Unmanned aerial vehicle autonomous navigation method based on dynamic interference suppression and topological feature mapping

Abstract

The invention provides an unmanned aerial vehicle autonomous navigation method based on dynamic interference suppression and topological feature mapping. The method comprises the following steps: acquiring scene images of a scene to be navigated of an unmanned aerial vehicle in continuous time in real time; determining an interference dynamic object in the to-be-navigated scene of the unmanned aerial vehicle based on the scene image, and performing interference dynamic object elimination processing on the scene image based on the interference dynamic object to obtain an interference-suppressed scene image; key feature point extraction is carried out on the scene image after interference suppression, and extracted scene key points are obtained; a topological graph matched with the to-be-navigated scene of the unmanned aerial vehicle is constructed based on the scene key points, graph nodes of the topological graph are the scene key points, and weights of edges in the topological graph are connection relations among the graph nodes; and determining a navigation path of the unmanned aerial vehicle in the to-be-navigated scene of the unmanned aerial vehicle based on the topological graph and a heuristic search algorithm. The autonomous navigation of the unmanned aerial vehicle in a complex environment is realized by dynamically suppressing an interference dynamic object and a topological graph.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) navigation technology, and in particular to an autonomous navigation method for UAVs based on dynamic interference suppression and topological feature mapping. Background Technology

[0002] In the field of autonomous navigation for unmanned aerial vehicles (UAVs), navigation in complex and dynamic environments, especially autonomous navigation in dynamically changing environments, is a major challenge for current technologies.

[0003] As known from relevant technologies, to address dynamic interference issues, traditional image processing methods or simple depth estimation techniques are typically employed to attempt to filter out some of the interfering objects. However, these methods struggle to cope with complex dynamic environments. Furthermore, most existing methods rely on geometric modeling for path planning, neglecting the extraction and utilization of key environmental topological features. This results in a relatively limited understanding of the environment by the navigation system, restricting the flexibility and accuracy of path planning. Summary of the Invention

[0004] This invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping, which improves the path planning capability of UAVs in complex environments, ensures safety and flight efficiency, and realizes autonomous navigation of UAVs in complex environments.

[0005] This invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. The method includes: real-time acquisition of scene images of the UAV's navigation scenario over a continuous time period; identifying interfering dynamic objects in the scene image based on the scene image, and removing these interfering dynamic objects to obtain an interference-suppressed scene image; extracting key feature points from the interference-suppressed scene image to obtain extracted scene key points, wherein the scene key points are points in the UAV's navigation scenario that affect the UAV's navigation path; constructing a topological graph matching the UAV's navigation scenario based on the scene key points, wherein the graph nodes of the topological graph are the scene key points, and the weights of the edges in the topological graph represent the connection relationships between the graph nodes; and determining the UAV's navigation path in the UAV's navigation scenario based on the topological graph and a heuristic search algorithm.

[0006] According to the present invention, an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping is provided. The step of determining interfering dynamic objects in the scene to be navigated by the UAV based on the scene image specifically includes: detecting the position information of dynamic objects in the scene to be navigated by the UAV using a target detection algorithm based on the scene image; determining the optical flow of pixels in the scene image based on the scene image at adjacent time points; and determining the interfering dynamic objects in the scene to be navigated by the UAV based on the optical flow and the position information.

[0007] According to the present invention, an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping is provided. The step of determining interfering dynamic objects in the UAV navigation scene based on the optical flow and the position information specifically includes: determining whether the motion trajectory of the dynamic object in the UAV navigation scene is consistent with the background motion trajectory in the camera's viewpoint, wherein the camera's viewpoint is the viewpoint of the camera acquiring the scene image; and determining that the dynamic object is an interfering dynamic object in the UAV navigation scene if the motion trajectory of the dynamic object in the UAV navigation scene is inconsistent with the background motion trajectory in the camera's viewpoint.

[0008] According to the present invention, an autonomous navigation method for unmanned aerial vehicles based on dynamic interference suppression and topological feature mapping is provided. The step of extracting key feature points from the scene image after interference suppression to obtain extracted scene key points specifically includes: extracting key feature points from the scene image after interference suppression based on a random sampling consensus algorithm to obtain extracted scene key points.

[0009] According to the present invention, an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping is provided. The method for determining the navigation path of the UAV in the navigation scenario based on the topological graph and a heuristic search algorithm specifically includes: determining the original distances between adjacent graph nodes and the weights of the edges connecting adjacent graph nodes based on the topological graph; minimizing the target value obtained by accumulating the product of the differences in distances and the weights of the corresponding edges as the optimization objective, and determining the optimized estimated distances between nodes, wherein the difference in distances is the difference between the original distances between nodes and the optimized estimated distances between nodes; and determining the navigation path of the UAV in the navigation scenario based on the optimized estimated distances between nodes and the heuristic search algorithm.

[0010] According to the present invention, an autonomous navigation method for unmanned aerial vehicles based on dynamic interference suppression and topological feature mapping is provided. The weight of the edge connecting adjacent graph nodes is determined in the following way: determining the distance between adjacent graph nodes; determining the weight value of obstacles between adjacent graph nodes; and determining the weight of the edge connecting adjacent graph nodes based on the distance between adjacent graph nodes and the weight value of obstacles between adjacent graph nodes.

[0011] According to the present invention, an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping, before determining the navigation path of the UAV in the navigation scenario based on the topological graph and a heuristic search algorithm, the method further includes: removing redundant graph nodes and redundant edges from the topological graph to obtain an optimized topological graph; the determination of the navigation path of the UAV in the navigation scenario based on the topological graph and the heuristic search algorithm specifically includes: determining the navigation path of the UAV in the navigation scenario based on the optimized topological graph and the heuristic search algorithm.

[0012] This invention also provides an autonomous navigation device for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. The device includes: an acquisition module for real-time acquisition of scene images of the UAV's navigation scenario over a continuous time period; a determination module for identifying interfering dynamic objects in the scene image based on the scene image, and performing interference dynamic object removal processing on the scene image based on the interfering dynamic objects to obtain an interference-suppressed scene image; an extraction module for extracting key feature points from the interference-suppressed scene image to obtain extracted scene key points, wherein the scene key points are points in the UAV's navigation scenario that affect the UAV's navigation path; a construction module for constructing a topological graph matching the UAV's navigation scenario based on the scene key points, wherein the graph nodes of the topological graph are the scene key points, and the weights of the edges in the topological graph represent the connection relationships between the graph nodes; and a generation module for determining the UAV's navigation path in the UAV's navigation scenario based on the topological graph and a heuristic search algorithm.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping as described above.

[0016] This invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. The method involves real-time acquisition of scene images of the UAV's navigation scenario over continuous time. Based on these images, interfering dynamic objects in the navigation scenario are identified, and the scene images are processed to remove these objects, resulting in an interference-suppressed scene image. Key feature points are extracted from the interference-suppressed scene image to obtain extracted scene key points. Based on these key points, a topological graph matching the UAV's navigation scenario is constructed, where the graph nodes are scene key points, and the weights of the edges represent the connectivity between nodes. Using the topological graph and a heuristic search algorithm, the navigation path of the UAV in the navigation scenario is determined. By dynamically suppressing interfering dynamic objects and mapping the topological graph, the path planning capability of the UAV in complex environments can be improved, ensuring safety and flight efficiency, and enabling autonomous navigation of the UAV in complex environments. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping provided by the present invention.

[0019] Figure 2 This is a schematic diagram of the process for identifying interfering dynamic objects in the scene to be navigated by a UAV based on scene images, provided by the present invention.

[0020] Figure 3 This is a flowchart illustrating the process of determining the navigation path of a UAV in a navigation scenario provided by the present invention, based on a topology graph and a heuristic search algorithm.

[0021] Figure 4 This is a schematic diagram of the process for determining the weight of the edge connecting adjacent graph nodes provided by the present invention.

[0022] Figure 5 This is a schematic diagram of the structure of the UAV autonomous navigation device based on dynamic interference suppression and topological feature mapping provided by the present invention.

[0023] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] This invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. This method can detect and filter dynamic interference in real time, and extract key topological feature points to construct an efficient topological map. By dynamically suppressing interference and optimizing the topological map structure, it can improve the UAV's path planning capability in complex environments, ensuring safety and flight efficiency. This invention aims to solve the above-mentioned problems by combining dynamic interference suppression and topological feature mapping to achieve autonomous navigation of UAVs in complex environments.

[0026] Figure 1 This is a flowchart illustrating the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping provided by the present invention.

[0027] The following will combine Figure 1 The process of the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping provided by the present invention is described.

[0028] In an exemplary embodiment of the present invention, combined with Figure 1 As can be seen, the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping can include steps 110 to 150, and each step will be described below.

[0029] In step 110, scene images of the scene to be navigated by the UAV are acquired in real time over a continuous period of time.

[0030] In one embodiment, the drone may be equipped with a high-definition camera or visual sensor to continuously acquire image sequences of the scene to be navigated at a fixed frame rate, such as acquiring scene images. .

[0031] In step 120, based on the scene image, interfering dynamic objects in the scene to be navigated by the UAV are identified, and the scene image is processed to remove interfering dynamic objects to obtain a scene image after interference suppression.

[0032] In one embodiment, the culling process is not limited to pixel-level image filtering, but also includes generating a corresponding effective depth flow or filtered depth scene map for input to subsequent topological feature extraction and path planning processes. This effective depth flow can be jointly generated based on continuous frame optical flow and depth estimation, thereby maintaining the stability of scene depth information even in the presence of visual dynamic interference.

[0033] In one embodiment, a deep learning-based object detection algorithm (such as YOLO) can be used to analyze each frame of the scene image to identify dynamic objects (such as fallen leaves, floating objects, etc.). This algorithm is pre-trained on a dataset containing common dynamic objects and can output the bounding boxes and categories of objects in real time. In another example, optical flow can be combined to detect moving regions to verify the presence of dynamic objects.

[0034] In another embodiment, if the identified dynamic object is determined to be an interfering dynamic object in the scene to be navigated by the UAV, the scene image can be processed to remove the interfering dynamic object to obtain the scene image after interference suppression. It can be understood that there is no interfering dynamic object in the scene image after interference suppression, thus laying the foundation for subsequent accurate navigation.

[0035] In step 130, key feature points are extracted from the scene image after interference suppression to obtain the extracted scene key points, where the scene key points are the points in the scene to be navigated by the UAV that affect the UAV's navigation path.

[0036] In another example embodiment of the present invention, key feature points are extracted from the scene image after interference suppression to obtain the extracted scene key points, which can be achieved in the following way: The Random Sampling Consensus (RANSAC) algorithm is used to extract key feature points from the scene image after interference suppression, resulting in extracted scene key points.

[0037] In one embodiment, the Random Sampling Consensus (RANSAC) algorithm can be used to extract key points from the scene image after interference suppression, thereby obtaining the extracted scene key points. The scene key points can be considered as points in the scene to be navigated by the UAV that affect the UAV's navigation path, such as wall edges, corners, door and window edges, etc.

[0038] In another embodiment, feature point extraction can be performed based on the RANSAC algorithm, which is an optimization problem and can be described by the following formula (1): (1) in, To fit the model, For actual data points, For the input points of the model, To identify points that fit the model, points that meet the criteria are extracted from the RANSAC results based on the above formula, such as wall edges, corner points, and door / window edges.

[0039] In step 140, a topology graph matching the scene to be navigated by the UAV is constructed based on scene key points. The graph nodes of the topology graph are scene key points, and the weights of the edges in the topology graph are the connection relationships between the graph nodes.

[0040] In another embodiment, each extracted scene keypoint can be treated as a graph node in the topology graph. The node coordinates correspond to the keypoint's position in the image. Edges can also be constructed based on the spatial relationships between scene keypoints. For example, the k-nearest neighbor algorithm (k-NN) can be used to connect each node to its k nearest neighbors (e.g., k=3). The edge weight is represented by the Euclidean distance between the nodes; the greater the distance, the higher the weight, indicating a higher travel cost. Furthermore, if there are obstacles (such as walls, dynamic objects, etc.) between nodes, the weight value increases, indicating a greater difficulty in traversing the path.

[0041] Based on the graph nodes and edges of the topological graph described above, a final topological graph can be formed. This topological graph is an undirected graph where nodes represent key points in the scene, edges represent feasible paths, and weights represent path costs. This graph captures the topological structure of the scene and is used for path planning.

[0042] In step 150, the navigation path of the UAV in the scenario to be navigated is determined based on the topology map and heuristic search algorithm.

[0043] In another embodiment, a heuristic search algorithm A* can be used to plan paths on a topology graph to determine the UAV's navigation path within the scenario to be navigated. In one example, the UAV's current and target positions can be mapped to the nearest nodes in the topology graph, serving as the start and target nodes. The A* algorithm is run on the topology graph to search for the minimum-cost path from the start to the target node. The path cost is determined based on the accumulated edge weights. Furthermore, the output path serves as the UAV's navigation path, consisting of a series of topology graph nodes. The UAV tracks this path using a control algorithm (such as a PID controller) to achieve autonomous navigation. In another embodiment, a topology graph is generated based on a topology graph optimization algorithm. This involves constructing a topology graph by using feature points as nodes and edge weights as connections. Nodes in the topology graph represent key locations in the scene, edges represent connections between these locations, and edge weights reflect connection costs. The objective function for determining the navigation path is given by formula (2): (2) in, Represents the original length of the edge (which can be compared to the original distance between adjacent graph nodes below). This represents the estimated path length after graph optimization (which corresponds to the optimized estimated distance between nodes below). The weights of the edges (which can be compared to the weights of the edges connecting adjacent graph nodes below) are used to complete path planning based on the A* algorithm and the topology graph, generating the final flight path to ensure that the UAV can navigate autonomously in complex and ever-changing environments.

[0044] This invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. The method involves real-time acquisition of scene images of the UAV's navigation scenario over continuous time. Based on these images, interfering dynamic objects in the navigation scenario are identified, and the scene images are processed to remove these objects, resulting in an interference-suppressed scene image. Key feature points are extracted from the interference-suppressed scene image to obtain extracted scene key points. Based on these key points, a topological graph matching the UAV's navigation scenario is constructed, where the graph nodes are scene key points, and the weights of the edges represent the connectivity between nodes. Using the topological graph and a heuristic search algorithm, the navigation path of the UAV in the navigation scenario is determined. By dynamically suppressing interfering dynamic objects and mapping the topological graph, the path planning capability of the UAV in complex environments can be improved, ensuring safety and flight efficiency, and enabling autonomous navigation of the UAV in complex environments.

[0045] Figure 2 This is a schematic diagram of the process for identifying interfering dynamic objects in the scene to be navigated by a UAV based on scene images, provided by the present invention.

[0046] The following will combine Figure 2 The process of identifying interfering dynamic objects in the scene to be navigated by a UAV based on scene images provided by the present invention will be described.

[0047] In an exemplary embodiment of the present invention, combined with Figure 2 As can be seen, determining the interfering dynamic objects in the scene to be navigated by the UAV based on the scene image can include steps 210 to 230, and each step will be described below.

[0048] In step 210, based on the scene image, the position information of dynamic objects in the scene to be navigated by the UAV is detected by the target detection algorithm.

[0049] In one embodiment, a pre-trained YOLO object detection model is used to process real-time acquired scene images to identify the categories of dynamic objects in the images and their bounding box coordinates; that is, for a single image... Corresponding object detection results ,in It contains the object's bounding box and category information.

[0050] In step 220, the optical flow of pixels in the scene image is determined based on the scene images at adjacent time points.

[0051] In one embodiment, the motion of each pixel between two frames can also be estimated, assuming the current frame is... The previous frame was The optical flow of each pixel is calculated based on the optical flow method. This can be expressed as formula (3): (3) in, It is a pixel Horizontal and vertical displacement between two frames of images.

[0052] In step 230, based on optical flow and the position information, interfering dynamic objects in the scene to be navigated by the UAV are identified.

[0053] In another embodiment, the positional information of dynamic objects in the scene obtained using the YOLO algorithm can be combined with optical flow judgment. If a dynamic object (such as a fallen leaf, floating object, etc.) exists and its motion is inconsistent with the background motion in the camera's viewpoint, it is considered invalid depth information, that is, an interfering dynamic object. Invalid depth information usually refers to false depth information caused by sensor errors or objects such as fallen leaves.

[0054] In yet another exemplary embodiment of the present invention, continuing with the previously described embodiments, the determination of interfering dynamic objects in the scene to be navigated by the UAV based on optical flow and position information can be achieved in the following manner: Based on optical flow and position information, it is determined whether the motion trajectory of dynamic objects in the scene to be navigated by the UAV is consistent with the motion trajectory of the background in the camera view. The camera view is the view of the camera that acquires the scene image. When the motion trajectory of a dynamic object in the scene to be navigated by the drone is inconsistent with the motion trajectory of the background in the camera's view, the dynamic object is identified as an interfering dynamic object in the scene to be navigated by the drone.

[0055] In one embodiment, since the drone's own motion (such as forward movement or rotation) causes a consistent motion pattern across the entire image background, this embodiment can employ an affine transformation model or homography matrix to fit this global motion. Alternatively, the RANSAC (Random Sample Consensus) algorithm can be used to robustly estimate the transformation matrix M describing the overall motion trajectory of the background from the optical flow points of the entire image. Here, the transformation matrix M represents the background motion trajectory caused by the movement of the camera (which can correspond to the drone) acquiring the image between adjacent frames. For a background pixel p, its expected motion trajectory (i.e., the motion caused by the drone's own movement) can be calculated using p' = M × p.

[0056] Multiple optical flow points can be sampled within the bounding box region (corresponding to location information) of the dynamic object. The actual motion vectors of these optical flow points are aggregated to obtain the actual observed motion trajectory T_object of the dynamic object. Then, for the bounding box region where the same dynamic object is located, the expected motion vectors of the same sampled points in the region are calculated using the estimated global motion model M. If the object is part of a stationary background, it should have the motion. These expected motion vectors are aggregated to obtain the expected motion trajectory T_background of the background in that region. The difference D between the actual observed motion trajectory T_object and the expected background motion trajectory T_background is then calculated. If the difference D is greater than the threshold ε, the motion trajectory of the dynamic object is determined to be inconsistent with the background motion trajectory in the camera's view. This indicates that the motion of the object cannot be explained by the motion of the UAV itself; it is an entity with independent motion and is therefore identified as an interfering dynamic object.

[0057] In yet another embodiment, a flag bit can be defined. This represents a flag indicating the detection result. For example, in the case of a dynamic object in the previous embodiment, if the detection result shows "moving object and inconsistent optical flow", otherwise The final depth flow output is shown in formula (4): (4) This embodiment can intelligently distinguish between true motion (the object's own movement) and false motion (caused by camera movement), significantly reducing the false alarm rate and making the identification results of interfering objects more accurate and reliable.

[0058] Figure 3 This is a flowchart illustrating the process of determining the navigation path of a UAV in a navigation scenario provided by the present invention, based on a topology graph and a heuristic search algorithm.

[0059] The following will combine Figure 3The process of determining the navigation path of a UAV in a navigation scenario provided by the present invention, based on a topology graph and a heuristic search algorithm, is described.

[0060] In an exemplary embodiment of the present invention, combined with Figure 3 As can be seen, determining the navigation path of a UAV in the scenario to be navigated, based on the topology map and heuristic search algorithm, can include steps 310 to 330. Each step will be described below.

[0061] In step 310, based on the topology graph, the original distance between adjacent graph nodes and the weight of the edges connecting adjacent graph nodes are determined.

[0062] In step 320, the optimization objective is to minimize the target value obtained by accumulating the product of the difference of each distance and the weight of each corresponding edge, and to determine the distance between each optimized estimated node. The difference of distance is the difference between the original distance between nodes and the distance between nodes after optimization estimation.

[0063] In step 330, based on the distances between nodes after optimization estimation, a heuristic search algorithm is used to determine the navigation path of the UAV in the scenario to be navigated.

[0064] In one embodiment, the original inter-node distance between adjacent graph nodes can be determined based on the topology graph. and the weights of the edges connecting adjacent graph nodes. Furthermore, the optimization objective is to minimize the product of the differences in distances and the weights of the corresponding edges, and then use a heuristic search algorithm to determine the distances between nodes after optimization estimation.

[0065] The objective function for optimization can be achieved using formula (2) mentioned above.

[0066] In another embodiment, the original distances between nodes on each edge of the topology graph can be replaced with the optimized estimated distances between nodes to form an optimized topology graph. In one embodiment, the redundancy removal and optimization steps of the topology graph can be implemented based on the open-source optimization library g2o (General Graph Optimization). This method achieves global consistency optimization of the topology structure by minimizing the weighted error between observed edges and estimated edges between nodes. Then, taking the nearest node corresponding to the current position of the UAV as the path start point and the nearest node corresponding to the target position as the path end point, based on the optimized topology graph and combined with a heuristic search algorithm, such as the A* algorithm, or alternatively Dijkstra's algorithm, D*Lite, or a learning-based dynamic programming algorithm, the navigation path of the UAV in the scenario to be navigated is determined by combining topological weights and heuristic distances. In this embodiment, the specific planning algorithm for determining the navigation path of the UAV in the scenario to be navigated is not limited.

[0067] In this embodiment, by incorporating edge weights into the distance estimation optimization process, the final distance value used for path search not only reflects geometric distance but also more accurately represents the overall travel cost in actual navigation. The generated paths avoid routes that, while geometrically short, are practically difficult to traverse, greatly enhancing their practical value. Furthermore, since the optimized distance between nodes already includes environmental complexity information, the A* algorithm can more accurately assess the actual cost of a path during the search. This not only improves search efficiency and reduces unnecessary node expansion but, more importantly, ensures that the found path is optimal in terms of overall travel cost, not just the shortest geometric distance.

[0068] Figure 4 This is a schematic diagram of the process for determining the weight of the edge connecting adjacent graph nodes provided by the present invention.

[0069] The following will combine Figure 4 The process of determining the weight of the edge connecting adjacent graph nodes provided by the present invention will be explained.

[0070] In an exemplary embodiment of the present invention, combined with Figure 4 As can be seen, determining the weight of the edge connecting adjacent graph nodes can include steps 410 to 430, which will be described in detail below.

[0071] In step 410, the distance between adjacent graph nodes is determined.

[0072] In step 420, the weight values ​​of obstacles between adjacent graph nodes are determined.

[0073] In step 430, the weight of the edge connecting adjacent graph nodes is determined based on the distance between adjacent graph nodes and the weight value of obstacles between adjacent graph nodes.

[0074] In calculating the weights between edges, the connectivity between nodes can be constructed based on the relative positions of these feature points in the image. If two feature points... and If they are close together and there are no obvious obstacles in between, they can be connected by an edge.

[0075] In another embodiment, the present invention further defines an adjustment strategy for path length and obstacle information weights. During application, the distance between adjacent graph nodes can be determined. And determine the weight values ​​of obstacles between adjacent graph nodes. .

[0076] Furthermore, based on the distance between adjacent graph nodes... And the weight values ​​of obstacles between adjacent graph nodes. Determine the weights of the edges connecting adjacent graph nodes. Among them, the weight value Representative node and The strength or cost of connections between adjacent graph nodes, and the weight of edges connecting adjacent graph nodes. It can be expressed using formula (5): (5) in, Representative node and The Euclidean distance between them This represents the weight of obstacles in the scene, that is, the weight value of the obstacles. , It is a coefficient that adjusts the path length and the degree of influence of obstacle information.

[0077] In this embodiment, by organically combining geometric distance with obstacle influence, the constructed edge weights can accurately characterize the comprehensive cost in actual navigation, including multiple dimensions such as flight distance, obstacle avoidance difficulty, and safety risks, providing a more reliable cost assessment basis for path planning.

[0078] Continuing with the example described above, before determining the navigation path of the UAV in the scenario to be navigated based on the topology graph and heuristic search algorithm (corresponding to step 150), the UAV autonomous navigation method based on dynamic interference suppression and topology feature mapping may further include the following steps: Remove redundant nodes and edges from the topology graph to obtain the optimized topology graph; The navigation path of the UAV in the scenario to be navigated, based on the topology graph and heuristic search algorithm, can be determined in the following ways: Based on the optimized topology map and heuristic search algorithm, the navigation path of the UAV in the scenario to be navigated is determined.

[0079] In one embodiment, redundant nodes and edges can be removed based on the g2o algorithm to construct an efficient topology map, thereby improving the graph's expressive power and efficiency, resulting in an optimized topology map. Furthermore, based on the optimized topology map and a heuristic search algorithm, the navigation path of the UAV in the scenario to be navigated is determined, thereby improving the accuracy of the navigation path.

[0080] As described above, this invention provides an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. The method involves real-time acquisition of scene images of the UAV's navigation scenario over continuous time. Based on these images, interfering dynamic objects in the navigation scenario are identified, and the scene images are processed to remove these objects, resulting in an interference-suppressed scene image. Key feature points are extracted from the interference-suppressed scene image to obtain extracted scene key points. Based on these key points, a topological graph matching the UAV's navigation scenario is constructed, where the graph nodes are scene key points, and the weights of the edges represent the connectivity between nodes. Based on the topological graph and a heuristic search algorithm, the navigation path of the UAV in the navigation scenario is determined. By dynamically suppressing interfering dynamic objects and mapping the topological graph, the path planning capability of the UAV in complex environments can be improved, ensuring safety and flight efficiency, and enabling autonomous navigation of the UAV in complex environments.

[0081] The UAV autonomous navigation device based on dynamic interference suppression and topology feature mapping provided by the present invention will be described below. The UAV autonomous navigation device based on dynamic interference suppression and topology feature mapping described below can be referred to in correspondence with the UAV autonomous navigation method based on dynamic interference suppression and topology feature mapping described above.

[0082] Figure 5 This is a schematic diagram of the structure of the UAV autonomous navigation device based on dynamic interference suppression and topological feature mapping provided by the present invention.

[0083] The following will combine Figure 5 The structure of the UAV autonomous navigation device based on dynamic interference suppression and topological feature mapping provided by the present invention will be described.

[0084] In an exemplary embodiment of the present invention, combined with Figure 5As can be seen, the UAV autonomous navigation device based on dynamic interference suppression and topological feature mapping may include a data acquisition module 510, a determination module 520, an extraction module 530, a construction module 540, and a generation module 550. Each module will be described in detail below.

[0085] The acquisition module 510 can be configured to acquire scene images of the drone's navigation scenario over a continuous period of time in real time; The determining module 520 can be configured to determine, based on the scene image, interfering dynamic objects in the scene to be navigated by the UAV, and perform interfering dynamic object removal processing on the scene image based on the interfering dynamic objects to obtain an interference-suppressed scene image. The extraction module 530 can be configured to extract key feature points from the scene image after interference suppression to obtain extracted scene key points, wherein the scene key points are points in the scene to be navigated by the UAV that affect the navigation path of the UAV. The construction module 540 can be configured to construct a topology graph that matches the scene to be navigated by the UAV based on the scene key points, wherein the graph nodes of the topology graph are the scene key points, and the weights of the edges in the topology graph are the connection relationships between the graph nodes; The generation module 550 can be configured to determine the navigation path of the UAV in the scenario to be navigated by the UAV based on the topology map and the heuristic search algorithm.

[0086] In an exemplary embodiment of the present invention, the determining module 520 may determine the interfering dynamic objects in the scene to be navigated by the UAV based on the scene image in the following manner: Based on the scene image, the position information of dynamic objects in the scene to be navigated by the UAV is detected by a target detection algorithm; Based on the scene images at adjacent time points, determine the optical flow of pixels in the scene images; Based on the optical flow and the location information, the interfering dynamic objects in the scene to be navigated by the UAV are identified.

[0087] In an exemplary embodiment of the present invention, the determining module 520 can determine the interfering dynamic objects in the scene to be navigated by the UAV based on the optical flow and the position information in the following manner: Based on the optical flow and the position information, it is determined whether the motion trajectory of the dynamic object in the scene to be navigated by the UAV is consistent with the background motion trajectory in the camera view, wherein the camera view is the view of the camera that acquires the scene image; If the motion trajectory of a dynamic object in the scene to be navigated by the UAV is inconsistent with the motion trajectory of the background in the camera's view, the dynamic object is determined to be an interfering dynamic object in the scene to be navigated by the UAV.

[0088] In an exemplary embodiment of the present invention, the extraction module 530 can extract key feature points from the interference-suppressed scene image in the following manner to obtain the extracted scene key points: The key feature points of the scene image after interference suppression are extracted based on the random sampling consensus algorithm to obtain the extracted scene key points.

[0089] In an exemplary embodiment of the present invention, the generation module 550 may determine the navigation path of the UAV in the scenario to be navigated by the UAV based on the topology map and the heuristic search algorithm in the following manner: Based on the topology graph, the original distance between adjacent graph nodes and the weight of the edges connecting adjacent graph nodes are determined. The optimization objective is to minimize the target value obtained by accumulating the product of the difference of each distance and the weight of each corresponding edge. The optimized estimated distance between nodes is then determined, wherein the difference of distance is the difference between the original distance between nodes and the optimized estimated distance between nodes. Based on the optimized estimated distances between nodes, a heuristic search algorithm is used to determine the navigation path of the UAV in the scenario to be navigated.

[0090] In an exemplary embodiment of the present invention, the generation module 550 may determine the weights of the edges connecting the adjacent graph nodes in the following manner: Determine the distance between adjacent graph nodes; Determine the weight values ​​of obstacles between adjacent graph nodes; The weights of the edges connecting adjacent graph nodes are determined based on the distance between adjacent graph nodes and the weights of obstacles between them.

[0091] In an exemplary embodiment of the present invention, the generation module 550 may further be configured to: Remove redundant nodes and edges from the topology graph to obtain the optimized topology graph. The generation module 550 can determine the navigation path of the UAV in the scenario to be navigated by the UAV based on the topology map and the heuristic search algorithm in the following way: Based on the optimized topology map and heuristic search algorithm, the navigation path of the UAV in the scenario to be navigated by the UAV is determined.

[0092] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630, and a communications bus 640, wherein the processor 610, the communications interface 620, and the memory 630 communicate with each other through the communications bus 640. The processor 610 can call logic instructions in the memory 630 to execute an autonomous navigation method for unmanned aerial vehicles (UAVs) based on dynamic interference suppression and topological feature mapping. This method includes: real-time acquisition of scene images of the UAV's navigation scenario over a continuous time period; identifying interfering dynamic objects in the scene image based on the scene image, and performing interference dynamic object removal processing on the scene image based on the interfering dynamic objects to obtain an interference-suppressed scene image; extracting key feature points from the interference-suppressed scene image to obtain extracted scene key points, wherein the scene key points are points in the UAV's navigation scenario that affect the UAV's navigation path; constructing a topological graph matching the UAV's navigation scenario based on the scene key points, wherein the graph nodes of the topological graph are the scene key points, and the weights of the edges in the topological graph represent the connection relationships between the graph nodes; and determining the UAV's navigation path in the UAV's navigation scenario based on the topological graph and a heuristic search algorithm.

[0093] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0094] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping provided by the above methods. The method includes: real-time acquisition of scene images of the UAV's navigation scene over a continuous time period; based on the scene images, identifying interfering dynamic objects in the UAV's navigation scene, and performing interference dynamic object removal processing on the scene images based on the interfering dynamic objects to obtain an interference-suppressed scene image; extracting key feature points from the interference-suppressed scene image to obtain extracted scene key points, wherein the scene key points are points in the UAV's navigation scene that affect the UAV's navigation path; constructing a topology graph matching the UAV's navigation scene based on the scene key points, wherein the graph nodes of the topology graph are the scene key points, and the weights of the edges in the topology graph are the connection relationships between the graph nodes; and determining the UAV's navigation path in the UAV's navigation scene based on the topology graph and a heuristic search algorithm.

[0095] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the UAV autonomous navigation method based on dynamic interference suppression and topological feature mapping provided by the above methods. The method includes: acquiring scene images of the UAV's navigation scene over a continuous time period in real time; identifying interfering dynamic objects in the UAV's navigation scene based on the scene images, and performing interference dynamic object removal processing on the scene images based on the interfering dynamic objects to obtain an interference-suppressed scene image; extracting key feature points from the interference-suppressed scene image to obtain extracted scene key points, wherein the scene key points are points in the UAV's navigation scene that affect the UAV's navigation path; constructing a topology graph matching the UAV's navigation scene based on the scene key points, wherein the graph nodes of the topology graph are the scene key points, and the weights of the edges in the topology graph are the connection relationships between the graph nodes; and determining the UAV's navigation path in the UAV's navigation scene based on the topology graph and a heuristic search algorithm.

[0096] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0097] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for autonomous navigation of unmanned aerial vehicles based on dynamic interference suppression and topology feature mapping, characterized in that, The method comprises: collecting a scene image of a scene to be navigated by a UAV in real time at a continuous time; based on the scene image, determining a dynamic interference object in the scene to be navigated by the UAV, and performing dynamic interference object culling processing on the scene image based on the dynamic interference object, to obtain a scene image after interference suppression; extracting key feature points from the scene image after interference suppression, to obtain extracted scene key points, wherein the scene key points are points in the scene to be navigated by the UAV that affect the direction of a UAV navigation path; based on the scene key points, constructing a topological graph matched with the scene to be navigated by the UAV, wherein the graph nodes of the topological graph are the scene key points, and the weights of the edges connecting adjacent graph nodes in the topological graph are the connection relationships between the graph nodes; based on the topological graph and a heuristic search algorithm, determining a navigation path of the UAV in the scene to be navigated by the UAV.

2. The method of claim 1, wherein, The method comprises: based on the scene image, detecting the position information of dynamic objects in the scene to be navigated by the UAV by using a target detection algorithm; based on the scene image at an adjacent time, determining the optical flow of the pixels in the scene image; based on the optical flow and the position information, determining the dynamic interference object in the scene to be navigated by the UAV.

3. The method of claim 2, wherein, The method comprises: based on the optical flow and the position information, determining whether the motion trajectory of the dynamic object in the scene to be navigated by the UAV is consistent with the background motion trajectory in the camera view, wherein the camera view is the camera view for collecting the scene image; in the case where the motion trajectory of the dynamic object in the scene to be navigated by the UAV is not consistent with the background motion trajectory in the camera view, determining that the dynamic object is the dynamic interference object in the scene to be navigated by the UAV.

4. The method of claim 1, wherein, The method comprises: based on the random sample consensus algorithm, extracting key feature points from the scene image after interference suppression, to obtain extracted scene key points.

5. The method of claim 1, wherein, The method comprises: based on the topological graph, determining the original node-to-node distance between adjacent graph nodes, and the weight of the edge connecting adjacent graph nodes; minimizing the target value obtained by accumulating the product of the distance difference and the weight of each corresponding edge as an optimization objective, to determine the optimized estimated node-to-node distance, wherein the distance difference is the difference between the original node-to-node distance and the optimized estimated node-to-node distance; based on the optimized estimated node-to-node distance, combining the heuristic search algorithm to determine the navigation path of the UAV in the scene to be navigated by the UAV.

6. The method of claim 5, wherein, The weight of the edge connecting adjacent graph nodes is determined in the following manner: determining the distance between adjacent graph nodes; determining the weight value of the obstacle existing between adjacent graph nodes; Determine the weight of the edge connecting the adjacent graph nodes based on the distance between the adjacent graph nodes and the weight value of the obstacle existing between the adjacent graph nodes.

7. The method of claim 1 to 6, wherein, Before determining the navigation path of the UAV in the scene to be navigated by the UAV based on the topological graph and the heuristic search algorithm, the method further comprises: Removing the redundant graph nodes and the redundant edges in the topological graph to obtain an optimized topological graph; The method of determining the navigation path of the UAV in the scene to be navigated by the UAV based on the topological graph and the heuristic search algorithm specifically comprises: Determine the navigation path of the UAV in the scene to be navigated by the UAV based on the optimized topological graph and the heuristic search algorithm.

8. An autonomous navigation apparatus for unmanned aerial vehicles based on dynamic interference suppression and topology feature mapping, characterized in that, The device comprises: The acquisition module is configured to acquire a scene image of the scene to be navigated by the UAV in real time; The determination module is configured to determine a dynamic interference object in the scene to be navigated by the UAV based on the scene image, and perform dynamic interference object removal processing on the scene image based on the dynamic interference object to obtain a scene image after interference suppression; The extraction module is configured to extract a key feature point from the scene image after interference suppression to obtain an extracted scene key point, wherein the scene key point is a point in the scene to be navigated by the UAV that affects the navigation path of the UAV; The construction module is configured to construct a topological graph matched with the scene to be navigated by the UAV based on the scene key point, wherein a graph node of the topological graph is the scene key point, and a weight of an edge in the topological graph is a connection relationship between the graph nodes. The generation module is configured to determine a navigation path of the UAV in the scene to be navigated by the UAV based on the topological graph and the heuristic search algorithm.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, The processor executes the computer program to implement the method for autonomous navigation of the UAV based on dynamic interference suppression and topological feature mapping according to any one of claims 1 to 7. 10.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method for autonomous navigation of the UAV based on dynamic interference suppression and topological feature mapping according to any one of claims 1 to 7.

Images