Application development method, application development device and XR equipment

Abstract

The invention discloses an application development method, an application development device and XR equipment, and relates to the technical field of augmented reality, and the method comprises the steps: receiving a preset XR scene component dragged by a user to an XR scene template through a visual editor; in response to a user request, calling an AIGC model to generate an XR content material, and inserting the XR content material into a specified position of a current XR scene; and receiving the configuration of the user on the interaction logic, converting the current XR scene and the interaction logic into an executable format of the target XR glasses through the adaptation layer, and outputting an application package adaptive to the target XR glasses. Visual development and AIGC automatic content generation are combined, so that the XR content creation threshold is reduced, the development efficiency is improved, and diversified application scenes are supported.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of extended reality, and in particular to an application development method, an application development device and an XR device. BACKGROUND

[0002] In recent years, the application of extended reality (XR) technology in the fields of education and training, industrial guidance, and travel and entertainment is becoming increasingly widespread, but the development of XR glasses applications still faces many technical bottlenecks. Traditional development methods are highly dependent on professional programming skills and three-dimensional art design capabilities. Developers need to manually write interaction logic using engines such as Unity and Unreal, and make or purchase 3D models, animations and other materials, resulting in long development cycles and high costs. Although some visualization tools attempt to lower the development threshold, their functions are limited to simple AR effect creation and cannot meet the development needs of complete XR applications, especially in terms of cross-device adaptation and complex interaction implementation.

[0003] Another major flaw in the existing XR development process is the efficiency bottleneck of content creation. High-quality 3D models, scene environments and animation effects usually require professional designers to spend time creating, and ordinary developers have difficulty quickly obtaining customized materials. Although generative artificial intelligence content (AIGC) technology can generate 3D content based on text or images, its integration with XR development tools is insufficient, and the generated models often need to be manually optimized in terms of topology and material parameters, and cannot be directly embedded into the development process. In addition, the writing of interaction logic still relies on traditional coding, and there is a lack of efficient means to automatically convert natural language intent into executable scripts, making it difficult for non-technical personnel to participate in development.

[0004] In terms of cross-platform adaptation, the hardware characteristics and running environment of different XR glasses differ significantly, and developers need to optimize the rendering pipeline, interaction logic and performance parameters for each type of device. Existing solutions usually require developers to manually adjust the code or switch development engines, which not only increases the workload but also easily introduces compatibility issues. Especially for the scenario of mixed deployment of WebXR and native applications, there is a lack of unified automatic adaptation mechanism, making it difficult to achieve one development, multiple deployment. SUMMARY

[0005] The application embodiments provide an application development method, an application development device and an XR device to solve the above technical problems.

[0006] In one aspect, the application embodiments provide an application development method, comprising: receiving, by a visual editor, a user dragging a preset XR scene component into an XR scene template; the XR scene component including a virtual object, interaction logic, and a UI component; in response to a user request, invoking an AIGC model to generate an XR content material, and inserting the XR content material into a specified location of a current XR scene; the XR content material including at least one of a three-dimensional model, an animation sequence, or an interaction script; receiving a user configuration of interaction logic, and converting, by an adaptation layer, the current XR scene and the interaction logic into an executable format of a target XR glasses, and outputting an application package adapted to the target XR glasses; the configuration including an interaction rule of a drag trigger or a natural language description.

[0007] In an implementation manner of the present application, in response to a user request, an AIGC model is invoked to generate an XR content material, and the XR content material is inserted into a specified location of a current XR scene, specifically including: receiving a user request in a case where a user required object does not exist in a preset basic object library; in a case where the XR content material to be generated is a three-dimensional model, parsing the user request to extract key feature parameters corresponding to the three-dimensional model; the user request including a natural language description, an image, or a video; in response to the user request, invoking an AIGC model based on the key feature parameters to generate a three-dimensional model file matching the natural language description; inserting the three-dimensional model file into the current XR scene, and performing real-time rendering on the three-dimensional model at a specified location of the current XR scene; receiving a user adjustment instruction for a rendering result to perform initial position adjustment and scaling adjustment on the rendered three-dimensional model.

[0008] In an implementation manner of the present application, further including: in a case where the XR content material to be generated is an interaction script, parsing a natural language sentence describing behavior logic of a user, and determining a spatial position relationship and an event trigger condition in the natural language sentence; generating a corresponding executable code segment by a large language model, and binding the executable code segment to an interaction event node of a specified virtual object in the visual editor to generate an interaction script; the executable code segment including a Unity C# script or a WebXR JavaScript script.

[0009] In an implementation manner of the present application, receiving a user configuration of interaction logic, specifically including: receiving, by the visual editor, a user connection operation of an event node and an action node to form an interaction logic chain. In the case that the user describes the interaction rule through natural language, the corresponding script code or visual logic node is generated through the AI model.

[0010] In an implementation manner of the present application, the current XR scene and the interaction logic are converted into an executable format of the target XR glasses through an adaptation layer, and an application package adapted to the target XR glasses is output, specifically including: detecting a running environment supported by the target XR glasses; the running environment includes WebXR and native application; in the case that the target XR glasses support WebXR, converting the current XR scene into code based on the WebXR standard, and encapsulating the corresponding code into an offline application package; the content based on the WebXR standard includes HTML5 code or WebGL code; in the case that the target XR glasses need a native application, calling an SDK provided by a Unity engine to generate corresponding engineering resources and an application package.

[0011] In an implementation manner of the present application, further including: starting a real-time preview mode in the editing process, and pushing the current XR scene to the target XR glasses in the case that it is detected that the user connects the target XR glasses; based on a local lightweight rendering engine, simulating a field of view angle parameter of the target XR glasses, and rendering the current XR scene; receiving the interactive test operation of the user, and dynamically updating the spatial coordinate mapping relationship in the case that the user adjusts the component position; in the case that it is detected that the user modifies the XR content material generated by the AIGC model, recording the feature description of the modified material, and retriggering the AIGC model to generate optimized XR content material; dynamically updating the preview picture according to the optimized XR content material, and feeding back the interaction delay parameter.

[0012] In an implementation manner of the present application, the real-time preview mode is started in the editing process, specifically including: connecting an actual target XR glasses device, and synchronizing head pose data; on the device screen of the target XR glasses, the rendering picture of the current XR scene is projected in real time, and the interactive error log of the user in the test process is recorded, and the abnormal node is highlighted.

[0013] In an implementation manner of the present application, before the current XR scene and the interaction logic are converted into an executable format of the target XR glasses through the adaptation layer, the method further includes: identify unused model vertex data in the current XR scene, and compress the unused model vertex data to grid simplify the three-dimensional model; reduce the sampling rate of high-resolution textures in the current XR scene according to a preset strategy to match the device computing power of the target XR glasses.

[0014] In another aspect, the embodiments of the present application also provide an XR device, which comprises: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the application development method as described above.

[0015] In another aspect, the embodiments of the present application also provide an application development device, which is characterized in that the device comprises: a visualization module configured to receive user dragging of a preset XR scene component into an XR scene template through a visualization editor; the XR scene component comprises a virtual object, an interaction logic, and a UI component; a model calling module configured to call an AIGC model to generate an XR content material in response to a user request, and insert the XR content material into a specified position of a current XR scene; the XR content material comprises at least one of a three-dimensional model, an animation sequence, or an interaction script; a scene conversion module configured to receive user configuration of an interaction logic, and convert the current XR scene and the interaction logic into an executable format of a target XR glasses through an adaptation layer, and output an application package adapted to the target XR glasses; the configuration comprises an interaction rule of a drag trigger or a natural language description.

[0016] The embodiments of the present application provide an application development method, an application development device, and an XR device, which at least have the following beneficial effects: By visualizing the drag-and-drop components and natural language interaction, users without professional programming skills can quickly build XR scenes and interaction logic, effectively solving the problem of relying on three-dimensional graphics programming and art design skills in traditional XR development; using AIGC models to automatically generate three-dimensional models, animation sequences, and interactive scripts, etc. XR content materials, the original manual modeling and coding work that takes several days or even weeks is shortened to minutes, greatly accelerating the conversion process from creativity to prototype; by analyzing natural language descriptions to automatically generate matching 3D models and interaction logic, developers can express creative requirements in a more intuitive way, and AI-generated materials can be directly embedded in the scene and adjusted in real time, avoiding the tedious steps of repeated modification and format conversion in the traditional process; through the adaptation layer, the scene content is automatically converted into an executable format for the target XR glasses, solving the ecological fragmentation problem of different XR devices. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings illustrated herein are used to provide further understanding of the present application, and constitute a part of the present application. The illustrative embodiments of the present application and their description serve to explain the present application. They do not constitute an improper limitation on the present application. In the drawings: Figure 1 An application scenario diagram of an application development method provided by an embodiment of the present application; Figure 2 A flowchart of an application development method provided by an embodiment of the present application; Figure 3 A flowchart of an XR content material generation method for calling an AIGC model provided by an embodiment of the present application; Figure 4 A flowchart of an application package output method for adapting to target XR glasses provided by an embodiment of the present application; Figure 5 An internal structure diagram of an XR device provided by an embodiment of the present application; Figure 6 An internal structure diagram of an application development device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described below in conjunction with specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0019] An application development method provided by an embodiment of the present application can be applied to, for example Figure 1The application environment shown. As Figure 1 As shown in the application environment, the application environment can include: a development terminal 101, an AIGC cloud service platform 102, an XR running engine 103, target XR glasses 104, and a resource management database 105.

[0020] The development terminal 101 is used as a main development tool to run a visual editor, provide a drag-and-drop XR scene construction interface and natural language interaction function, and realize real-time preview through a local lightweight rendering engine.

[0021] The AIGC cloud service platform 102 is connected with the development terminal 101, integrates various generative AI models including a 3D model generator, an animation sequence generator, and a code script generator, and is used to automatically generate XR content materials according to the request of the development terminal.

[0022] The XR running engine 103 is used as an adaptation layer core component to convert the edited XR scene into a format executable by the target device, and supports the conversion of WebXR standard and mainstream XR engine SDK interfaces.

[0023] The target XR glasses 104 are connected with the development terminal 101 through wired or wireless mode, used to receive the deployed application package and run the XR scene, and feedback the device parameters and interaction data to the development terminal.

[0024] The resource management database 105 stores and manages the XR scene template library, the prefabricated component library, the AI generated content, and the user project data, supports version control and team collaborative development. The components are interconnected through high-speed network to form a complete XR application development, generation, adaptation, and deployment closed loop.

[0025] The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0026] Figure 2 A flowchart of an application development method provided by an embodiment of the present application.

[0027] The analysis method related to the embodiments of the present application can be a terminal device or a server, and the present application does not make special limitations thereon. For the convenience of understanding and description, the following embodiments are described in detail taking the server as an example.

[0028] It should be noted that the server can be a single device or a system composed of multiple devices, i.e., a distributed server, and the present application does not make specific limitations thereon.

[0029] As Figure 2 The application development method provided by an embodiment of the present application includes: Step 201, receiving, by a visual editor, a user dragging a preset XR scene component into an XR scene template.

[0030] It should be noted that the XR scene component in the embodiments of the present application includes a virtual object, interaction logic, and a UI component.

[0031] XR refers to the collective term for all immersive reality technologies, including Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR). XR emphasizes the fusion and interaction of virtual digital content and real environment.

[0032] VR refers to a technology that generates an immersive virtual digital environment through a computer, allowing users to completely isolate from the real world and immerse in a virtual scene. AR refers to a technology that superimposes digital virtual content on the real world, allowing users to perceive both the real environment and the virtual information superimposed on it. MR refers to the fusion of AR and VR, where virtual content can be superimposed on reality and interact with real objects in real time, with closer integration of virtual and real.

[0033] AIGC refers to content generated automatically by artificial intelligence (especially generative models), such as images, three-dimensional models, animations, and text scripts. AIGC technology can automatically create the required materials or code based on user-provided input, such as text descriptions or sketches, by AI models.

[0034] The present application discloses an XR glasses application development platform based on AIGC. The platform adopts a C / S architecture and includes a user front end and a cloud service, with both local running capabilities. The user front end runs on the developer's computer or tablet, and its main function is to provide a visual editing interface and local preview function based on a visual editor. The cloud service provides AI content generation and complex scene rendering support, as well as cloud storage and collaboration of project resources.

[0035] In the present embodiment, a graphical development interface is provided through the development terminal, including a WYSIWYG scene editing view and a component dragging panel. Users can construct applications by dragging pre-made XR interaction components into the scene without writing code, such as UI interface controls, virtual objects, and interaction triggers. At the same time, users are allowed to describe their requirements in natural language, and the system will then analyze the intent and recommend appropriate components or settings, realizing a humanized interaction design process.

[0036] It should be noted that the visual editor is a GUI application directly interacted by the developer, which has built-in scene editing window, component panel, property setting panel and event flow editor and other sub-interfaces. After the developer starts the visual editor, the developer first selects the target XR glasses device type, and selects the required XR scene template in the visual editor. Then, in the scene window, the virtual object, UI control and interaction logic are dragged to the selected XR scene template by using the mouse, keyboard or gesture device, and the position and attribute of the dragged virtual object or UI control are adjusted, and a simple path or area trigger is drawn. The visual editor can capture the user operation in real time, and feedback the display effect, for example, display the added three-dimensional model, and simulate a simple interaction trigger.

[0037] In the embodiment, first, a project is created and a scene is initialized. The user starts the visual editor, selects "new XR project", inputs the project name and specifies the target device type, for example, a certain model of AR glasses. The editor loads the corresponding device preset (such as field of view angle, resolution parameter) and default empty scene according to this. The user can select a suitable scene template from the built-in template library, such as "indoor display scene" or "outdoor street scene", and the platform loads the pre-established basic scene environment into the editor for the user to modify.

[0038] Then, objects and interface components are added. The user drags the required XR object into the scene through the component panel. For example, in an AR teaching application, the user drags in a "machine device" object as the teaching subject; or in a game, several "3D character" objects are dragged in as interactive characters. The platform provides a basic object library (basic geometry, light source, camera, UI panel, etc.) for selection.

[0039] Step 202, in response to a user request, calling an AIGC model to generate XR content material, and inserting the XR content material into a specified position of a current XR scene.

[0040] It should be noted that the XR content material in the embodiment of the application includes at least one of a three-dimensional model, an animation sequence or an interaction script.

[0041] Figure 3 A flowchart of a method for generating XR content material by calling an AIGC model provided in the embodiment of the application is shown in FIG. 3. Figure 3 As shown in FIG. 3, the method for generating XR content material by calling an AIGC model provided in the embodiment of the application specifically includes the following steps: Step 301, in the case where a user required object does not exist in a preset basic object library, receiving a user request; At step 302, in the case that the XR content material to be generated is a three-dimensional model, the user request is parsed to extract the key feature parameters corresponding to the three-dimensional model; the user request includes a natural language description, an image, or a video; At step 303, in response to the user request, an AIGC model is called based on the key feature parameters to generate a three-dimensional model file matching the natural language description; At step 304, the three-dimensional model file is inserted into the current XR scene, and the three-dimensional model is rendered in real time at the specified location in the current XR scene. At step 305, the user receives adjustment instructions for the rendering result to make initial position adjustment and scaling adjustment to the rendered three-dimensional model.

[0042] In the present embodiment, the AIGC content generation system adopts an intelligent multi-modal request processing mechanism. It can be understood that when the user searches the basic object library in the visual editor, the system will monitor the matching degree of the search keywords and the resources in the library in real time. It should be noted that the basic object library is stored in a graph database structure, and the relevance of user input and existing resources is calculated by a semantic similarity algorithm. Specifically, when the user searches for a "retro phone booth model", the system will first search for an exact match in the ontology library, and if there is no result, it will expand the search to synonyms such as "old telephone room", and finally confirm that there is no match in the basic object library and activate the AIGC generation process. For example, when it is detected that the user searches for no matching result for three times in a row or the user actively clicks the AI generation button. It can be understood that the request analysis module adopts a multi-modal input processing architecture, which specifically includes a text semantic parser, an image feature extractor, and a video key frame analyzer.

[0043] For example, when developing an AR guide application in the field of travel and tourism, when the user needs to generate a "Paris subway entrance three-dimensional model with Art Nouveau style", first, the natural language analysis module will deconstruct the description sentence and identify "Art Nouveau style" as an art genre feature and "Paris subway entrance" as the main object. Second, if the user supplements the upload of historical photo references, the image analysis submodule will extract visual features such as ironwork curves and colored glass. Finally, the feature parameter set includes main structure parameters, decoration feature parameters, and historical style parameters, such as arch size, staircase layout, vine pattern density, material reflectivity, and 1900s typical color scheme.

[0044] It should be noted that the AIGC model calling adopts a hierarchical generation strategy. Specifically, the system first converts the abstract features into control parameters of the generation model, generates the basic geometric shape using the latent diffusion model, refines the surface details through the neural radiance field, and finally optimizes the material performance with the physical renderer. It can be understood that the generation process maintains real-time interactivity, and the user can adjust the parameter weight at any time. For example, in the industrial training scene, when generating a model of the internal structure of a numerical control machine tool, the proportion of transparent parts can be dynamically adjusted to better show the internal mechanical structure.

[0045] In this embodiment, the model insertion and adjustment process adopts scene perception technology. For example, when developing an anatomy AR application in the field of education and training, the generated "heart and blood vessel system" model automatically adapts to the proportion of the dissection table in the current scene and inherits the physical interaction parameters preset by the scene, such as the peelable layer setting. Specifically, the system determines the best placement position through spatial semantic analysis, identifies the "teaching demonstration area" spatial marker in the scene, calculates the topological relationship between the model and the surrounding anatomical models, and automatically adjusts the initial orientation to facilitate teaching observation. It can be understood that the user can make secondary adjustments to the model through gesture control or parameter panel, and the system will provide real-time feedback on collision detection results to ensure that the adjusted model does not penetrate the existing objects in the scene.

[0046] In this embodiment, the intelligent generation of interaction scripts adopts a semantic and code bidirectional mapping mechanism. It should be noted that this mechanism includes three core components: intent understanding module, spatial relationship parser, and code generator. It can be understood that when the user describes "display the operation menu when the handle is close to the control panel", the intent understanding module determines that it is an interaction mode of touch trigger information display, the spatial relationship parser calculates the threshold distance of proximity, and finally the code generator outputs a complete script containing collision detection and UI control logic, realizing platform-adapted script output.

[0047] The code generation process adopts a context-aware constraint generation strategy. Specifically, when generating the script "play the maintenance animation after the technician looks at the equipment fault point for 3 seconds" for the industrial site guidance application, if the target platform is Unity, the generated C# script will inherit the project's basic interaction class, and if the target platform is WebXR, an asynchronous loading mode for performance optimization will be adopted. For example, the generated code will contain detailed comments explaining the original natural language description corresponding to each logic block.

[0048] It can be understood that the code binding mechanism adopts a visual node and script mixed programming mode. When developing an AR guide application in the field of marketing and tourism interaction, when the user describes "automatically play the historical story when the tourist walks in front of the exhibit", the system first locoses the "exhibit" virtual object in the scene graph, then creates a new event listener node on the interaction component, and finally associates the generated script logic with the node. It should be noted that this binding process is kept in two-way synchronization, and when the user adjusts the node connection in the visual editor, the underlying script will also be updated accordingly, and vice versa. When the script is modified, the node relationship will also be automatically refreshed.

[0049] For example, in the personal creative game development scenario, when the developer describes "open the hidden level when the player collects three gems at the same time", the system intelligently identifies that this is a composite condition event and automatically generates a complete script architecture containing a state manager. Specifically, the generated code not only contains basic condition detection logic, but also adds auxiliary code segments such as particle effect triggering and sound playing to enhance immersion according to the project style. It can be understood that this context-based code completion capability greatly reduces the implementation burden of the developer.

[0050] In this embodiment, the AIGC cloud service platform is deployed on a cloud server cluster, integrates multiple AIGC models, and interacts with the front-end editor through a unified interface. The AIGC cloud service platform includes an image / texture generation submodule, a three-dimensional model generation submodule, an animation generation submodule, and a code / script generation submodule. It should be noted that the image / texture generation submodule generates scene maps or skybox images based on diffusion models; the three-dimensional model generation submodule uses a trained 3D generation model to generate simple 3D models based on text or reference images, and outputs common format files such as GLB / OBJ; the animation generation submodule generates character skeleton animations or object motion curves; and the code / script generation submodule generates corresponding script code segments such as Unity C# scripts or WebXR JS code segments based on large language models according to user language descriptions.

[0051] When the user issues a content generation request in the visual editor, such as the text description "generate a tall oak tree model", the front-end sends the request to the cloud AI service. After the AI service calls the corresponding model to generate the content, it returns the model file, image, or script text result data to the front-end editor. The visual editor presents the generated content in the current XR scene for the user to preview and confirm after receiving the result data. If the user is not satisfied with the generated result, the user can adjust the description and generate again, or manually edit the details. The entire AI generation calling process is transparent to the user, and the user does not need to understand the complexity of the underlying AI model. The generated content can be directly applied to the scene in the visual editor, greatly reducing the workload of manually producing materials.

[0052] In this embodiment, when the required object does not exist in the library, the user can describe the desired object in natural language, which is created by the AI content generation module. For example, the user inputs: "generate a red industrial water pump model with a pipe interface", and the AI model will generate a 3D model of the water pump according to the description, and the editor will place the model in the scene after receiving it. For interface interaction elements, the user can also drag in UI components such as buttons, progress bars, text prompts, etc., and place them in the field of view interface for display on the XR glasses.

[0053] Step 203, receiving user configuration of interaction logic, and converting the current XR scene and interaction logic into an executable format for the target XR glasses through the adaptation layer, outputting an application package adapted to the target XR glasses.

[0054] It should be noted that the configuration in the embodiments of the present application includes the drag trigger or natural language description of the interaction rule.

[0055] In this embodiment, the interaction logic configuration system adopts a dual-modal input fusion architecture. It can be understood that this architecture converts different input methods into standardized interaction logic descriptions through a unified semantic understanding intermediate layer. It should be noted that the event nodes in the visual editor adopt a color coding classification system, for example, blue represents input events, green represents processing logic, and red represents output actions. This design significantly improves the user's configuration efficiency in complex scenarios.

[0056] Specifically, in the industrial training application development scenario, when the user drags the "gesture recognition" event node and the "device disassembly animation" action node to establish a connection, the system will automatically analyze the parameter compatibility between the nodes and intelligently insert the necessary conversion logic nodes, such as adding a mapping relationship between gesture intensity and animation playback speed.

[0057] For example, for professional fields such as precision instrument maintenance training, the system will automatically supplement safety operation verification logic based on the domain knowledge base, such as detecting whether the user has performed a power-off operation before starting the disassembly step. It can be understood that this intelligent node connection assistance mechanism not only retains the flexibility of configuration, but also ensures the safety of operations in professional scenarios.

[0058] The natural language to logic node conversion process adopts a progressive parsing strategy. It is noted that when the developer in the medical training field inputs "vibrate to prompt when the trainee incorrectly operates the surgical instrument and record the number of errors", the system will first decompose the compound sentence into discrete interaction elements, such as the trigger condition "incorrect operation", the feedback action "vibration prompt", and the data record "error number", and then clarify ambiguous expressions through multiple rounds of dialogue, such as confirming the intensity level of the "vibration prompt". Specifically, the generated logic node will retain the semantic label of the original description, supporting subsequent modification and adjustment through natural language.

[0059] For example, in the development of a travel and tourism interactive application, the user can first describe the basic rule through natural language, "display the introduction when the visitor looks at the exhibit", and then supplement the details through the visual editor, dragging the "gaze time" parameter to the "fade-in speed" control node of the introduction panel. It can be understood that this mixed editing mode fully takes advantage of the strengths of both input methods, with natural language quickly building a framework and visual adjustments accurately adjusting details. It is noted that the system will maintain consistency between the two representations in real time, and any modification by one party will be immediately synchronized to the other representation.

[0060] In this embodiment, the user selects a virtual object in the current XR scene and adjusts various properties of the virtual object in the property panel. For example, position coordinates, scaling ratio, material map, etc. In addition, the user can also use AI to assist in setting complex properties, such as generating a texture map for the wall based on the description "this wall needs some industrial-style graffiti posters" and automatically applying it. In terms of behavior configuration, the platform supports adding interaction components to objects, such as "collision body", "grabable object", "timed rotation", and other preset behaviors. The user can assign the corresponding interaction characteristics to the object by simply checking or setting parameters, without the need to write code manually.

[0061] If more complex or custom behavior logic is required, the user can generate a sub-module through a script based on the natural language description to automatically generate. For example, in a VR puzzle game, it is desired that a certain prop triggers the door to open after being placed in the correct position. The user describes the logic "when prop A is placed in position B, trigger the open door animation of object C" in natural language, and the AI will generate the corresponding script code or visual script node configuration and attach it to the event of the relevant virtual object. The user can view or fine-tune the generated logic. In this way, even users who are not familiar with programming can create complex interactive behaviors with the help of AI.

[0062] Figure 4 A flowchart of an application package output method for adapting to a target XR glasses provided in an embodiment of the present application. As shown in Figure 4As shown, the application embodiment provides an application package output method suitable for target XR glasses, which specifically includes the following steps: Step 401, detecting a running environment supported by the target XR glasses; the running environment includes WebXR and native application; Step 402, in the case that the target XR glasses support WebXR, converting the current XR scene into code based on the WebXR standard, and encapsulating the corresponding code into an offline application package; the content based on the WebXR standard includes HTML5 code or WebGL code; Step 403, in the case that the target XR glasses need a native application, calling an SDK provided by a Unity engine to generate corresponding engineering resources and application packages.

[0063] In the embodiment, the cross-platform adaptation system adopts an intelligent environment detection and adaptive conversion architecture. It can be understood that the architecture dynamically determines the optimal packaging strategy through the combination of a device feature library and real-time performance analysis. It should be noted that the environment detection module collects multi-dimensional device parameters, such as the GPU model, memory capacity, sensor configuration, and API features supported by the system of the XR glasses, and these data are standardized by device fingerprint technology.

[0064] Specifically, when outputting to a WebXR environment, the system first converts scene resources into the glTF2.0 standard format to ensure cross-browser compatibility, secondly automatically injects device-specific performance tuning parameters, such as optimizing WebGL shaders for Magic Leap browsers, and finally generates service worker scripts that can achieve complete offline running capabilities. For example, in a travel guide application, the system intelligently identifies panoramic video resources, automatically converts them into a 360-degree video player component suitable for WebXR, and maintains consistent interaction with other 3D objects in the scene.

[0065] The native application packaging process adopts a modular SDK integration scheme. It should be noted that different SDK plug-in packages are dynamically loaded according to the characteristics of the target device. For Hololens2, a Windows MR device, the Mixed Reality Toolkit core module is automatically integrated; and for an Android-based AR glasses, ARCore extended functions are preferentially used. Specifically, the conversion process maintains bidirectional traceability, and developers can at any time fall back to the visual editor to adjust the original scene, and all modifications will be automatically synchronized to the native engineering file.

[0066] For example, in the development of an industrial remote assistance application, when the target device is detected as a Vuzix M400, the system automatically enables the waveguide display calibration parameters of the device, adjusts the contrast settings of the UI elements to adapt to the industrial environment lighting conditions, and maps the key interaction instructions to the physical buttons on the side of the device. It can be understood that this deep device adaptation capability enables the generated application to fully exploit the hardware features of various XR glasses, without the developer having to manually handle complex platform differences.

[0067] In this embodiment, during editing, the user can click the "preview" button at any time to simulate the current scene. The editor will switch to preview mode, render the scene through the XR engine, and simulate the perspective and interaction of the XR glasses. If the XR glasses device is connected, it can also be directly previewed on the device. The user can test the interaction in the preview, such as clicking the button, moving the perspective to view the overlay effect, etc. For problems found in preview, such as inappropriate object position, insensitive interaction trigger, the user can directly adjust the scene content or parameter configuration after exiting the preview, and immediately preview and verify again. This fast WYSIWYG cycle enables the developer to efficiently iterate and optimize the experience details.

[0068] In this embodiment, when the developer is satisfied with the preview effect, the application can be deployed to the actual XR glasses for testing or publishing through the project management module. The platform provides a "one-click publishing" function, and after the user selects the target device or distribution channel, the system automatically completes the remaining processes. For example, for WebXR deployment, the platform will generate a URL link for the application on the cloud, and the user can run it by accessing it through the browser with the XR glasses; for the installation package, the platform will generate an installation file that adapts to the system of the device and push it to the glasses or provide a download. During deployment, the system will also optimize resources, such as compressing textures and cutting unused model parts, to ensure smooth operation on the limited computing power of the XR glasses. At the same time, the platform records this build version for future updates and iterations. Thus, a complete XR glasses application development process is completed, and an actual experienceable XR application is produced.

[0069] In this embodiment, real-time preview achieves frame synchronization state sharing between the development environment and the XR glasses through the establishment of a bidirectional data channel. It should be noted that when the target XR glasses is connected, the system will automatically load the device-specific display configuration file, such as the field of view distortion parameters, the pupil distance adaptation range, and the screen gamut characteristics. These parameters will be applied to the shader program of the local lightweight rendering engine in real time.

[0070] Specifically, an inverse projection matrix algorithm is used to accurately reproduce the optical distortion effects of XR glasses on the development side. For example, when developing medical AR applications, the system will specifically simulate the special optical paths of microscope-style XR glasses to ensure that the preview image seen by developers is exactly the same as when doctors actually use them. This precise simulation avoids content position deviation problems caused by optical differences, which is particularly important for surgical navigation applications that require millimeter-level positioning accuracy.

[0071] It should be noted that when the user adjusts the position of the virtual object during the preview process, the system will establish a spatial anchor point association relationship, which not only maintains the original coordinates of the object in the scene logic, but also records the offset relative to the user's current perspective. Specifically, this dual coordinate system enables the same set of interaction logic to adapt to XR devices of different specifications without the need for developers to manually adjust position parameters. For example, in an industrial maintenance guidance application, the tooltip panel will automatically adjust the suspension distance according to the field of view of different XR glasses, and always maintain the most comfortable reading position.

[0072] It is understandable that when the user adjusts the AI-generated 3D model, the system will record the modification vectors, such as scale changes, material parameter adjustments, etc., and use these feedback data as conditional inputs to re-trigger the generation process. It should be noted that this reinforcement learning mechanism based on human feedback enables the AI ​​model to continuously adapt to the artistic style and technical requirements of the current project. Specifically, in the development of educational applications, when the teacher adjusts the transparency of the organs of the anatomical model, the system will automatically learn this preference and preset the same transparency parameters on other anatomical models generated subsequently.

[0073] For example, a heat map is used to display the rendering load distribution, and a timing curve chart is used to show the changes in interaction delay. This intuitive feedback mechanism helps developers quickly identify performance bottlenecks, which is particularly important when developing industrial operation guidance applications that require strict real-time response. Specifically, when gesture recognition delay is detected to be excessive, the system will intelligently recommend reducing the number of bones in the hand model or optimizing the collision detection algorithm to maintain overall interaction smoothness.

[0074] In this embodiment, the device-level real-time preview system uses a low-latency data synchronization architecture. High-speed USB or Wi-Fi 6 connections are used to transmit head posture data, and millimeter-wave wireless screen projection technology is used to ensure real-time image quality. It should be noted that posture data synchronization uses a predictive compensation algorithm. In scenarios with extremely high precision requirements, such as medical training, the system dynamically calculates the full-link delay from sensor sampling to image update and compensates for it through spatiotemporal interpolation technology.

[0075] Specifically, when developing an industrial equipment maintenance guide application, the system automatically calibrates the spatial mapping relationship in real time by comparing the virtual guide markers with the key reference points of the actual equipment when an engineer tests the disassembly process while wearing XR glasses. For example, for large mechanical maintenance scenarios, the system intelligently identifies the feature points of the equipment to establish an accurate correspondence between the world coordinate system and the glasses SLAM system. It can be understood that this spatial calibration mechanism based on the actual equipment can effectively solve the misalignment problem between virtual content and real objects.

[0076] It should be noted that the system structurally records three types of key data, including user behavior sequences (such as head movement trajectories, gesture operations), system response events (such as rendering frame rate, script execution status), and environmental parameters (such as light changes, spatial anchor point stability). Specifically, when an anomaly is detected, the system constructs a causal relationship diagram. In a travel guide application, if the user repeatedly attempts to trigger the explanation before the painting without success, the system will mark the possible interactive hot zone setting problem and automatically suggest expanding the trigger range or increasing visual prompts.

[0077] It can be understood that in the testing of education and training applications, when a student makes a mistake, a semi-transparent warning box and a three-dimensional arrow pointer will be superimposed on the XR glasses screen, and the development side will also display the call stack of the error node synchronously. Specifically, this two-way feedback mechanism enables developers to quickly understand the context in which the problem occurs, rather than just seeing the final error state. It should be noted that the system can intelligently distinguish between occasional errors and systematic defects, and for frequently occurring interactive problems, it can automatically generate optimization suggestions, such as adjusting the size of the collision body or increasing haptic feedback.

[0078] In this embodiment, before converting the current XR scene and interaction logic into an executable format for the target XR glasses through the adaptation layer, the system accurately identifies the resource objects that can be optimized by establishing a dependency graph of the three-dimensional scene. It should be noted that by using a strategy combining view frustum culling and occlusion detection, the system not only analyzes the visible faces in static scenes, but also predicts the visible areas that may appear in dynamic interaction processes.

[0079] Specifically, in the development of industrial equipment training applications, when dealing with complex mechanical assembly models, the system identifies the geometric structures that are completely occluded by the shell in embedded parts and automatically removes the vertex data of these invisible faces; at the same time, for visible but distant components, the system uses a progressive simplification algorithm to retain the contour features. For example, for standard parts such as screws and gaskets, the system calls a preset optimization template to significantly reduce the model complexity while ensuring functional demonstration. It can be understood that this semantic-based optimization method can better maintain the accuracy of teaching guidance compared to simple geometric simplification.

[0080] It should be noted that the system analyzes the content characteristics of texture images, maintaining high-precision sampling for areas containing important details (such as equipment nameplates and instrument panel scales), and using block compression for large areas of solid color or gradients. Specifically, in the development of medical anatomy applications, organ textures are dynamically adjusted based on the viewing distance, using medium-precision maps at normal viewing distances and automatically loading high-definition detail maps when the user focuses on the object. This adaptive hierarchical loading mechanism, for example, ensures visual quality while significantly reducing memory usage.

[0081] When the target device is replaced with a higher-performance XR glasses, the system can restore the original high-precision resources with one click, avoiding repeated imports and settings. It should be noted that all optimization operations are recorded in the version control system, and developers can compare the differences in effects before and after optimization at any time to ensure that key visual elements are not affected. Specifically, in cultural and tourism guide applications, for models of important cultural relics exhibits, the system will establish a "protective optimization" whitelist to ensure that these core display objects always maintain the best visual effects.

[0082] In one embodiment of the present application, in the field of education and training, a vocational training institution hopes to create a practical teaching application that runs on AR glasses to train trainees to repair complex equipment. Through this platform, training teachers do not need to program, they only need to select the "Industrial Equipment Training" scene template, adjust the position of the equipment model, and use the AIGC module to generate 3D part models and fault animations for specific models. Then configure the step instructions and gesture prompts by dragging and dropping interactive components. Finally, the generated AR teaching application can be deployed on the XR glasses worn by trainees, realizing teaching with real equipment superimposed on digital guidance, improving training efficiency and safety.

[0083] In one embodiment of the present application, in an industrial field guidance scenario, a company provides XR glasses to field engineers for equipment inspection and maintenance guidance. This platform can help the company quickly generate customized inspection applications. Engineers import the CAD model of the factory workshop in the visual editor or directly use AI to describe the platform to generate a 3D environment for the workshop, add inspection point marks, and AI writes inspection item descriptions based on existing documents. By dragging and dropping logic components, the interaction of displaying maintenance steps when the engineer's line of sight focuses on a certain component can be set. With almost no manual coding, an on-site guidance AR application can be generated to help novice engineers complete complex equipment inspections independently.

[0084] In one embodiment of the present application, in a marketing and cultural tourism interactive scenario, a scenic spot plans an AR treasure hunt game to enhance the interactive experience of tourists. Planners use this platform to select a game rule template, use AIGC to generate virtual treasure models and character animations related to the scenic spot culture, and place these digital elements in corresponding locations on the scenic spot map. Through platform settings, when tourists wear AR glasses provided by the scenic spot at specific locations, virtual treasures will appear and trigger puzzles. The content and logic design of the entire game can be completed in a short time, without the need for a professional program development team, so that marketing ideas can be quickly implemented.

[0085] In one embodiment of this application, an independent creator wanted to develop an immersive puzzle game for VR glasses, specifically for personal creative game development. Using this platform, they first selected a VR interior scene template. Then, they had the AI ​​generate a castle-style interior environment and mechanism model. They dragged and dropped components to set up object interactions, such as picking up keys to unlock mechanisms, and instructed the AI ​​to generate some plot dialogue scripts. The creator completed a prototype of the game within a few days and directly exported the application for a specific VR headset for testing, significantly reducing the difficulty of individual VR game development.

[0086] The above is an embodiment of the method proposed in this application. Based on the same inventive concept, this application embodiment also provides an XR device, whose structure is as follows Figure 5 shown.

[0087] Figure 5 This is a schematic diagram of the internal structure of an XR device provided in an embodiment of the present application. Figure 5 As shown, the equipment includes: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: The user drags the preset XR scene components into the XR scene template through the visual editor; the XR scene components include virtual objects, interaction logic and UI components; In response to a user request, the AIGC model is called to generate XR content material, and the XR content material is inserted into a specified position of the current XR scene; the XR content material includes at least one of a 3D model, an animation sequence, or an interactive script; Receive the user's configuration of the interaction logic, convert the current XR scene and interaction logic into the executable format of the target XR glasses through the adaptation layer, and output the application package adapted to the target XR glasses; the configuration includes drag triggers or interaction rules described in natural language.

[0088] like Figure 6As shown, the embodiments of the present specification also provide an application development device. In one or more embodiments of the present specification, an application development device, the device 600 includes: Figure 6 As shown, the embodiments of the present specification also provide an application development device. In one or more embodiments of the present specification, an application development device, the device 600 includes: a visualization module 601, configured to receive a user dragging a preset XR scene component into an XR scene template through a visualization editor; the XR scene component includes a virtual object, interaction logic, and a UI component; a model calling module 602, configured to, in response to a user request, call an AIGC model to generate an XR content material, and insert the XR content material into a specified position of a current XR scene; the XR content material includes at least one of a three-dimensional model, an animation sequence, or an interaction script; a scene conversion module 603, configured to receive a user configuration of interaction logic, and convert the current XR scene and the interaction logic into an executable format of a target XR glasses through an adaptation layer, and output an application package adapted to the target XR glasses; the configuration includes an interaction rule of a drag trigger or a natural language description.

[0089] In some embodiments, in response to a user request, calling an AIGC model to generate an XR content material, and inserting the XR content material into a specified position of a current XR scene, specifically includes: in the case that a user required object does not exist in a preset basic object library, receiving a user request; in the case that the XR content material to be generated is a three-dimensional model, parsing the user request to extract key feature parameters corresponding to the three-dimensional model; the user request includes a natural language description, an image, or a video; in response to the user request, calling the AIGC model based on the key feature parameters to generate a three-dimensional model file matching the natural language description; inserting the three-dimensional model file into the current XR scene, and performing real-time rendering on the three-dimensional model at a specified position of the current XR scene; receiving a user adjustment instruction for the rendering result to perform initial position adjustment and scaling adjustment on the rendered three-dimensional model.

[0090] In some embodiments, further including: in the case that the XR content material to be generated is an interaction script, parsing a natural language sentence describing behavior logic of the user, and determining a spatial position relationship and an event trigger condition in the natural language sentence; generating a corresponding executable code segment through a large language model, and binding the executable code segment to an interaction event node of a specified virtual object in the visualization editor to generate the interaction script; the executable code segment includes a Unity C# script or a WebXR JavaScript script.

[0091] In some embodiments, the user receives the configuration of the interaction logic, specifically including: The user receives the connection operation of the event node and the action node through the visual editor to form the interaction logic chain; In the case where the user describes the interaction rule through natural language, the corresponding script code or visual logic node is generated through the AI model.

[0092] In some embodiments, the current XR scene and the interaction logic are converted into an executable format of the target XR glasses through the adaptation layer, and an application package adapted to the target XR glasses is output, specifically including: Detect the running environment supported by the target XR glasses; the running environment includes WebXR and native application; In the case where the target XR glasses support WebXR, the current XR scene is converted into code based on the WebXR standard, and the corresponding code is packaged into an offline application package; the content based on the WebXR standard includes HTML5 code or WebGL code; In the case where the target XR glasses require a native application, the SDK provided by the Unity engine is called to generate corresponding engineering resources and application packages.

[0093] In some embodiments, it also includes: Start the real-time preview mode during the editing process, and push the current XR scene to the target XR glasses in the case where the user is detected to connect the target XR glasses; Based on the local lightweight rendering engine, the field of view angle parameters of the target XR glasses are simulated, and the current XR scene is rendered; Receive the user's interaction test operation, and dynamically update the spatial coordinate mapping relationship in the case where the user adjusts the component position; In the case where the user modifies the XR content material generated by the AIGC model, record the modified material feature description, and retrigger the AIGC model to generate optimized XR content material; According to the optimized XR content material, dynamically update the preview picture, and feedback the interaction delay parameter.

[0094] In some embodiments, the real-time preview mode is started during the editing process, specifically including: Connect the actual target XR glasses device, and synchronize the head pose data; On the device screen of the target XR glasses, the rendering picture of the current XR scene is projected in real time, and the interaction error log of the user in the test process is recorded, and the abnormal node is highlighted.

[0095] In some embodiments, before converting the current XR scene and interaction logic into an executable format for the target XR glasses through the adaptation layer, the method further comprises: identifying model vertex data that is not used in the current XR scene, and compressing the unused model vertex data to perform mesh simplification on the three-dimensional model; reducing the sampling rate of high-resolution textures in the current XR scene according to a preset strategy to match the device computing power of the target XR glasses.

[0096] Each of the embodiments in the present application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0097] The device and medium provided by the embodiments of the present application correspond to the method, and therefore, the device and medium also have similar beneficial technical effects as the method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the device and medium will not be described here.

[0098] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0099] The present application is described with reference to flowcharts and / or block diagrams according to the method, device (system), and computer program product of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The function of the device specified in one flow or multiple flows and / or blocks Figure 1 The function of the device specified in one flow or multiple flows and / or blocks

[0100] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0101] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0102] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0103] The memory can include non-persistent memory and / or volatile memory, such as a random access memory (RAM) including a cache area for the temporary storage of data. The memory can also include non-volatile memory, such as a read only memory (ROM), EPROM, EEPROM, or flash memory. The memory can be another form of computer-readable media.

[0104] Computer-readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media, such as modulated data signals and carrier waves.

[0105] It should also be noted that the terms "comprising", "comprises" or other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0106] The above embodiments are only used to illustrate the present application, but not to limit it. Instead of the above, various modifications and changes can be made to the application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall fall into the scope of the claims of the application.

Claims

1. An application development method, characterized in that: The method comprises: Receive a user drag and drop of preset XR scene components into an XR scene template through a visual editor; the XR scene components include virtual objects, interaction logic, and UI components; In response to a user request, calling the AIGC model to generate XR content material, and inserting the XR content material into a specified location of the current XR scene; the XR content material includes at least one of a three-dimensional model, an animation sequence, or an interactive script; Receive the user's configuration of the interaction logic, convert the current XR scene and interaction logic into an executable format for the target XR glasses through the adaptation layer, and output an application package adapted to the target XR glasses; the configuration includes drag triggers or interaction rules described in natural language.

2. An application development method according to claim 1, characterized in that: In response to a user request, the AIGC model is called to generate XR content material, and the XR content material is inserted into a specified location of the current XR scene, specifically including: When the object required by the user does not exist in the preset basic object library, receiving the user request; When the XR content material to be generated is a three-dimensional model, parsing the user request to extract key feature parameters corresponding to the three-dimensional model; the user request includes a natural language description, image, or video; In response to the user request, calling the AIGC model based on the key feature parameters to generate a three-dimensional model file that matches the natural language description; Inserting the 3D model file into the current XR scene, and rendering the 3D model in real time at a specified location in the current XR scene; Receive user adjustment instructions for the rendering result to adjust the initial position and scale of the rendered three-dimensional model.

3. An application development method according to claim 2, characterized in that: The method further comprises: When the XR content material to be generated is an interactive script, parsing the natural language sentence of the user describing the behavior logic, and determining the spatial position relationship and event triggering conditions in the natural language sentence; Generate a corresponding executable code segment through the large language model, bind the executable code segment to the interactive event node of the specified virtual object in the visual editor to generate an interactive script; the executable code segment includes a Unity C# script or a WebXR JavaScript script.

4. An application development method according to claim 1, characterized in that: Receive user configuration of interaction logic, including: Receive, through the visual editor, a user's connection operation on an event node and an action node to form an interactive logic chain; When users describe the interaction rules in natural language, the corresponding script code or visual logic nodes are generated through the AI ​​model.

5. An application development method according to claim 1, characterized in that: The adaptation layer converts the current XR scene and interaction logic into an executable format for the target XR glasses, and outputs an application package adapted to the target XR glasses, specifically including: Detecting the operating environment supported by the target XR glasses; the operating environment includes WebXR and native applications; If the target XR glasses support WebXR, convert the current XR scene into code based on the WebXR standard, and encapsulate the corresponding code into an offline application package; the content based on the WebXR standard includes HTML5 code or WebGL code; When the target XR glasses require native applications, the SDK provided by the Unity engine is called to generate corresponding engineering resources and application packages.

6. An application development method according to claim 1, characterized in that: The method further comprises: Start the real-time preview mode during the editing process, and push the current XR scene to the target XR glasses when it is detected that the user is connected to the target XR glasses; Based on a local lightweight rendering engine, simulate the field of view parameters of the target XR glasses and render the current XR scene; receiving an interactive test operation of the user, and dynamically updating a spatial coordinate mapping relationship when the user adjusts a component position; When detecting a modification instruction of the user on the XR content material generated by the AIGC model, recording the modified material feature description, and re-triggering the AIGC model to generate optimized XR content material; Based on the optimized XR content material, the preview screen is dynamically updated and the interaction delay parameters are fed back.

7. An application development method according to claim 6, characterized in that: Enable real-time preview mode during editing, including: Connect to the actual target XR glasses device and synchronize head posture data; On the device screen of the target XR glasses, a rendering image of the current XR scene is projected in real time, and the user's interaction error log during the test is recorded, highlighting abnormal nodes.

8. An application development method according to claim 1, characterized in that: Before converting the current XR scene and interaction logic into an executable format for the target XR glasses through the adaptation layer, the method further includes: Identifying unused model vertex data in the current XR scene and compressing the unused model vertex data to perform mesh simplification on the three-dimensional model; The sampling rate of the high-resolution texture in the current XR scene is reduced according to a preset strategy to match the device computing power of the target XR glasses.

9. An XR device, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute an application development method according to any one of claims 1 to 8.

10. An application development device, characterized in that: The device comprises: The visualization module is used to receive preset XR scene components dragged by users into the XR scene template through the visual editor; XR scene components include virtual objects, interaction logic and UI components; A model calling module is configured to call the AIGC model to generate XR content materials in response to a user request, and insert the XR content materials into a specified location of the current XR scene; the XR content materials include at least one of a 3D model, an animation sequence, or an interactive script; The scene conversion module is used to receive the user's configuration of the interaction logic, and convert the current XR scene and interaction logic into the executable format of the target XR glasses through the adaptation layer, and output the application package adapted to the target XR glasses; the configuration includes drag triggers or interaction rules described in natural language.

Images