A vision twin system suitable for discrete manufacturing industry

Abstract

The application relates to the technical field of production line visiting and training, and particularly relates to a visual number twin system suitable for production line visiting and training, which comprises a resource storage module, a feature matching module, a coordinate mapping module, a model resource module, a beat acquisition module, a beat adaptation module, a real-time monitoring module and an application display module. Through image feature extraction and matching, three-dimensional coordinate mapping, equipment action beat adaptation and other technologies, the system realizes the fusion display of production line static video, a three-dimensional model and real-time data, and supports action simulation, blind area processing and signal interaction functions; the application can provide a high-precision virtual display and training environment, improve visiting and training effects, and simultaneously support homomorphism simulation of equipment action and historical data analysis.

Description

Technical Field

[0001] This invention relates to the fields of industrial automation and digitalization, and in particular to a visual digital twin system suitable for discrete manufacturing. Background Technology

[0002] Currently, production line visits and training methods include hands-on demonstrations and explanations, or the creation of virtual scenarios based on digital twin technology to recreate the production process. Digital twin technology, which combines a 3D engine-driven equipment model with multi-source data such as equipment status and video monitoring, can achieve a high degree of simulation of the production line and is widely used in external presentations and internal training. With the increasing demand for production line flexibility in discrete manufacturing and the widespread adoption of multi-product switching, the role of digital twin technology in production line visits and training is becoming increasingly important.

[0003] For existing digital twin technology applications, it is usually necessary to adjust the form of the 3D model according to the actual production line layout. Especially in scenarios where production lines are frequently adjusted, the system needs to continuously update the 3D model to match the actual production environment. The above methods have certain limitations when dealing with high-frequency production line changes, especially in terms of the realism of virtual simulation restoration, data security requirements, and system construction costs. These factors further highlight the challenges of existing technologies in practical applications. Summary of the Invention

[0004] This invention addresses the problems of low realism and low data security requirements in existing virtual simulation technologies. It provides a visual twin system suitable for production line visits and training, which can adapt to production line layout adjustments without frequent updates to the 3D model, reducing system maintenance costs. Furthermore, it ensures data integrity through blind spot handling and scan surface stitching, and further enhances system stability and robustness by combining a standard environmental model trained with historical data. Therefore, this invention adopts the following technical solution:

[0005] A visual-digital twin system suitable for production line visits and training includes a resource storage module, a feature matching module, a coordinate mapping module, a model resource module, a cycle time acquisition module, a cycle time adaptation module, a real-time monitoring module, and an application demonstration module. The resource storage module stores static video files of the production line and images of the equipment's exterior. The static video files record the operating status of the production equipment from a fixed viewpoint. The feature matching module analyzes the equipment exterior images using image feature extraction algorithms and extracts key feature values. Simultaneously, it samples the static video files at time intervals to obtain keyframes and matches these keyframes with the equipment exterior images. The key feature values ​​of the image are matched and calculated, and the output parameter set a = (PID, DID, T, C1(X1,Y1)), where PID is the production line unique identifier, DID is the device unique identifier, T is the playback duration of the device in the static video, and C1(X1,Y1) is the coordinate of the center pixel point within the feature range of the device in the keyframe; the coordinate mapping module obtains parameter set A from the feature matching module, converts the pixel coordinates C1(X1,Y1) into world coordinates C2(X2,Y2,Z2) within the 3D engine, and outputs parameter set B = (PID, DID, T, C2(X1,Y1)). Z2); Model Resource Module: Stores component-level 3D models corresponding to equipment appearance images. Each component 3D model contains hierarchical structure information and motion description information, defined as parameter group M = (DID, SubID, Level(), Att()), where SubID is the unique identifier of the component, level() is the positional relationship of the component in the equipment hierarchy, and Att() is the set of motion descriptions for the component; Cycle Time Acquisition Module: Connects to the production equipment control system or centralized acquisition system, acquires equipment motion cycle time status data and transmits it to the cycle time adaptation module; Cycle Time Adaptation Module: Acquires equipment motion cycle time status data from the cycle time acquisition module using a periodic scanning method, adapts it with the parameter group M provided by the model resource module, updates the start / stop status in the motion description set Att(), and outputs the real-time parameter group Mreal = (DID, SubID, Level(), Attreal()), where Attreal() is the updated set of motion descriptions; Real-time Monitoring Module: Acquires real-time video streams from the production line and transmits them to the application display module; Application Display Module: Receives data from the coordinate mapping module, cycle time adaptation module, and real-time monitoring module to implement specific application functions.

[0006] The resource storage module includes a static video storage unit and an equipment image storage unit. The static video storage unit records the production line operation status according to a preset viewpoint and speed, while the equipment image storage unit captures and stores images of the equipment's appearance from the same viewpoint. The feature matching module uses a feature extraction algorithm to extract key feature values ​​from the equipment images and extracts keyframes from the static video files using a time interval sampling method. It then uses a similarity calculation method to match the keyframes with the equipment images, outputting parameter group A. The coordinate mapping module completes the conversion from pixel coordinates to 3D coordinates by mapping screen coordinates to normalized device coordinates (NDC), converting NDC coordinates to camera space coordinates using an inverse projection matrix, and converting camera space coordinates to world coordinates using an inverse view matrix.

[0007] The model resource module stores component-level 3D models containing hierarchical structure information: Level() = (num, Direction(x,y,z)), where num represents the specific level of the component in the equipment hierarchy, and Direction(x,y,z) are spatial vector description parameters. The action description set Att() contains multiple independent action descriptions: Att() = (StaticDes, StartID, StopID, animation), where StaticDes is the static description parameter, StartID is the action start signal identifier, StopID is the action stop signal identifier, and animation is a pre-made animation script. The beat acquisition module obtains equipment action beat status data, including start signals, stop signals, and action execution status, by connecting to the equipment control system or a centralized acquisition system.

[0008] The beat adaptation module periodically scans the beat acquisition module to obtain device action beat status data, adapts it to the parameter set M provided by the model resource module, updates the StopID and StopID status in the action description set Att(), and outputs the real-time parameter set Mreal. The application display module receives the parameter set B from the coordinate mapping module and generates anchor icons in the 3D engine based on parameters T and C2(X2,Y2,Z2) for user click operation. Using the static information in the Mreal parameter set received from the beat adaptation module, the static parameters of the device indicated by the anchor icons are displayed as needed. Each component 3D model identified by SubID can be moved step-by-step using Level() in the Mreal parameter set to form an exploded view of the device component, or the script animation in the animation parameter can be demonstrated using the Attreal() parameter set based on StartIDreal and StopIDreal, and displayed on the same screen as the real-time video obtained by the real-time monitoring module.

[0009] The application demonstration module also includes an adjustment parameter group, Adjust = (Δt1, Δt2, Vn), where Δt1 adjusts the motion simulation start time, Δt2 adjusts the motion simulation stop time, and Vn adjusts the animation playback speed. For equipment components with slow-moving rhythms, increasing Δt1 makes the simulation effect more compact; for equipment components with faster operating speeds, Δt2 adjusts the consistency between simulated and real motions; for equipment components with high-frequency motions, Vn reduces the playback speed to approximate the effect observed by the human eye. The application demonstration module can store the Mreal parameter group and use the stored data to play synchronously with static video, achieving homomorphic simulation of equipment component movements when the production equipment is not in operation.

[0010] The beat adaptation module also includes a blind zone filtering step, which retrieves uncovered areas from the model resource module as blind zone ranges. Feedback data within these blind zone ranges is then substituted into the environmental standard model to determine consistency. If consistent, the blind zone range is removed for subsequent overlay steps. It also includes a scan surface stitching step, which stitches the scan area from the model resource module with the area processed in the blind zone filtering step to obtain a complete detection area. Finally, it includes a signal interaction module, which controls the master device to send a call command when external auxiliary equipment needs to be invoked, and the external auxiliary equipment then forwards the command to the secondary device.

[0011] Historical data includes expert experience and historical datasets. The expert experience and historical datasets are mixed in a fixed ratio to form a training set and a validation set. The scene model is trained using the training set to obtain a preliminary model, and then the preliminary model is validated using the validation set to obtain a standard environmental model.

[0012] The advantages and beneficial effects of this invention are as follows: Through the collaboration of resource storage, feature matching, and coordinate mapping modules, precise association between static videos, equipment images, and 3D models is achieved, ensuring that the equipment positions in the virtual scene are highly consistent with the actual production line. This provides a high-precision virtual environment for visits and training, significantly improving the realism of the display. The component-level 3D models stored in the model resource module contain hierarchical structure and motion description information. Combined with the processing of real-time data by the beat acquisition and adaptation module, homomorphic simulation of equipment movements can be achieved. It supports exploded views of component deconstruction and script animation display, facilitating in-depth understanding of equipment structure and operating principles, and enhancing training effectiveness. In addition, the application display module supports real-time video and virtual model display on the same screen, and can flexibly adapt to the motion rhythm of different equipment by adjusting parameter groups, solving the problem of displaying fast and slow movements. At the same time, it can store real-time parameters to realize the motion simulation of equipment in non-working states, improving the applicability of the system. It can cope with production line layout adjustments without frequent updates to the 3D model, reducing system maintenance costs. Furthermore, data integrity is ensured through blind zone processing and scan surface stitching. Combined with the environmental standard model trained with historical data, the stability and robustness of the system are improved. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0014] Figure 2 This is an exploded view of the equipment in this invention; Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] It should be noted that when a component is described as "fixed to" another component, it can be directly on the other component or may have a component in between. When a component is considered "connected to" another component, it can be directly connected to the other component or may have a component in between. When a component is considered "set on" another component, it can be directly set on the other component or may have a component in between. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0018] The embodiments of the present invention will be further described in detail below with reference to the accompanying drawings:

[0019] Due to limitations in handling high-frequency production line changes, particularly regarding the realism of virtual simulation, data security requirements, and system construction costs, existing technologies face challenges in practical applications. Therefore, this invention relates to a visual digital twin system suitable for production line visits and training, the specific implementation of which is as follows: Figure 1As shown, the system architecture includes a resource storage module, a feature matching module, a coordinate mapping module, a model resource module, a beat acquisition module, a beat adaptation module, a real-time monitoring module, and an application display module. These modules communicate and collaborate via data flow. The resource storage module comprises a static video storage unit and a device image storage unit. The static video storage unit records the production line's operating status at preset angles and speeds and generates static video files. The device image storage unit captures and stores images of the device's appearance from the same angle. The feature matching module analyzes the device appearance images using image feature extraction algorithms to extract key feature values. It then samples the static video files at time intervals to obtain keyframes. The keyframes are then matched with the key feature values ​​of the device appearance images, outputting a parameter set A = (PID, DID, T, C1(X1,Y1)), where PID is the production line's unique identifier, DID is the device's unique identifier, T is the playback duration of the device in the static video, and C1(X1,Y1) is the coordinates of the center pixel within the device's feature range in the keyframe.

[0020] The feature matching module is connected to the coordinate mapping module. The coordinate mapping module obtains parameter set A from the feature matching module and converts the NDC coordinates to camera space coordinates through the mapping from screen coordinates to normalized device coordinates (NDC), the inverse projection matrix to convert the NDC coordinates to camera space coordinates, and the inverse view matrix to convert the camera space coordinates to world coordinates. This completes the conversion process from pixel coordinates C1(X1,Y1) to world coordinates C2(X2,Y2,Z2) in the 3D engine. Finally, the parameter set B = (PID, DID, T, C2(X2,Y2,Z2)) is output. The model resource module stores component-level 3D models corresponding to the device's appearance images. Each component 3D model contains hierarchical structure information Level() = (num, Direction(x,y,z)) and action description information Att(), where num represents the specific level of the component in the device's hierarchical structure, Direction(x,y,z) is a spatial vector description parameter, and the action description information Att() contains multiple independent action descriptions Att() = (StaticDes, StopID, StopID, animation), where StaticDes is a static description parameter, StartID is the action start signal identifier code, StopID is the action stop signal identifier code, and animation is a pre-made animation script.

[0021] The cycle time acquisition module connects to the production equipment control system or centralized acquisition system to obtain equipment action cycle time status data, including start signals, stop signals, and action execution status, and transmits the data to the cycle time adaptation module. The cycle time adaptation module uses a periodic scanning method to obtain the equipment action cycle time status data from the cycle time acquisition module and adapts it with the parameter set M = (DID, SubID, Level(), Att()) provided by the model resource module. It updates the StartID and StopID states in the action description set Att() and outputs the real-time parameter set Mreal = (DID, SubID, Level(), Attreal()), where Attreal() is the updated action description set. The cycle time adaptation module also includes a blind zone screening step and a scan surface stitching step. The blind zone screening step retrieves the uncovered area from the model resource module as the blind zone range and uses the environmental standard model to determine whether the feedback data within the blind zone range is consistent. If consistent, the blind zone range is removed for subsequent overlay steps. The scan surface stitching step stitches the scan area in the model resource module with the area processed by the blind zone screening step to obtain a complete detection area.

[0022] The real-time monitoring module acquires real-time video streams from the production line and transmits the data to the application display module. The application display module receives parameter group B from the coordinate mapping module and generates anchor point icons within the 3D engine based on parameters T and C2(X2,Y2,Z2) for user clicks. Using static information from the Mreal parameter group received from the beat adaptation module, the static parameters of the equipment indicated by the anchor point icons are displayed as needed. Each component's 3D model, identified by SubID, can be progressively moved using the Level() function in the Mreal parameter group to form an exploded view of the equipment component's structure, or the script animation in the animation parameters can be displayed simultaneously with the real-time video acquired by the real-time monitoring module using the Attreal() parameter set based on StartIDreal and StopIDreal. Figure 2 As shown, the application demonstration module also includes an adjustment parameter group Adjust = (Δt1, Δt2, Vn), where Δt1 adjusts the motion simulation start time, Δt2 adjusts the motion simulation stop time, and Vn adjusts the animation playback speed. For equipment components with slow motion rhythm, increasing Δt1 makes the simulation effect more compact; for equipment components with faster operating speed, Δt2 adjusts the consistency between simulated and real motion; for equipment components with high-frequency motion, Vn reduces the playback speed to approximate the effect observed by the human eye. The application demonstration module can store the Mreal parameter group and use the stored data to play synchronously with static video, achieving homomorphic simulation of equipment component motion when the production equipment is not in operation.

[0023] In specific application scenarios, the resource storage module first records the production line operation status and generates static video files through the static video storage unit, and simultaneously captures images of the equipment appearance through the equipment image storage unit. The feature matching module extracts image features and key feature values ​​from the equipment appearance images, while simultaneously sampling the static video files at time intervals to obtain keyframes. These keyframes are then matched with the key feature values ​​of the equipment appearance images, outputting parameter set A. The coordinate mapping module, after obtaining parameter set A from the feature matching module, converts the pixel coordinates C1(X1,Y1) to world coordinates C2(X2,Y2,Z2) within the 3D engine through a series of coordinate transformation operations, outputting parameter set B. The model resource module stores component-level 3D models containing hierarchical structure information and motion description information, providing basic data support for subsequent exploded views of equipment components and script animations. The beat acquisition module acquires equipment motion beat status data by connecting to the production equipment control system or centralized acquisition system and transmits the data to the beat adaptation module. The beat adaptation module acquires the device action beat status data through periodic scanning, adapts it with the parameter group M provided by the model resource module, updates the start and stop status in the action description set Att(), and outputs the real-time parameter group Mreal.

[0024] The real-time monitoring module acquires real-time video streams from the production line and transmits the data to the application display module. The application display module receives parameter group B from the coordinate mapping module and generates anchor point icons within the 3D engine based on parameters T and C2(X2,Y2,Z2) for user clicks. Using static information from the Mreal parameter group received from the beat adaptation module, it displays the static parameters of the equipment indicated by the anchor point icons as needed. Each component's 3D model, identified by SubID, can be progressively moved using Level() in the Mreal parameter group to form an exploded view of the equipment component's structure, or the script animation in the animation parameters can be displayed using the Attreal() parameter set based on StartIDreal and StopIDreal, showing the animation alongside the real-time video acquired by the real-time monitoring module. The application display module also includes an adjustment parameter group Adjust = (Δt1, Δt2, Vn), which allows for the adjustment of animation playback speed and time, thus meeting the display needs of different scenarios.

[0025] To enable those skilled in the art to fully understand and implement this invention, the specific implementation principles of this invention are further supplemented below with a specific application scenario.

[0026] In practical applications such as production line visits and training, system operation begins with the resource storage module. The static video storage unit records the production line operation status according to a preset viewpoint and speed, generating static video files; the equipment image storage unit captures and stores images of the equipment's appearance from the same viewpoint. Subsequently, the feature matching module analyzes the equipment appearance images using an image feature extraction algorithm to extract key feature values, and performs time-interval sampling on the static video files to obtain keyframes. The keyframes are then matched with the key feature values ​​of the equipment appearance images, outputting a parameter set A = (PID, DID, T, C1(X1, Y1)). During this process, the image feature extraction algorithm employs edge detection and corner recognition technologies to ensure the accuracy of key feature value extraction, while the time-interval sampling method ensures the representativeness of the keyframes.

[0027] After obtaining parameter set A from the feature matching module, the coordinate mapping module performs a series of coordinate transformation operations to convert pixel coordinates C1(X1,Y1) into world coordinates C2(X2,Y2,Z2) within the 3D engine, outputting parameter set B = (PID, DID, T, C2(X2,Y2,Z2)). Specifically, screen coordinates are mapped using Normalized Device Coordinates (NDC), converted to camera space coordinates using an inverse projection matrix, and converted to world coordinates using an inverse view matrix, completing a precise 2D to 3D conversion. This process ensures the consistency between the device model position in the virtual scene and the actual production environment.

[0028] The model resource module stores component-level 3D models corresponding to the equipment's appearance images. Each component 3D model contains hierarchical structure information Level() = (num, Direction(x,y,z)) and motion description information Att(). The hierarchical structure information defines the relative positional relationships between equipment components, while the motion description information provides the motion logic support for the equipment components. For example, for a rotating machinery component, its motion description information includes the start signal identifier StartID, the stop signal identifier StopID, and the pre-made animation script animation. This data lays the foundation for subsequent exploded views of the equipment components and the display of scripted animations.

[0029] The cycle time acquisition module connects to the production equipment control system or centralized acquisition system to acquire real-time equipment action cycle time status data, including start signals, stop signals, and action execution status, and transmits the data to the cycle time adaptation module. The cycle time adaptation module uses a periodic scanning method to acquire the equipment action cycle time status data from the cycle time acquisition module and adapts it to the parameter set M = (DID, SubID, Level(), Att()) provided by the model resource module. It updates the StartID and StopID states in the action description set Att() and outputs the real-time parameter set Mreal = (DiD, SubID, Level(), Attreal()). During this process, the blind zone screening step retrieves the uncovered area from the model resource module as the blind zone range and uses the environmental standard model to determine if the feedback data within the blind zone range is consistent. If consistent, the blind zone range is removed for subsequent overlay steps. The scan area stitching step stitches the scan area from the model resource module with the area processed by the blind zone screening step to obtain the complete detection area.

[0030] The real-time monitoring module acquires real-time video streams from the production line and transmits the data to the application display module. The application display module receives parameter set B from the coordinate mapping module and generates anchor point icons within the 3D engine based on parameters T and C2(X2,Y2,Z2) for user interaction. When a user clicks an anchor point icon, the system displays the static parameters of that device as needed, based on the static information in the Mreal parameter set received from the beat adaptation module. Furthermore, each component's 3D model, identified by SubID, can be progressively moved using the Level() function in the Mreal parameter set to create an exploded view of the device component, or the script animation in the animation parameters can be displayed using the Attreal() parameter set based on StartIDreal and StopIDreal, simultaneously displayed with the real-time video acquired by the real-time monitoring module. This functionality enables synchronized display of the virtual scene and actual production, significantly enhancing the realism of visits and training.

[0031] like Figure 2 As shown, the application demonstration module also includes the adjustment parameter group Adjust = (Δt1, Δt2, Vn). Here, Δt1 adjusts the motion simulation start time, Δt2 adjusts the motion simulation stop time, and Vn adjusts the animation playback speed. For device components with slow-moving rhythms, increasing Δt1 makes the simulation effect more compact; for device components with faster operating speeds, Δt2 adjusts the consistency between simulated and real actions; for device components with high-frequency actions, Vn reduces the playback speed to approximate the effect observed by the human eye. This mechanism can meet the display needs of different scenarios, thereby improving the system's flexibility and adaptability.

[0032] In the above process, the system also has the ability to learn from historical data. Historical data includes expert experience and historical datasets. The expert experience and historical datasets are mixed in a fixed ratio to form a training set and a validation set. The scenario model is trained using the training set to obtain a preliminary model, and then the preliminary model is validated using the validation set to obtain a standard environmental model. This process ensures the robustness and accuracy of the system in complex scenarios.

[0033] Through the above steps, this system can achieve a high degree of simulation for production line visits and training, especially in situations where production line layouts are frequently adjusted, maintaining a highly realistic virtual scene display without the need for frequent updates to the 3D model. This technical solution effectively solves the limitations of existing digital twin technology in scenarios with high-frequency production line changes, while also taking into account factors such as the realism of virtual simulation, data security requirements, and system construction costs, providing an efficient and flexible solution for discrete manufacturing industries.

[0034] All content not described in detail in this specification is prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited; conventional equipment can be used. Electrical control components not mentioned in this technical solution are not shown in the figures because they are prior art, and will not be described further here.

[0035] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A visual digital twin system suitable for discrete manufacturing, characterized in that: include The resource storage module stores static videos of the production line and images of the equipment's appearance from a fixed perspective as basic visual data. The feature matching module extracts key feature values ​​and keyframes from the basic visual data, performs matching calculations based on the key feature values ​​and keyframes, and outputs a parameter set containing the identifier code, playback duration, and keyframe pixel coordinates. ; The feature matching module uses a feature extraction algorithm to extract key feature values ​​from the device image and extracts key frames from the static video file using a time interval sampling method. It then uses a similarity calculation method to match the key frames with the device image, outputting a parameter set. ; The coordinate mapping module maps the parameter set. The keyframe pixel coordinates are converted to world coordinates within the 3D engine, and the output includes playback duration and parameters in world coordinates. ; The model resource module stores component-level 3D models corresponding to the device's appearance images. Each component 3D model contains hierarchical structure information and motion description information. It combines the identifier code, hierarchical structure information, and motion description information to output parameter sets. ; The component-level 3D models stored in the model resource module contain hierarchical structure information. ,in This indicates the specific level of the component within the equipment hierarchy. The parameters are spatial vector descriptions, and the action description information contains multiple independent action descriptions. ,in For static description parameters, This is the action start signal identifier code. This is the action stop signal identifier code. For pre-made animation scripts; The beat acquisition module connects to the equipment control system to obtain action beat status data; The beat adaptation module acquires device action beat status data from the beat acquisition module through periodic scanning and compares it with the parameter set provided by the model resource module. Adapt and update the action description information. and The status output includes a set of parameters based on the updated motion scan information. ; The real-time monitoring module acquires real-time video streams from the production line. Application demonstration module, receiving parameter groups , It can generate anchor point icons for real-time video streams, display device parameters, exploded views of components or animations, and display them on the same screen as the real-time video.

2. The visual digital twin system for discrete manufacturing according to claim 1, characterized in that: The resource storage module includes a static video storage unit and an equipment image storage unit. The static video storage unit records the production line operation status according to a preset angle and speed, while the equipment image storage unit takes and stores images of the equipment appearance from the same angle.

3. The visual digital twin system for discrete manufacturing according to claim 1, characterized in that: The coordinate mapping module achieves the conversion from pixel coordinates to three-dimensional coordinates by mapping screen coordinates to standardized device coordinates, converting standardized device coordinates to camera space coordinates using an inverse projection matrix, and converting camera space coordinates to world coordinates using an inverse view matrix.

4. The visual digital twin system suitable for discrete manufacturing according to claim 1, characterized in that: The beat adaptation module also includes a blind spot screening step and a scan surface stitching step; The blind zone screening step involves retrieving the uncovered area in the model resource module as the blind zone range, substituting the feedback data within the blind zone range into the preset environmental standard model to determine whether they are consistent. If they are consistent, the blind zone range is removed. The scanning area stitching step stitches together the scanning area in the model resource module with the area processed by the blind area screening step to obtain a complete detection area.

5. The visual digital twin system suitable for discrete manufacturing according to claim 4, characterized in that: The application demonstration module also includes an adjustment parameter group, which is used to adjust the motion simulation start time, the motion simulation stop time, and the animation playback speed.

6. The visual digital twin system for discrete manufacturing according to claim 5, characterized in that: The application demonstration module can store parameter sets and use the stored data to play static videos synchronously, realizing the homomorphic simulation of the actions of equipment components when the production equipment is not working.

7. The visual digital twin system for discrete manufacturing according to claim 6, characterized in that: The beat adaptation module also includes a signal interaction module. When it is necessary to call an external auxiliary device, the main device is controlled to send a call command, and then the external auxiliary device forwards the command to the secondary device.

8. The visual digital twin system for discrete manufacturing according to claim 7, characterized in that: The beat adaptation module also includes an environmental standard model construction step. The steps for constructing the environmental standard model involve mixing expert experience with historical datasets in a fixed ratio to form a training set and a validation set. The scenario model is trained using the training set to obtain a preliminary model, and then the preliminary model is validated using the validation set to obtain the environmental standard model.

Images