FDM printing control method, control device and control system

Abstract

The present invention discloses an FDM printing control method, control device, and control system. The method includes: controlling a printing mechanism to perform a current layer printing operation based on current device parameters to obtain a current layer image; analyzing and processing the current layer image to obtain a current layer point cloud model corresponding to the current layer image; performing an error assessment based on the current layer image and the current layer point cloud model to obtain an error assessment result; if the error assessment result meets a preset error condition, obtaining target device parameters, updating the target device parameters to the current device parameters, and repeatedly performing the current layer printing operation based on the current device parameters to obtain a current layer image; if the error assessment result does not meet the preset error condition, stopping the printing operation. This FDM printing control method ensures that the printed workpiece has high precision and achieves the purpose of device parameters reflecting the workpiece printing status in real time.

Description

Technical Field

[0001] The present invention relates to the field of FDM printing control, and in particular to an FDM printing control method, control equipment and control system. Background Art

[0002] During the additive manufacturing process, deviations in workpiece shape and geometry, including even processing failure, can occur due to issues such as misaligned print platforms, nozzle blockages, motor slippage, consumable exhaustion, and vibration or impact on the equipment. Existing full-process monitoring systems primarily monitor various signals and parameters during the additive manufacturing process in real time through devices such as piezoelectric sensors and signal sensors. However, these device parameters cannot accurately reflect the workpiece printing status and topographical accuracy in real time, leading to problems such as low workpiece printing accuracy. While computer vision-based monitoring methods improve workpiece printing accuracy, they require high-precision cameras and depth cameras to acquire data, resulting in expensive equipment and high production costs. Summary of the Invention

[0003] The embodiments of the present invention provide an FDM printing control method, a control device and a control system to solve the problem of low workpiece printing precision.

[0004] An FDM printing control method, comprising:

[0005] Control the printing mechanism to perform the current layer printing operation based on the current device parameters and obtain the current layer image;

[0006] Analyzing and processing the current layer image to obtain a current layer point cloud model corresponding to the current layer image;

[0007] Perform error evaluation based on the current layer image and the current layer point cloud model to obtain an error evaluation result;

[0008] If the error evaluation result satisfies the preset error condition, obtaining target device parameters, updating the target device parameters to current device parameters, and repeatedly controlling the printing mechanism to perform the current layer printing operation based on the current device parameters to obtain the current layer image;

[0009] If the error evaluation result does not meet the preset error condition, the printing operation is stopped.

[0010] Preferably, the current layer image includes at least two current layer top views shot at different focal lengths; and analyzing and processing the current layer image to obtain the current layer point cloud model corresponding to the current layer image includes:

[0011] Performing clarity analysis on at least two top views of the current layer to obtain a target top view with the highest clarity;

[0012] Perform point cloud conversion processing on the target top view to obtain a current layer point cloud model corresponding to the target top view.

[0013] Preferably, performing clarity analysis on at least two top views of the current layer to obtain a target top view with optimal clarity includes:

[0014] Using an energy gradient function, respectively performing energy gradient calculations on at least two top views of the current layer, and determining current gradient values ​​corresponding to the at least two top views of the current layer;

[0015] The top view of the current layer with the largest current gradient value is determined as the target top view with the highest clarity.

[0016] Preferably, performing point cloud conversion processing on the target top view to obtain a current layer point cloud model corresponding to the current layer image includes:

[0017] Performing fitting and segmentation on the target top view to obtain a workpiece plane image;

[0018] Noise points are filtered out of the workpiece plane image to obtain a current layer point cloud model corresponding to the workpiece plane image.

[0019] Preferably, the current layer image includes at least two current layer main views taken at different times; performing error assessment based on the current layer image and the current layer point cloud model to obtain an error assessment result includes:

[0020] Performing error analysis on at least two of the current layer main views to obtain a measured view error;

[0021] Performing point cloud registration on the current layer point cloud model to obtain a point cloud measured error;

[0022] The error evaluation result is obtained based on the view measured error and the point cloud measured error.

[0023] Preferably, performing point cloud registration on the current layer point cloud model to obtain the point cloud measured error includes:

[0024] Registering the current layer point cloud model with the current standard point cloud model to obtain a first registration error;

[0025] The overall point cloud model is registered with the overall standard point cloud model to obtain a second registration error, wherein the overall point cloud model is a model obtained by superimposing the current layer point cloud model and the historical layer point cloud model.

[0026] Preferably, obtaining an error evaluation result based on the view measured error and the point cloud measured error includes:

[0027] comparing the view measured error with the first view error, and comparing the point cloud measured error with the first point cloud error;

[0028] If the view measured error is smaller than the first view error, and the point cloud measured error is smaller than the first point cloud error, obtaining an error evaluation result that meets the preset error condition;

[0029] If the view measured error is not less than the first view error, or the point cloud measured error is not less than the first point cloud error, an error evaluation result that does not meet the preset error condition is obtained.

[0030] Preferably, the acquiring target device parameters includes:

[0031] comparing the view measured error with the second view error, and comparing the point cloud measured error with the second point cloud error;

[0032] If the view measured error is smaller than the second view error, and the point cloud measured error is smaller than the second point cloud error, determining the current device parameters as the target device parameters;

[0033] If the view measured error is not less than the second view error, or the point cloud measured error is not less than the second point cloud error, optimization processing is performed based on the current device parameters, the view measured error and the point cloud measured error to obtain the target device parameters.

[0034] Preferably, the optimization process is performed based on the current device parameters, the view measurement error, and the point cloud measurement error to obtain the target device parameters, including:

[0035] Inputting the current device parameters, the view measured error, and the point cloud measured error into a learning network to determine the optimized device parameters;

[0036] Using a learning network to optimize the parameters of the optimization device, and outputting a view prediction error and a point cloud prediction error, wherein the point cloud prediction error includes a first registration prediction error and a second registration prediction error;

[0037] If the view prediction error is smaller than the second view error, and the point cloud prediction error is smaller than the second point cloud error, the optimized device parameters are determined as target device parameters.

[0038] A control device comprises a memory, a processor, and a control program stored in the memory and executable on the processor, wherein the processor implements any one of the above-mentioned FDM printing control methods when executing the control program.

[0039] An FDM printing control system, comprising the aforementioned control device, printing platform, printing mechanism and camera device;

[0040] The printing mechanism is arranged on the printing platform and is used for performing a printing operation of the current layer image;

[0041] The camera device is arranged on the printing platform and is used to obtain the current layer image;

[0042] The control device is connected to the printing mechanism and the camera device, and is used to control the printing mechanism to perform a current layer printing operation according to current device parameters.

[0043] Preferably, the camera device includes a first camera and a second camera;

[0044] The first camera is arranged on the printing platform in a direction perpendicular to the printing platform, and is used to obtain at least two top views of the current layer with different shooting focal lengths;

[0045] The second camera is arranged on the printing platform in a direction parallel to the printing platform, and is used to obtain at least two current layer main views with different shooting times.

[0046] In the aforementioned FDM printing control method, control device, and control system, the printing mechanism performs printing operations using current device parameters to obtain an image of the current layer. This image is then converted to a point cloud to obtain a point cloud model of the current layer. The control device then performs an error assessment based on the current layer image and the current layer point cloud model, obtaining an error assessment result. Based on a comparison of the error assessment result with a preset error condition, the control device determines whether to repeat the printing operation for the current layer. By performing error assessment on the image-to-point cloud model conversion, printing accuracy is ensured, avoiding the cost pressures associated with expensive equipment. Furthermore, by comparing the error assessment results for each current layer image, the printing status reflected by the current device parameters can be observed in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 is a flow chart of an FDM printing control method according to an embodiment of the present invention;

[0049] Figure 2 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0050] Figure 3 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0051] Figure 4 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0052] Figure 5 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0053] Figure 6 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0054] Figure 7 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0055] Figure 8 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0056] Figure 9 is another flow chart of the FDM printing control method according to one embodiment of the present invention;

[0057] Figure 10 FIG. 1 is a schematic diagram of an apparatus for an FDM printing control method according to an embodiment of the present invention.

[0058] Among them, 1. printing mechanism; 2. printing platform; 3. camera equipment; 31. first camera; 32. second camera. DETAILED DESCRIPTION

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0060] An FDM printing control method provided by an embodiment of the present invention can be applied in an FDM printing control system. The FDM printing control system includes a control device that can execute the FDM printing control method. The control device can perform real-time monitoring during the printing control process to ensure the workpiece processing accuracy.

[0061] In one embodiment, if Figure 1 As shown, a FDM printing control method is provided, which is described by taking the application of the method in a control device as an example. The method includes the following steps:

[0062] S101: Controlling the printing mechanism 1 to perform a current layer printing operation based on current device parameters to obtain a current layer image;

[0063] S102: Analyze and process the current layer image to obtain the current layer point cloud model corresponding to the current layer image;

[0064] S103: performing error evaluation based on the current layer image and the current layer point cloud model to obtain an error evaluation result;

[0065] S104: If the error evaluation result meets the preset error condition, the target device parameters are obtained, the target device parameters are updated to the current device parameters, and the printing mechanism 1 is repeatedly controlled based on the current device parameters to perform the current layer printing operation to obtain the current layer image;

[0066] S105: If the error evaluation result is that the preset error condition is not met, the printing operation is stopped.

[0067] The current device parameters refer to the various parameter values ​​of the device at the current moment. The printing mechanism 1 refers to the mechanism for performing printing operations according to the current device parameters. The current layer image refers to the image printed by the printing mechanism 1 on the current layer at the current moment.

[0068] As an example, in step S101, the control device controls the printing mechanism 1 to print the current layer based on current device parameters, thereby acquiring an image of the current layer. Current device parameters include, but are not limited to, nozzle movement speed, nozzle temperature, retraction speed, fan speed, print platform 2 temperature, adhesion type, infill method, fill rate, overhang length, overhang angle, overhang distance, layer thickness, and wall thickness. Based on the acquired image of the current layer, the control device can accurately understand the differences between the current layer images printed using different device parameters, and can subsequently adjust the device parameters in real time to clearly reflect the workpiece topography.

[0069] The current layer point cloud model refers to the point cloud model obtained after performing a point cloud conversion operation on the current layer image.

[0070] As an example, in step S102, the control device performs point cloud analysis on the current layer image to determine the point cloud distribution state of the current layer image, thereby obtaining the current layer point cloud model after the current layer image is processed. The current layer point cloud model can reflect the shape of the workpiece more clearly and three-dimensionally, ensure the accuracy of the workpiece, and at the same time does not require expensive equipment to support visual reconstruction technology, thereby reducing production costs.

[0071] The error evaluation result refers to the result used to evaluate the error size of the printed workpiece.

[0072] As an example, in step S103, the control device performs an error assessment on the current layer image and the current layer point cloud model, determining the magnitude of the error between the current layer image and the current layer point cloud model, and thereby obtaining a corresponding error assessment result. The control device uses this error assessment result to ensure workpiece printing accuracy and avoid costly errors caused by excessive errors.

[0073] The preset error condition refers to a preset acceptable error threshold, which is used to evaluate whether the error evaluation result of the workpiece meets the error threshold for repeatedly executing the current layer printing operation.

[0074] The target device parameter refers to a target parameter value used to control the printing mechanism 1 to repeatedly perform the current layer printing operation.

[0075] As an example, in step S104, if the control device identifies that the error assessment result meets the preset error conditions, it determines the current device parameters of printing mechanism 1 as target device parameters. The control device then continues to use the target device parameters as the current device parameters and repeatedly controls printing mechanism 1 to print the current layer using these current device parameters, repeatedly acquiring an image of the current layer of the workpiece. By comparing the error assessment result with the preset error conditions and determining that the error assessment result meets the preset error conditions, the accuracy of workpiece printing can be guaranteed, preventing excessive errors that could affect workpiece topography and reducing cost.

[0076] As an example, in step S105, if the control device recognizes that the error evaluation result does not meet the preset error condition, it determines that the current device parameter error is too large, and the control device will judge that the workpiece printing operation has failed. Therefore, the control device will control the printing mechanism 1 to stop the printing operation to avoid the error being too large and affecting the printed workpiece morphology, thereby reducing cost waste.

[0077] In this example, before acquiring the current layer image, the control device controls and completes self-calibration. This self-calibration establishes correspondence between image points and establishes correspondence between images. This provides high flexibility, thereby improving printing efficiency and ensuring the accuracy of the current layer image data acquired by printing mechanism 1. After completing self-calibration, the printing platform 2 of printing mechanism 1 is determined, that is, the image of the printing area is segmented, resulting in a printing platform 2 containing multiple reference blocks. This ensures the accuracy of subsequent data acquisition and facilitates the collection of corresponding current layer image data based on changes in workpiece topography.

[0078] In this embodiment, the control device controls the printing mechanism 1 to perform the current layer printing operation based on the current device parameters, obtains the current layer image, and then performs point cloud analysis on the current layer image to obtain the current layer point cloud model, so that the workpiece morphology can be clearly reflected based on the point cloud model, and the workpiece accuracy can be accurately reflected. The current layer image and the current layer point cloud model are evaluated to obtain an error assessment result. The control device then compares the error assessment result with the preset error condition and determines whether the current layer printing result meets the printing requirements based on the comparison result. At the same time, it is also convenient to clearly observe the status of the workpiece printing and monitor in real time the impact of the current device parameters on the workpiece printing status and morphological accuracy. This ensures that the printing operation on the workpiece does not require the support of expensive equipment, thereby reducing production costs.

[0079] In one embodiment, if Figure 2 As shown, the current layer image includes at least two current layer top views shot at different focal lengths;

[0080] Step S102: Analyze and process the current layer image to obtain the current layer point cloud model corresponding to the current layer image, including:

[0081] S201: Perform clarity analysis on at least two top views of the current layer to obtain a target top view with the highest clarity;

[0082] S202: Performing point cloud conversion processing on the target top view to obtain a current layer point cloud model corresponding to the target top view.

[0083] The current layer top view refers to the top view printed by the printing mechanism 1 at the current moment. The target top view refers to the top view of the current layer with the highest definition printed by the printing mechanism 1 at the current moment.

[0084] As an example, in step S201, the current layer image acquired by the control device includes at least two top views of the current layer captured at different focal lengths. The top views of the current layer are captured by the control device above the printing plane, with the printing plane as the reference. The control device compares the top views of the current layer captured at different focal lengths and determines the top view with the highest clarity, thereby ensuring the accuracy of workpiece printing. The control device performs a clarity analysis on the at least two top views of the current layer captured at different focal lengths to ensure that the selected top view best reflects the workpiece's topography. After analysis, the target top view with the highest clarity is determined.

[0085] As an example, in step S202, after acquiring a target top view, the control device performs point cloud conversion on the target top view, improving acquisition accuracy and reducing processing time for the target top view. This allows the control device to obtain a point cloud model of the current layer corresponding to the target top view. Based on the acquired point cloud model, the control device can present the workpiece's topography more clearly and three-dimensionally.

[0086] In this embodiment, the control device obtains at least two top views of the current layer with different shooting focal lengths, from which the top view of the current layer with the highest clarity can be obtained, and it is determined as the target top view. The target top view is then subjected to point cloud conversion processing to improve the accuracy of the data. After the conversion, the current layer point cloud model is obtained, thereby achieving accurate reflection of the workpiece morphology and ensuring the accuracy of the printed workpiece.

[0087] In one embodiment, if Figure 3 As shown, step S201, performing clarity analysis on at least two current layer top views to obtain a target top view with optimal clarity, includes:

[0088] S301: Using an energy gradient function, respectively performing energy gradient calculations on at least two top views of the current layer, and determining current gradient values ​​corresponding to the at least two top views of the current layer;

[0089] S302: Determine the top view of the current layer with the largest current gradient value as the target top view with the highest clarity.

[0090] The current gradient value refers to the gradient value corresponding to each top view of the current layer at the current moment.

[0091] As an example, in step S301, the control device uses the energy gradient function to calculate the energy gradient of at least two current layer top views, and obtains the current gradient value corresponding to each current layer top view. Since the shooting focal length of each current layer top view is different, the corresponding current gradient value is also different. The larger the current gradient value, the clearer the corresponding current layer top view. The control device obtains the current gradient value to ensure that the selected current layer top view has the highest clarity, thereby ensuring the high precision of the printed workpiece. In this example, a pixel point in the current layer top view is (x, y), and the grayscale value corresponding to the pixel point is f(x, y). Then, the grayscale values ​​of all pixel points in the current layer top view are summed to obtain the current gradient value F(x, y) of the current layer top view. The formula for calculating the current gradient value using the energy gradient function can be:

[0092]

[0093] As an example, in step S302, after the control device obtains the current gradient values ​​of at least two current layer top views, it selects the current layer top view with the largest current gradient value as the target top view with the highest clarity. The larger the current gradient value of the current layer top view, the higher the clarity. The control device selects the current layer top view with the largest current gradient value to accurately reflect the workpiece topography.

[0094] In this embodiment, the control device uses an energy gradient function to calculate the energy gradient of at least two current layer top views. Since the current gradient values ​​of different current layer top views are inconsistent, after obtaining the current gradient value of each current layer top view, the current layer top view with the highest clarity is determined as the target top view. The larger the current gradient value, the clearer the current layer top view, and the more it can reflect the workpiece morphology and printing status. At the same time, by comparing the current gradient values, it can be ensured that the printing effect produced by the selected target top view is the best, thereby reducing errors.

[0095] In one embodiment, if Figure 4 As shown, step S202, performing point cloud conversion processing on the target top view to obtain the current layer point cloud model corresponding to the current layer image, includes:

[0096] S401: performing fitting and segmentation on the target top view to obtain a plane image of the workpiece;

[0097] S402: Filter out noise points on the workpiece plane image to obtain a current layer point cloud model corresponding to the workpiece plane image.

[0098] The workpiece plane image refers to the surface image of the workpiece in the target top view.

[0099] As an example, in step S401, the control device performs planar segmentation on the target top view and fits the target top view into two planes. This allows the acquisition of a workpiece plane image and a surface image of the printing platform 2. However, since the surface image of the printing platform 2 is not required for subsequent operations, only the workpiece plane image is retained after fitting. In this example, the control device can use a random sampling algorithm (RANSAC) to fit and segment the target top view. Since the random sampling algorithm can process data sets containing a large number of outliers, the use of this algorithm ensures that the two resulting plane images have high precision and greater reliability. Fitting and segmenting the target top view can accurately reflect the workpiece morphology and achieve accurate recognition of the workpiece shape.

[0100] As an example, in step S402, because external interference can cause pixel value variations in an image, noise points are filtered from the workpiece planar image to improve image quality and clarity, ensuring that the resulting workpiece planar image is easier to observe, analyze, and process. After the control device removes noise points from the workpiece planar image, it can obtain the current layer point cloud model corresponding to the workpiece planar image. The current layer point cloud model can accurately and realistically display the workpiece planar image, thereby improving the accuracy of workpiece printing and enabling the reproduction of workpiece topography without the need for external equipment.

[0101] In this embodiment, the control device fits and segments the target top view to obtain a workpiece plane image, and then filters out noise points in the workpiece plane image to ensure that the obtained workpiece plane image is easier to observe, analyze and process. After filtering, the current layer point cloud model is formed, which accurately and realistically displays the workpiece plane image, thereby improving the printing accuracy of the workpiece.

[0102] In one embodiment, if Figure 5 As shown, the current layer image includes at least two current layer main views with different shooting times;

[0103] Step S103, i.e., performing error evaluation based on the current layer image and the current layer point cloud model to obtain the error evaluation result, includes:

[0104] S501: Perform error analysis on at least two current layer main views to obtain measured view errors;

[0105] S502: performing point cloud registration on the current layer point cloud model to obtain the point cloud measured error;

[0106] S503: Obtain an error evaluation result based on the view measured error and the point cloud measured error.

[0107] The current layer main view refers to the main view printed by the printing mechanism 1 at the current layer at the current moment. The view measured error refers to the error result obtained after error analysis of the current layer main view, which is used to reflect the error size of the current layer main view.

[0108] As an example, in step S501, the control device performs error analysis on at least two current layer main views with different shooting times, obtains the measured view error, determines the error size of the current layer main view, and facilitates subsequent adaptive adjustments based on the measured view error size to ensure that the workpiece morphology can be accurately reflected. In this example, the main view Hu moment of the current layer main view can be obtained by inputting the code data of the current layer main view, and the error size of the current layer main view can be reflected by comparing the Hu moment sizes of at least two adjacent main views. Since the Hu moment is invariant to rotation, translation, and scaling, and the Hu moment can accurately describe the image shape, it can achieve image matching and alignment. Therefore, by comparing the main view Hu moment of the current layer main view, the error size of the current layer main view can be accurately determined.

[0109] The measured error of the point cloud refers to the error result obtained after error analysis of the current layer point cloud model, which is used to reflect the error size of the current layer point cloud model.

[0110] As an example, in step S502, the control device selects a point cloud registration intelligent algorithm to perform point cloud registration on the current layer point cloud model. After calculation, the measured error of the point cloud can be obtained, thereby determining the error size of the current layer point cloud model. In this example, after registering the current layer point cloud model, the registered error value can be obtained, that is, the RMSE (root mean square) error. By determining the RMSE (root mean square) error, it is convenient to clearly know the deviation value of the current layer point cloud model, so as to facilitate subsequent adaptive adjustments based on the measured error size of the point cloud, and ensure that the workpiece morphology can be accurately reflected.

[0111] As an example, in step S503, the control device may determine the error assessment results for the current layer primary view and the current layer point cloud model in the current layer image based on the acquired view measured error and point cloud measured error. The acquired error assessment results may facilitate subsequent execution of different operations based on the error assessment results and facilitate subsequent adaptive adjustments based on the error assessment results.

[0112] In this embodiment, the control device performs error analysis on at least two current layer main views to obtain the view measured error. The error size of the main view Hu moment can be determined through the view measured error, and the point cloud model of the current layer is aligned to obtain the point cloud measured error. The RMSE (root mean square) error size of the point cloud model can be determined through the view measured error. Subsequently, the control device can obtain the error evaluation result based on the view measured error and the point cloud measured error, and then perform adaptive adjustments according to the error evaluation result to ensure that the workpiece morphology is accurately reflected and the workpiece printing accuracy is improved.

[0113] In one embodiment, if Figure 6 As shown, step S502 is to perform point cloud registration on the current layer point cloud model to obtain the point cloud measured error, including:

[0114] S601: Register the current layer point cloud model with the current standard point cloud model to obtain a first registration error;

[0115] S602: Register the entire point cloud model with the entire standard point cloud model to obtain a second registration error. The entire point cloud model is a model obtained by superimposing the current layer point cloud model and the historical layer point cloud model.

[0116] The current standard point cloud model refers to a pre-set model used to determine the size of the registration error of the current layer point cloud model. The first registration error refers to the error result obtained after the point cloud registration of the current layer point cloud model.

[0117] As an example, in step S601, after the control device selects a point cloud registration method, it registers the current point cloud model with the current standard point cloud model and determines the registration error of the current point cloud model, i.e., obtains a first registration error. Point cloud registration methods include, but are not limited to, traditional point cloud registration algorithms such as ICP, GICP, NDT, and FPFH, as well as swarm intelligence algorithms such as genetic algorithms and simulated annealing algorithms. By determining the first registration error, the accuracy of the current point cloud model can be clearly understood, facilitating subsequent adjustments based on the first registration error.

[0118] The historical point cloud model refers to the point cloud model formed by superimposing the point cloud models of each layer before the current point cloud model. The overall standard point cloud model refers to a pre-set model used to determine the size of the overall point cloud model registration error. Specifically, it is the overall point cloud model containing multiple layers formed by superimposing the current standard point cloud model and the historical standard point cloud model. The second registration error refers to the error result obtained after performing point cloud registration on the overall point cloud model.

[0119] As an example, in step S602, the control device overlays the current layer point cloud model with the historical layer point cloud model to obtain an overall point cloud model. Through the overlay process, the control device can clearly obtain an overall point cloud model composed of at least two layers, thereby more comprehensively and three-dimensionally displaying the printing status and appearance of the workpiece. Subsequently, the control device aligns the overall point cloud model with the overall standard point cloud model. After the alignment calculation, a second alignment error can be obtained. The second alignment error determined by the control device after aligning the overall standard point cloud model can reflect the overall printing status of the workpiece, and can clearly understand the accuracy of the current layer point cloud model, which is also convenient for subsequent adjustments based on the second alignment error.

[0120] In this embodiment, the control device registers the current layer point cloud model with the current standard point cloud model to obtain a first registration error, and then registers the entire point cloud model with the entire standard point cloud model to obtain a second registration error. Based on the first and second registration errors, the error in the workpiece printing can be determined, thereby achieving the goal of improving the workpiece printing accuracy and facilitating subsequent adjustments based on the error size.

[0121] In one embodiment, as shown in FIG7 , step S503 , obtaining an error evaluation result based on the view measured error and the point cloud measured error, includes:

[0122] S701: comparing the view measured error with the first view error, and comparing the point cloud measured error with the first point cloud error;

[0123] S702: If the view measured error is less than the first view error, and the point cloud measured error is less than the first point cloud error, obtaining an error evaluation result that meets a preset error condition;

[0124] S703: If the view measured error is not less than the first view error, or the point cloud measured error is not less than the first point cloud error, then obtaining an error evaluation result that does not meet the preset error condition.

[0125] The first view error refers to a preset error value used to determine whether the view error measured meets the preset conditions. The first point cloud error refers to a preset error value used to determine whether the point cloud error measured meets the preset conditions.

[0126] As an example, in step S701, the control device compares the measured view error with the first view error, and compares the measured point cloud error with the first point cloud error. In this example, the first view error and the first point cloud error may have the same value, for example, 0.5, or may be different. Based on the first view error and the first point cloud error, the control device can clearly understand the specific error conditions of the measured view error and the measured point cloud error, facilitating subsequent determination of required operations based on the comparison results and reducing unnecessary resource waste.

[0127] The preset error condition refers to a pre-set condition for evaluating whether the view measured error and the point cloud measured error meet the preset error.

[0128] As an example, in step S702, when the control device recognizes that the measured view error is less than the first view error and the measured point cloud error is less than the first point cloud error, it determines that the measured view error and the measured point cloud error meet the preset error condition, indicating that the current workpiece printing accuracy is high and reflecting a high degree of workpiece topography matching. Furthermore, when the measured point cloud error includes a first registration error and a second registration error, the measured point cloud error only meets the preset error condition if both the first registration error and the second registration error are less than the first point cloud error.

[0129] As an example, in step S703, upon recognizing that the measured view error is not less than the first view error, or the measured point cloud error is not less than the first point cloud error, the control device determines that the measured view error or the measured point cloud error does not meet the preset error condition, indicating that the current workpiece printing error is large, the printed workpiece topography has a low degree of match, and printing has failed. Furthermore, if the measured point cloud error includes a first registration error and a second registration error, the measured point cloud error does not meet the preset error condition if at least one of the first registration error and the second registration error is not less than the first point cloud error.

[0130] In this embodiment, the control device compares the view measured error with the first view error, and compares the point cloud measured error with the first point cloud error, determines the accuracy of workpiece printing based on the comparison results, and determines the matching degree of the workpiece morphology. At the same time, based on the comparison results, the workpiece printing status reflected by the current device parameters can be clearly understood, so as to achieve real-time monitoring.

[0131] In one embodiment, if Figure 8 As shown, step S104, obtaining target device parameters, includes:

[0132] S801: Compare the view measured error with the second view error, and compare the point cloud measured error with the second point cloud error;

[0133] S802: If the view measured error is less than the second view error, and the point cloud measured error is less than the second point cloud error, then determine the current device parameters as the target device parameters;

[0134] S803: If the view measured error is not less than the second view error, or the point cloud measured error is not less than the second point cloud error, optimization processing is performed based on the current device parameters, the view measured error, and the point cloud measured error to obtain the target device parameters.

[0135] The second view error refers to a preset error value used to determine whether the view error measured meets the preset conditions. The second point cloud error refers to a preset error value used to determine whether the point cloud error measured meets the preset conditions.

[0136] As an example, in step S801, the control device compares the measured view error with the second view error, and compares the measured point cloud error with the second point cloud error. In this example, the second view error and the second point cloud error can have the same value, for example, both 0.01, or they can be different. Based on the first view error and the first point cloud error, the control device can clearly understand the specific error conditions of the measured view error and the measured point cloud error, facilitating subsequent determination of required operations based on the comparison results and reducing unnecessary resource waste.

[0137] As an example, in step S802, if the control device identifies that the measured view error is less than the second view error, and the measured point cloud error is less than the second point cloud error, this indicates that the workpiece printing accuracy is high, the printed workpiece topography matches well, and the current device parameters do not require optimization. Therefore, the control device uses the current device parameters as the target device parameters, ensuring that the workpiece printing status is reflected in real time based on the current device parameters. Furthermore, if the measured point cloud error includes a first registration error and a second registration error, the first and second registration errors must both be less than the second point cloud error for the measured point cloud error to meet the preset condition.

[0138] As an example, in step S803, when the control device recognizes that the view measured error is not less than the second view error, or the point cloud measured error is not less than the second point cloud error, it indicates that the printing operation still needs to be optimized. Therefore, the control device optimizes the current device parameters, the view measured error, and the point cloud measured error, and determines the optimized device parameters as the target device parameters, ensuring that the printing status of the workpiece is reflected in real time according to the current device parameters and that the printed workpiece shape has a high degree of matching. Furthermore, when the point cloud measured error includes a first registration error and a second registration error, if at least one of the first registration error and the second registration error is not less than the second point cloud error, the point cloud measured error does not meet the preset conditions.

[0139] In this embodiment, the control device compares the measured view error with the second view error, and also compares the measured point cloud error with the second point cloud error. Based on the comparison results, the control device determines whether to set the current device parameters as the target device parameters or to optimize the current device parameters again. This comparison result allows the current device parameters to reflect the workpiece's printing status in real time, and can also indicate the degree of topographical matching of the printed workpiece, enabling real-time monitoring.

[0140] In one embodiment, if Figure 9 As shown, in S803, optimization processing is performed based on the current device parameters, the view measured error, and the point cloud measured error to obtain the target device parameters, including:

[0141] S901: Input the current device parameters, the view measured error, and the point cloud measured error into the learning network to determine the optimized device parameters;

[0142] S902: Optimize the parameters of the optimization device using a learning network, and output the view prediction error and the point cloud prediction error;

[0143] S903: If the view prediction error is smaller than the second view error, and the point cloud prediction error is smaller than the second point cloud error, the optimized device parameters are determined as target device parameters.

[0144] The optimized device parameters refer to the parameters that need to be optimized among the current device parameters, and the learning network refers to the structure used to optimize the current device parameters.

[0145] As an example, in step S901, the control device inputs the current device parameters, view measured error and point cloud measured error into the learning network. The learning network will store the current device parameters, view measured error and point cloud measured error as a training set, so as to facilitate the subsequent precise optimization based on the stored multiple training sets. In this example, the learning network uses the Swin Transformer deep learning network to perform multiple orthogonal experimental learning. Swin Transformer is an emerging image recognition model, which is mainly used in natural language processing, image recognition and target detection and other fields, greatly reducing the complexity of data processing. The Swin Transformer deep learning network can quickly and accurately find the current device parameters that need to be optimized based on the view measured error and the point cloud measured error, reduce time costs, and simplify the complexity of optimizing the current device parameters.

[0146] The view prediction error refers to the view error value predicted by the learning network based on the optimized device parameters. The point cloud prediction error refers to the point cloud error value predicted by the learning network based on the optimized device parameters.

[0147] As an example, in step S902, the learning network optimizes the parameters of the optimized device. The range of the parameter optimization is within the training set of the learning network, and the numerical value of the parameter closest to the target device parameter is randomly varied. The specific parameter optimization range depends on the specific situation. After the learning network completes the parameter optimization, it calculates and outputs the view prediction error and point cloud prediction error corresponding to the optimized device parameters to the control device. Adaptive parameter adjustment is performed through the learning network to reduce time costs and improve printing efficiency and printing accuracy. Furthermore, when the point cloud prediction error includes a first registration prediction error and a second registration prediction error, it is necessary to satisfy that the first registration prediction error and the second registration prediction error are both less than the second point cloud error, and the point cloud prediction error meets the preset conditions.

[0148] As an example, in step S903, if the control device recognizes that the view prediction error output by the learning network is less than the second view error, and the point cloud prediction error is less than the second point cloud error, the control device determines that the optimized device parameters are highly accurate, therefore no further optimization is required, and determines the optimized device parameters as the target device parameters. If the control device recognizes that the view prediction error output by the learning network is not less than the second view error, or the point cloud prediction error is not less than the second point cloud error, the control device determines that the optimized device parameters are not sufficiently accurate and further optimization is required. Therefore, the control device continues to send the optimized device parameters to the learning network, causing the learning network to continue optimizing until the view prediction error output by the learning network is less than the second view error, and the point cloud prediction error is less than the second point cloud error. The optimization operation is then terminated. Furthermore, when the point cloud prediction error includes a first registration prediction error and a second registration prediction error, if at least one of the first registration prediction error and the second registration prediction error is not less than the second point cloud error, the point cloud prediction error does not meet the preset condition.

[0149] In this embodiment, the control device inputs the current device parameters, the measured view error, and the measured point cloud error into the learning network. After determining the optimized device parameters, the learning network performs parameter optimization to output the view prediction error and the point cloud prediction error. Only when the control device recognizes that the view prediction error is less than the second view error and the point cloud prediction error is less than the second point cloud error will the optimized device parameters be determined as the target device parameters. Using the learning network to optimize the current device parameters ensures that the target device parameters produce optimal results during the printing process, resulting in highly accurate workpieces that accurately reflect the workpiece's topography. Furthermore, the adaptive optimization performed by the learning network reduces printing time and improves printing efficiency.

[0150] In one embodiment, an FDM printing control system is provided, comprising a control device, a printing platform 2, a printing mechanism 1, and a camera device 3;

[0151] The printing mechanism 1 is arranged on the printing platform 2 and is used to perform the printing operation of the current layer image;

[0152] The camera device 3 is set on the printing platform 2 and is used to obtain the current layer image;

[0153] The control device is connected to the printing mechanism 1 and the camera device 3, and is used to control the printing mechanism 1 to perform the current layer printing operation according to the current device parameters.

[0154] The control device is a device for controlling the printing mechanism 1. The camera device 3 is a device for taking pictures of the workpiece.

[0155] As an example, the printing mechanism 1 is arranged on the printing platform 2, so that the printing mechanism 1 can print the current layer image of the workpiece. At the same time, the printing mechanism 1 and the printing platform 2 can be connected by a slide rail to ensure that the printing platform 2 can move within the printing mechanism 1.

[0156] As an example, the camera device 3 is positioned perpendicular to and parallel to the printing platform 2, and is movable. In this example, when capturing an image of the current layer, the printing platform 2 can remain stationary while the camera device 3 moves, allowing the printing mechanism 1 to complete the printing operation of the current layer image. Alternatively, when the camera device 3 is not moving, the printing platform 2 can be moved, allowing the printing mechanism 1 to complete the printing operation of the current layer image. This ensures that the camera device 3 can capture the image of the current layer of the workpiece and also ensures the accuracy of the captured image of the current layer.

[0157] As an example, the control device is in communication with the printing mechanism 1 and the imaging device 3, allowing the control device to control the printing mechanism 1 and the imaging device 3 to perform corresponding operations. Specifically, the entire processing process can be controlled through an application programming interface (API), including but not limited to adjusting the current layer device parameters, sending data and instructions, and monitoring the device operation status. At the same time, the control device will control the printing mechanism 1 to print the current layer image according to the current device parameters, ensuring the orderly execution of the printing operation.

[0158] The camera device 3 includes a first camera 31 and a second camera 32;

[0159] The first camera 31 is arranged on the printing platform 2 in a direction perpendicular to the printing platform 2, and is used to obtain at least two top views of the current layer with different shooting focal lengths;

[0160] The second camera 32 is disposed on the printing platform 2 in a direction parallel to the printing platform 2 , and is used to obtain at least two current layer main views with different shooting times.

[0161] The first camera 31 is a camera for capturing a top view of the current layer, and the second camera 32 is a camera for capturing a main view of the current layer.

[0162] As an example, the first camera 31 is positioned perpendicular to the print platform 2 and connected to the printing mechanism 1 via a leadscrew. Specifically, it is positioned above the print platform 2, with the print platform 2 as the reference point. In this example, the first camera 31 is a top-mounted floating depth camera. By sliding the first camera 31 within the printing mechanism 1 (in this example, the range of movement is 50 mm), it can capture at least two top-down views of the current layer with different focal lengths.

[0163] As an example, the second camera 32 is set in a direction parallel to the printing platform 2 and can be connected to the printing platform 2 through a synchronous belt. The other side of the second camera 32 can be connected to the printing mechanism 1 through a screw. When the first camera 31 cannot slide, the second camera 32 and the printing platform 2 will slide up and down the printing mechanism 1 to shoot at least two current layer main views at different times.

[0164] In one embodiment, a control device is provided, including a memory, a processor, and a control program stored in the memory and executable on the processor, wherein the processor implements the FDM printing control method in the above embodiment when executing the control program, for example Figure 1 The FDM printing control method shown in S101 to S105.

[0165] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0166] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, which can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0167] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0168] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. An FDM printing control method, characterized in that: include: Control the printing mechanism to perform the current layer printing operation based on the current device parameters to obtain the current layer image; Analyzing and processing the current layer image to obtain a current layer point cloud model corresponding to the current layer image; Perform error evaluation based on the current layer image and the current layer point cloud model to obtain an error evaluation result; If the error evaluation result satisfies the preset error condition, obtaining target device parameters, updating the target device parameters to current device parameters, and repeatedly controlling the printing mechanism to perform the current layer printing operation based on the current device parameters to obtain the current layer image; If the error evaluation result is that the preset error condition is not met, the printing operation is stopped; The current layer image includes at least two current layer top views shot at different focal lengths; and analyzing and processing the current layer image to obtain a current layer point cloud model corresponding to the current layer image includes: Performing clarity analysis on at least two top views of the current layer to obtain a target top view with the highest clarity; Performing point cloud conversion processing on the target top view to obtain a current layer point cloud model corresponding to the target top view; The current layer image includes at least two current layer main views taken at different times; performing error evaluation based on the current layer image and the current layer point cloud model to obtain an error evaluation result includes: Performing error analysis on at least two of the current layer main views to obtain a measured view error; Performing point cloud registration on the current layer point cloud model to obtain a point cloud measured error; The error evaluation result is obtained based on the view measured error and the point cloud measured error.

2. The FDM printing control method according to claim 1, wherein: The performing clarity analysis on at least two top views of the current layer to obtain a target top view with optimal clarity includes: Using an energy gradient function, respectively performing energy gradient calculations on at least two top views of the current layer, and determining current gradient values ​​corresponding to the at least two top views of the current layer; The top view of the current layer with the largest current gradient value is determined as the target top view with the highest clarity.

3. The FDM printing control method according to claim 1, wherein: The performing point cloud conversion processing on the target top view to obtain a current layer point cloud model corresponding to the current layer image includes: Performing fitting and segmentation on the target top view to obtain a workpiece plane image; Noise points are filtered out of the workpiece plane image to obtain a current layer point cloud model corresponding to the workpiece plane image.

4. The FDM printing control method according to claim 1, wherein: The performing point cloud registration on the current layer point cloud model to obtain the point cloud measured error includes: Registering the current layer point cloud model with the current standard point cloud model to obtain a first registration error; The overall point cloud model is registered with the overall standard point cloud model to obtain a second registration error, wherein the overall point cloud model is a point cloud model obtained by superimposing the current layer point cloud model and the historical layer point cloud model.

5. The FDM printing control method according to claim 4, wherein: Based on the measured view error and point cloud error, error assessment results are obtained, including: comparing the view measured error with the first view error, and comparing the point cloud measured error with the first point cloud error; If the view measured error is smaller than the first view error, and the point cloud measured error is smaller than the first point cloud error, obtaining an error evaluation result that meets the preset error condition; If the view measured error is not less than the first view error, or the point cloud measured error is not less than the first point cloud error, an error evaluation result that does not meet the preset error condition is obtained.

6. The FDM printing control method according to claim 3, wherein: The obtaining of target device parameters includes: comparing the view measured error with the second view error, and comparing the point cloud measured error with the second point cloud error; If the view measured error is smaller than the second view error, and the point cloud measured error is smaller than the second point cloud error, determining the current device parameters as the target device parameters; If the view measured error is not less than the second view error, or the point cloud measured error is not less than the second point cloud error, optimization processing is performed based on the current device parameters, the view measured error and the point cloud measured error to obtain the target device parameters.

7. The FDM printing control method according to claim 6, wherein: The optimization process is performed based on the current device parameters, the view measured error, and the point cloud measured error to obtain the target device parameters, including: Inputting the current device parameters, the view measured error, and the point cloud measured error into a learning network to determine the optimized device parameters; Using a learning network to optimize the parameters of the optimization device, and outputting a view prediction error and a point cloud prediction error; If the view prediction error is smaller than the second view error, and the point cloud prediction error is smaller than the second point cloud error, the optimized device parameters are determined as target device parameters.

8. A control device, characterized in that: The method comprises a memory, a processor, and a control program stored in the memory and executable on the processor, wherein the processor implements the FDM printing control method according to any one of claims 1 to 7 when executing the control program.

9. An FDM printing control system, characterized in that: comprising the control device, printing platform, printing mechanism and camera device as claimed in claim 8; The printing mechanism is arranged on the printing platform and is used for performing a printing operation of the current layer image; The camera device is arranged on the printing platform and is used to obtain the current layer image; The control device is connected to the printing mechanism and the camera device, and is used to control the printing mechanism to perform a current layer printing operation according to current device parameters.

10. The FDM printing control system according to claim 9, wherein: The camera device includes a first camera and a second camera; The first camera is arranged on the printing platform in a direction perpendicular to the printing platform, and is used to obtain at least two top views of the current layer with different shooting focal lengths; The second camera is arranged on the printing platform in a direction parallel to the printing platform, and is used to obtain at least two current layer main views with different shooting times.

Images