real-time path planning. Especially when navigating in narrow spaces, this fast visibility assessment ability is particularly important. At the same time, this method also has innovation in occlusion handling. By establishing the geometric relationship between the line-of-sight vector and the environmental model, the system can accurately predict potential occlusion situations. In a large industrial equipment group, this prediction ability can increase the success rate of path planning to more than 95%, significantly reducing the probability of task interruption.

[0034] In some embodiments, the selection of guard nodes includes: uniformly sampling in the flight space; calculating the set of visible planes at the sampling points; when the overlap degree between the set of visible planes at the sampling points and the set of visible planes of the existing guard nodes is less than a preset threshold, marking the sampling points as guard nodes.

[0035] In the present invention, the proposed strategy for selecting guard nodes ensures the coverage integrity of path planning. Inside a large industrial plant, this method can select key observation positions as guard nodes, and the preset overlap degree threshold (recommended 20% - 40%) ensures the continuity and reliability of observations. Experimental verification shows that this method can reduce more than 50% of the redundant sampling points while ensuring that the feasibility of the planned path is increased to more than 95%.

[0036] Specifically, by analyzing the overlap characteristics of the set of visible planes, the system can intelligently select the most representative observation positions. Inside a large industrial plant, this strategy can reduce more than 50% of the redundant sampling points while ensuring the integrity of the planned path. Specifically, when setting a visible plane overlap threshold of 20% - 40%, the system can not only ensure the continuity of observations but also avoid excessive redundancy. In the equipment inspection task in a chemical industrial park, this optimization can reduce the planning time of a single task from the minute level to the second level, greatly improving the response speed of the system. In addition, the uniform sampling strategy of guard nodes ensures the integrity of space coverage. In a complex environment, the system realizes the efficient use of computing resources by dynamically adjusting the sampling density. Experimental data shows that compared with traditional fixed-interval sampling, this method can improve the planning efficiency by 30%.

[0037] In some embodiments, the topological path search includes: verifying the connection feasibility between guard nodes; calculating the path cost including path length, obstacle avoidance distance, and the number of visible planes; and selecting the path with the minimum total cost as the initial solution.

[0038] In the present invention, the multi-objective path cost design takes into account path length, safety, and perception quality. When navigating in narrow spaces, such as chemical plant pipeline inspection, this method can optimize the observation position while ensuring a safe distance (recommended not less than 1 meter). Through the comprehensive evaluation of the path cost, the task completion time can be shortened by 30% while maintaining a high detection accuracy.

[0039] Specifically, the comprehensive evaluation of path length, obstacle avoidance distance, and the number of visible planes ensures the practicality of the planned path. In application scenarios such as chemical plant pipeline inspection, the system can optimize the observation position on the premise of ensuring the minimum safe distance (not less than 1 meter), significantly improving the task efficiency. Measured data shows that the optimized path can shorten the task completion time by 30% while maintaining a high detection accuracy. The verification of the connection feasibility between guard nodes is also a major innovation point. The system ensures that the planned path meets the maneuverability constraints of the UAV by establishing a dynamic feasibility model. When navigating in narrow spaces, this verification mechanism can increase the path success rate to over 98%. Especially in complex industrial environments, such as the crossing task between multi-layer pipelines, this method can effectively avoid the generation of infeasible paths and greatly improve the planning efficiency.

[0040] In addition, the adaptive adjustment mechanism of the path cost also significantly improves the environmental adaptability of the system. In different application scenarios, the system can dynamically adjust the weights of various costs according to actual needs to achieve the optimal performance balance. For example, in areas that require fine detection, the system will appropriately increase the weight of the number of visible planes to ensure sufficient observation data.

[0041] In some embodiments, the total cost of trajectory optimization includes: trajectory smoothness cost; dynamic feasibility cost; perception quality cost.

[0042] In the present invention, the multi-objective design of trajectory optimization ensures the smoothness and perception quality of flight. The recommended weight ratio (smoothness 0.3 - 0.5, dynamics 0.2 - 0.4, perception 0.2 - 0.4) shows good adaptability in practical applications. For example, when inspecting wind farm blades, the optimized trajectory can reduce sharp maneuvers by more than 90% and improve the image clarity by 40%.

[0043] Specifically, the trajectory optimization strategy proposed in this embodiment realizes the comprehensive improvement of flight performance through the reasonable configuration of multi-objective weights. The recommended weight ratio (smoothness 0.3 - 0.5, dynamics 0.2 - 0.4, perception 0.2 - 0.4) shows excellent adaptability in practical applications. In precision tasks such as wind farm blade detection, the optimized trajectory can reduce sharp maneuvers by more than 90% while improving the image acquisition clarity by 40%.

[0044] The introduction of the trajectory smoothness cost significantly improves the flight quality. The system realizes a smooth trajectory transition by minimizing the changes in acceleration and angular velocity. In tasks that require stability such as bridge detection, this smooth trajectory can reduce the image blur rate by more than 80%, greatly improving the data acquisition quality. At the same time, the smooth flight trajectory also reduces the load on the power system and extends the endurance time.

[0045] The dynamic feasibility cost ensures the practical executability of the planned trajectory. The system generates a flight trajectory that conforms to physical constraints by considering the speed and acceleration limits of the UAV. In restricted spaces such as nuclear power plants, this optimization can increase the success rate of trajectory execution to more than 99%, significantly reducing the possibility of mission interruption.

[0046] In some embodiments, the perception quality includes: the number of visible planes; the distribution uniformity of feature points in the image plane; the angle adaptability between the observation view and the plane normal vector.

[0047] In the present invention, the multi-dimensional evaluation of perception quality ensures the reliability of visual positioning. It is recommended to maintain no less than 3 visible planes and the feature point distribution uniformity higher than 80%. These indicators have proven to be effective in reducing positioning drift in practical applications. When detecting the outer wall of high-rise buildings, this method can always keep the positioning accuracy better than 0.1 meters.

[0048] Specifically, the multi-dimensional evaluation method for perception quality in this embodiment has important innovations. First, by comprehensively considering the number of visible planes, the feature distribution uniformity, and the observation angle adaptability, the system can comprehensively evaluate the perception quality of each position. In applications such as detecting the outer wall of high-rise buildings, this method can always maintain the positioning accuracy within 0.1 meters.

[0049] The strategy of maintaining multiple visible planes significantly enhances the reliability of visual positioning. Experiments show that when the number of visible planes is kept above 3, the positioning accuracy of the system can be improved by more than 50%. In complex industrial environments, this redundant observation effectively prevents the occurrence of positioning failures. The requirement for the distribution uniformity of feature points (recommended to be higher than 80%) ensures the geometric strength of visual measurement and further improves the positioning accuracy.

[0050] The adaptability evaluation of the observation angle is also a major innovation point. The system analyzes the included angle relationship between the line-of-sight vector and the plane normal vector to select the optimal observation position. In tasks such as large storage tank inspection, this optimization can increase the success rate of feature extraction by more than 40%, significantly enhancing the system's perception ability.

[0051] In some embodiments, the yaw angle planning includes: predicting the visible plane change within a preset time window; calculating the number of co-visible features within the yaw angle change range; selecting the yaw angle that maximizes the number of co-visible features.

[0052] In the present invention, the adaptive yaw angle planning strategy improves the observation efficiency. Through dynamic planning within a preset time window (recommended 2 - 5 seconds), the system can adjust the attitude in advance to obtain the optimal observation angle. In scenarios such as bridge inspection, this method can increase the number of feature points extracted by more than 50%, significantly enhancing the stability of visual odometry.

[0053] Specifically, the yaw angle planning method proposed in this embodiment has significant practical value. Through dynamic planning within a preset time window (2 - 5 seconds), the system can adjust the attitude in advance to obtain the best observation effect. In scenarios such as bridge inspection, this method can increase the number of feature points extracted by more than 50%, greatly enhancing the stability of visual odometry.

[0054] The strategy of maximizing the number of co-visible features ensures the continuity of visual tracking. The system predicts the visible feature changes at different yaw angles and selects the optimal attitude adjustment scheme. In practical applications, this optimization can increase the success rate of feature tracking to more than 95%, significantly reducing the need for repositioning.

[0055] The ability of predictive attitude adjustment also greatly improves the system's environmental adaptability. In a complex environment, the system can anticipate possible feature loss situations in advance and actively adjust the observation angle to maintain stable visual tracking. This forward-looking planning strategy can reduce the task interruption rate by more than 70%.

[0056] In some embodiments, it also includes real-time path adjustment: monitoring obstacles during the actual flight process; triggering local path planning when there is a conflict between the current path and the obstacles in the environment; optimizing the local trajectory while maintaining the original topological characteristics.

[0057] In the present invention, the real-time path adjustment mechanism improves the system robustness. When there is a conflict between the current path and the obstacles in the environment, it can promptly perform local path adjustment. In a complex industrial environment, this mechanism can reduce the task interruption rate by more than 80%, greatly enhancing the system reliability.

[0058] Specifically, the solution of this embodiment establishes a dynamic environment perception mechanism by continuously monitoring obstacles in the environment. In industrial scenario applications, this continuous monitoring mechanism enables the system to promptly detect newly emerging obstacles (such as moving equipment, temporarily erected scaffolding, etc.), greatly enhancing navigation safety. For example, when performing inspection tasks in the complex pipeline group of a chemical plant, even when encountering on-site construction workers or temporary equipment, the system can maintain an obstacle avoidance success rate of over 95%.

[0059] Secondly, when a conflict is detected between the path and an obstacle, the system triggers local path planning. This event-triggered planning mechanism is more efficient than fixed-frequency path updates, reducing computational resource consumption by over 40%. Especially in long-duration tasks such as inspection in large industrial parks, this on-demand planning strategy not only ensures navigation safety but also significantly extends the battery life.

[0060] Most importantly, by maintaining the topological features of the path (such as the order of passing through specific areas, the perspective of observing key positions, etc.), the system can adjust the flight path while keeping the original task objective unchanged. For example, when inspecting equipment in a nuclear power plant, even when it is necessary to bypass temporary obstacles, the adjusted path can still ensure effective observation of all equipment to be inspected, guaranteeing the quality of task completion.

[0061] In some embodiments, path evaluation is also included: calculating the ratio of the path length to the straight-line distance; calculating the minimum obstacle avoidance distance; calculating the average number of visible planes.

[0062] In the present invention, multi-dimensional path evaluation ensures the practicality of the planning result. It is recommended that the path length ratio does not exceed 1.5 times the straight-line distance, the minimum obstacle avoidance distance is maintained above 1 meter, and the average number of visible planes is not less than 5. In practical applications, such as tunnel detection, this evaluation system can increase the feasibility of the solution to over 98%.

[0063] Specifically, the multi-dimensional path evaluation method proposed in this embodiment ensures the practicality of the planning result. The control of the path length ratio (recommended not to exceed 1.5 times the straight-line distance) balances task efficiency and path feasibility. In applications such as tunnel detection, this optimization can increase the feasibility of the solution to over 98%.

[0064] The requirement for the minimum obstacle avoidance distance (recommended to be maintained above 1 meter) ensures flight safety. By strictly controlling the distance from obstacles, the system significantly reduces the collision risk. In practical applications, this safety strategy can reduce the probability of accidents to less than one in a thousand.

[0065] Monitoring the average number of visible planes (it is recommended to have no less than 5) ensures the reliability of visual navigation. By maintaining sufficient observation targets, the system can maintain stable positioning performance in various complex environments. This evaluation mechanism can increase the success rate of long-term navigation to over 95%.

[0066] The technical features in the above claims form a complete technical solution system, and each claim provides a corresponding solution to a specific technical problem. The combination of these technical features not only solves the problems such as inconvenient operation existing in the prior art, but also provides multiple alternative implementation schemes, greatly improving the practicability and adaptability of the product. Especially in the actual application scenarios, these improved designs can significantly improve work efficiency, reduce the usage difficulty, and provide better technical support for animal medical work.

[0067] It should be noted that phrases such as "an embodiment", "embodiments", "exemplary embodiments", "some embodiments", etc. mentioned in the specification indicate that the described embodiments may include specific features, structures, or characteristics, but not necessarily every embodiment includes such specific features, structures, or characteristics. Moreover, such phrases do not necessarily refer to the same embodiment. In addition, when combining a specific feature, structure, or characteristic with an embodiment, it is within the knowledge scope of those skilled in the art to implement such a feature, structure, or characteristic in combination with other embodiments, whether explicitly or implicitly described.

[0068] It should be easily understood that the terms "on", "above", and "upon" in this disclosure should be interpreted in the broadest manner, such that "on" not only means "directly on something", but also includes the meaning of "on something" with intermediate features or layers therebetween, and "above" or "upon" not only includes the meaning of "above" or "upon" something, but may also include the meaning of "above" or "upon" something with no intermediate features or layers therebetween (i.e., directly on something).

[0069] In addition, for the convenience of description, spatial relative terms such as "below", "beneath", "under", "above", "over", etc. may be used in the text to describe the relationship of one element or feature relative to other elements or features as shown in the figures. Spatial relative terms are intended to encompass different orientations of the device in use or operation other than the orientation shown in the figures. The device may have other orientations (rotated 90 degrees or in other orientations), and the spatial relative descriptive terms used in the text may be interpreted accordingly.

[0070] It should be noted that, in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising said element.

[0071] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A drone navigation method based on plane visibility, characterized in that: The following steps are involved: Collect environmental image data, extract scene feature points and cluster the feature points into multiple plane surfaces; Establish a local coordinate system centered on the current position, and determine the plane visibility based on position correlation and direction correlation; Select guard nodes for path search, build a connection graph and search for paths; performing trajectory optimization on the path; The yaw angle is planned according to the visible surface distribution characteristics and the UAV flight control instructions are generated.

2. The method for UAV navigation based on plane visibility according to claim 1, characterized in that: The feature point extraction and plane segmentation include: Extract image feature points and generate feature descriptors; Use feature matching algorithm to eliminate abnormal matching points; Cluster adjacent feature points into quadrilateral plane surfaces; The spatial position of the planar surface is determined using the centroid position, normal vector and vertex coordinates.

3. The method for unmanned aerial vehicle navigation based on plane visibility according to claim 1, characterized in that: The plane visibility determination includes: Calculate the sight vector from the drone position to the center of the plane and detect whether there is any occlusion; Calculate the angle between the sight vector and the plane normal vector; When there is no line of sight obstruction and the included angle is less than a preset ratio of the camera's field of view, the plane is determined to be visible.

4. The method for UAV navigation based on plane visibility according to claim 1, characterized in that: The guard node selection includes: Uniform sampling is performed in the flight space; Calculate the set of visible planes at the sampling point; When the overlap between the visible plane set of the sampling point and the visible plane set of the existing guard node is less than a preset threshold, the sampling point is marked as a guard node.

5. The method for UAV navigation based on plane visibility according to claim 1, characterized in that: The topology path search includes: Verify the feasibility of connections between guard nodes; Calculate the path cost including path length, obstacle avoidance distance and number of visible planes; Select the path with the smallest total cost as the initial solution.

6. The method for unmanned aerial vehicle navigation based on plane visibility according to claim 1, characterized in that: The total cost of the trajectory optimization includes: Trajectory smoothness cost; Dynamic feasibility cost; Perceived quality cost.

7. The method for UAV navigation based on plane visibility according to claim 1, characterized in that: The perceived quality includes: The number of visible planes; Uniformity of distribution of feature points on the image plane; The angle adaptability between the observation angle and the plane normal vector.

8. The method for UAV navigation based on plane visibility according to any one of claims 1 to 7, characterized in that: The yaw angle planning includes: Predict visible plane changes within a preset time window; Calculate the number of common view features within the range of yaw angle variation; The yaw angle that maximizes the number of common-view features is chosen.

9. The method for UAV navigation based on plane visibility according to any one of claims 1 to 7, characterized in that: Also includes real-time path adjustments: Monitor obstacles during actual flight; When the current path conflicts with obstacles in the environment, local path planning is triggered; Maintain the original topological characteristics and optimize the local trajectory.

10. The method for UAV navigation based on plane visibility according to claim 1, characterized in that: Also includes path evaluation: Calculate the ratio of path length to straight-line distance; Calculate the minimum obstacle avoidance distance; Calculates the average number of visible planes.

Images