An intelligent partition slicing method applied to 3D printing and a 3D printing method

Abstract

The application relates to an intelligent partition slicing method applied to 3D printing and a 3D printing method, and relates to the technical field of 3D printing. The application aims to improve the problems that the printing precision and the printing efficiency cannot be considered simultaneously and that the problem of different slicing thickness partition continuous integrated forming printing cannot be solved in the prior art. Technical points: the intelligent partition slicing method comprises the following steps: S100, model acquisition and feature extraction: acquiring a three-dimensional model to be printed, and extracting a geometric complexity feature of the three-dimensional model; S200, artificial intelligence partition: dividing the three-dimensional model into multiple partitions with geometric consistency, and each partition corresponds to a complexity score; S300, printing sequence recombination: according to the complexity scores of the partitions, different slice numbers are allocated to each partition; S400, automatic control and slicing configuration: through an automatic script, a 3D printing software is controlled, corresponding partition model data is loaded and slicing parameters are configured according to a printing batch and the slice numbers of the partitions, and printing instructions of the partitions are generated.

Description

Technical Field

[0001] This invention relates to the field of 3D printing technology, specifically to an intelligent partitioning and slicing method and a 3D printing method applied to 3D printing. Background Technology

[0002] The core step in 3D printing is to discretize a three-dimensional digital model into a series of two-dimensional layers using slicing software, and then generate a processing path that the printer can execute. The slicing parameters, especially the layer thickness, directly determine the printing efficiency and forming accuracy of the final model.

[0003] Currently, mainstream 3D printing slicing methods generally use globally fixed slice layer thicknesses. This approach presents a significant technical contradiction in practical applications: when the model to be printed contains complex geometry, to ensure printing quality and avoid the "step effect," a smaller layer thickness is needed, i.e., increasing the number of slice layers, but this significantly increases printing time; conversely, if a larger layer thickness is used to improve printing efficiency, it will lead to severe loss of complex details in the model, resulting in a step effect and severely reducing the quality of the finished product.

[0004] To address the aforementioned contradictions, existing technologies have proposed a partitioned slicing method, which divides the model into different regions based on its geometric complexity and applies differentiated slice thicknesses to different regions. However, existing partitioned slicing schemes first print each region independently, and then assemble the pieces into a complete model through methods such as gluing. For example, the user manually divides the model, prints the slices, and then uses adhesives for post-processing assembly. This method not only prolongs the production cycle due to the gluing operation, but also results in a significantly lower connection strength and overall integrity compared to one-piece printing, leading to a decrease in the reliability of the final product.

[0005] Among the existing related patent technologies, the invention patent with publication number CN111739147A discloses a method for continuous layer slicing of a three-dimensional data model, which can realize slice thickness adjustment and color texture output, but does not involve model space partitioning and intelligent complexity determination; the invention patent with publication number CN116638767A discloses an intelligent 3D printing path planning method, which optimizes the single-layer printing path through concave polygon decomposition and tabu search algorithm, but does not realize three-dimensional model space partitioning and adaptive slicing; the invention patent with publication number CN117601434A discloses a biological three-dimensional model printing path planning method, which adjusts the layer height according to the model height and structural information, but does not use artificial intelligence for spatial complexity partitioning and differentiated continuous slicing.

[0006] In summary, the following problems urgently need to be addressed in the existing technology:

[0007] First, there is a lack of intelligent partitioning mechanisms based on the geometric complexity of the model. Existing slicing methods either use a globally fixed layer thickness or rely on manual model segmentation, failing to automatically identify fine structures and rough areas in the model, making it difficult to balance printing accuracy and efficiency.

[0008] Second, there is a lack of continuous printing solutions that support varying slice thicknesses. Existing partitioned slicing methods use a "block printing, post-assembly and bonding" approach, which cannot achieve integrated printing of partitions with different slice densities. This not only increases post-processing time but also compromises the structural integrity of the printed parts.

[0009] Third, there is a lack of automated control methods that can be deeply integrated with the existing slicing software ecosystem. Existing technologies mostly remain at the theoretical level of algorithms, lacking engineering solutions that can drive commercial slicing software and automate the entire process from partitioning and parameter configuration to printing control. Summary of the Invention

[0010] The purpose of this invention is to provide an intelligent partitioning and slicing method and a 3D printing method for 3D printing, aiming to improve the problems in the prior art that cannot balance printing accuracy and printing efficiency, and cannot achieve continuous integral molding printing of differentiated slice thickness partitions.

[0011] This invention is implemented as follows:

[0012] According to a first aspect of the present invention, the present invention provides an intelligent partitioning and slicing method for 3D printing, comprising the following steps:

[0013] S100, Model Acquisition and Feature Extraction: Acquire the 3D model to be printed, preprocess the 3D model, and extract the geometric complexity features of the 3D model;

[0014] S200, Artificial Intelligence Partitioning: The extracted geometric complexity features are input into a pre-trained deep learning network model. The deep learning network model determines the spatial complexity of the 3D model and divides the 3D model into multiple geometrically consistent partitions based on the determination results. Each partition corresponds to a complexity score.

[0015] S300, Printing Sequence Reorganization: Based on the complexity score of each partition, a different number of slice layers are assigned to each partition; and the printing sequence of each partition is reorganized along the direction perpendicular to the printing platform to form a phased printing batch, wherein each partition in the same printing batch is within the same preset height tolerance range.

[0016] S400, Automated Control and Slicing Configuration: The 3D printing software is controlled by an automated script to load the corresponding partition model data and configure the slicing parameters in sequence according to the printing batch and the number of slicing layers of each partition, and to generate printing instructions for each partition.

[0017] Furthermore, the geometric complexity features include one or more of curvature estimation, vertex density, and normal vector variation.

[0018] Furthermore, in step S300, the reordering of printing specifically includes:

[0019] S310. Along the Z-axis, use the partition with the smallest current height as the base partition for the current printing batch.

[0020] S320. For other partitions, calculate the height of the currently printed portion.

[0021] S330. When the difference between the height of other partitions and the height of the reference partition is within the preset height tolerance range, the other partitions are included in the current printing batch.

[0022] S340. When the height of other partitions exceeds the height of the reference partition and the excess is greater than the preset height tolerance range, the excess part is divided so that the divided part is at the same height as the reference partition and is included in the current printing batch, and the remaining part is reserved for subsequent batches.

[0023] Furthermore, the automation script is based on a graphical user interface automation library to achieve automated control of the 3D printing software.

[0024] Furthermore, in step S400, the specific method of controlling the 3D printing software through an automated script includes: fixing the window size of the 3D printing software, obtaining the resolution of the current device, and proportionally converting the coordinates of the preset interactive elements according to the ratio between the resolution of the current device and the preset standard resolution, so as to accurately locate and operate the interactive elements, and then sequentially loading the corresponding partition model data and configuring the slicing parameters.

[0025] Furthermore, the 3D printing software controlled by automated scripts also includes:

[0026] Position the mouse cursor over the slice count input box and enter the preset slice value;

[0027] Move the mouse to the print button and perform a simulated click to start the print job for the current partition;

[0028] After the current partition is printed, move the mouse to the stop printing button and perform a simulated click to end the current printing task.

[0029] According to a second aspect of the present invention, the present invention provides a 3D printing method, including the steps of the above-described intelligent partitioning and slicing method, and further including the following steps:

[0030] S500, Print Control: Based on the print instructions for each partition, execute the print task for each partition in sequence.

[0031] Furthermore, in step S500, after the current partition printing task is completed, the partition model data and corresponding slice parameters of the next partition are automatically retrieved, and the printing task is repeated until all partitions are printed.

[0032] Furthermore, the S500 and printing control specifically include:

[0033] S510, Trigger Print: Triggers a print task for the current partition;

[0034] S520 Status Monitoring: Monitors the completion status of the current partition print task;

[0035] S530, End Printing: End the printing task for the current partition;

[0036] S540, Safety Lift: Before the print head moves to the next partition start point, control the print head to lift to a preset safety Z-axis height, which is a preset value higher than the highest point of the currently printed model.

[0037] Furthermore, step S510 includes: using an automated script to move the mouse to the print button and perform a simulated click operation to start the print job;

[0038] Step S520 includes: recognizing the print completion pop-up window through image recognition function, or monitoring by setting an estimated waiting time;

[0039] Step S530 includes: using an automated script to move the mouse to the stop printing button and perform a simulated click operation to end the current printing task.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention extracts the geometric complexity features of the three-dimensional model, uses a deep learning network to determine the spatial complexity and divide it into multiple partitions, assigns different number of slice layers according to the complexity score of each partition, and reorganizes the printing order to form a phased printing batch. While ensuring the printing accuracy of complex areas, it significantly improves the overall printing efficiency and realizes integrated continuous printing of partitions with different slice densities. Attached Figure Description

[0041] Figure 1This is a flowchart of the intelligent partitioning and slicing method for 3D printing provided by the present invention;

[0042] Figure 2 This is a detailed flowchart of step S300 in the intelligent partitioning and slicing method for 3D printing provided by the present invention;

[0043] Figure 3 This is a flowchart of the 3D printing method provided by the present invention;

[0044] Figure 4 This is a flowchart of step S500 in the 3D printing method provided by the present invention. Detailed Implementation

[0045] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0046] The following description, in conjunction with the accompanying drawings and specific embodiments, provides further details:

[0047] Example 1

[0048] This embodiment provides a smart partitioning and slicing method for 3D printing, such as... Figure 1 and Figure 2 As shown, the specific steps are as follows:

[0049] S100, Model Acquisition and Feature Extraction

[0050] First, the 3D model to be printed is obtained. In this embodiment, the 3D model uses an STL (Standard Triangle Language) format file, which describes the surface geometry of a 3D object through a series of triangular facets. The STL file is imported into the system of this invention, and the system parses the file, reading the vertex coordinates and normal vector data of all triangular facets.

[0051] After completing the data reading, extract the feature information used to characterize the geometric complexity of the model, specifically including the following three features:

[0052] (1) Curvature estimation: For each vertex of the model surface, calculate the Gaussian curvature or average curvature of its local region. The larger the curvature value, the higher the degree of curvature of the region and the more complex the geometry. For example, the curvature of a sphere surface is uniform and small, while the curvature of the face region of a finely sculpted statue varies drastically and the local curvature value is large.

[0053] (2) Vertex density: The number of vertices per unit volume. In STL models, regions with complex geometries usually require more triangular facets to accurately represent them, and therefore the vertex density is correspondingly higher. For example, the vertex density of finely textured regions in the model is significantly higher than that of flat surface regions.

[0054] (3) Normal vector change: Calculate the angle between the normal vectors of adjacent triangular facets. Areas with drastic changes in the angle usually correspond to abrupt changes in the edges, sharp corners or surface orientation of the model. These areas are also complex parts with rich details.

[0055] S200, Artificial Intelligence Partition

[0056] The extracted curvature estimates, vertex density values, and normal vector change angles are fused to form a feature vector for each local region. This feature vector is then input into a pre-trained deep learning network model.

[0057] In this embodiment, the deep learning network employs a 3D point cloud segmentation network based on the PointNet++ architecture. This network has been trained on a large dataset of manually annotated 3D models, with each model in the training data annotated with regions of varying complexity. Through learning, the network is able to establish a mapping relationship between geometric features and spatial complexity.

[0058] The deep learning network performs a region-by-region analysis of the 3D model, outputting a spatial complexity heatmap. This heatmap identifies the complexity score of each region of the model (e.g., a score range of 0-100, with higher scores indicating higher complexity). Then, based on preset partitioning thresholds (e.g., 0-30 points for simple regions, 31-70 points for medium regions, and 71-100 points for complex regions), the system divides the 3D model into multiple geometrically consistent partitions. Each partition is assigned a unique identifier (e.g., Part_001, Part_002) and its geometric boundary information is recorded (including the vertex coordinates of the minimum bounding box, the indices of all triangles within the partition, etc.), along with the corresponding complexity score.

[0059] For example, for a vase model with intricate carvings, a deep learning network would divide the smooth curved surface at the neck of the bottle into simple partitions, the intricate patterned area in the middle of the bottle into complex partitions, and the basic support structure at the bottom of the bottle into simple partitions.

[0060] S300, Print Sequence Reorganization

[0061] Based on the complexity score of each partition, a different number of slice layers are assigned to each partition. Specifically, in this embodiment, the base slice layer thickness is preset to 0.3mm, the slice layer thickness for complex partitions is 0.1mm (i.e., the number of layers is increased by 3 times), and the slice layer thickness for simple partitions is 0.3mm. For example, a complex partition with a height of 30mm will be assigned 300 slice layers, while a simple partition of the same height will only be assigned 100 slice layers.

[0062] After the slice layer allocation is completed, the height increments of each partition in the Z-axis direction are no longer synchronized (complex partitions increase by only 0.1mm per layer, and simple partitions increase by 0.3mm per layer), making it impossible to use the traditional layer-by-layer printing sequence. Therefore, the printing sequence needs to be reorganized to avoid collisions between the print head and the already formed parts when moving across partitions.

[0063] The specific recombination steps in this embodiment are as follows:

[0064] S310. Scan all partitions from bottom to top along the Z-axis, and select the partition with the smallest height among the currently printed portions as the base partition for the current printing batch. Initially, all partitions have a height of 0, so the partition with the smallest lowest point along the Z-axis is selected as the first base partition.

[0065] S320. For other partitions that have not yet finished printing, calculate the height of the currently printed portion (initially 0).

[0066] S330. In this embodiment, the preset height tolerance range is ±0.05mm. When the difference between the height of other partitions and the height of the reference partition is within this tolerance range, the other partitions are included in the current printing batch. For example, if the current height of the reference partition is 2.0mm and the current height of another partition is 2.03mm, the difference between the two is 0.03mm, which is less than 0.05mm, then the entire partition is included in the current batch for synchronous printing.

[0067] S340. When the height of other partitions exceeds the height of the reference partition, and the excess is greater than the preset height tolerance range (i.e., greater than 0.05mm), the excess portion of the partition is virtually segmented. Specifically, the system cuts the partition model at a position where its height equals the current height of the reference partition, making the lower part of the segment the same height as the reference partition, and includes this lower part in the current printing batch; the remaining upper part of the partition is reserved for processing in subsequent batches, and the height of its currently printed portion is updated to the height of the reference partition.

[0068] For example, if the current height of the baseline partition is 2.0mm and the current height of another partition is 2.5mm, the difference between the two is 0.5mm, which is greater than 0.05mm. The system will virtually split the other partition at a height of 2.0mm, allocating the 0-2.0mm portion to the current batch and reserving the 2.0-2.5mm portion for subsequent batches.

[0069] After completing the division of the current batch, the system outputs the print data for the current batch, including: the model data of each partition in the batch (original STL file or cut STL file), the number of slice layers corresponding to each partition, and the coordinates of the printing start point of each partition in the batch. Then, the current height of each partition is updated, and steps S310 to S340 above are repeated until the entire height of all partitions has been printed.

[0070] S400, Automation Control and Slice Configuration

[0071] In this embodiment, the automation script is written in Python and uses the PyAutoGUI graphical user interface automation library to control the UP Studio 3D printing software.

[0072] First, launch the UP Studio software. Use a script to call the Windows API or PyAutoGUI's window control functions to maximize and bring the software window to the top. Once the window size is fixed, the relative positions of all interactive elements (such as the print button, stop button, and slice count input box) on the screen are also fixed, avoiding coordinate drift issues caused by changes in window size.

[0073] During the development phase, the developers recorded the coordinates of each interactive element on the screen using a computer with a standard resolution (1920×1080 in this embodiment). For example, the preset coordinates of the slice count input box are (800, 500), the preset coordinates of the print button are (900, 600), and the preset coordinates of the stop button are (900, 650).

[0074] When the automation script runs on the target computer, it first calls the `pyautogui.size()` function to obtain the current device's screen resolution (e.g., an actual resolution of 2560×1440). Then, it calculates the ratio of the actual resolution to the standard resolution: width ratio = 2560 / 1920 = 1.333, height ratio = 1440 / 1080 = 1.333. The preset coordinates of interactive elements are then proportionally converted according to this ratio: actual X coordinate = preset X coordinate × width ratio, actual Y coordinate = preset Y coordinate × height ratio. For example, the actual coordinates of the slice count input box are (800 × 1.333 = 1066, 500 × 1.333 = 666). This adaptive coordinate conversion ensures the script's accurate positioning capability on devices with different resolutions.

[0075] After the coordinate conversion is complete, the script reads the print configuration file generated in step S300. This file records detailed data for each print batch in sequence. The script then performs the following operations on each batch:

[0076] (1) Call the pyautogui.moveTo() function to move the mouse to the actual coordinates of the slice count input box;

[0077] (2) Call the pyautogui.write() function to input the slice value corresponding to the partition (for example, input 300 for complex partitions and 100 for simple partitions).

[0078] (3) Call the pyautogui.moveTo() function to move the mouse to the actual coordinates of the print button;

[0079] (4) Call the pyautogui.click() function to perform a simulated click operation and start the printing task of the current batch;

[0080] (5) Wait for the current batch to finish printing (monitor or estimate the time through image recognition);

[0081] (6) Call the pyautogui.moveTo() function to move the mouse to the actual coordinates of the stop printing button;

[0082] (7) Call the pyautogui.click() function to perform a simulated click operation and end the current printing task;

[0083] (8) Retrieve the data for the next printing batch and repeat steps (1) to (7) above until all batches are processed.

[0084] Through the above steps, step S400 finally generates printing instructions for each partition, which are then sent to the 3D printer for execution.

[0085] Example 2

[0086] This embodiment provides a 3D printing method, such as... Figure 3 and Figure 4 As shown, the steps of the intelligent partitioning and slicing method for 3D printing provided in Example 1 and the printing control steps (S500) are included to achieve full automation from slicing configuration to physical printing.

[0087] I. Implementing the intelligent partitioning and slicing method

[0088] First, following steps S100 to S400 in Example 1, the intelligent partitioning of the 3D model to be printed, the reorganization of the printing order, the configuration of slicing parameters, and the generation of printing instructions are completed. At this point, the system has obtained complete data for multiple printing batches arranged in sequence, each batch including: partitioned model file, number of slice layers, coordinates of printing start point, etc.

[0089] II. Printing Control (Step S500)

[0090] S510, trigger printing:

[0091] Based on the coordinate adaptive function implemented in step S400 of Example 1, the script accurately locates the print button. Specifically, the script calls the pyautogui.moveTo() function to move the mouse to the actual converted coordinates of the print button, and calls the pyautogui.click() function to perform a simulated click operation, starting the print task for the current partition.

[0092] For example, when printing the first partition, the script first configures the slicing parameters of the partition (layer thickness 0.1mm, 300 layers in total), and then simulates clicking the print button, and the printer begins to print the partition layer by layer.

[0093] S520, Condition Monitoring:

[0094] After the current partition print job starts, the script needs to monitor the print completion status. This embodiment provides two monitoring methods, which can be selected according to the actual situation:

[0095] Method 1 (Image Recognition Monitoring): The script calls image recognition libraries such as OpenCV to perform real-time screenshot analysis of a specific area of ​​the UP Studio software interface. When printing is complete, the software usually pops up a "Printing Complete" or similar notification window. The script detects the appearance of this pop-up window through template matching or OCR text recognition technology, thus determining that the current printing task has been completed.

[0096] Method 2 (Estimated Printing Time Monitoring): The script calculates the total estimated printing time for the current partition based on the total number of layers and the estimated printing time for each layer (e.g., the printing time for each layer can be estimated using slicing software or obtained from historical data statistics). The script calls the `time.sleep()` function or loops to wait for this estimated time before determining that the printing task is complete. To ensure reliability, a 10% redundancy time can be added to the estimated time.

[0097] In this embodiment, it is preferable to use two monitoring methods simultaneously: using estimated time as the primary criterion and image recognition as an auxiliary verification. Once the estimated time is reached, the script confirms the printing completion by displaying a pop-up window via image recognition before proceeding to the next step, thus preventing time errors caused by printing anomalies.

[0098] S530, End printing:

[0099] After confirming that the current partition has finished printing, the script calls the `pyautogui.moveTo()` function to move the mouse to the actual calculated coordinates of the "Stop Printing" button, and then calls the `pyautogui.click()` function to perform a simulated click operation to end the current printing task. Simultaneously, the script calls `pyautogui.click()` to simulate clicking the "Remove Printout" or "Clean Platform" button (if necessary) to prepare platform space for printing the next partition.

[0100] S540, Safety Lift:

[0101] Before the printing task in the current partition ends and the print head needs to move to the starting point of the next partition, the script sends G-code instructions to the 3D printer to control the print head to rise along the Z-axis to a preset safe height. In this embodiment, the safe height is set to "5mm higher than the highest point of the currently printed model".

[0102] For example, if the highest point of the currently printed model (including previously completed sections) is 15.0 mm, the print head will rise to a height of 20.0 mm before moving horizontally. This mechanism effectively prevents the print head from colliding with the already formed parts when moving across sections, ensuring the safety of the printing process.

[0103] After the safe lifting is completed, the script controls the print head to move to the starting point coordinates of the next partition (which has been recorded in the S300 reassembly step).

[0104] III. Execute repeatedly until completion.

[0105] The loop control in step S500: After the current partition printing task is completed, the system automatically retrieves the partition model data and corresponding slice parameters of the next partition from the print configuration file. The script determines whether there are any unprocessed partitions. If so, it repeats steps S510 to S540, namely: configure the slice parameters of the next partition → trigger printing → monitor the completion status → end printing → safe lift → move to the starting point of the next partition.

[0106] For example, for a vase model containing three partitions, the execution order is:

[0107] (1) Print partition 1 (simple partition at the bottom of the bottle, 0.3mm layer thickness) → Raise to a safe height → Move to the starting point of partition 2;

[0108] (2) Load partition 2 (complex partition of bottle body, 0.1mm layer thickness) → Configure slicing parameters → Trigger printing → Monitor completion → End printing → Raise safety height → Move to the starting point of partition 3;

[0109] (3) Load partition 3 (simple partition at the bottle mouth, 0.3mm layer thickness) → Configure slicing parameters → Trigger printing → Monitor completion → End printing.

[0110] Once all partitions have been printed, the script will automatically terminate and send a printing completion notification to the user.

[0111] IV. Fault Tolerance

[0112] This embodiment also incorporates a fault-tolerance mechanism into the automated script. The script monitors the software interface in real time using image recognition. When an abnormal pop-up window is detected (such as "insufficient consumables," "printing failure," "abnormal temperature," etc.), the script performs corresponding operations according to preset abnormality handling rules: for minor abnormalities that can be automatically recovered (such as confirmation prompts), the script automatically simulates clicking the "OK" button; for serious abnormalities (such as printing failure), the script pauses the printing task and sends an alarm notification to the user via email or message, while recording the current printing progress so that the user can resume printing after resolving the problem.

[0113] Compared with the prior art, the present invention has the following technical effects:

[0114] First, printing efficiency is improved: a larger slice layer thickness is used for areas with simple geometry, and a smaller slice layer thickness is used for complex areas. While ensuring print quality, the overall printing time can be reduced by about 30%.

[0115] Secondly, it can be printed in one piece: through the printing sequence reorganization algorithm and the safety lifting mechanism, it realizes continuous one-piece printing of partitions with different slice densities, avoiding the cumbersome steps of bonding after printing in sections, and ensuring the structural strength of the printed parts.

[0116] Third, it has a high degree of intelligence: it uses deep learning networks to automatically identify the geometric complexity features of the model and partition them, reducing human intervention and reliance on personal experience.

[0117] Fourth, it has a high degree of automation: through Python scripts and the PyAutoGUI library, it achieves fully automated control of commercial slicing software, including window management, coordinate adaptation, parameter configuration, print start and stop, etc., forming a closed-loop automated printing process.

[0118] In summary, this invention extracts the geometric complexity features of a 3D model, uses a deep learning network to determine spatial complexity and divide it into multiple partitions, assigns different number of slice layers based on the complexity score of each partition, and reassembles the printing sequence to form a phased printing batch. This solves the problem in existing technologies where it is difficult to balance printing accuracy and printing efficiency, and it is impossible to achieve continuous integrated printing of partitions with different slice thicknesses. While ensuring printing accuracy in complex areas, it significantly improves overall printing efficiency and achieves integrated continuous printing of partitions with different slice densities.

[0119] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A smart partitioning and slicing method for 3D printing, characterized in that, Includes the following steps: S100, Model Acquisition and Feature Extraction: Acquire the 3D model to be printed, preprocess the 3D model, and extract the geometric complexity features of the 3D model; S200, Artificial Intelligence Partitioning: The extracted geometric complexity features are input into a pre-trained deep learning network model. The deep learning network model determines the spatial complexity of the 3D model and divides the 3D model into multiple geometrically consistent partitions based on the determination results. Each partition corresponds to a complexity score. S300, Printing Sequence Reorganization: Based on the complexity score of each partition, a different number of slice layers are assigned to each partition; and the printing sequence of each partition is reorganized along the direction perpendicular to the printing platform to form a phased printing batch, wherein each partition in the same printing batch is within the same preset height tolerance range. S400, Automated Control and Slicing Configuration: The 3D printing software is controlled by an automated script to load the corresponding partition model data and configure the slicing parameters in sequence according to the printing batch and the number of slicing layers of each partition, and to generate printing instructions for each partition.

2. The intelligent partitioning and slicing method for 3D printing according to claim 1, characterized in that, The geometric complexity features include one or more of curvature estimation, vertex density, and normal vector variation.

3. The intelligent partitioning and slicing method for 3D printing according to claim 1, characterized in that, In step S300, the reordering of printing specifically includes: S310. Along the Z-axis, use the partition with the smallest current height as the base partition for the current printing batch. S320. For other partitions, calculate the height of the currently printed portion. S330. When the difference between the height of other partitions and the height of the reference partition is within the preset height tolerance range, the other partitions are included in the current printing batch. S340. When the height of other partitions exceeds the height of the reference partition and the excess is greater than the preset height tolerance range, the excess part is divided so that the divided part is at the same height as the reference partition and is included in the current printing batch, and the remaining part is reserved for subsequent batches.

4. The intelligent partitioning and slicing method for 3D printing according to claim 1, characterized in that, The automation scripts are based on a graphical user interface automation library to automate the control of 3D printing software.

5. The intelligent partitioning and slicing method for 3D printing according to claim 1, characterized in that, In step S400, the specific method of controlling the 3D printing software through an automated script includes: fixing the window size of the 3D printing software, obtaining the resolution of the current device, and proportionally converting the coordinates of the preset interactive elements according to the ratio between the resolution of the current device and the preset standard resolution, so as to accurately locate and operate the interactive elements, and then sequentially loading the corresponding partition model data and configuring the slicing parameters.

6. The intelligent partitioning and slicing method for 3D printing according to claim 5, characterized in that, The 3D printing software controlled by automated scripts also includes: Position the mouse cursor over the slice count input box and enter the preset slice value; Move the mouse to the print button and perform a simulated click to start the print job for the current partition; After the current partition is printed, move the mouse to the stop printing button and perform a simulated click to end the current printing task.

7. A 3D printing method, characterized in that, The intelligent partitioning and slicing method according to any one of claims 1 to 6 further includes the following steps: S500, Print Control: Based on the print instructions for each partition, execute the print task for each partition in sequence.

8. The 3D printing method according to claim 7, characterized in that, In step S500, after the current partition printing task is completed, the partition model data and corresponding slice parameters of the next partition are automatically retrieved, and the printing task is repeated until all partitions are printed.

9. The 3D printing method according to claim 7, characterized in that, The S500 and printing control specifically include: S510, Trigger Print: Triggers a print task for the current partition; S520 Status Monitoring: Monitors the completion status of the current partition print task; S530, End Printing: End the printing task for the current partition; S540, Safety Lift: Before the print head moves to the next partition start point, control the print head to lift to a preset safety Z-axis height, which is a preset value higher than the highest point of the currently printed model.

10. The 3D printing method according to claim 9, characterized in that, Step S510 includes: using an automated script to move the mouse to the print button and perform a simulated click operation to start the print job; Step S520 includes: recognizing the print completion pop-up window through image recognition function, or monitoring by setting an estimated waiting time; Step S530 includes: using an automated script to move the mouse to the stop printing button and perform a simulated click operation to end the current printing task.

Images