3D printing dynamic visualization interaction system and method based on spatial unconstraint

Abstract

The invention discloses a 3D printing dynamic visualization interaction system and method based on spatial unconstraint, and the method comprises the steps: S1, collecting and preprocessing environment data, and generating a camera pose, a projection carrier normal direction and an anchor point set; s2, generating timestamps and writing the timestamps into a pose channel and an anchor point channel; s3, the dynamic modeling engine receives the input three-dimensional model, and semantic layering and slicing processing is executed; s4, establishing a double-buffer asynchronous rendering architecture; s5, on the basis of a zero-interruption displacement mechanism, when it is detected that the posture changes and the projection sites are switched, continuous output of foreground buffering is kept, new full-precision frames are rendered through background buffering, and foreground buffering and background buffering are switched at the vertical synchronization moment after rendering is completed; s6, receiving a parameter updating instruction input and analyzed by the user; s7, outputting a synchronous rendering frame and an interactive updating result; and S8, outputting the application interface. Free projection and zero-interruption interaction in the 3D printing process are achieved.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction and intelligent manufacturing, and in particular to a dynamic visualization interactive system and method for 3D printing based on unconstrained space. Background Technology

[0002] With the continuous development of additive manufacturing technology, 3D printing has gradually expanded from simple prototype manufacturing to applications in industrial production, education and training, and interdisciplinary research. Along with the popularization of the technology, the demand for visualization and interactivity in the printing process is also constantly increasing. Most existing 3D printing visualization solutions rely on the coordinate system of the physical printing equipment for display and demonstration. While this method can meet basic observation needs to a certain extent, it suffers from significant spatial constraints. Because the projection and display process is confined within the printer's cabin space, the user's observation angle and position are limited by the fixed physical environment, hindering rapid switching between different scenarios in teaching demonstrations and making it difficult to adapt to the cross-workstation, multi-user collaborative needs of industrial reviews. This spatially limited characteristic makes current visualization methods lack flexibility and scalability.

[0003] Besides spatial limitations, existing visualization solutions suffer from process interruptions during interaction. Currently, if a user needs to adjust the viewing angle or switch the projection position, the running printing animation process often needs to be temporarily paused until the position is updated and then reloaded, resulting in a significant interruption to the operation. Experiments show that when performing such operations on mixed reality devices, the average interruption time reaches several seconds. This fragmented interactive experience not only disrupts the continuity and immersion of the operation but also significantly reduces the efficiency of teaching and training processes and the user's understanding of the entire printing workflow. Especially in complex industrial verification and collaborative scenarios, frequent process interruptions can significantly reduce the efficiency of cross-team communication and review.

[0004] On the other hand, existing 3D printing visualization systems are mostly one-way passive observation tools, lacking the ability to intervene in key process parameters in real time during printing. Users can often only see the forming process of the virtual model, but cannot actively modify the infill rate, temperature, or path parameters during the printing process, let alone simulate possible actual failures during demonstrations. This lack of interactive depth makes the system unable to meet the needs of fault teaching and process optimization training in vocational education, and is also not conducive to the real-time verification of parameter adjustment effects in engineering review scenarios.

[0005] Therefore, how to provide a spatially unconstrained 3D printing dynamic visualization interactive system and method is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a spatially unconstrained 3D printing dynamic visualization interactive system and method. This invention possesses spatially decoupled projection, zero-interruption process maintenance, and real-time parameter intervention capabilities. By constructing a virtual coordinate system independent of the physical device, this system achieves dynamic visualization projection of the printing process on any geometric surface carrier, fundamentally eliminating the problem of limited viewing angles in existing solutions. Through the introduction of a double-buffered asynchronous rendering architecture and a zero-interruption displacement mechanism, continuous image output is maintained even when the projection carrier switches and pose changes, avoiding process interruptions and image flickering. The interactive control module introduces multimodal inputs including gestures, voice, and six-degree-of-freedom tools, combined with a parameter thermal control unit, enabling users to modify process parameters, trigger fault simulations, and perform local profile analysis in real time during the printing process.

[0007] The 3D printing dynamic visualization interaction method based on spatially unconstrained space according to an embodiment of the present invention includes the following steps:

[0008] S1. Collect and preprocess environmental data, and generate camera pose, projection carrier normal and anchor point set through SLAM and geometric plane recognition technology, and establish a virtual coordinate system;

[0009] S2. Generate a timestamp for the camera pose, projection carrier normal, and anchor point set, and write it into the pose channel and anchor point channel.

[0010] S3. The dynamic modeling engine receives the input 3D model, performs semantic layering and slicing processing in combination with the timestamp, generates a layer-by-layer printing animation stream, and calls the fault model of the defect simulation library when the preset fault triggering conditions are met. The animation stream and the fault model are written into the animation stream channel together.

[0011] S4. The process preservation and rendering subsystem module receives data from the pose channel, anchor point channel and animation stream channel, and establishes a double-buffered asynchronous rendering architecture.

[0012] S5. Based on the zero-interruption displacement mechanism, when a change in pose and a switch in the projection point are detected, the foreground buffer continues to output continuously, while the background buffer renders a new full-precision frame. After rendering is completed, the foreground and background buffers are switched at the moment of vertical synchronization.

[0013] S6. Receive user gesture input, voice input, and six-degrees-of-freedom tool input, and parse them into parameter update instructions;

[0014] S7. Distribute the parameter update command to the dynamic modeling engine and the process persistence and rendering subsystem module simultaneously, and output synchronous rendering frames and interactive update results;

[0015] S8. Transmit the synchronous rendering frame and interactive update result to the application interface for output.

[0016] Optionally, step S1 specifically includes:

[0017] S11. Collect color image data, depth image data and inertial sensor data of the environment, and synchronize the color image data, depth image data and inertial sensor data in time.

[0018] S12. Perform distortion correction and grayscale normalization processing on the color image data, and perform hole filling and filtering processing on the depth image data;

[0019] S13. Extract corner points and edge features from the processed color image, generate feature descriptors, and perform feature matching between adjacent frames to obtain an initial set of matching pairs.

[0020] S14. Based on the inertial sensor data, perform pre-integration and output motion prior to initialize the simultaneous localization and mapping state vector.

[0021] S15. Calculate the current camera pose using the initial matching pair set and motion prior, perform local binding adjustment, and update the camera pose;

[0022] S16. Use keyframes and depth information to construct a dense point cloud, and perform geometric surface fitting on the dense point cloud region to generate a set of geometric surfaces.

[0023] S17. In the set of geometric surfaces, candidate projection carriers are selected based on area threshold and normal consistency, and the normal vector and fitting residual of the projection carrier are calculated.

[0024] S18. Generate a candidate set of anchor points on the candidate carrier according to the grid spacing, and filter the anchor points that intersect with the camera view frustum to obtain the anchor point set;

[0025] S19. Establish a virtual coordinate system and calculate the homogeneous transformation matrix of the camera pose in the virtual coordinate system. Output the camera pose, the normal of the projection carrier, and the set of anchor points.

[0026] Optionally, S2 specifically includes:

[0027] S21. Receive the camera pose, the normal of the projection carrier, and the set of anchor points;

[0028] S22. Call the global unified clock to generate corresponding timestamps for the camera pose, projection carrier normal and anchor point set;

[0029] S23. Encapsulate the camera pose and timestamp into a pose message packet and write it into the pose channel;

[0030] S24. Encapsulate the projection carrier normal and anchor point set and the timestamp of the projection carrier normal and anchor point set into an anchor point message packet and write it into the anchor point channel.

[0031] Optionally, S3 specifically includes:

[0032] S31. The dynamic modeling engine receives the input 3D model file and parses it into geometric mesh data, which includes vertex coordinates, normal vectors, topological relationships and material identifiers.

[0033] S32. Perform semantic layering processing on the geometric mesh data, identify and label the outer contour layer, filling structure layer and supporting structure layer of the model as different categories, and assign a unique semantic label to each category;

[0034] S33. Based on the semantic tags, the geometric mesh is sliced ​​according to the set layer thickness parameters to generate layer-by-layer two-dimensional contour slices. Each slice contains a closed contour, path nodes and layer height index.

[0035] S34. Load the material physical parameters corresponding to the three-dimensional model from the material database, including thermal conductivity, melting temperature and cooling shrinkage coefficient;

[0036] S35. Based on the material physical parameters, perform thermophysical simulation on the layer-by-layer slices, calculate the interlayer deformation and correct the printing path, and embed the temperature field distribution and deformation vector field information into the animation clip to generate a layered printing animation clip.

[0037] S36. Monitor printing process parameters in real time. When a preset fault triggering condition is detected, call the defect simulation library to insert the corresponding fault model. The fault triggering conditions include material flow interruption, nozzle temperature surge, edge curling, interlayer peeling, hot bed temperature difference alarm, coordinate axis offset, step-like fracture, and positioning inaccuracy.

[0038] S37. Write the layer-by-layer printing animation stream and the fault model together into the animation stream channel, and attach the timestamp information to output the animation stream channel data.

[0039] Optionally, S4 specifically includes:

[0040] S41. Receive pose channel and anchor point channel data, as well as the animation stream channel data;

[0041] S42. Establish a dual-buffered asynchronous rendering architecture. The front-end buffer is used to display simplified geometric models in real time, and the back-end buffer is used to render complete animation frames based on full-precision geometric data.

[0042] S43. Load simplified geometric placeholder frames into the foreground buffer and bind the pose message packet to achieve stable scene display with continuous low latency;

[0043] S44. In the background buffer, full-precision rendering is performed based on the virtual coordinate system, anchor point message packet and animation stream channel data. The full-precision rendering includes texture mapping, lighting calculation and layer-by-layer animation composition to form a target high-precision rendering frame.

[0044] S45. During background rendering, continuously output progress prompts and buffer status information, and maintain stable output of the foreground buffer while background rendering is not completed to prevent screen flickering.

[0045] S46. When the background rendering is detected to be complete, perform a foreground-background buffer swap within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving seamless screen switching.

[0046] S47. After performing buffer swapping, release the old foreground buffer resources, mark the background buffer as the new foreground buffer, and start a new background rendering thread to prepare for the next frame, thus achieving continuous loop rendering.

[0047] Optionally, the zero-interruption displacement mechanism specifically includes:

[0048] S51. When the user makes a drag gesture, the runtime scheduling and data bus module detects the change in pose and the switching of the projection point, and starts the coordinate validity detection.

[0049] S52. When the detection result is invalid, the system outputs an error message; when the detection result is valid, the system proceeds to render the simplified static framework in the foreground.

[0050] S53. While outputting the front-end simplified static framework, start the background full-precision rendering thread;

[0051] S54. During the background rendering process, continuously generate and update progress prompts to provide real-time feedback on the background rendering status.

[0052] S55. When it is detected that the background rendering is not complete, continue to maintain the output of the front-end simplified static framework and update the progress prompt in a loop.

[0053] S56. When the background rendering is detected to be complete, the foreground and background buffers are swapped within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving zero-interruption displacement and continuous visualization.

[0054] S57. After completing the buffer swap, release the background buffer resources and mark the existing output buffer as the new foreground buffer, preparing for the next background rendering loop.

[0055] Optionally, S6 specifically includes:

[0056] S61. Receive the user's multimodal input signal, wherein the user's multimodal input signal includes gesture input, voice input and six degrees of freedom tool input;

[0057] S62. Perform hand joint detection, trajectory tracking, ray collision detection and shader freezing on the gesture input. Through multi-angle observation and freezing of the current layer profile analysis, identify the user's five-finger rotation and fist holding operation, and generate corresponding operation instructions.

[0058] S63. Perform automatic speech recognition on the voice input, extract keywords, and map the keywords into parameter control commands;

[0059] S64. Input the position coordinates, attitude angle and button trigger signal of the six-degree-of-freedom tool and generate the corresponding spatial operation instructions;

[0060] S65. The operation instructions, parameter control instructions and spatial operation instructions are uniformly encapsulated into an interactive instruction set;

[0061] S66. Call the parameter thermal control unit to generate parameter update instructions according to the interactive instruction set.

[0062] Optionally, S7 specifically includes:

[0063] S71. Receive the parameter update instruction;

[0064] S72. Generate a timestamp for the parameter update instruction according to the global unified clock, and perform event sorting to ensure that the timing is consistent with the animation stream channel and pose channel;

[0065] S73. Distribute the parameter update command to the dynamic modeling engine, apply it to the slice parameters and simulation parameters, and generate the adjusted layer-by-layer printing animation clip;

[0066] S74. The parameter update instruction is simultaneously distributed to the process persistence and rendering subsystem module, which applies to the rendering resolution, lighting parameters and dynamic display parameters to generate the adjusted rendering frame.

[0067] S75. After the dynamic modeling engine module and the process maintenance and rendering subsystem module have completed the update, perform synchronization processing on the update results and output the synchronized rendering frame and interactive update results.

[0068] S76. The synchronous rendering frame and interactive update results are displayed on the user interface in real time to realize the hot-effect of parameter modification and visual confirmation.

[0069] Optionally, S8 specifically includes:

[0070] S81. Input the synchronous rendering frame and interactive update result into the application interface;

[0071] S82. Perform format adaptation processing on the synchronous rendering frame, including resolution scaling, color space conversion and frame rate adjustment, to match the hardware specifications of different display terminals.

[0072] S83. In educational training scenarios, the synchronized rendering frames after format adaptation are combined with teaching prompts and output to the augmented reality terminal to support visualization of the printing process, fault simulation demonstration and multi-angle observation.

[0073] S84. In industrial review scenarios, the synchronized rendering frames after format adaptation and review annotation information are output to the collaborative work platform in a synchronized manner to support online review by multiple users, parameter comparison and anomaly reproduction.

[0074] S85. In the product design verification scenario, the synchronous rendering frame and the CAD nominal model are spatially registered and overlaid, and the deformation deviation heat map and geometric tolerance deviation table are output. The version number, material parameters and slice parameter snapshots are written into the design verification report and pushed to the design review workstation or PLM system through the application interface module.

[0075] According to an embodiment of the present invention, a 3D printing dynamic visualization interactive system based on spatial unconstraint includes:

[0076] The data acquisition and preprocessing module is used to acquire color images, depth images, and inertial sensor data, and perform preprocessing.

[0077] The virtual-real fusion projection module is used to detect and identify the target projection carrier through SLAM and geometric plane recognition technology, and to establish a virtual coordinate system independent of the physical device coordinate system. It maintains the pose transformation and anchor point data for rendering and continuously outputs the perception results to the rendering end.

[0078] The runtime scheduling and data bus module is used to receive camera pose, projection carrier normal and anchor point set, generate timestamps, write them into the pose channel and anchor point channel, and uniformly manage the global clock, message channel and event sequence.

[0079] The dynamic modeling engine module receives input 3D model files, performs semantic layering, slicing operations and material physics simulation, generates layer-by-layer printing animation streams, and calls the defect simulation library to write fault models when preset conditions are met.

[0080] The process persistence and rendering subsystem module is used to establish a double-buffered asynchronous rendering architecture, output simplified geometric placeholder frames in the foreground, generate full-precision rendering frames in the background, and achieve vertical synchronous switching under a zero-interruption displacement mechanism.

[0081] The interactive control module is used to receive and parse gestures, voice and six degrees of freedom tool input, generate interactive instructions and parameter update instructions, and realize hot-effect through parameter thermal control unit;

[0082] The parameter synchronization and feedback module is used to synchronously distribute the parameter update instructions to the dynamic modeling engine and the rendering subsystem, perform feedback processing, and output synchronous rendering frames and interactive results.

[0083] The application interface module is used to receive synchronous rendering frames and interactive update results, perform format adaptation, and output the results to educational training, industrial review, and product design verification application scenarios.

[0084] The beneficial effects of this invention are:

[0085] This invention provides a 3D printing dynamic visualization system and method with spatially decoupled projection, zero-interruption process maintenance, and real-time parameter intervention capabilities. In terms of spatial visualization, this invention introduces SLAM and semantic geometric plane recognition technology to achieve dynamic projection of the printing process onto any physical surface, freeing it from the spatial constraints of traditionally bound printer housings. Actual measurements show that the projection accuracy error does not exceed 1.2mm, while traditional solutions generally have an error greater than 5mm. Furthermore, it supports free switching between at least six independent projection points, significantly improving the flexibility and adaptability of visualization across multiple scenarios.

[0086] Regarding process persistence, this invention proposes a dual-buffered asynchronous rendering architecture and a zero-interruption displacement mechanism. When the user switches the viewing position or the projection medium, the foreground buffer continuously outputs a simplified geometric image, while the background buffer asynchronously generates a full-precision animation, completing a seamless switch within the vertical synchronization cycle. This mechanism reduces the displacement switching latency to less than 182ms, far lower than the 2200ms of traditional solutions, effectively eliminating screen flickering and interactive lag, and ensuring a continuous visual experience during the printing process.

[0087] In terms of interaction depth, this invention introduces a multimodal interaction method combining gestures, voice, and a six-degrees-of-freedom tool, along with a parameter hot-update unit, enabling immediate effect of parameter modifications. Users can directly modify core parameters such as print fill rate and temperature without interrupting the process, or perform operations such as local profile freezing and virtual pen annotation. Testing shows that the response delay for gesture parameter adjustment is no more than 0.15 seconds, and the response time for voice control is no more than 0.3 seconds, far lower than the more than 2 seconds of traditional solutions, resulting in a qualitative improvement in interaction efficiency and immersion.

[0088] This invention achieves breakthroughs in three major aspects: spatial decoupling projection, zero-interruption process maintenance, and real-time interactive intervention. It significantly improves the visualization freedom, interactive fluency, and teaching application value of the 3D printing process. It can effectively solve the technical bottlenecks of existing technologies, such as limited space, fragmented processes, and rigid interaction, and has broad prospects for industrial promotion and educational application. Attached Figure Description

[0089] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0090] Figure 1 This is an overall flowchart of the 3D printing dynamic visualization interaction method based on spatial unconstraint proposed in this invention;

[0091] Figure 2 This is a schematic diagram of the core module structure of the spatially unconstrained 3D printing dynamic visualization interaction method proposed in this invention;

[0092] Figure 3 This is a flowchart of the zero-interruption displacement process for the spatially unconstrained 3D printing dynamic visualization interaction method proposed in this invention. Detailed Implementation

[0093] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0094] refer to Figure 1-3 A spatially unconstrained 3D printing dynamic visualization and interaction method includes the following steps:

[0095] S1. Collect and preprocess environmental data, and generate camera pose, projection carrier normal and anchor point set through SLAM and geometric plane recognition technology, and establish a virtual coordinate system;

[0096] S2. Generate a timestamp for the camera pose, projection carrier normal, and anchor point set, and write it into the pose channel and anchor point channel.

[0097] S3. The dynamic modeling engine receives the input 3D model, performs semantic layering and slicing processing in combination with the timestamp, generates a layer-by-layer printing animation stream, and calls the fault model of the defect simulation library when the preset fault triggering conditions are met. The animation stream and the fault model are written into the animation stream channel together.

[0098] S4. The process preservation and rendering subsystem module receives data from the pose channel, anchor point channel and animation stream channel, and establishes a double-buffered asynchronous rendering architecture.

[0099] S5. Based on the zero-interruption displacement mechanism, when a change in pose and a switch in the projection point are detected, the foreground buffer continues to output continuously, while the background buffer renders a new full-precision frame. After rendering is completed, the foreground and background buffers are switched at the moment of vertical synchronization.

[0100] S6. Receive user gesture input, voice input, and six-degrees-of-freedom tool input, and parse them into parameter update instructions;

[0101] S7. Distribute the parameter update command to the dynamic modeling engine and the process persistence and rendering subsystem module simultaneously, and output synchronous rendering frames and interactive update results;

[0102] S8. Transmit the synchronous rendering frame and interactive update result to the application interface for output.

[0103] This invention begins with data acquisition, using a virtual-real fusion projection module to obtain color images, depth information, and inertial data. It then utilizes SLAM and geometric surface detection to construct an independent virtual coordinate system and generate camera pose, projection carrier normal, and anchor point sets. The runtime scheduling and data bus modules uniformly timestamp these data and write them to the channel to ensure timing consistency. Subsequently, the dynamic modeling engine completes the semantic layering, slicing, and simulation of the 3D model, generating a layer-by-layer printing animation stream and inserting fault models under trigger conditions. The process persistence and rendering subsystem employs a double-buffered asynchronous architecture to achieve seamless switching between foreground and background views, thus maintaining continuous visualization even during pose changes or projection switching. Finally, the multimodal interaction module, combined with the parameter thermal control unit, provides real-time feedback on user input, outputting to the application interface module. This process forms a complete closed loop from data acquisition and dynamic modeling to interactive visualization.

[0104] In this embodiment, S1 specifically includes:

[0105] S11. Collect color image data, depth image data and inertial sensor data of the environment, and synchronize the color image data, depth image data and inertial sensor data in time.

[0106] S12. Perform distortion correction and grayscale normalization processing on the color image data, and perform hole filling and filtering processing on the depth image data;

[0107] S13. Extract corner points and edge features from the processed color image, generate feature descriptors, and perform feature matching between adjacent frames to obtain an initial set of matching pairs.

[0108] S14. Based on the inertial sensor data, perform pre-integration and output motion prior to initialize the simultaneous localization and mapping state vector.

[0109] S15. Calculate the current camera pose using the initial matching pair set and motion prior, perform local binding adjustment, and update the camera pose;

[0110] S16. Use keyframes and depth information to construct a dense point cloud, and perform geometric surface fitting on the dense point cloud region to generate a set of geometric surfaces.

[0111] S17. In the set of geometric surfaces, candidate projection carriers are selected based on area threshold and normal consistency, and the normal vector and fitting residual of the projection carrier are calculated.

[0112] S18. Generate a candidate set of anchor points on the candidate carrier according to the grid spacing, and filter the anchor points that intersect with the camera view frustum to obtain the anchor point set;

[0113] S19. Establish a virtual coordinate system and calculate the homogeneous transformation matrix of the camera pose in the virtual coordinate system. Output the camera pose, the normal of the projection carrier, and the set of anchor points.

[0114] This invention extracts corner and edge features through the simultaneous acquisition and preprocessing of color images, depth data, and inertial data. It then generates motion priors by combining inertial pre-integration with localized binding adjustments to optimize camera pose. Based on keyframes and depth point clouds, a geometric surface is fitted, a carrier is selected, and an anchor point set is generated. Finally, a virtual coordinate system is established, forming a stable spatial reference.

[0115] In this embodiment, S2 specifically includes:

[0116] S21. Receive the camera pose, the normal of the projection carrier, and the set of anchor points;

[0117] S22. Call the global unified clock to generate corresponding timestamps for the camera pose, projection carrier normal and anchor point set;

[0118] S23. Encapsulate the camera pose and timestamp into a pose message packet and write it into the pose channel;

[0119] S24. Encapsulate the projection carrier normal and anchor point set and the timestamp of the projection carrier normal and anchor point set into an anchor point message packet and write it into the anchor point channel.

[0120] This invention receives pose and projection carrier parameters through a runtime scheduling module, calls a unified clock to generate a timestamp, encapsulates the camera pose into a pose message packet and writes it into the pose channel, and encapsulates the projection normal and anchor point set into an anchor point message packet and writes it into the anchor point channel, thereby achieving global temporal consistency and traceability of multi-channel data.

[0121] In this embodiment, S3 specifically includes:

[0122] S31. The dynamic modeling engine receives the input 3D model file and parses it into geometric mesh data, which includes vertex coordinates, normal vectors, topological relationships and material identifiers.

[0123] S32. Perform semantic layering processing on the geometric mesh data, identify and label the outer contour layer, filling structure layer and supporting structure layer of the model as different categories, and assign a unique semantic label to each category;

[0124] S33. Based on the semantic tags, the geometric mesh is sliced ​​according to the set layer thickness parameters to generate layer-by-layer two-dimensional contour slices. Each slice contains a closed contour, path nodes and layer height index.

[0125] S34. Load the material physical parameters corresponding to the three-dimensional model from the material database, including thermal conductivity, melting temperature and cooling shrinkage coefficient;

[0126] S35. Based on the material physical parameters, perform thermophysical simulation on the layer-by-layer slices, calculate the interlayer deformation and correct the printing path, and embed the temperature field distribution and deformation vector field information into the animation clip to generate a layered printing animation clip.

[0127] S36. Monitor printing process parameters in real time. When a preset fault triggering condition is detected, call the defect simulation library to insert the corresponding fault model. The fault triggering conditions include material flow interruption, nozzle temperature surge, edge curling, interlayer peeling, hot bed temperature difference alarm, coordinate axis offset, step-like fracture, and positioning inaccuracy.

[0128] S37. Write the layer-by-layer printing animation stream and the fault model together into the animation stream channel, and attach the timestamp information to output the animation stream channel data.

[0129] This invention parses the 3D model into mesh data, performs semantic layering and slicing operations, and combines material physics parameters to perform thermophysical simulation, generating a layer-by-layer printing animation embedding temperature fields and deformation vectors. When conditions such as flow interruption, warping, or layer misalignment are detected, the corresponding model is inserted by calling the fault simulation library, and the animation stream and fault model are output to the animation channel in a unified manner.

[0130] In this embodiment, S4 specifically includes:

[0131] S41. Receive pose channel and anchor point channel data, as well as the animation stream channel data;

[0132] S42. Establish a dual-buffered asynchronous rendering architecture. The front-end buffer is used to display simplified geometric models in real time, and the back-end buffer is used to render complete animation frames based on full-precision geometric data.

[0133] S43. Load simplified geometric placeholder frames into the foreground buffer and bind the pose message packet to achieve stable scene display with continuous low latency;

[0134] S44. In the background buffer, full-precision rendering is performed based on the virtual coordinate system, anchor point message packet and animation stream channel data. The full-precision rendering includes texture mapping, lighting calculation and layer-by-layer animation composition to form a target high-precision rendering frame.

[0135] S45. During background rendering, continuously output progress prompts and buffer status information, and maintain stable output of the foreground buffer while background rendering is not completed to prevent screen flickering.

[0136] S46. When the background rendering is detected to be complete, perform a foreground-background buffer swap within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving seamless screen switching.

[0137] S47. After performing buffer swapping, release the old foreground buffer resources, mark the background buffer as the new foreground buffer, and start a new background rendering thread to prepare for the next frame, thus achieving continuous loop rendering.

[0138] This invention achieves process persistence through a dual-buffered asynchronous rendering architecture. The foreground buffer outputs simplified geometric frames in real time to ensure low latency, while the background buffer renders full-precision animation frames in a virtual coordinate system. After the background rendering is completed, the foreground and background are switched during the vertical synchronization cycle, achieving seamless screen switching and continuous output, avoiding flickering and stuttering.

[0139] In this embodiment, the zero-interruption displacement mechanism specifically includes:

[0140] S51. When the user makes a drag gesture, the runtime scheduling and data bus module detects the change in pose and the switching of the projection point, and starts the coordinate validity detection.

[0141] S52. When the detection result is invalid, the system outputs an error message; when the detection result is valid, the system proceeds to render the simplified static framework in the foreground.

[0142] S53. While outputting the front-end simplified static framework, start the background full-precision rendering thread;

[0143] S54. During the background rendering process, continuously generate and update progress prompts to provide real-time feedback on the background rendering status.

[0144] S55. When it is detected that the background rendering is not complete, continue to maintain the output of the front-end simplified static framework and update the progress prompt in a loop.

[0145] S56. When the background rendering is detected to be complete, the foreground and background buffers are swapped within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving zero-interruption displacement and continuous visualization.

[0146] S57. After completing the buffer swap, release the background buffer resources and mark the existing output buffer as the new foreground buffer, preparing for the next background rendering loop.

[0147] This invention proposes a zero-interruption displacement mechanism. When a user triggers a pose change or projection switch, the system performs coordinate validity checks, maintains a simplified model output in the foreground, and renders full-precision frames in parallel in the background, generating real-time progress prompts. After background rendering is complete, buffer swapping is completed within a single vertical synchronization cycle, achieving uninterrupted continuous visualization and interaction.

[0148] In this embodiment, S6 specifically includes:

[0149] S61. Receive the user's multimodal input signal, wherein the user's multimodal input signal includes gesture input, voice input and six degrees of freedom tool input;

[0150] S62. Perform hand joint detection, trajectory tracking, ray collision detection and shader freezing on the gesture input. Through multi-angle observation and freezing of the current layer profile analysis, identify the user's five-finger rotation and fist holding operation, and generate corresponding operation instructions.

[0151] S63. Perform automatic speech recognition on the voice input, extract keywords, and map the keywords into parameter control commands;

[0152] S64. Input the position coordinates, attitude angle and button trigger signal of the six-degree-of-freedom tool and generate the corresponding spatial operation instructions;

[0153] S65. The operation instructions, parameter control instructions and spatial operation instructions are uniformly encapsulated into an interactive instruction set;

[0154] S66. Call the parameter thermal control unit to generate parameter update instructions according to the interactive instruction set.

[0155] This invention's interactive control module integrates gestures, voice, and six-degree-of-freedom tool input. Gesture recognition enables multi-angle observation and profile freezing; voice recognition converts it into parameter control commands; and tool input is parsed to parse spatial position and posture to generate operation commands. All multimodal inputs are uniformly encapsulated into an interactive command set, which is then updated in real-time via a parameter thermal control unit.

[0156] In this embodiment, S7 specifically includes:

[0157] S71. Receive the parameter update instruction;

[0158] S72. Generate a timestamp for the parameter update instruction according to the global unified clock, and perform event sorting to ensure that the timing is consistent with the animation stream channel and pose channel;

[0159] S73. Distribute the parameter update command to the dynamic modeling engine, apply it to the slice parameters and simulation parameters, and generate the adjusted layer-by-layer printing animation clip;

[0160] S74. The parameter update instruction is simultaneously distributed to the process persistence and rendering subsystem module, which applies to the rendering resolution, lighting parameters and dynamic display parameters to generate the adjusted rendering frame.

[0161] S75. After the dynamic modeling engine module and the process maintenance and rendering subsystem module have completed the update, perform synchronization processing on the update results and output the synchronized rendering frame and interactive update results.

[0162] S76. The synchronous rendering frame and interactive update results are displayed on the user interface in real time to realize the hot-effect of parameter modification and visual confirmation.

[0163] The runtime scheduling module of this invention receives parameter update commands and timestamps them to ensure synchronization with the animation stream and pose channel. Update commands are distributed to the modeling engine and rendering subsystem; the former adjusts slicing and simulation parameters to generate new animations, while the latter modifies rendering and display parameters to generate new rendering frames. The synchronization results are fed back to the interactive interface, enabling parameters to take effect hot-swappably.

[0164] In this embodiment, S8 specifically includes:

[0165] S81. Input the synchronous rendering frame and interactive update result into the application interface;

[0166] S82. Perform format adaptation processing on the synchronous rendering frame, including resolution scaling, color space conversion and frame rate adjustment, to match the hardware specifications of different display terminals.

[0167] S83. In educational training scenarios, the synchronized rendering frames after format adaptation are combined with teaching prompts and output to the augmented reality terminal to support visualization of the printing process, fault simulation demonstration and multi-angle observation.

[0168] S84. In industrial review scenarios, the synchronized rendering frames after format adaptation and review annotation information are output to the collaborative work platform in a synchronized manner to support online review by multiple users, parameter comparison and anomaly reproduction.

[0169] S85. In the product design verification scenario, the synchronous rendering frame and the CAD nominal model are spatially registered and overlaid, and the deformation deviation heat map and geometric tolerance deviation table are output. The version number, material parameters and slice parameter snapshots are written into the design verification report and pushed to the design review workstation or PLM system through the application interface module.

[0170] This invention adapts the format of rendered frames and interactive results in the application interface module, adjusting resolution, color, and frame rate to ensure compatibility with different terminals. In educational scenarios, it outputs to augmented reality terminals for printed demonstrations and fault teaching; in industrial scenarios, it outputs to collaborative platforms, supporting multi-user review, parameter comparison, and anomaly reproduction, thus enhancing system applicability.

[0171] A spatially unconstrained 3D printing dynamic visualization and interactive system includes:

[0172] The data acquisition and preprocessing module is used to acquire color images, depth images, and inertial sensor data, and perform preprocessing.

[0173] The virtual-real fusion projection module is used to detect and identify the target projection carrier through SLAM and geometric plane recognition technology, and to establish a virtual coordinate system independent of the physical device coordinate system. It maintains the pose transformation and anchor point data for rendering and continuously outputs the perception results to the rendering end.

[0174] The runtime scheduling and data bus module is used to receive camera pose, projection carrier normal and anchor point set, generate timestamps, write them into the pose channel and anchor point channel, and uniformly manage the global clock, message channel and event sequence.

[0175] The dynamic modeling engine module receives input 3D model files, performs semantic layering, slicing operations and material physics simulation, generates layer-by-layer printing animation streams, and calls the defect simulation library to write fault models when preset conditions are met.

[0176] The process persistence and rendering subsystem module is used to establish a double-buffered asynchronous rendering architecture, output simplified geometric placeholder frames in the foreground, generate full-precision rendering frames in the background, and achieve vertical synchronous switching under a zero-interruption displacement mechanism.

[0177] The interactive control module is used to receive and parse gestures, voice and six degrees of freedom tool input, generate interactive instructions and parameter update instructions, and realize hot-effect through parameter thermal control unit;

[0178] The parameter synchronization and feedback module is used to synchronously distribute the parameter update instructions to the dynamic modeling engine and the rendering subsystem, perform feedback processing, and output synchronous rendering frames and interactive results.

[0179] The application interface module is used to receive synchronous rendering frames and interactive update results, perform format adaptation, and output the results to educational training, industrial review, and product design verification application scenarios.

[0180] Example 1:

[0181] To verify the feasibility of this invention in practice, it was applied to an additive manufacturing course at a vocational school, aiming to achieve safe and efficient 3D printing operation training. The implementation process used an XR head-mounted display with a 52-degree field of view and gesture tracking accuracy of ±1.5 mm. The interactive tool was a 6DoF light pen with 2048 levels of pressure sensitivity and 0.3 mm positioning accuracy. The computing workstation was equipped with an AMD-EPYC-9654P processor, an NVIDIA-RTX-6000-Ada graphics card, and 256GB of memory.

[0182] The software environment consists of three core modules: the semantic layering engine uses version v2.3.1 and supports layer thickness sectioning of 0.05-0.3 mm; the fault simulation library uses version v1.8.0 and has 12 pre-built industrial fault models; the SLAM positioning module is developed based on Intel-RealSense-SDK-2.5 and the positioning error is controlled within 1.2 mm.

[0183] The implementation process begins with environmental initialization: the training classroom is scanned using SLAM technology to create a spatial map and identify available geometric surfaces: a 1200×800 mm desk plane with a scanning accuracy of ±3 mm. Subsequently, a gear teaching model (STL format, 100×100×50 mm) is loaded, and a semantic layering engine performs layer processing, setting the layer thickness to 0.1 mm and the material type to PLA.

[0184] Virtual-to-real projection stage: The teacher drags the virtual model to the specified coordinates on the desk plane (x=1.2 meters, y=0.8 meters, z=0 meters) using gestures. At this time, the zero-interruption displacement mechanism is triggered: the foreground immediately displays a low-precision static frame (LOD level 0, number of triangles not exceeding 500), and the background asynchronously renders the remaining 520 layers of printing animation (resolution 1920×1080@60fps).

[0185] Troubleshooting tutorial: After injecting the material blockage command, the system displays a real-time red alert indicating material flow interruption and nozzle temperature rising to 245℃. Students adjust the nozzle temperature parameters (from 200℃ to 220℃) using gestures and test interlayer bonding force using a virtual scraper tool. The system simultaneously generates a stress distribution cloud map. The final output is an operation report, including the time taken for fault diagnosis (2 minutes and 18 seconds) and parameter optimization (3 minutes and 45 seconds).

[0186] The preparation of the specialized material involves a thermosensitive developing coating: containing 60 wt% thermochromic microcapsules (for temperature visualization from 50-300℃), 35 wt% waterborne polyurethane resin (substrate adhesive), and 5 wt% nano-silica (for enhanced wear resistance). The preparation process includes 60℃ water bath mixing (300 rpm / 30 min), ultrasonic dispersion (40 kHz / 15 min), and spray application (film thickness 50 ± 5 μm).

[0187] To verify the advantages of this invention in the dynamic visualization method of 3D printing, a comparative experiment was set up as shown in Table 1:

[0188] Table 1. Comparison of the effects of 3D printing training methods in vocational education

[0189] Comparison Dimensions Method of the present invention Traditional methods Improvement effect Projection accuracy (mm) Positioning error ≤ 1.2 Positioning error ≥ 5 Improved positioning accuracy by more than 4 times Process hold (ms) The dual-buffered asynchronous rendering architecture achieves a switching latency of ≤182ms and zero interruptions. The camera angle needs to be switched, and the animation is interrupted. The average interruption time is about 2200±300 seconds, resulting in a lack of smooth operation. Significantly improved coherence Response latency (s) Gestures ≤ 0.15, speech ≤ 0.3, 6DoF ≤ 0.08 Gestures ≥ 2.0, voice control not supported. Latency reduced by an order of magnitude Parameter modification The parameters are thermally applied and applied to the printed layer. The changes will not take effect immediately and the process needs to be restarted. Improve practical training efficiency

[0190] As shown in Table 1, the method of this invention significantly outperforms traditional methods in several indicators. The comparison results show that, in terms of projection accuracy, the error of traditional methods is typically above 5 mm, leading to significant deviations in the observed results. The projection error of the system of this invention is controlled within 1.2 mm, representing an accuracy improvement of more than four times, making teaching demonstrations more intuitive and reliable. Meanwhile, traditional systems require process interruption when switching perspectives, with an average interruption time exceeding 2 seconds, severely compromising operational continuity. This invention, however, employs a dual-buffered asynchronous rendering architecture, achieving smooth switching through a zero-interruption displacement mechanism, reducing latency to within 182 milliseconds, ensuring continuous visualization output.

[0191] Therefore, the spatially unconstrained 3D printing dynamic visualization system proposed in this invention is significantly superior to traditional solutions in terms of spatial freedom, projection accuracy, process preservation, and interaction modes. It provides an efficient, immersive, and scalable solution for 3D printing training in vocational education, greatly enhancing the practicality and reliability of teaching.

[0192] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A spatially unconstrained 3D printing dynamic visualization and interactive method, characterized in that, Includes the following steps: S1. Collect and preprocess environmental data, and generate camera pose, projection carrier normal and anchor point set through SLAM and geometric plane recognition technology, and establish a virtual coordinate system; S2. Generate a timestamp for the camera pose, projection carrier normal, and anchor point set, and write it into the pose channel and anchor point channel. S3. The dynamic modeling engine receives the input 3D model, performs semantic layering and slicing processing in combination with the timestamp, generates a layer-by-layer printing animation stream, and calls the fault model of the defect simulation library when the preset fault triggering conditions are met. The animation stream and the fault model are written into the animation stream channel together. S4. The process preservation and rendering subsystem module receives data from the pose channel, anchor point channel and animation stream channel, and establishes a double-buffered asynchronous rendering architecture. S5. Based on the zero-interruption displacement mechanism, when a change in pose and a switch in the projection point are detected, the foreground buffer continues to output continuously, while the background buffer renders a new full-precision frame. After rendering is completed, the foreground and background buffers are switched at the moment of vertical synchronization. S6. Receive user gesture input, voice input, and six-degrees-of-freedom tool input, and parse them into parameter update instructions; S7. Distribute the parameter update command to the dynamic modeling engine and the process persistence and rendering subsystem module simultaneously, and output synchronous rendering frames and interactive update results; S8. Transmit the synchronous rendering frame and interactive update result to the application interface for output.

2. The 3D printing dynamic visualization interaction method based on spatially unconstrained design according to claim 1, characterized in that, S1 specifically includes: S11. Collect color image data, depth image data and inertial sensor data of the environment, and synchronize the color image data, depth image data and inertial sensor data in time. S12. Perform distortion correction and grayscale normalization processing on the color image data, and perform hole filling and filtering processing on the depth image data; S13. Extract corner points and edge features from the processed color image, generate feature descriptors, and perform feature matching between adjacent frames to obtain an initial set of matching pairs. S14. Based on the inertial sensor data, perform pre-integration and output motion prior to initialize the simultaneous localization and mapping state vector. S15. Calculate the current camera pose using the initial matching pair set and motion prior, perform local binding adjustment, and update the camera pose; S16. Use keyframes and depth information to construct a dense point cloud, and perform geometric surface fitting on the dense point cloud region to generate a set of geometric surfaces. S17. In the set of geometric surfaces, candidate projection carriers are selected based on area threshold and normal consistency, and the normal vector and fitting residual of the projection carrier are calculated. S18. Generate a candidate set of anchor points on the candidate carrier according to the grid spacing, and filter the anchor points that intersect with the camera view frustum to obtain the anchor point set; S19. Establish a virtual coordinate system and calculate the homogeneous transformation matrix of the camera pose in the virtual coordinate system. Output the camera pose, the normal of the projection carrier, and the set of anchor points.

3. The 3D printing dynamic visualization interaction method based on spatially unconstrained 3D printing according to claim 1, characterized in that, S2 specifically includes: S21. Receive the camera pose, the normal of the projection carrier, and the set of anchor points; S22. Call the global unified clock to generate corresponding timestamps for the camera pose, projection carrier normal and anchor point set; S23. Encapsulate the camera pose and timestamp into a pose message packet and write it into the pose channel; S24. Encapsulate the projection carrier normal and anchor point set and the timestamp of the projection carrier normal and anchor point set into an anchor point message packet and write it into the anchor point channel.

4. The 3D printing dynamic visualization interaction method based on spatial unconstraint as described in claim 1, characterized in that, S3 specifically includes: S31. The dynamic modeling engine receives the input 3D model file and parses it into geometric mesh data, which includes vertex coordinates, normal vectors, topological relationships and material identifiers. S32. Perform semantic layering processing on the geometric mesh data, identify and label the outer contour layer, filling structure layer and supporting structure layer of the model as different categories, and assign a unique semantic label to each category; S33. Based on the semantic tags, the geometric mesh is sliced ​​according to the set layer thickness parameters to generate layer-by-layer two-dimensional contour slices. Each slice contains a closed contour, path nodes and layer height index. S34. Load the material physical parameters corresponding to the three-dimensional model from the material database, including thermal conductivity, melting temperature and cooling shrinkage coefficient; S35. Based on the material physical parameters, perform thermophysical simulation on the layer-by-layer slices, calculate the interlayer deformation and correct the printing path, and embed the temperature field distribution and deformation vector field information into the animation clip to generate a layered printing animation clip. S36. Monitor printing process parameters in real time. When a preset fault triggering condition is detected, call the defect simulation library to insert the corresponding fault model. The fault triggering conditions include material flow interruption, nozzle temperature surge, edge curling, interlayer peeling, hot bed temperature difference alarm, coordinate axis offset, step-like fracture, and positioning inaccuracy. S37. Write the layer-by-layer printing animation stream and the fault model together into the animation stream channel, and attach the timestamp information to output the animation stream channel data.

5. The 3D printing dynamic visualization interaction method based on spatial unconstraint as described in claim 1, characterized in that, S4 specifically includes: S41. Receive pose channel and anchor point channel data, as well as the animation stream channel data; S42. Establish a dual-buffered asynchronous rendering architecture. The front-end buffer is used to display simplified geometric models in real time, and the back-end buffer is used to render complete animation frames based on full-precision geometric data. S43. Load simplified geometric placeholder frames into the foreground buffer and bind the pose message packet to achieve stable scene display with continuous low latency; S44. In the background buffer, full-precision rendering is performed based on the virtual coordinate system, anchor point message packet and animation stream channel data. The full-precision rendering includes texture mapping, lighting calculation and layer-by-layer animation composition to form a target high-precision rendering frame. S45. During background rendering, continuously output progress prompts and buffer status information, and maintain stable output of the foreground buffer while background rendering is not completed to prevent screen flickering. S46. When the background rendering is detected to be complete, perform a foreground-background buffer swap within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving seamless screen switching. S47. After performing buffer swapping, release the old foreground buffer resources, mark the background buffer as the new foreground buffer, and start a new background rendering thread to prepare for the next frame, thus achieving continuous loop rendering.

6. The 3D printing dynamic visualization interaction method based on spatial unconstraint as described in claim 1, characterized in that, The zero-interruption displacement mechanism specifically includes: S51. When the user makes a drag gesture, the runtime scheduling and data bus module detects the change in pose and the switching of the projection point, and starts the coordinate validity detection. S52. When the detection result is invalid, the system outputs an error message; when the detection result is valid, the system proceeds to render the simplified static framework in the foreground. S53. While outputting the front-end simplified static framework, start the background full-precision rendering thread; S54. During the background rendering process, continuously generate and update progress prompts to provide real-time feedback on the background rendering status. S55. When it is detected that the background rendering is not complete, continue to maintain the output of the front-end simplified static framework and update the progress prompt in a loop. S56. When the background rendering is detected to be complete, the foreground and background buffers are swapped within a single vertical synchronization cycle, so that the background rendering result replaces the foreground output, achieving zero-interruption displacement and continuous visualization. S57. After completing the buffer swap, release the background buffer resources and mark the existing output buffer as the new foreground buffer, preparing for the next background rendering loop.

7. The 3D printing dynamic visualization interaction method based on spatial unconstraint as described in claim 1, characterized in that, S6 specifically includes: S61. Receive the user's multimodal input signal, wherein the user's multimodal input signal includes gesture input, voice input and six degrees of freedom tool input; S62. Perform hand joint detection, trajectory tracking, ray collision detection and shader freezing on the gesture input. Through multi-angle observation and freezing of the current layer profile analysis, identify the user's five-finger rotation and fist holding operation, and generate corresponding operation instructions. S63. Perform automatic speech recognition on the voice input, extract keywords, and map the keywords into parameter control commands; S64. Input the position coordinates, attitude angle and button trigger signal of the six-degree-of-freedom tool and generate the corresponding spatial operation instructions; S65. The operation instructions, parameter control instructions and spatial operation instructions are uniformly encapsulated into an interactive instruction set; S66. Call the parameter thermal control unit to generate parameter update instructions according to the interactive instruction set.

8. The 3D printing dynamic visualization interaction method based on spatially unconstrained 3D printing according to claim 1, characterized in that, Specifically, S7 includes: S71. Receive the parameter update instruction; S72. Generate a timestamp for the parameter update instruction according to the global unified clock, and perform event sorting to ensure that the timing is consistent with the animation stream channel and pose channel; S73. Distribute the parameter update command to the dynamic modeling engine, apply it to the slice parameters and simulation parameters, and generate the adjusted layer-by-layer printing animation clip; S74. The parameter update instruction is simultaneously distributed to the process persistence and rendering subsystem module, which applies to the rendering resolution, lighting parameters and dynamic display parameters to generate the adjusted rendering frame. S75. After the dynamic modeling engine module and the process maintenance and rendering subsystem module have completed the update, perform synchronization processing on the update results and output the synchronized rendering frame and interactive update results. S76. The synchronous rendering frame and interactive update results are displayed on the user interface in real time to realize the hot-effect of parameter modification and visual confirmation.

9. The 3D printing dynamic visualization interaction method based on spatial unconstraint as described in claim 1, characterized in that, S8 specifically includes: S81. Input the synchronous rendering frame and interactive update result into the application interface; S82. Perform format adaptation processing on the synchronous rendering frame, including resolution scaling, color space conversion and frame rate adjustment, to match the hardware specifications of different display terminals. S83. In educational training scenarios, the synchronized rendering frames after format adaptation are combined with teaching prompts and output to the augmented reality terminal to support visualization of the printing process, fault simulation demonstration and multi-angle observation. S84. In industrial review scenarios, the synchronized rendering frames after format adaptation and review annotation information are output to the collaborative work platform in a synchronized manner to support online review by multiple users, parameter comparison and anomaly reproduction. S85. In the product design verification scenario, the synchronous rendering frame and the CAD nominal model are spatially registered and overlaid, and the deformation deviation heat map and geometric tolerance deviation table are output. The version number, material parameters and slice parameter snapshots are written into the design verification report and pushed to the design review workstation or PLM system through the application interface module.

10. A spatially unconstrained 3D printing dynamic visualization interaction system, used to execute the spatially unconstrained 3D printing dynamic visualization interaction method according to any one of claims 1 to 9, characterized in that, include: The data acquisition and preprocessing module is used to acquire color images, depth images, and inertial sensor data, and perform preprocessing. The virtual-real fusion projection module is used to detect and identify the target projection carrier through SLAM and geometric plane recognition technology, and to establish a virtual coordinate system independent of the physical device coordinate system. It maintains the pose transformation and anchor point data for rendering and continuously outputs the perception results to the rendering end. The runtime scheduling and data bus module is used to receive camera pose, projection carrier normal and anchor point set, generate timestamps, write them into the pose channel and anchor point channel, and uniformly manage the global clock, message channel and event sequence. The dynamic modeling engine module receives input 3D model files, performs semantic layering, slicing operations and material physics simulation, generates layer-by-layer printing animation streams, and calls the defect simulation library to write fault models when preset conditions are met. The process persistence and rendering subsystem module is used to establish a double-buffered asynchronous rendering architecture, output simplified geometric placeholder frames in the foreground, generate full-precision rendering frames in the background, and achieve vertical synchronous switching under a zero-interruption displacement mechanism. The interactive control module is used to receive and parse gestures, voice and six degrees of freedom tool input, generate interactive instructions and parameter update instructions, and realize hot-effect through parameter thermal control unit; The parameter synchronization and feedback module is used to synchronously distribute the parameter update instructions to the dynamic modeling engine and the rendering subsystem, perform feedback processing, and output synchronous rendering frames and interactive results. The application interface module is used to receive synchronous rendering frames and interactive update results, perform format adaptation, and output the results to educational training, industrial review, and product design verification application scenarios.

Images