Long-range ar spatial positioning method, system, and medium based on natural visual anchor points

Abstract

This invention discloses a long-range AR spatial positioning method, system, and medium based on natural visual anchors. The method includes deploying natural visual anchors at key locations in the target space; recording the image data of each natural visual anchor into an AR tracking database and assigning a unique identifier index to each anchor; during runtime, capturing real-time images through the AR device's camera to perform image tracking and detection of the natural visual anchors; in response to the identification of any natural visual anchor, obtaining the absolute pose of that anchor in the world coordinate system; using the absolute pose of that anchor as an independent positioning reference, instantiating or updating the spatial position of virtual content bound to the unique identifier index of that anchor, thus decoupling the spatial position of the virtual content from any positioning information prior to that anchor. This invention achieves non-intrusive environmental deployment and high accuracy of AR positioning in long-distance mobile scenarios, while also eliminating the propagation of cumulative errors across anchors from an architectural perspective.

Description

Technical Field

[0001] This invention relates to the field of augmented reality technology, and in particular to a long-distance AR spatial positioning method, system and medium based on natural visual anchor points. Background Technology

[0002] With the increasing application of augmented reality technology in fields such as smart navigation, industrial inspection, and digital exhibitions, continuous and accurate spatial positioning of AR content in long-distance mobile scenarios has become a core requirement. In these scenarios, users typically need to move hundreds of meters or even further along a preset path, and the AR virtual content must maintain precise alignment with the physical space throughout the entire journey.

[0003] Currently, the mainstream AR spatial positioning solutions mainly include the following three types: Single-image tracking scheme: This method identifies pre-defined visual markers in the environment, places AR content at the marker's location, and tracks it. This method offers high positioning accuracy within the marker's visible range, but is limited by the camera's field of view. If the user moves and the marker moves out of view, the AR content is lost. Alternatively, it switches to visual inertial odometry for relative positioning, but this results in drift over time, making it impossible to maintain spatial consistency over long distances.

[0004] Pure SLAM solution: This solution achieves six degrees of freedom tracking through visual inertial odometry without relying on external physical markers. While flexible and requiring no pre-deployed markers, pose estimation errors accumulate with increasing distance, leading to a significant decrease in alignment accuracy between distant AR content and the physical space, making it difficult to meet the high-precision positioning requirements of long-distance scenarios.

[0005] Multi-marker dense deployment scheme: Densely deploy manual markers along the movement path to frequently correct drift. This scheme can maintain long-distance accuracy to a certain extent, but a large number of manual markers intrude into the physical environment, destroying the visual immersion, and the deployment and long-term maintenance costs of the markers are high.

[0006] In summary, existing solutions cannot simultaneously achieve positioning accuracy, environmental integration, and deployment convenience in long-distance mobile scenarios, and a new AR spatial positioning method is urgently needed to solve the above technical problems. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the existing technology. To achieve the above objective, a long-distance AR spatial positioning method, system and medium based on natural visual anchor points are adopted to solve the problems mentioned in the background technology.

[0008] Firstly, a long-range AR spatial positioning method based on natural visual anchor points is provided, including the following steps: Natural visual anchors are deployed at key locations in the target space. The natural visual anchors are existing planar visual elements in the target space that are fused with environmental semantics. The image data of each natural visual anchor is recorded into the AR tracking database, and a unique identifier index is assigned to each anchor. During operation, real-time images are captured through the camera of the AR device, and image tracking and detection are performed on the natural visual anchor points to obtain the recognition results of the natural visual anchor points. In response to the recognition result indicating that any of the natural visual anchor points has been identified, the absolute pose of the anchor point in the world coordinate system is obtained; Using the absolute pose of the anchor point as an independent positioning reference, the spatial position of the virtual content bound to the unique identifier index of the anchor point is instantiated or updated, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor point, thereby achieving precise alignment between the virtual content and the corresponding anchor point at each natural visual anchor point passed along the path.

[0009] Furthermore, the step of instantiating or updating the spatial position of the virtual content bound to the unique identifier index of the anchor point, using the absolute pose of the anchor point as an independent positioning reference, specifically includes: when the natural visual anchor point is first identified, the virtual content is instantiated based on the configuration information bound to the anchor point, using the spatial position after applying a preset offset to the absolute pose as a reference; when the natural visual anchor point that has been removed from the field of view is identified again, the spatial position of the virtual content is directly reset using the latest acquired absolute pose, so as to correct the pose drift accumulated during the period when the anchor point is removed from the field of view.

[0010] Furthermore, the method also includes: in response to the identified natural visual anchor point moving out of the camera's field of view, stopping the update of the anchor point's pose, but keeping the current position of the instantiated virtual content in space unchanged, and maintaining its spatial persistence by SLAM until the anchor point is identified again.

[0011] Furthermore, the method also includes a global drift diagnosis and correction step: using the first identified natural visual anchor point as the spatial reference origin, its world coordinate pose is recorded; when subsequent natural visual anchor points are identified, the deviation vector between the theoretical pose of the subsequent anchor point based on the reference origin and its actual detected pose is calculated as the cumulative drift amount; after the cumulative drift amount is subjected to amplitude limiting and smoothing processing, it is applied to the spatial pose of all instantiated virtual content to achieve global spatial consistency correction.

[0012] Furthermore, each natural visual anchor point is independently bound to its own configuration information in the tracking database. The configuration information includes: the virtual content prefab to be rendered, the offset vector of the virtual content relative to the center of the anchor point, and the persistent behavior strategy of the virtual content when tracking is lost. The persistent behavior strategy includes continuous display, automatic destruction after a period of time, automatic destruction after exceeding a set distance, destruction after identifying the next anchor point, and setting custom strategies according to different scenarios.

[0013] Secondly, a long-range AR spatial positioning system based on natural visual anchors is provided, comprising: an anchor management module for storing image data and unique identifier indexes of natural visual anchors deployed in a target space, wherein the natural visual anchors are existing planar visual elements in the target space that are fused with environmental semantics; an image tracking module for acquiring real-time images through the camera of an AR device, performing image tracking detection on the natural visual anchors, and obtaining the absolute pose of any natural visual anchor in the world coordinate system when it is identified; a SLAM module for maintaining the continuity of spatial positioning during movement between anchors; an anchoring control module connected to the image tracking module for controlling the instantiation or spatial position update of virtual content using the absolute pose of the identified anchor as an independent positioning reference, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor; and a content rendering module connected to the anchoring control module for rendering the virtual content according to the instructions of the anchoring control module.

[0014] Furthermore, the anchoring control module is also used to: when the natural visual anchor point is first identified, trigger the content rendering module to instantiate the virtual content based on the configuration information bound to the anchor point and the spatial position after applying a preset offset to the absolute pose; when the natural visual anchor point that has been removed from the field of view is identified again, directly reset the spatial position of the virtual content using the latest acquired absolute pose to correct the pose drift accumulated by the SLAM module during the period when the anchor point is removed from the field of view.

[0015] Furthermore, the anchoring control module is also used to: stop updating the pose of the anchor point in response to the identified natural visual anchor point moving out of the camera's field of view, but keep the current position of the instantiated virtual content in space unchanged, and maintain its spatial persistence by the SLAM module until the anchor point is identified again.

[0016] Furthermore, the system also includes a global drift diagnosis module, which uses the first identified natural visual anchor point as the spatial reference origin to calculate the deviation vector between the theoretical pose and the actual detected pose of the subsequently identified natural visual anchor points based on the reference origin, and outputs it as the cumulative drift amount; the anchoring control module is also used to apply the cumulative drift amount to the spatial pose of all instantiated virtual content after amplitude limiting and smoothing, so as to achieve global spatial consistency correction.

[0017] Thirdly, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps of the method as described in any of the first aspects.

[0018] The technical solution provided by this invention has at least the following beneficial effects compared with the prior art: Non-intrusive environmental deployment: It uses existing natural visual elements in the environment as positioning anchors, without the need to add any physical markers, without compromising the spatial visual integrity and user immersion, and at the same time saving the installation and maintenance costs of artificial markers.

[0019] Long-distance high-precision positioning: The absolute pose of each anchor point is obtained through image tracking, achieving centimeter-level accurate positioning; the relative movement between anchor points is maintained by SLAM to maintain spatial continuity, and the cumulative drift is controlled to a very small range under reasonable anchor point spacing.

[0020] Error propagation chain interruption: The virtual content of each anchor point is independently anchored to the absolute pose of that anchor point itself. The positioning accuracy of subsequent anchor points does not depend on the pose transmission of previous anchor points, thus eliminating the accumulation and propagation of errors across anchor points from the architectural level.

[0021] Robustness to tracking loss: When the anchor point leaves the field of view, the virtual content is maintained in space by SLAM; when the anchor point re-enters the field of view, it is automatically calibrated with the latest absolute pose, achieving seamless and jump-free tracking recovery.

[0022] Flexible configuration and diagnostics: The virtual content of each anchor point can be independently configured with prefabs, offsets and persistence strategies; it also supports quantitative monitoring and selective correction of global drift with the first anchor point as a reference. Attached Figure Description

[0023] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings: Figure 1 This is a schematic diagram illustrating the steps of the spatial positioning method according to an embodiment of this application; Figure 2 This is a flowchart illustrating a spatial positioning method according to an embodiment of this application. Detailed Implementation

[0024] 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 some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Please refer to Figure 1 and Figure 2 In this embodiment of the invention, a long-distance AR spatial positioning method based on natural visual anchor points is applied to long-distance mobile scenarios such as AR navigation and industrial inspection. Specific steps include: Step S1: Register natural visual anchor points; Deploy natural visual anchors at key locations in the target space. The natural visual anchors are existing planar visual elements in the target space that are integrated with environmental semantics. Each natural visual anchor point is independently bound to its own configuration information in the tracking database. The configuration information includes: the virtual content prefab to be rendered, the offset vector of the virtual content relative to the center of the anchor point, and the persistent behavior strategy of the virtual content when tracking is lost. The persistent behavior strategy includes continuous display, automatic destruction after a period of time, automatic destruction after exceeding a set distance, destruction after identifying the next anchor point, and setting custom strategies according to different scenarios.

[0026] Specifically, the target space refers to the actual physical space where the user conducts the AR experience, such as enclosed or semi-enclosed places like rooms, exhibition halls, and corridors.

[0027] Key location deployment refers to setting anchor points in suitable locations within the target space, such as walls, display case surfaces, and column surfaces, where visual content can be placed or already exists.

[0028] Natural visual anchors refer to existing planar visual elements in the target space that are integrated with the semantics of the environment. These elements are part of the environmental design themselves, rather than markers specifically added for AR tracking purposes.

[0029] Two-dimensional visual elements refer to image information in space that has rich visual features and can be reliably recognized by an image tracking engine, such as display boards, posters, advertising images, decorative paintings, and signs.

[0030] In this embodiment, existing posters, display boards, signs, and decorative paintings in the exhibition hall are used. Image data of each natural visual anchor point is recorded in the AR tracking database, and a unique identifier index is assigned to each anchor point, along with its physical size information. Each anchor point is bound to corresponding AR content configuration information, including a content prefab, an offset vector relative to the anchor point center, and a persistence strategy in case of tracking loss.

[0031] The unique identifier index refers to the unique number assigned to each registered anchor point in the AR tracking database. It is used by the system to distinguish and retrieve different anchor points and their bound configuration information at runtime.

[0032] Step S2: Hybrid tracking and content anchoring; During operation, real-time images are captured through the camera of the AR device, and image tracking and detection are performed on the natural visual anchor points to obtain the recognition results of the natural visual anchor points. S21, (Anchor point detection): When the image tracking engine recognizes any natural visual anchor point, it obtains the pose of the anchor point in the world coordinate system, including the position vector and rotation quaternion. S22, (Content Instantiation and Positioning): Based on the configuration information corresponding to the anchor point, instantiate the AR content and position it in the spatial position after the anchor point pose is offset, so that the content is precisely aligned with the anchor point. S23, (Tracking Update): When the pose of the identified anchor point is continuously updated, the spatial pose of the corresponding AR content is updated in real time to maintain the precise tracking of the content and the anchor point.

[0033] Real-time images refer to the current environment images continuously captured at a certain frame rate by the front-facing camera of the AR device during runtime, which are used by the image tracking engine for real-time feature matching and anchor point detection.

[0034] Absolute pose refers to the precise three-dimensional position and orientation of the identified anchor point in the world coordinate system. It is usually represented in the form of a position vector and a rotation quaternion and is used to determine the placement of virtual content in physical space.

[0035] Step S3: SLAM transition and persistence maintenance; In response to the recognition result indicating that any of the natural visual anchor points has been identified, the absolute pose of the anchor point in the world coordinate system is obtained; S31, (Anchor point detachment): When the identified anchor point moves out of the camera's field of view, the anchor point pose is stopped from being updated, but the current position of the instantiated AR content in space remains unchanged, and its spatial persistence is maintained by visual inertial odometry. S32, (Anchor Reentry): When an anchor point that has left the field of view is re-identified, the spatial position of the corresponding AR content is recalibrated using the latest acquired anchor point pose to correct the pose drift accumulated during the tracking detachment.

[0036] Step S4: Spatial consistency across anchor points; Using the absolute pose of the anchor point as an independent positioning reference, the spatial position of the virtual content bound to the unique identifier index of the anchor point is instantiated or updated, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor point, thereby achieving precise alignment between the virtual content and the corresponding anchor point at each natural visual anchor point passed along the path.

[0037] Specifically, it includes: When the natural visual anchor point is first identified, the virtual content is instantiated based on the configuration information bound to the anchor point and the spatial position after applying a preset offset to the absolute pose. When a natural visual anchor point that has been out of sight is re-identified, the spatial position of the virtual content is directly reset using the newly acquired absolute pose to correct the pose drift accumulated during the time the anchor point was out of sight. Direct reset means that the AR content of each anchor point is independently anchored to its own anchor point pose, and the current content positioning is updated after re-identification.

[0038] In a specific implementation, when a user sequentially passes through multiple spatially distributed natural visual anchor points along the AR navigation path: The AR content of each anchor point is independently anchored to its own anchor point pose, and the relative spatial relationship between anchor points is maintained by SLAM during the movement. When a subsequent anchor point is identified, the AR content of that anchor point will be accurately located based on the current image tracking results, without relying on the cumulative pose transfer of prior anchor points. This step also includes: In response to the identified natural visual anchor point moving out of the camera's field of view, the pose of the anchor point is stopped from being updated, but the current position of the instantiated virtual content in space remains unchanged, and its spatial persistence is maintained by SLAM until the anchor point is identified again.

[0039] And including global drift diagnosis and correction steps: Using the first-ever identified natural visual anchor point as the spatial reference origin, its world coordinate pose is recorded; When a subsequent natural visual anchor point is identified, the deviation vector between the theoretical pose of the subsequent anchor point based on the reference origin and its actual detected pose is calculated as the cumulative drift amount. The accumulated drift is then subjected to amplitude limiting and smoothing processing and applied to the spatial pose of all instantiated virtual content to achieve global spatial consistency correction. Amplification refers to setting a threshold limit on the vector magnitude.

[0040] Example 2 This embodiment also provides a long-range AR spatial positioning system based on natural visual anchor points, including: The anchor point management module is used to store image data and unique identifier indexes of natural visual anchor points deployed in the target space. The natural visual anchor points are existing planar visual elements in the target space that are integrated with environmental semantics, such as posters, display boards, signs, and decorative paintings.

[0041] The anchor point management module also stores the physical dimensions of each anchor point, as well as the configuration information for each anchor point's independent binding. This configuration information includes: the content prefab to be rendered, the offset vector of the virtual content relative to the anchor point's center, and the persistence strategy for the virtual content in case of tracking loss.

[0042] The image tracking module is used to acquire real-time images through the camera of the AR device, perform image tracking and detection on the natural visual anchor points, and when any natural visual anchor point is identified, obtain the absolute pose of the anchor point in the world coordinate system, which includes a position vector and a rotation quaternion.

[0043] The SLAM module is used to maintain the continuity of spatial positioning during movement between anchor points. When an identified anchor point moves out of the camera's field of view, the SLAM module maintains the current position of the instantiated virtual content in space, ensuring visual continuity for the user during anchor point switching.

[0044] An anchoring control module, connected to the image tracking module, is used to control the instantiation or spatial position update of virtual content with the absolute pose of the identified anchor point as an independent positioning reference, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor point.

[0045] In this embodiment, the connection relationships and data flow between the modules are as follows: The anchor point management module, serving as the system's data storage hub, maintains a dictionary data structure to store the unique identifier index and positional attribute information of each registered natural visual anchor point. The image tracking module connects to the anchor point management module. When the image tracking module detects and identifies a natural visual anchor point through the AR device's camera, it obtains the absolute pose of that anchor point in the world coordinate system and saves this absolute pose as positional attribute information to the anchor point management module's dictionary, completing the initial input of anchor point information.

[0046] The anchoring control module is connected to both the image tracking module and the anchor point management module. When the image tracking module detects that an anchor point is within the camera's recognition range, it triggers an anchor point update event. The anchoring control module responds to this event by synchronously updating the anchor point's position attribute information in the anchor point management module's dictionary. When the image tracking module detects that an anchor point has moved out of the recognition range, it stops updating, and the anchoring control module retains the last recorded information for that anchor point in its dictionary.

[0047] The SLAM module connects to the anchoring control module and provides the device's own pose estimation information to the anchoring control module when the anchor points are not visible, in order to maintain the spatial persistence of the instantiated virtual content. The anchoring control module uses the absolute pose of the identified anchor points stored in the dictionary as an independent positioning reference to calculate the target spatial pose of the virtual content and transmits the processed pose information to the content rendering module.

[0048] The content rendering module is connected to the anchoring control module. After receiving the pose information, it transmits data through the graphics interface with the display hardware to update the spatial pose of the virtual object on the screen, thereby achieving precise alignment between the virtual content and the physical anchor point.

[0049] Specifically, when a natural visual anchor point is first identified, the anchoring control module, based on the configuration information bound to that anchor point and using the spatial position after applying a preset offset to the absolute pose as a reference, triggers the content rendering module to instantiate virtual content. When a natural visual anchor point that has been removed from the field of view is identified again, the anchoring control module uses the newly acquired absolute pose to directly reset the spatial position of the virtual content, thereby correcting the pose drift accumulated by the SLAM module during the time the anchor point was removed from the field of view.

[0050] When a recognized natural visual anchor point moves out of the camera's field of view, the anchoring control module stops updating the pose of the anchor point, but keeps the current position of the instantiated virtual content in space unchanged. The SLAM module maintains its spatial persistence until the anchor point is recognized again.

[0051] As the user passes through multiple natural visual anchor points along the path, the anchoring control module independently anchors the virtual content of each anchor point to the absolute pose of that anchor point. The virtual content of subsequent anchor points is accurately located based on the current image tracking results, without relying on the cumulative pose transfer of prior anchor points.

[0052] The content rendering module, connected to the anchoring control module, is used to render the virtual content according to the instructions of the anchoring control module. When an instantiation instruction is received, the content rendering module loads the corresponding content prefab and places it in the specified spatial position; when a position update instruction is received, the content rendering module updates the spatial pose of the corresponding virtual content.

[0053] The system in this embodiment also includes a global drift diagnosis module. This module uses the first identified natural visual anchor point as the spatial reference origin and records its world coordinate pose. When subsequent natural visual anchor points are identified, the global drift diagnosis module compares the deviation vector between the theoretical pose of the subsequent anchor point based on the reference origin and the actual detected pose, and outputs it as the SLAM cumulative drift amount.

[0054] The anchoring control module is also used to apply the cumulative drift amount output by the global drift diagnosis module, after amplitude limiting and smoothing, to the spatial pose of all instantiated virtual content to achieve global spatial consistency correction.

[0055] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method as described in any one of the above embodiments.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention. The scope of the invention is defined by the appended claims and their equivalents, all of which should be included within the scope of protection of the invention.

Claims

1. A long-range AR spatial positioning method based on natural visual anchor points, characterized in that, Includes the following steps: Natural visual anchors are deployed at key locations in the target space. The natural visual anchors are existing planar visual elements in the target space that are fused with environmental semantics. The image data of each natural visual anchor is recorded into the AR tracking database, and a unique identifier index is assigned to each anchor. During operation, real-time images are captured through the camera of the AR device, and image tracking and detection are performed on the natural visual anchor points to obtain the recognition results of the natural visual anchor points. In response to the recognition result indicating that any of the natural visual anchor points has been identified, the absolute pose of the anchor point in the world coordinate system is obtained; Using the absolute pose of the anchor point as an independent positioning reference, the spatial position of the virtual content bound to the unique identifier index of the anchor point is instantiated or updated, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor point, thereby achieving precise alignment between the virtual content and the corresponding anchor point at each natural visual anchor point passed along the path.

2. The method according to claim 1, characterized in that, The step of instantiating or updating the spatial position of virtual content bound to the unique identifier index of the anchor point, using the absolute pose of the anchor point as an independent positioning reference, specifically includes: When the natural visual anchor point is first identified, the virtual content is instantiated based on the configuration information bound to the anchor point and the spatial position after applying a preset offset to the absolute pose. When the natural visual anchor point that has been out of sight is re-identified, the spatial position of the virtual content is directly reset using the newly acquired absolute pose to correct the pose drift that accumulated during the period when the anchor point was out of sight.

3. The method according to claim 2, characterized in that, Also includes: In response to the identified natural visual anchor point moving out of the camera's field of view, the pose of the anchor point is stopped from being updated, but the current position of the instantiated virtual content in space remains unchanged, and its spatial persistence is maintained by SLAM until the anchor point is identified again. Among them, SLAM maintains spatial persistence by fixing the virtual content in its current position in the world coordinate system, and indirectly maintains the visual static nature of the content by relying on SLAM to track the pose of the device itself.

4. The method according to claim 1, characterized in that, It also includes global drift diagnosis and correction steps: Using the first-ever identified natural visual anchor point as the spatial reference origin, its world coordinate pose is recorded; When a subsequent natural visual anchor point is identified, the deviation vector between the theoretical pose of the subsequent anchor point based on the reference origin and its actual detected pose is calculated as the cumulative drift amount. The accumulated drift amount, after being limited and smoothed, is applied to the spatial pose of all instantiated virtual content to achieve global spatial consistency correction.

5. The method according to claim 1, characterized in that, Each natural visual anchor point is independently bound to its own configuration information in the tracking database. The configuration information includes: the virtual content prefab to be rendered, the offset vector of the virtual content relative to the center of the anchor point, and the persistent behavior strategy of the virtual content when tracking is lost. The persistent behavior strategy includes continuous display, automatic destruction after a period of time, automatic destruction after exceeding a set distance, destruction after identifying the next anchor point, and setting custom strategies according to different scenarios.

6. A long-range AR spatial positioning system based on natural visual anchor points, characterized in that, include: Anchor point management module is used to store image data and unique identifier index of natural visual anchor points deployed in the target space. The natural visual anchor points are existing planar visual elements in the target space that are fused with environmental semantics. The image tracking module is used to acquire real-time images through the camera of the AR device, perform image tracking and detection on the natural visual anchor points, and obtain the absolute pose of the anchor point in the world coordinate system when any natural visual anchor point is identified. The SLAM module is used to maintain the continuity of spatial positioning during movement between anchor points; An anchoring control module, connected to the image tracking module, is used to control the instantiation or spatial position update of virtual content with the absolute pose of the identified anchor point as an independent positioning reference, so that the spatial position of the virtual content is decoupled from any positioning information before the anchor point. The content rendering module is connected to the anchoring control module and is used to render the virtual content according to the instructions of the anchoring control module.

7. The system according to claim 6, characterized in that, The anchoring control module is also used for: When the natural visual anchor point is first identified, the content rendering module is triggered to instantiate the virtual content based on the configuration information bound to the anchor point and the spatial position after applying a preset offset to the absolute pose. When the natural visual anchor point that has been out of sight is re-identified, the spatial position of the virtual content is directly reset using the newly acquired absolute pose to correct the pose drift accumulated by the SLAM module during the time the anchor point was out of sight.

8. The system according to claim 7, characterized in that, The anchoring control module is also used for: In response to the identified natural visual anchor point moving out of the camera's field of view, the pose of the anchor point is stopped from being updated, but the current position of the instantiated virtual content in space remains unchanged, and its spatial persistence is maintained by the SLAM module until the anchor point is identified again.

9. The system according to claim 6, characterized in that, It also includes a global drift diagnosis module, which uses the first identified natural visual anchor point as the spatial reference origin to calculate the deviation vector between the theoretical pose of the subsequently identified natural visual anchor points based on the reference origin and their actual detected pose, and outputs the cumulative drift amount as the deviation vector. The anchoring control module is also used to apply the accumulated drift amount, after amplitude limiting and smoothing, to the spatial pose of all instantiated virtual content to achieve global spatial consistency correction.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images