Vision-guided drone rapid dense recovery method, device and equipment

Abstract

The present application relates to a visual guidance-based unmanned aerial vehicle rapid dense recovery method, device and equipment. The method comprises the following steps: deploying large-scale mark units and small-scale mark units in a recovery area; the unmanned aerial vehicle is equipped with a variable field of view imaging device, and a multi-stage visual guidance process is used for recovery positioning: in the long-distance guidance stage, the variable field of view imaging device captures the large-scale mark units in a large field of view, and solves the initial pose of the unmanned aerial vehicle relative to the recovery area; in the medium-distance guidance stage, the variable field of view imaging device adjusts the field of view size to capture the small-scale mark units, identifies the recovery position in combination with the topological structure in the small-scale mark units, and positions the coarse pose of the unmanned aerial vehicle; in the near-distance guidance stage, sub-pixel level feature extraction is performed on the small-scale mark units, then the accurate pose of the unmanned aerial vehicle is calculated, and the unmanned aerial vehicle is guided to land in the recovery area. The present application can realize continuous and dense recovery of unmanned aerial vehicles.

Description

Technical Field

[0001] This invention relates to the field of drone recovery technology, and in particular to a method, apparatus, and equipment for rapid and intensive drone recovery based on vision guidance. Background Technology

[0002] With the widespread application of drone technology in logistics, agricultural plant protection, and emergency rescue, the problem of easy takeoff but difficult recovery has become increasingly prominent. Traditional recovery methods mainly include manual remote-controlled landing, parachute recovery, net-crash recovery, and robotic arm grasping. However, these methods all reveal significant drawbacks in scenarios with multiple drones and limited space: low site utilization, parachute landing or net-crash recovery requires a large safety buffer zone and cannot support dense recovery; robotic arm recovery requires reserved space for mechanical movement, resulting in long intervals between single-drone recovery; limited recovery accuracy, manual remote control relies on operator experience and is easily affected by ambient light, GPS signal obstruction, or electromagnetic interference, leading to a large landing dispersion radius; complex equipment and high cost, with robotic arms and parachute systems requiring additional launch or recovery devices, increasing platform weight and maintenance costs.

[0003] Existing visual-assisted landing methods mostly employ independent GPS / RTK positioning, which poses risks of frequency interference and path crossing when multiple drones approach simultaneously. While visual guidance technologies attempt to improve positioning accuracy through planar cooperative targets such as AprilTag and ArUco, the following bottlenecks remain: fixed-focal-length cameras struggle to balance large-scale capture and small-scale recognition due to the conflict between field of view and resolution during the transition from long-range search to close-range precision landing; cooperative targets are often static reflectors or passive markers, easily obscured in complex backgrounds such as strong light or nighttime, resulting in high false recognition rates; existing visual constraint equations typically rely on camera focal length parameters, requiring recalculation of focal lengths when drone swarms use multi-focal-length or zoom cameras, leading to poor system versatility; and the lack of dynamic marker management strategies linked to the occupancy status of the recovery area interferes with subsequent landings. Summary of the Invention

[0004] Therefore, it is necessary to provide a vision-guided method, apparatus, and equipment for rapid and dense drone retrieval that can quickly complete the intensive retrieval of drones in a limited space, addressing the aforementioned technical problems.

[0005] A method for rapid and intensive recovery of unmanned aerial vehicles (UAVs) based on vision guidance, the method comprising: Large-scale marker units and small-scale marker units are deployed in the recycling area, wherein the small-scale marker units are located within the large-scale marker units; The drone is equipped with a variable field of view imaging device and performs recovery and positioning through a multi-stage visual guidance process: wherein the multi-stage visual guidance process includes a long-distance guidance stage, a medium-distance guidance stage and a short-distance guidance stage; During the long-range guidance phase, the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area. Based on the initial pose, the UAV's heading is adjusted, and it enters the mid-range guidance phase; in the mid-range guidance phase, the variable field of view imaging device adjusts the field of view size to capture small-scale marker units, and identifies the retrieval position by combining the topological structure in the small-scale marker units, thereby locating the coarse pose of the UAV. Based on the coarse pose, the UAV's heading is adjusted, and it enters the close-range guidance phase. During the close-range guidance phase, sub-pixel-level features are extracted from the small-scale marker units, and then the precise pose of the UAV is calculated. Based on the precise pose, the UAV is guided to land in the recovery area.

[0006] On the other hand, a vision-guided rapid and intensive drone recovery device is also provided, the device comprising: A recycling area deployment module is used to deploy large-scale marker units and small-scale marker units in the recycling area, wherein the small-scale marker units are located within the large-scale marker units; An airborne vision module is used for UAVs to carry variable field-of-view imaging devices and perform recovery and positioning through a multi-stage visual guidance process: wherein the multi-stage visual guidance process includes a long-range guidance stage, a medium-range guidance stage and a short-range guidance stage. The flight control module is used for the following purposes: During the long-range guidance phase, the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area; based on the initial pose, the UAV's heading is adjusted, and it enters the mid-range guidance phase; During the long-range guidance phase, the variable field-of-view imaging device adjusts the field of view size to capture small-scale marker units, identifies the recovery position by combining the topological structure in the small-scale marker units, and locates the coarse pose of the UAV; Based on the coarse pose, the UAV's heading is adjusted, and it enters the close-range guidance phase; During the close-range guidance phase, sub-pixel-level features are extracted from the small-scale marker units, and then the precise pose of the UAV is calculated, and the UAV is guided to land in the recovery area based on the precise pose.

[0007] On the other hand, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described vision-guided rapid and intensive drone recovery method.

[0008] Compared with existing technologies, the vision-guided rapid and intensive drone recovery method, apparatus, and equipment provided by this invention have the following advantages: 1. By using a hierarchical layout of large-scale marker units nested with small-scale marker units, combined with a collaborative design of progressive pose positioning in three stages (far, medium, and near), after adjusting the field of view size at medium distance, each UAV locks onto a small-scale recovery position within its own target large-scale unit. By distinguishing exclusive landing positions through topological structure, UAVs can achieve phased, high-precision pose guidance in limited spaces, significantly reducing the safety distance required for single-UAV recovery, supporting multiple UAVs to land continuously at a higher density, and greatly improving site utilization compared to traditional recovery methods.

[0009] 2. During the long-range phase, the large field of view of the variable field-of-view imaging device can quickly capture large-scale marker units, and the initial pose can be directly calculated based on the large-scale marker units, which can significantly shorten the preparation time from entering the recovery airspace to starting the landing. Combined with the field of view transformation and pose progression from medium range to close range, the efficiency of the overall recovery process can be improved and rapid positioning can be achieved.

[0010] 3. By using a progressive solution logic of initial pose, coarse pose, and precise pose, the risk of pose deviation can be reduced. Even when the distance between retrieval positions is small, collisions caused by inaccurate positioning can be avoided, providing a safety guarantee for dense retrieval. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention, and those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0012] Figure 1 This is an application environment diagram of the vision-guided rapid and dense drone recovery method provided in Example 1; Figure 2 This is a schematic diagram of the recycling area deployment provided in Example 1; Figure 3 This is a schematic diagram of the viewing angle switching of the variable field-of-view imaging device provided in Example 1; Figure 4 This is a structural block diagram of the vision-guided rapid and dense drone recovery device provided in Example 2; Figure 5 This is an internal structural diagram of the computer device provided in Example 3.

[0013] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0015] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a specific posture. If the specific posture changes, the directional indication will also change accordingly.

[0016] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0017] It is understood that the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0018] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0019] Example 1 like Figure 1 As shown, this embodiment provides a vision-guided method for rapid and intensive drone recovery, including the following steps: Step 201: Deploy large-scale marker units and small-scale marker units in the recycling area, wherein the small-scale marker units are located within the large-scale marker units; Step 202: The UAV is equipped with a variable field of view imaging device and performs recovery and positioning through a multi-stage visual guidance process: the multi-stage visual guidance process includes a long-distance guidance stage, a medium-distance guidance stage and a short-distance guidance stage.

[0020] Step 203: In the long-range guidance phase, the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area.

[0021] Step 204: Adjust the heading of the UAV based on the initial pose and enter the mid-range guidance phase; in the range guidance phase, the variable field of view imaging device adjusts the field of view size to capture small-scale marker units, and identifies the recovery position by combining the topological structure in the small-scale marker units to locate the coarse pose of the UAV.

[0022] Step 205: Adjust the heading of the UAV based on the coarse pose and enter the close-range guidance phase; in the close-range guidance phase, extract sub-pixel-level features from small-scale marker units, then calculate the precise pose of the UAV, and guide the UAV to land in the recovery area based on the precise pose.

[0023] In the specific implementation of step 201, such as Figure 2 As shown, the drone recovery area is first surveyed on-site. Based on the actual area, shape and size of the drone, the recovery area is divided into several independent sub-recovery areas, using the horizontal and vertical directions as a reference. The number of sub-recovery areas can be flexibly adjusted according to the size of the site, and the division density can be adapted to the size of the drone.

[0024] For each sub-recycling area, set a first cooperative objective (cooperative objective) at its edge corner. ,in, , The number of the first cooperative target is determined by the large-scale marker unit. The first cooperative target uses a programmable point light source, which is installed on a collapsible platform or embedded in the ground surface of the recovery area to prevent accidental damage to the UAV.

[0025] Subsequently, within each sub-recycling area, a second cooperative objective is deployed according to a pre-defined topology. ,in, , (The number of second cooperative targets) yields several small-scale marker units. The second cooperative target is also a programmable point light source, and its emission frequency is distinct from that of the first cooperative target to avoid mutual interference. It is embedded in the ground surface of the recovery area to prevent accidental damage to the drone. It is worth noting that the programmable point light source exhibits brightness variations over time.

[0026] Finally, the three-dimensional coordinates of all first and second cooperative targets in the local coordinate system of the recovery area are measured using a total station or RTK, and the coordinate data is stored in the database of the UAV's onboard vision module to ensure that the spatial position reference of all marker units is consistent.

[0027] It is understandable that by using a hierarchical layout of large-scale units containing small-scale units, the recovery area is divided into finely defined independent recovery positions, significantly reducing the safety distance required for single-unit recovery and laying a spatial foundation for dense multi-unit recovery. The programmable point light source's design, which allows it to be tilted or embedded, not only avoids collision damage to cooperative targets when the UAV lands but also provides hardware support for subsequent anti-interference identification. The unified local coordinate system ensures the consistency of the reference for subsequent pose calculations, avoiding positioning deviations caused by coordinate confusion.

[0028] In the specific implementation of step 202, the variable field-of-view imaging device is composed of a variable focal length camera or a multi-focal length camera (a multi-camera system composed of cameras with different focal lengths), such as... Figure 3 As shown, by dynamically adjusting the focal length or a combination of telephoto, medium telephoto, and short telephoto, the field of view can be adjusted to achieve variable field of view observation, thereby enabling high-resolution and clear imaging of cooperative targets at different distances.

[0029] First, the variable field of view imaging device is mounted on the underside of the drone. The imaging unit (camera) of the variable field of view imaging device is then calibrated for internal and external parameters. Internal parameter calibration requires determining the camera's focal length, principal point coordinates, distortion coefficients, and other core parameters. External parameter calibration requires determining the position and attitude relationship of the imaging unit relative to the drone's body coordinate system. After calibration, the parameters are stored in the drone's visual guidance system as the basis for subsequent pose calculation.

[0030] Next, the specific stages of the multi-stage visual guidance process are defined, namely the long-distance guidance stage, the medium-distance guidance stage, and the close-distance guidance stage. The field of view adjustment rules for each stage must match the type of imaging device: if it is a variable focal length camera, the field of view is adjusted by dynamically adjusting the focal length; if it is a multi-focal length camera, the variable field of view observation is achieved by switching between sub-cameras with different focal lengths (a combination of telephoto, medium telephoto, and short telephoto).

[0031] Finally, the triggering conditions for field of view adjustment are set: the visual guidance system will automatically determine whether to enter the next guidance stage based on the UAV pose data calculated in the previous stage, and trigger the variable field of view imaging device to switch to the corresponding field of view, so as to ensure that high-resolution and clear imaging of ground cooperative targets can be achieved at different distances.

[0032] It is understood that the variable field-of-view imaging device of the present invention uses a variable focal length camera or a multi-focal length camera, which solves the problems of small field of view at long distances and low resolution at close distances of traditional fixed focal length cameras, and takes into account the needs of large-scale search and small-scale recognition. The calibration of intrinsic and extrinsic parameters provides an accurate parameter basis for subsequent pose calculation, avoiding positioning errors caused by missing parameters. The division of the multi-stage guidance process and the setting of the field-of-view adjustment trigger mechanism enable the visual guidance process to be dynamically adjusted according to the distance between the UAV and the recovery area without manual intervention, which improves the automation of the recovery process and adapts to the high-efficiency needs of multiple UAVs landing continuously in dense recovery areas.

[0033] In the specific implementation of step 203, when the drone to be recovered enters the airspace of the recovery area, when the distance between the drone and the recovery area is detected to reach the threshold of the long-distance guidance stage, the variable field of view imaging device is activated and adjusted to the large field of view mode to capture large-scale marker units.

[0034] The variable field-of-view imaging device continuously captures ground images of the recovery area with a large field of view, processes the acquired images, and identifies the first cooperative target at the edge corners of large-scale marker units. After successfully capturing the first cooperative target, its two-dimensional coordinates in the image coordinate system are extracted, and its pose is calculated using a visual constraint model independent of the variable field-of-view imaging device's focal length to obtain the initial pose of the UAV. The expression for the visual constraint model is: ; ; In the formula, This represents the intrinsic parameter matrix of the camera corresponding to the captured cooperative target; This represents the extrinsic rotation matrix of the camera corresponding to the captured cooperative target; This indicates the three-dimensional position coordinates of the camera corresponding to the captured cooperative target in the UAV's body coordinate system; This represents the three-dimensional position coordinates of the corresponding cooperative target in the local coordinate system of the recovery area; , This indicates the position and attitude of the drone relative to the recovery area; The image coordinates representing the cooperative objective.

[0035] Based on the initial pose, the drone adjusts its course, descends towards the sub-recovery area, and enters the mid-range guidance phase.

[0036] Understandably, image acquisition and target recognition in the wide field-of-view mode enable rapid localization of the recovery area, avoiding prolonged hovering and searching by the UAV over the recovery area and shortening the recovery process time. The designed visual constraint model, independent of the focal length of the variable field-of-view imaging device, allows the UAV to directly utilize multi-focal length images for localization without repeated calibration. The combined strategy of long-range wide field-of-view search and short-range high-resolution precise localization ensures a high success rate for marker capture, achieving full-process visual guidance from coarse alignment to precise ground contact, thus improving the efficiency and versatility of pose calculation. The initial pose calculation and flight control adjustments ensure that the UAV does not deviate from the recovery area before entering the mid-range phase, providing a directional reference for subsequent precise locking of the recovery position and avoiding the risk of collisions due to path intersections between multiple UAVs at long distances.

[0037] In the specific implementation of step 204, the UAV flies down and approaches the sub-recovery area. When the descent reaches the threshold of the mid-range guidance phase, the variable field of view imaging device adjusts the field of view size to capture small-scale marker units.

[0038] The variable field-of-view imaging device adjusts the field of view to capture small-scale marker units inside large-scale marker units with a small field of view. Since both the first and second cooperative targets have optical features, the device dynamically adjusts the optical features of the first and second cooperative targets to generate interference-filtered images. During the mid-range guidance phase, the UAV pose is calculated based on the interference-filtered images.

[0039] Specifically, a pre-set program controls the first and second cooperative targets to maintain opposite brightness relationships. By acquiring images of the bright and dark states of the target and subtracting them, background interference is filtered out, enabling accurate identification of cooperative targets in complex environments. For example, the first and second cooperative targets alternate between "dark" and "bright" states. The variable field-of-view imaging device simultaneously acquires bright and dark images composed of the first and second cooperative targets. Differential operations are used to filter out background interference such as strong ground light and shadows, resulting in a filtered image. It is worth noting that other brightness relationships can also be used, and the example is not limited to that described in this embodiment.

[0040] The system identifies second cooperative targets in the interference-filtered image. If the optical state of a second cooperative target is detected as constantly lit or constantly off, it indicates that the recovery position has been occupied and the recovery position is automatically excluded. For unoccupied recovery positions, the system analyzes the topology of the second cooperative target and matches it with the preset recovery position number topology in the airborne database to determine the unique available target recovery position number.

[0041] After successfully capturing the second cooperative target, subpixel-level coordinate extraction is performed on the second cooperative target at the recovery position to obtain higher-precision cooperative target image coordinates. These coordinates, along with the preset three-dimensional coordinates of the second cooperative target, are then substituted into a visual constraint model that is independent of the focal length of the variable field-of-view imaging device to calculate the coarse pose of the UAV relative to the target recovery position.

[0042] Based on the recovery position number and coarse pose data, the attitude of the UAV is adjusted to control it to move towards the recovery position and enter the close-range guidance phase.

[0043] Understandably, the small field-of-view mode improves the imaging resolution of small-scale marker units, and combined with sub-pixel-level feature extraction, provides a high-precision feature foundation for subsequent precise positioning. Bright-dark image differential operations effectively eliminate background interference in complex environments, effectively eliminating interference from strong light, nighttime, or dust, ensuring accurate marker identification even under complex conditions, guaranteeing the continuity and reliability of the recovery process, and improving the robustness of UAV recovery in all-weather, complex environments. The exclusion of occupied recovery positions and topology recognition ensure that multiple UAVs can each lock onto their dedicated recovery positions at mid-range, avoiding target contention and collisions, and meeting the needs of multi-UAV collaborative intensive recovery.

[0044] In the specific implementation of step 205, the variable field-of-view imaging device maintains a small field-of-view mode and continuously focuses on the successfully captured second cooperative target. Subsequently, subpixel-level coordinate extraction is performed on the second cooperative target at the recovery position to further obtain higher-precision cooperative target image coordinates. These coordinates, along with the preset three-dimensional coordinates of the second cooperative target, are substituted into a visual constraint model independent of the focal length of the variable field-of-view imaging device to calculate the precise pose of the UAV relative to the target recovery position.

[0045] Based on the recovery position number and precise pose data, the drone's attitude is corrected, and the drone is controlled to descend toward the recovery position until it is successfully guided to land at the recovery position and complete the recovery.

[0046] Understandably, sub-pixel-level feature extraction significantly improves the coordinate positioning accuracy of the second cooperative target. Combined with a focal length-independent visual constraint model, it can dynamically correct the UAV's pose deviation, improve landing accuracy, avoid multi-UAV collisions, and provide a foundation for dense recovery. The continuous use of the small field-of-view mode and image enhancement processing ensures that the features of the second cooperative target are clearly distinguishable at close range, solving the problem of traditional visual guidance being easily obscured by the aircraft body and having blurred features at close range. This adapts to dense recovery scenarios where multiple UAVs land continuously, improving the reliability of the recovery process.

[0047] It should be understood that, although this embodiment Figure 1The steps are shown sequentially as indicated by the arrows, but they are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are performed; they can be executed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0048] Example 2 Based on the vision-guided rapid and dense drone recovery method in Embodiment 1, this embodiment discloses a vision-guided rapid and dense drone recovery device, such as... Figure 4 As shown, the vision-guided rapid and dense drone recovery device includes: a recovery area deployment module 401, an airborne vision module 402, and a flight control module 403, wherein: The recycling area deployment module 401 is used to deploy large-scale marker units and small-scale marker units in the recycling area, wherein the small-scale marker units are located within the large-scale marker units.

[0049] The airborne vision module 402 is used for UAVs to carry variable field of view imaging devices and to perform recovery and positioning through a multi-stage visual guidance process: the multi-stage visual guidance process includes a long-range guidance stage, a medium-range guidance stage and a short-range guidance stage.

[0050] The flight control module 403 is used in the long-range guidance phase, where the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area; based on the initial pose, the UAV's heading is adjusted, and it enters the mid-range guidance phase; in the long-range guidance phase, the variable field-of-view imaging device adjusts the field of view size to capture small-scale marker units, and identifies the recovery position by combining the topological structure in the small-scale marker units, thus locating the coarse pose of the UAV; based on the coarse pose, the UAV's heading is adjusted, and it enters the short-range guidance phase; in the short-range guidance phase, sub-pixel-level features are extracted from the small-scale marker units, and then the precise pose of the UAV is calculated, and the UAV is guided to land in the recovery area based on the precise pose.

[0051] In this embodiment, the specific working process and working principle of the recovery area deployment module 401, the airborne vision module 402, and the flight control module 403 are the same as those in Embodiment 1, and therefore will not be described again in this embodiment. Each unit module can be implemented entirely or partially through software, hardware, or a combination thereof. Each unit module can be embedded in or independent of the processor in the computer device in hardware form, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above unit modules.

[0052] Example 3 like Figure 5 The diagram illustrates a terminal device disclosed in this embodiment, comprising a transmitter, a receiver, a memory, and a processor. The transmitter transmits instructions and data, the receiver receives instructions and data, the memory stores computer-executed instructions, and the processor executes the computer-executed instructions stored in the memory to implement the method described in Embodiment 1 above.

[0053] It is important to note that the aforementioned memory can be either standalone or integrated with the processor. When the memory is set up independently, the terminal device also includes a bus for connecting the memory and the processor.

[0054] Example 4 This embodiment discloses a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the method in Embodiment 1 above.

[0055] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0056] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0057] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for rapid and intensive recovery of unmanned aerial vehicles (UAVs) based on vision guidance, characterized in that, The method includes: Large-scale marker units and small-scale marker units are deployed in the recycling area, wherein the small-scale marker units are located within the large-scale marker units; The drone is equipped with a variable field of view imaging device and performs recovery and positioning through a multi-stage visual guidance process: wherein the multi-stage visual guidance process includes a long-distance guidance stage, a medium-distance guidance stage and a short-distance guidance stage; During the long-range guidance phase, the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area. Based on the initial pose, the UAV's heading is adjusted, and it enters the mid-range guidance phase; in the mid-range guidance phase, the variable field of view imaging device adjusts the field of view size to capture small-scale marker units, and identifies the retrieval position by combining the topological structure in the small-scale marker units, thereby locating the coarse pose of the UAV. Based on the coarse pose, the UAV's heading is adjusted, and it enters the close-range guidance phase. During the close-range guidance phase, sub-pixel-level features are extracted from the small-scale marker units, and then the precise pose of the UAV is calculated. Based on the precise pose, the UAV is guided to land in the recovery area.

2. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 1, characterized in that, Large-scale and small-scale marker units are deployed in the recycling area, including: The recycling area is divided into several sub-recycling areas based on the horizontal and vertical directions. A first cooperative target is set at the edge corner of each sub-recycling area to obtain a large-scale marker unit. In each sub-recycling region, a second cooperative objective is set according to a preset topology, resulting in several small-scale marker units.

3. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 1 or 2, characterized in that, In the multi-stage visual guidance process, each stage uses a visual constraint model that is independent of the focal length of the variable field of view imaging device to calculate the pose.

4. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 3, characterized in that, The expression for the visual constraint model is: ； ； In the formula, This represents the intrinsic parameter matrix of the camera corresponding to the captured cooperative target; This represents the extrinsic rotation matrix of the camera corresponding to the captured cooperative target; This indicates the three-dimensional position coordinates of the camera corresponding to the captured cooperative target in the UAV's body coordinate system; This represents the three-dimensional position coordinates of the corresponding cooperative target in the local coordinate system of the recovery area; , This indicates the position and attitude of the drone relative to the recovery area; The image coordinates representing the cooperative objective.

5. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 2, characterized in that, The first cooperative target and the second cooperative target have optical characteristics; By dynamically adjusting the optical characteristics of the first and second cooperative targets, a filtered image is generated; during the mid-range guidance phase, the pose of the UAV is calculated based on the filtered image.

6. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 5, characterized in that, By dynamically adjusting the optical characteristics of the first cooperative target and the second cooperative target, an interference-filtered image is generated, including: Simultaneously acquire bright and dark images of the first and second cooperative targets under preset optical features; wherein, the second cooperative target corresponding to the already occupied recovery position is set to a specific optical state; A difference operation is performed on the bright image and the dark image to obtain a filtered image.

7. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to claim 6, characterized in that, Both the first cooperative target and the second cooperative target adopt programmable point light sources, and the brightness of the programmable point light sources changes over time.

8. The method for rapid and intensive recovery of unmanned aerial vehicles based on vision guidance according to any one of claims 4 to 7, characterized in that, The internal and external parameters of the variable field of view imaging device are calibrated, and the field of view of the variable field of view imaging device is adjusted based on the distance threshold between the UAV and the recovery area.

9. A vision-guided rapid and intensive drone recovery device, characterized in that, The device includes: A recycling area deployment module is used to deploy large-scale marker units and small-scale marker units in the recycling area, wherein the small-scale marker units are located within the large-scale marker units; An airborne vision module is used for UAVs to carry variable field-of-view imaging devices and perform recovery and positioning through a multi-stage visual guidance process: wherein the multi-stage visual guidance process includes a long-range guidance stage, a medium-range guidance stage and a short-range guidance stage. The flight control module is used for the following purposes: During the long-range guidance phase, the variable field-of-view imaging device captures large-scale marker units with a large field of view and solves the initial pose of the UAV relative to the recovery area; based on the initial pose, the UAV's heading is adjusted, and it enters the mid-range guidance phase; During the long-range guidance phase, the variable field-of-view imaging device adjusts the field of view size to capture small-scale marker units, identifies the recovery position by combining the topological structure in the small-scale marker units, and locates the coarse pose of the UAV; Based on the coarse pose, the UAV's heading is adjusted, and it enters the close-range guidance phase; During the close-range guidance phase, sub-pixel-level features are extracted from the small-scale marker units, and then the precise pose of the UAV is calculated, and the UAV is guided to land in the recovery area based on the precise pose.

10. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the vision-guided rapid and intensive drone recovery method according to any one of claims 1 to 7.

Images