Intelligent manufacturing three-dimensional visualization system based on digital twinning

Abstract

The invention discloses an intelligent manufacturing three-dimensional visualization system based on digital twinning, relates to the technical field of intelligent manufacturing supervision, and improves the instantaneity of intelligent manufacturing supervision. According to the method, after pixels in a texture detection area are removed and filled, a corresponding feature texture distribution curve is divided from each piece of historical processing image data, and then a historical product processing dynamic structure model is segmented into a historical equipment dynamic change model and a historical product dynamic model according to the distribution of the feature texture distribution curve; setting a plurality of track identification points in the historical product processing dynamic structure model, judging whether the track identification points among the production supervision time points coincide or not, generating a standard processing track fluctuation range according to a judgment result, generating a real-time processing interaction track according to the real-time processing video data and the real-time laser video data, and outputting the real-time processing interaction track. And judging whether the real-time processing interaction track is within the quasi-processing fluctuation range or not, and judging whether the corresponding processing equipment has production abnormality or not.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing supervision technology, and in particular to an intelligent manufacturing three-dimensional visualization system based on digital twins. Background Art

[0002] Intelligent manufacturing is the future trend of the manufacturing industry. It integrates advanced information technology, automation technology and artificial intelligence technology to realize the intelligent, flexible and efficient production process. The goal of intelligent manufacturing is to improve production efficiency, reduce costs, improve product quality, and enhance the market competitiveness of enterprises. It is mainly used in laser processing and the production of complex parts, and puts forward higher requirements for real-time monitoring and precise control during the processing process.

[0003] The existing intelligent manufacturing visualization technology generally has the problem of low accuracy of digital twin models, which directly affects its application effect. The existing model building technology lacks sufficient details or real-time performance and cannot accurately reflect the state and behavior of physical entities. In addition, due to the complicated and varied steps of intelligent manufacturing, the location of production anomalies cannot be accurately displayed during the manufacturing visualization process.

[0004] Therefore, how to improve the details of model construction while realizing the immediate discovery and processing of production anomalies is a difficulty in the existing technology. For this purpose, a smart manufacturing 3D visualization system based on digital twins is provided. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a three-dimensional visualization system for intelligent manufacturing based on digital twins.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: The digital twin-based intelligent manufacturing three-dimensional visualization system includes a control terminal, which is communicatively connected to a manufacturing process data acquisition module, a production line model building module, a manufacturing visualization module, and a production supervision module; The manufacturing process data acquisition module is used to set up several product production plans, establish a product-equipment process tree according to the product production plan, and collect historical processing video data and historical laser video data when each processing equipment on the production line executes different product production plans; The production line model building module is used to divide the historical processed video data into a number of historical processed image data, and divide a number of texture detection areas from each historical processed image data, remove and fill pixels in the texture detection area, overlap and splice each texture detection area, and divide the corresponding characteristic texture distribution curve in each historical processed image data; Establish a corresponding historical product processing dynamic appearance model according to each historical processing image data, and establish a historical product processing dynamic structure model according to the historical laser video data, and then divide the historical product processing dynamic structure model into a historical equipment dynamic change model and a historical product dynamic model according to the characteristic texture distribution curve distribution on the historical product processing dynamic appearance model; The manufacturing visualization module is used to set a number of trajectory identification points in the historical product processing dynamic structure model according to the historical equipment dynamic change model and the historical product dynamic model distribution, set the production supervision time point, and then determine whether there is overlap between the trajectory identification points between the various production supervision time points, generate the historical processing interaction trajectory according to the judgment result, and obtain the corresponding standard processing trajectory fluctuation range according to the historical processing interaction trajectory generated corresponding to different historical laser video data; The production supervision module is used to obtain real-time processing video data and real-time laser video data when executing the product production plan, and then generate a real-time processing interaction trajectory based on the real-time processing video data and the real-time laser video data. By judging whether the real-time processing interaction trajectory is within the quasi-processing fluctuation range, it is judged whether there is a production anomaly in the corresponding processing equipment, and corresponding abnormal repair operations are performed.

[0007] Furthermore, the acquisition process of the historical processed video data and the historical laser video data includes: A high-speed camera and a laser sensor are set for each processing equipment, and several product production plans are set, wherein the product production plan includes a product name, a name of an associated processing equipment, and processing steps; According to the processing steps in each product production plan and the corresponding associated processing equipment names, as well as the distribution of various processing equipment on the production line, a product-equipment flow tree of each product production plan is traversed on the production line, wherein the product-equipment flow tree is composed of a number of equipment nodes; Each product production plan is executed m times on the production line, and every time the product passes through a processing device to perform a corresponding processing step, the high-speed camera and laser sensor on the corresponding processing device collect the historical processing video data and historical laser video data of the production interaction between the processing equipment and the product, where m is a natural number greater than 0.

[0008] Furthermore, the process of removing pixels in the texture detection area includes: Generate a number of historical processed image data according to the historical processed video data of each product production plan, and obtain the pixel value of each pixel in each historical processed image data; Select multiple i*i texture detection areas starting from the first pixel in the upper left corner in the historical processed image data until all pixels are traversed, where i=3, ..., and i is less than or equal to the number of pixels of the smallest side of the historical processed image data; One pixel in the adjacent row and column is removed respectively, and then one pixel in the adjacent row and column of the next pixel is removed respectively, until the last pixel in the lower right corner of the historical processed image data is reached.

[0009] Furthermore, the process of filling pixels in the texture detection area includes: After the pixel removal operation in the texture detection area is completed, a number of 2*2 pixel areas are selected from the texture detection area, and then the average value of the pixel difference between adjacent pixels in the pixel area is obtained, and smoothing pixels are set between each adjacent pixel in the pixel area; A pixel difference threshold is set, and the pixel difference between adjacent pixels in each texture detection area is obtained, and then the pixel difference between adjacent pixels is compared with the pixel difference threshold, and a number of appearance textures are divided in each texture detection area according to the comparison result.

[0010] Furthermore, the process of acquiring the characteristic texture distribution curve includes: A two-dimensional coordinate system is established, and according to the corresponding position of each texture detection area in the historical processed image data, each historical processed image data after performing the filling and smoothing pixel operation is overlapped and mapped in the two-dimensional coordinate system; Since each texture detection area corresponds to the same historical processed image data, the distribution and spreading trend of the appearance texture in each texture detection area must have the same parts; Then, according to the overlapping results of each texture detection area, each appearance texture is spliced ​​to obtain the characteristic texture distribution curve corresponding to the historical processed image data.

[0011] Furthermore, the process of dividing the historical product processing dynamic structure model into the historical equipment dynamic change model and the historical product dynamic model includes: According to each historical processing image data, the corresponding product-equipment interaction model is established, and overlapping and splicing are performed along the characteristic texture distribution curve to obtain the historical product processing dynamic appearance model between the corresponding processing equipment and the type of product, and the historical product processing dynamic structure model is generated according to the historical laser video data; Establish a three-dimensional coordinate system, and simultaneously map the historical product processing dynamic structure model and the historical product processing dynamic appearance model in the three-dimensional coordinate system, split the historical product processing dynamic structure model and the historical product processing dynamic appearance model into a number of static structure models and static appearance models in chronological order, and then divide the static structure model into a plurality of static product component models and static equipment component models according to the characteristic texture distribution curve distribution on the static appearance model at the same time point; According to the positions of each static product component model and static equipment component model at each time point, the positions are distributed on the historical product processing dynamic structure model, and then the historical equipment dynamic change model and the historical product dynamic model are divided on the historical product processing dynamic structure model, and bound to the corresponding equipment node in the corresponding product-equipment process tree.

[0012] Furthermore, the process of dividing the product processing area includes: The manufacturing visualization module maps all historical product processing dynamic structure models corresponding to the same equipment node and type of product in the product-equipment process tree into the same three-dimensional coordinate system; According to the execution time length of each processing step, several production supervision time points are set, and then the model parts where the historical product processing dynamic structure model, the historical equipment dynamic change model and the historical product dynamic model interact at each production supervision time point are obtained and recorded as product processing areas.

[0013] Furthermore, the process of obtaining the fluctuation range of the standard machining trajectory includes: A plurality of interactive identification points are respectively set at edge positions of model parts corresponding to the historical equipment dynamic change model and the historical product dynamic model in the product processing area; Before all production supervision time points have passed, whenever there is an interaction between a pair of production supervision time points, the interaction identification points on the historical equipment dynamic change model and the historical product dynamic model interact, the overlapping interaction identification points are recorded as trajectory identification points, and the corresponding historical processing interaction trajectory is generated according to the spatial position changes of the trajectory identification points between the corresponding production supervision time points; The historical processing interaction trajectories between the same processing steps, the same positions and production supervision time points with different execution times but corresponding to the same product production plan are mapped to the same three-dimensional coordinate system, and the m historical processing interaction trajectories are split into several coordinate points. The coordinate points at the same time are normally distributed, and the data interval at the center of the normal distribution result is selected as the data fluctuation range at the corresponding time. The data fluctuation ranges at each time are connected in sequence to obtain the standard processing trajectory fluctuation range at the corresponding position.

[0014] Furthermore, the process of determining whether there is production abnormality in the processing equipment includes: During the execution of the product production plan, every time the product passes through a processing device, the high-speed camera and laser sensor on the corresponding processing device collect real-time laser video data and real-time processing video data, and establish a corresponding real-time processing visualization model based on the real-time laser video data and real-time processing video data; Whenever a production supervision time point ends, the real-time processing visualization model generated between the previous production supervision time point and the corresponding historical equipment dynamic change model will be overlapped and mapped, and the process of establishing the standard processing trajectory fluctuation range will be adopted to obtain the real-time processing interaction trajectory of the real-time processing visualization model between the corresponding production supervision time points; Determine whether the real-time processing interaction trajectory is within the standard processing trajectory fluctuation range. If so, do not do anything. If not, determine that there is a production anomaly between the corresponding production supervision time points. If there are more than three production supervision time points where production anomalies are judged, then at the end of the corresponding production step, the real-time processing visualization model is compared with the historical equipment dynamic change model to obtain the model parts where there are differences between the two, and a product repair decision is generated based on the model parts where there are differences; According to the product repair decision, the corresponding processing equipment sends the product to the processing equipment of the same level for repair via the conveyor belt. If the processing equipment of the same level is already working, the corresponding processing equipment will execute the product repair decision. Otherwise, it will be sent to the processing equipment of the same level to execute the product repair decision, and the corresponding processing equipment will process the next product.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. After removing and filling the pixels in the texture detection area, the present invention divides the corresponding characteristic texture distribution curve in each historical processing image data, and then divides the historical product processing dynamic structure model into a historical equipment dynamic change model and a historical product dynamic model according to the distribution of the characteristic texture distribution curve, thereby realizing the supervision of minor changes in the production process and improving the precision and accuracy of the visualization model.

[0016] 2. The present invention sets a number of trajectory identification points in the dynamic structure model of historical product processing, determines whether there is any overlap between the trajectory identification points between various production supervision time points, generates a standard processing trajectory fluctuation range according to the judgment result, generates a real-time processing interaction trajectory according to the real-time processing video data and the real-time laser video data, and determines whether the real-time processing interaction trajectory is within the quasi-processing fluctuation range. It determines whether there is a production anomaly in the corresponding processing equipment by judging whether the real-time processing interaction trajectory is within the quasi-processing fluctuation range, thereby realizing instant discovery and processing of production anomalies, greatly shortening the time for anomaly processing, and improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0018] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0019] like Figure 1 As shown, the intelligent manufacturing three-dimensional visualization system based on digital twins includes a control terminal, which is communicatively connected to a manufacturing process data acquisition module, a production line model building module, a manufacturing visualization module, and a production supervision module; The manufacturing process data acquisition module is used to set up several product production plans, establish a product-equipment process tree according to the product production plan, and collect historical processing video data and historical laser video data when each processing equipment on the production line executes different product production plans; The production line model building module is used to divide the historical processed video data into a number of historical processed image data, and divide a number of texture detection areas from each historical processed image data, remove and fill pixels in the texture detection area, overlap and splice each texture detection area, and divide the corresponding characteristic texture distribution curve in each historical processed image data; Establish a corresponding historical product processing dynamic appearance model according to each historical processing image data, and establish a historical product processing dynamic structure model according to the historical laser video data, and then divide the historical product processing dynamic structure model into a historical equipment dynamic change model and a historical product dynamic model according to the characteristic texture distribution curve distribution on the historical product processing dynamic appearance model; The manufacturing visualization module is used to set a number of trajectory identification points in the historical product processing dynamic structure model according to the historical equipment dynamic change model and the historical product dynamic model distribution, set the production supervision time point, and then determine whether there is overlap between the trajectory identification points between the various production supervision time points, generate the historical processing interaction trajectory according to the judgment result, and obtain the corresponding standard processing trajectory fluctuation range according to the historical processing interaction trajectory generated corresponding to different historical laser video data; The production supervision module is used to obtain real-time processing video data and real-time laser video data when executing the product production plan, and then generate a real-time processing interaction trajectory based on the real-time processing video data and the real-time laser video data. By judging whether the real-time processing interaction trajectory is within the quasi-processing fluctuation range, it is judged whether there is a production anomaly in the corresponding processing equipment, and corresponding abnormal repair operations are performed.

[0020] Further, the working principle of the present invention is specifically described below through embodiments: The production line is equipped with several processing equipment, such as material handling devices, laser cutting machines, 3D printers, welding machines, etc.; The manufacturing process data acquisition module sets a number h1, h2, h3, ... h for each processing equipment n , where n is a natural number greater than 0; It should be noted that on the production line, each processing equipment is connected by two conveyor belts running in opposite directions, and each processing equipment is hierarchically connected and distributed in a tree structure. Each level of processing equipment can simultaneously carry out two-way allocation of processed products with its upper and lower levels and the same level of processing equipment; A high-speed camera and a laser sensor are set for each processing equipment, and several product production plans are set. The product production plan includes a product name, a name of an associated processing equipment, and processing steps. The processing steps are, for example, cutting, turning, product packaging, and part printing. It should be further explained that there are multiple processing steps that are executed progressively or in parallel, and the data collection range and collection angle of the high-speed camera and the laser sensor are always the same; According to the processing steps in each product production plan and the corresponding associated processing equipment names, as well as the distribution of various processing equipment on the production line, a product-equipment flow tree of each product production plan is traversed on the production line, wherein the product-equipment flow tree is composed of a plurality of equipment nodes, and arrow connecting lines are provided between each equipment node; Each product production plan is executed m times on the production line, and every time the product passes through a processing device to perform a corresponding processing step, the high-speed camera and laser sensor on the corresponding processing device collect the historical processing video data and historical laser video data of the production interaction between the processing equipment and the product, where m is a natural number greater than 0.

[0021] Furthermore, the manufacturing process data acquisition module sends the historical processing video data and historical laser video data generated after each execution of all product production plans to the production line model building module; The production line model building module generates a plurality of historical processed image data according to the historical processed video data of each product production plan, and obtains the pixel value of each pixel in each historical processed image data; Select multiple i*i texture detection areas starting from the first pixel in the upper left corner in the historical processed image data until all pixels are traversed, where i=3, ..., and i is less than or equal to the number of pixels of the smallest side of the historical processed image data; For any texture detection area, start from the first pixel in the upper left corner, remove one pixel in the adjacent row and column respectively, and then remove one pixel in the adjacent row and column of the next pixel respectively, until the last pixel in the lower right corner of the historical processed image data; It should be noted that the pixel that has been removed from adjacent rows and adjacent columns will not be removed; When the pixel removal operation of the texture detection area is completed, several 2*2 pixel areas are selected from the texture detection area, and then the average value of the pixel difference between adjacent pixels in the pixel area is obtained, and smooth pixels are set between each adjacent pixel in the pixel area, and the average value of the pixel difference and the minimum pixel value of the corresponding two adjacent pixels are accumulated, and the accumulated value is assigned to the smooth pixel as its pixel value; It should be noted that if the texture detection area is divided into pixel areas and the remaining pixels are insufficient to form a 2*2 pixel area, the remaining pixels will be incorporated into the nearest pixel area; Repeat the above-mentioned operation of filling smooth pixels in each texture detection area, set a pixel difference threshold, and obtain the pixel difference between adjacent pixels in each texture detection area, and then compare the pixel difference between each adjacent pixel with the pixel difference threshold; If the pixel difference is greater than the pixel difference threshold, no operation is performed; If the pixel difference is less than or equal to the pixel difference threshold, the same label is set for the corresponding adjacent pixels; Connect adjacent pixels with the same annotation in each texture detection area in sequence, and then divide a number of appearance textures in each texture detection area; A two-dimensional coordinate system is established, and according to the corresponding position of each texture detection area in the historical processed image data, each historical processed image data after performing the filling and smoothing pixel operation is overlapped and mapped in the two-dimensional coordinate system; Since each texture detection area corresponds to the same historical processed image data, the distribution and spreading trend of the appearance texture in each texture detection area must have the same parts; Then, according to the overlapping results of each texture detection area, each appearance texture is spliced ​​to obtain the characteristic texture distribution curve corresponding to the historical processed image data.

[0022] Furthermore, since each historical processed image data is generated corresponding to the same historical processed video data, the characteristic texture distribution curves of each historical processed image data also have related or identical parts; Then, a corresponding product-equipment interaction model is established according to each historical processing image data, and a product-equipment interaction model is generated corresponding to the same historical processing video data, and overlapped and spliced ​​along the characteristic texture distribution curve, so as to obtain a historical product processing dynamic appearance model between the corresponding processing equipment and the type of product; At the same time, since the historical processing video data is collected by a high-speed camera with a fixed collection angle, the historical product processing dynamic appearance model generated according to the historical processing video data only displays one dimensional plane of the interaction between the processing equipment and the product, making it difficult to distinguish the model parts corresponding to the processing equipment and the product in the historical product processing dynamic appearance model; Since the historical laser video data is generated by laser sensors emitting laser signals to processing equipment and products, the historical product processing dynamic structure model is generated based on the historical laser video data; A three-dimensional coordinate system is established, and the historical product processing dynamic structure model and the historical product processing dynamic appearance model are simultaneously mapped in the three-dimensional coordinate system. Since the data acquisition range and acquisition angle of the laser sensor are consistent with those of the high-speed camera, the historical product processing dynamic structure model and the historical product processing dynamic appearance model are synchronously associated, that is, the characteristic texture distribution curve on the historical product processing dynamic appearance model represents the interaction dividing line between the processing equipment and the product, and represents the component dividing line of the processing equipment and the product respectively; The historical product processing dynamic structure model and the historical product processing dynamic appearance model are split into several static structure models and static appearance models in chronological order, and then the static structure model is divided into multiple static product component models and static equipment component models according to the characteristic texture distribution curve distribution on the static appearance model at the same time point; According to the positions of each static product component model and static equipment component model at each time point, the positions are distributed on the historical product processing dynamic structure model, and then the historical equipment dynamic change model and the historical product dynamic model are divided on the historical product processing dynamic structure model; Repeat the above historical equipment dynamic change model and historical product dynamic model division process to obtain the historical equipment dynamic change model and historical product dynamic model of each product when executing different processing steps, and bind them to the corresponding equipment node in the corresponding product-equipment process tree.

[0023] Furthermore, the production line model building module sends the entire product-equipment process tree to the manufacturing visualization module; The manufacturing visualization module maps all historical product processing dynamic structure models corresponding to the same equipment node and type of product in the product-equipment process tree into the same three-dimensional coordinate system; According to the execution time length of each processing step, several production supervision time points are set, and then the model parts where the historical product processing dynamic structure model, the historical equipment dynamic change model and the historical product dynamic model interact at each production supervision time point are obtained, and recorded as product processing areas; Starting from the first production supervision time point, multiple interactive identification points are set at edge positions of model parts corresponding to the historical equipment dynamic change model and the historical product dynamic model in the product processing area; Before all production supervision time points have passed, whenever there is an interaction between a pair of production supervision time points, the interaction identification points on the historical equipment dynamic change model and the historical product dynamic model interact, the overlapping interaction identification points are recorded as trajectory identification points, and the corresponding historical processing interaction trajectory is generated according to the spatial position changes of the trajectory identification points between the corresponding production supervision time points; The historical processing interaction trajectories with different execution times but corresponding to the same product production plan, the same processing steps, the same position and the same production supervision time point are mapped to the same three-dimensional coordinate system, and the m historical processing interaction trajectories are split into several coordinate points. The coordinate points at the same time are normally distributed, and the data interval at the center of the normal distribution result is selected as the data fluctuation range at the corresponding time. Connect the data fluctuation ranges at each moment in sequence to obtain the standard processing trajectory fluctuation range at the corresponding position, repeat the above operation, and then obtain multiple standard processing trajectory fluctuation ranges for each processing step between different production supervision time points, and mark each standard processing trajectory fluctuation range at the corresponding position of the historical equipment dynamic change model bound to the equipment node in the corresponding product-equipment process tree.

[0024] Furthermore, the manufacturing visualization module sends the entire product-equipment process tree to the production supervision module; The staff uploads the product production plan to be executed to the production supervision module, and then the production supervision module matches the corresponding product-equipment process tree according to the product production plan, and judges whether the processing equipment on the current production line meets the requirements for executing the product production plan according to the corresponding equipment types and quantities in the product-equipment process tree; If not, a prompt that the execution cannot be performed will be sent to the staff; If the conditions are met, the corresponding processing equipment will be dispatched to execute the product production plan; During the execution of the product production plan, every time the product passes through a processing device, the high-speed camera and laser sensor on the corresponding processing device collect real-time laser video data and real-time processing video data, and establish a corresponding real-time processing visualization model based on the real-time laser video data and real-time processing video data; Whenever a production supervision time point ends, the real-time processing visualization model generated between the previous production supervision time point and the corresponding historical equipment dynamic change model will be overlapped and mapped, and the process of establishing the standard processing trajectory fluctuation range will be adopted to obtain the real-time processing interaction trajectory of the real-time processing visualization model between the corresponding production supervision time points; Determine whether the real-time processing interaction trajectory is within the standard processing trajectory fluctuation range. If so, do not perform any operation. If not, determine that there is a production anomaly between the corresponding production supervision time points. If there are more than three production supervision time points where production anomalies are judged, then at the end of the corresponding production step, the real-time processing visualization model is compared with the historical equipment dynamic change model to obtain the model parts where there are differences between the two, and a product repair decision is generated based on the model parts where there are differences; According to the product repair decision, the corresponding processing equipment sends the product to the processing equipment of the same level for repair through the conveyor belt. If the processing equipment of the same level is already working, the corresponding processing equipment will execute the product repair decision. Otherwise, it will be sent to the processing equipment of the same level to execute the product repair decision, and the corresponding processing equipment will process the next product. Repeat the above production supervision operation until all processing steps of the product are completed.

[0025] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A three-dimensional visualization system for intelligent manufacturing based on digital twins, including a control terminal, characterized in that: The control terminal is communicatively connected to a manufacturing process data acquisition module, a production line model building module, a manufacturing visualization module, and a production supervision module; The manufacturing process data acquisition module is used to set up several product production plans, establish a product-equipment process tree according to the product production plan, and collect historical processing video data and historical laser video data when each processing equipment on the production line executes different product production plans; The production line model building module is used to divide the historical processed video data into a number of historical processed image data, and divide a number of texture detection areas from each historical processed image data, remove and fill pixels in the texture detection area, overlap and splice each texture detection area, and divide the corresponding characteristic texture distribution curve in each historical processed image data; Establish a corresponding historical product processing dynamic appearance model according to each historical processing image data, and establish a historical product processing dynamic structure model according to the historical laser video data, and then divide the historical product processing dynamic structure model into a historical equipment dynamic change model and a historical product dynamic model according to the distribution status of the characteristic texture distribution curve on the historical product processing dynamic appearance model; The manufacturing visualization module is used to set a number of trajectory identification points in the historical product processing dynamic structure model according to the historical equipment dynamic change model and the historical product dynamic model distribution, set the production supervision time point, and then determine whether there is overlap between the trajectory identification points between the various production supervision time points, generate the historical processing interaction trajectory according to the judgment result, and obtain the corresponding standard processing trajectory fluctuation range according to the historical processing interaction trajectory generated corresponding to different historical laser video data; The production supervision module is used to obtain real-time processing video data and real-time laser video data when executing the product production plan, and then generate a real-time processing interaction trajectory based on the real-time processing video data and the real-time laser video data. By judging whether the real-time processing interaction trajectory is within the quasi-processing fluctuation range, it is judged whether there is a production anomaly in the corresponding processing equipment, and corresponding abnormal repair operations are performed.

2. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 1 is characterized in that: The acquisition process of the historical processed video data and the historical laser video data includes: A high-speed camera and a laser sensor are set for each processing equipment, and several product production plans are set, and a product-equipment process tree of each product production plan is traversed on the production line, wherein the product-equipment process tree is composed of several equipment nodes; Each product production plan is executed m times on the production line, and every time the product passes through a processing device to perform a corresponding processing step, the high-speed camera and laser sensor on the corresponding processing device collect the historical processing video data and historical laser video data of the production interaction between the processing equipment and the product, where m is a natural number greater than 0.

3. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 2 is characterized in that: The process of removing pixels in the texture detection area includes: Generate several pieces of historical processed image data according to the historical processed video data, and obtain the pixel value of each pixel in each piece of historical processed image data, select multiple i*i texture detection areas starting from the first pixel in the upper left corner in the historical processed image data until all pixels are traversed, where i=3, 4, ..., and i is less than or equal to the number of pixels of the smallest side of the historical processed image data; One pixel in the adjacent row and column is removed respectively, and then one pixel in the adjacent row and column of the next pixel is removed respectively, until the last pixel in the lower right corner of the historical processed image data is reached.

4. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 3 is characterized in that: The process of filling pixels in the texture detection area includes: After the pixel removal operation in the texture detection area is completed, a number of 2*2 pixel areas are selected from the texture detection area, and smoothing pixels are set between adjacent pixels in the pixel area; A pixel difference threshold is set, and the pixel difference between adjacent pixels in each texture detection area is obtained, and then the pixel difference between adjacent pixels is compared with the pixel difference threshold, and a number of appearance textures are divided in each texture detection area according to the comparison result.

5. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 4 is characterized in that: The process of acquiring the characteristic texture distribution curve includes: A two-dimensional coordinate system is established. According to the corresponding position of each texture detection area in the historical processed image data, the historical processed image data after the filling and smoothing pixel operation are overlapped and mapped in the two-dimensional coordinate system. According to the overlapping results of each texture detection area, each appearance texture is spliced ​​to obtain the characteristic texture distribution curve of the corresponding historical processed image data.

6. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 5 is characterized in that: The process of splitting the historical product processing dynamic structure model into the historical equipment dynamic change model and the historical product dynamic model includes: A product-equipment interaction model is established based on each historical processing image data, and overlapping and splicing are performed along the characteristic texture distribution curve to obtain a historical product processing dynamic appearance model between corresponding processing equipment and types of products, and a historical product processing dynamic structure model is generated based on historical laser video data; Establish a three-dimensional coordinate system, split the historical product processing dynamic structure model and the historical product processing dynamic appearance model into several static structure models and static appearance models in chronological order, and then divide the static structure model into multiple static product component models and static equipment component models according to the characteristic texture distribution curve distribution on the static appearance model at the same time point; According to the positions of each static product component model and static equipment component model at each time point, the positions are distributed on the historical product processing dynamic structure model, and then the historical equipment dynamic change model and the historical product dynamic model are divided on the historical product processing dynamic structure model, and bound to the corresponding equipment node in the corresponding product-equipment process tree.

7. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 6 is characterized in that: According to the execution time length of each processing step, several production supervision time points are set, and the model parts where the historical product processing dynamic structure model, the historical equipment dynamic change model and the historical product dynamic model interact at each production supervision time point are obtained and recorded as the product processing area.

8. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 7 is characterized in that: The process of obtaining the fluctuation range of the standard machining trajectory includes: A plurality of interactive identification points are respectively set at edge positions of model parts corresponding to the historical equipment dynamic change model and the historical product dynamic model in the product processing area; Before all production supervision time points have passed, whenever there is an interaction between a pair of production supervision time points and the interaction identification points on the historical equipment dynamic change model and the historical product dynamic model, the overlapping interaction identification points are recorded as trajectory identification points, and the corresponding historical processing interaction trajectory is generated according to the spatial position changes of the trajectory identification points between the corresponding production supervision time points; The historical processing interaction trajectories are mapped to the same three-dimensional coordinate system, and then the m historical processing interaction trajectories are split into several coordinate points. The coordinate points at the same time are normally distributed, and the fluctuation range of the standard processing trajectory at the corresponding position is obtained according to the normal distribution results.

9. The intelligent manufacturing three-dimensional visualization system based on digital twin according to claim 8, characterized in that: The process of determining whether there is production anomaly in the processing equipment includes: During the execution of the product production plan, every time the product passes through a processing device, real-time laser video data and real-time processing video data are collected, and a corresponding real-time processing visualization model is established based on the real-time laser video data and real-time processing video data; Whenever a production supervision time point ends, the real-time processing visualization model generated between the previous production supervision time point and the corresponding historical equipment dynamic change model will be overlapped and mapped, and the process of establishing the standard processing trajectory fluctuation range will be adopted to obtain the real-time processing interaction trajectory of the real-time processing visualization model between the corresponding production supervision time points; Determine whether the real-time processing interaction trajectory is within the standard processing trajectory fluctuation range, generate and execute product repair decisions based on the judgment results.

Images