A feedback optimization method and system for 3D printing process based on vision inspection

Abstract

This invention provides a 3D printing process feedback optimization method and system based on visual inspection. By using visual inspection technology, it collects part processing image data at various stages of the 3D printing process, providing an image basis for further part inspection. Based on standard processing data and combined with image processing technology, it analyzes the part processing image data to obtain printing defect data, enabling real-time monitoring and analysis of the processed parts. Based on the correlation between printing defect data and printing process data, it determines adjustment parameters for the initial printing process data, enabling analysis of printing defects and deviations based on monitoring data, providing adjustment data, and inputting the adjustment parameters into a feedback control device to automatically control parameter adjustments during the 3D printing process. Based on automatic adjustment of printing parameters, dynamic optimization control is achieved. This scheme improves printing accuracy and quality, reduces manual intervention, and is suitable for the precision manufacturing of complex parts.

Description

Technical Field

[0001] This invention relates to the field of 3D printing technology, and in particular to a method and system for feedback optimization of the 3D printing process based on visual inspection. Background Technology

[0002] With the rapid development of the manufacturing industry, the demand for high-precision and complex structural parts is also increasing. For example, the aerospace, automotive, medical, electronics and mold manufacturing industries have high precision requirements for parts. However, due to the limitations of traditional manufacturing methods in certain aspects, such as the inadequacy of complex structures, material utilization and production cycle, there are certain limitations in the precision manufacturing of complex parts.

[0003] The core value of 3D printing complex parts lies in breaking through the geometric limitations of traditional manufacturing and achieving a unity of "design freedom" and "manufacturing freedom". Its application has moved from prototype manufacturing to direct production, becoming a key enabling technology in fields such as high-end equipment manufacturing, precision medicine, and green manufacturing.

[0004] However, in the process of 3D printing complex parts, how to combine image detection technology to accurately control the production process data is a key link in achieving high-precision production based on 3D printing. Summary of the Invention

[0005] This invention provides a visual inspection-based feedback optimization method and system for 3D printing process to solve the problems mentioned in the background art.

[0006] A feedback optimization method for 3D printing process based on visual inspection, comprising:

[0007] S1: Based on visual inspection technology, collect part processing image data at each stage of the 3D printing process;

[0008] S2: Based on standard part image data, combined with image processing technology, the part processing image data is analyzed to obtain printing defect data;

[0009] S3: Based on the correlation between printing defect data and printing process data, determine the adjustment parameters for the initial printing process data;

[0010] S4: Input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

[0011] Preferably, in step S1, based on visual inspection technology, image data of part processing at each stage of the 3D printing process are collected, including:

[0012] Visual inspection devices are installed at each stage of the 3D printing process;

[0013] Image data of part processing at each stage are collected based on a vision inspection device;

[0014] The part processing image data is transmitted to the server.

[0015] Preferably, in step S2, based on standard part image data and combined with image processing technology, the part processing image data is analyzed to obtain printing defect data, including:

[0016] The part processing image data is sequentially subjected to grayscale conversion, noise reduction, and contrast enhancement to obtain a preprocessed image;

[0017] Based on the positional information of the preprocessed image, the preprocessed image is fused to obtain a comprehensive part processing image;

[0018] Contour and texture features are extracted from the full part machining images to obtain contour and texture features;

[0019] The standard contour features and standard texture features of the standard part image data are compared with the contour features and texture features respectively to obtain the contour differences and texture differences;

[0020] Printing defect data is obtained based on the contour differences and texture differences.

[0021] Preferably, after obtaining the printing defect data, the method further includes:

[0022] The printing defect data is first divided according to the different stages of the 3D printing process to obtain stage defect data. From the stage defect data, the stage batch defect data of all processed parts in each stage is obtained.

[0023] Single contour differences are obtained from stage batch defect data, and a contour difference sequence is established. Single texture differences are obtained from stage batch defect data, and a texture difference sequence is established.

[0024] The first processing deviation is determined based on the average contour difference of all single contour differences, the second processing deviation is determined based on the average texture difference of all single texture differences, and the processing deviation trend is determined based on the sequence trends of the contour difference sequence and the texture difference sequence.

[0025] Based on the first machining deviation, the second machining deviation, and the machining deviation trend, the overall machining part deviation at each stage is determined;

[0026] The second division is performed based on the defect data of each part at all stages of the 3D printing process, to obtain the full-stage defect data of each processed part.

[0027] Based on the full-stage defect data of the part, the cumulative defect characteristics of the machined part are determined, and each extreme defect is marked to obtain the cumulative defect marking characteristics.

[0028] Based on a neural network model, the accumulated features of the defect markers are learned to determine the influence features of the difference between the previous stage and the subsequent stage in the processing. Based on the influence features of all stages, the eliminateable deviations of the processed parts caused by the processing sequence are determined.

[0029] The printing defect data is supplemented based on the comprehensive machining part deviations at each stage and the eliminable deviations caused by the machining sequence of the parts.

[0030] Preferably, the printing defect data is supplemented based on the comprehensive machining part deviations at each stage and the eliminable deviations of the machining parts due to the machining sequence, including:

[0031] Incorporate both machining part deviations and eliminable deviations into the printing defect data;

[0032] The difference between the overall machining part deviation and the eliminable deviation is taken as the part deviation to be eliminated and added to the printing defect data.

[0033] Preferably, in step S3, determining adjustment parameters for the initial printing process data based on the correlation between printing defect data and printing process data includes:

[0034] From the initial contour features and initial texture features corresponding to the printed defect data, and based on the deviation of the part to be eliminated, the initial contour features and initial texture features are eliminated to obtain the target contour features and target texture features, and the defect type corresponding to the target contour features and target texture features is obtained.

[0035] Calculate the correlation value between each defect type and the printing process data of each 3D printing device, and sort the printing process data based on the correlation value to determine the correlation sequence between defect types and multiple printing process data.

[0036] Printing process data with a correlation value greater than a preset value are obtained from the relevant sequences and used as initial running data related to the defect type.

[0037] Historical operating data and historical defect data related to the initial operating data are obtained, and a linear relationship between the historical operating data and historical defect data is learned based on a machine model to construct a defect association identification model.

[0038] Based on the defect association identification model and combined with relevant sequences, a secondary association analysis is performed on the printing defect data and printing process data to obtain the target operating data related to the printing defect data, and to determine the degree of difference between the target operating data and the standard operating data.

[0039] Adjustment parameters for the target operating data are determined based on the degree of difference.

[0040] Preferably, based on the defect association identification model and combined with relevant sequences, a secondary association analysis is performed on the printing defect data and printing process data to obtain target operational data related to the printing defect data, including:

[0041] The data type of the initial running data related to the defect type is determined based on the relevant sequence.

[0042] Based on the defect association identification model, the correlation between defect data corresponding to the defect type and the initial running data is identified. Based on the correlation, the target running data related to the printed defect data is obtained.

[0043] Preferably, the defect types corresponding to the target contour features and target texture features are obtained, including:

[0044] The target contour features and target texture features are compared with the standard contour features and standard texture features of the standard part image data to obtain the difference contour features and difference texture features;

[0045] Obtain the defect types corresponding to the differential contour features and differential texture features from the defect type feature library.

[0046] Preferably, in step S4, inputting the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process includes:

[0047] The adjustment parameters are automatically input into the feedback control device;

[0048] The feedback control device automatically generates adjustment strategies based on the adjustment parameters to automatically control the adjustment of parameters during the 3D printing process.

[0049] A visual inspection-based 3D printing process feedback optimization system includes:

[0050] The data acquisition module is used to collect part processing image data at each stage of the 3D printing process based on visual inspection technology;

[0051] The data analysis module is used to analyze part processing image data based on standard part image data and combined with image processing technology to obtain printing defect data;

[0052] The parameter determination module is used to determine the adjustment parameters for the initial printing process data based on the correlation between printing defect data and printing process data.

[0053] The parameter adjustment module is used to input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

[0054] Compared with the prior art, the present invention has achieved the following beneficial effects:

[0055] By using visual inspection technology, image data of part processing at each stage of 3D printing is collected, providing an image basis for further inspection of the parts. Based on standard processing data and combined with image processing technology, the part processing image data is analyzed to obtain printing defect data, enabling real-time monitoring and analysis of the processed parts. This provides a basis for feedback optimization of the 3D printing process. Based on the correlation between printing defect data and printing process data, adjustment parameters for the initial printing process data are determined, enabling analysis of printing defects and deviations based on monitoring data and providing adjustment data. The adjustment parameters are input into the feedback control device to automatically control the adjustment of parameters during the 3D printing process. Based on the automatic adjustment of printing parameters, dynamic optimization control is achieved. This scheme improves printing accuracy and quality, reduces manual intervention, and is suitable for the precision manufacturing of complex parts.

[0056] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0057] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0058] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0059] Figure 1 This is a flowchart of a 3D printing process feedback optimization method based on visual detection in an embodiment of the present invention;

[0060] Figure 2 This is a flowchart illustrating the process of obtaining printing defect data in an embodiment of the present invention;

[0061] Figure 3 This is a structural diagram of a visual inspection-based 3D printing process feedback optimization system according to an embodiment of the present invention. Detailed Implementation

[0062] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0063] Example 1:

[0064] This invention provides a feedback optimization method for the 3D printing process based on visual inspection, such as... Figure 1 As shown, it includes:

[0065] S1: Based on visual inspection technology, collect part processing image data at each stage of the 3D printing process;

[0066] S2: Based on standard part image data, combined with image processing technology, the part processing image data is analyzed to obtain printing defect data;

[0067] S3: Based on the correlation between printing defect data and printing process data, determine the adjustment parameters for the initial printing process data;

[0068] S4: Input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

[0069] In this embodiment, each stage of the 3D printing process is determined based on the structure of the complex part being processed, thereby determining the processing level and the manufacturing process for each layer.

[0070] In this embodiment, visual inspection techniques include, for example, camera inspection and laser scanning inspection.

[0071] In this embodiment, standard part image data and data such as the size and shape of the part when it is qualified are included. Each stage corresponds to a set of standard part image data.

[0072] In this embodiment, a stage includes multiple part processing image data, which are data from various orientations.

[0073] In this embodiment, image processing techniques include, for example, image preprocessing and image comparison.

[0074] In this embodiment, the printing defect data includes, for example, bumps, depressions, dimensional deviations, etc.

[0075] In this embodiment, the correlation between printing defect data and printing process data is, for example, that dimensional deviation is related to the printing distance in the printing process data, and that protrusions and depressions are related to the temperature and air pressure in the printing process data.

[0076] In this embodiment, a feedback control device is pre-installed in the 3D printing equipment to achieve automatic feedback adjustment of the 3D printing process.

[0077] The beneficial effects of the above design scheme are as follows: By collecting part processing image data at each stage of the 3D printing process based on visual inspection technology, image data is collected to provide a basis for further inspection of the parts. Based on standard processing data and combined with image processing technology, the part processing image data is analyzed to obtain printing defect data, realizing real-time monitoring and analysis of the processed parts, providing a basis for feedback optimization of the 3D printing process. Based on the correlation between printing defect data and printing process data, adjustment parameters for the initial printing process data are determined, realizing the analysis of printing defects and deviations based on monitoring data and providing adjustment data. The adjustment parameters are input into the feedback control device to automatically control the adjustment of parameters during the 3D printing process. Based on the automatic adjustment of printing parameters, dynamic optimization control is achieved. This scheme improves printing accuracy and quality, reduces manual intervention, and is suitable for the precision manufacturing of complex parts.

[0078] Example 2:

[0079] Based on Embodiment 1, this embodiment of the invention provides a 3D printing process feedback optimization method based on visual inspection. In step S1, based on visual inspection technology, image data of part processing at each stage of the 3D printing process are collected, including:

[0080] Visual inspection devices are installed at each stage of the 3D printing process;

[0081] Image data of part processing at each stage are collected based on a vision inspection device;

[0082] The part processing image data is transmitted to the server.

[0083] In this embodiment, the acquisition time for collecting part processing image data at each stage based on the visual inspection device is flexibly set according to the specific situation.

[0084] The beneficial effects of the above design scheme are: by setting up visual inspection devices at each stage of the 3D printing process, collecting part processing image data at each stage based on the visual inspection devices, and transmitting the part processing image data to the server, the acquisition and acquisition of part processing image data is realized, providing an image basis for further inspection of the parts.

[0085] Example 3:

[0086] Based on Example 1, this embodiment of the invention provides a feedback optimization method for the 3D printing process based on visual detection, such as... Figure 2 As shown, in step S2, based on standard part image data and combined with image processing technology, the part processing image data is analyzed to obtain printing defect data, including:

[0087] The part processing image data is sequentially subjected to grayscale conversion, noise reduction, and contrast enhancement to obtain a preprocessed image;

[0088] Based on the positional information of the preprocessed image, the preprocessed image is fused to obtain a comprehensive part processing image;

[0089] Contour and texture features are extracted from the full part machining images to obtain contour and texture features;

[0090] The standard contour features and standard texture features of the standard part image data are compared with the contour features and texture features respectively to obtain contour differences and texture differences. Based on the contour differences and texture differences, printing defect data is obtained.

[0091] In this embodiment, each stage corresponds to a set of printing defect data.

[0092] In this embodiment, image fusion is performed on the preprocessed image to obtain a comprehensive part processing image, which enriches the data information of the part processing image and makes the obtained comprehensive part processing image have all-round features.

[0093] The beneficial effects of the above design scheme are as follows: by extracting contour and texture features from the comprehensive part processing images, contour features and texture features are obtained. The standard contour features and standard texture features of the standard part image data are compared with the contour features and texture features respectively to obtain contour differences and texture differences. Based on the contour differences and texture differences, printing defect data is obtained, realizing real-time monitoring and analysis of the processed parts, and providing a basis for feedback optimization of the 3D printing process.

[0094] Example 4:

[0095] Based on Example 1, this embodiment of the invention provides a 3D printing process feedback optimization method based on visual inspection, which, after obtaining printing defect data, further includes:

[0096] The printing defect data is first divided according to the different stages of the 3D printing process to obtain stage defect data. From the stage defect data, the stage batch defect data of all processed parts in each stage is obtained.

[0097] Single contour differences are obtained from stage batch defect data, and a contour difference sequence is established. Single texture differences are obtained from stage batch defect data, and a texture difference sequence is established.

[0098] The first processing deviation is determined based on the average contour difference of all single contour differences, the second processing deviation is determined based on the average texture difference of all single texture differences, and the processing deviation trend is determined based on the sequence trends of the contour difference sequence and the texture difference sequence.

[0099] Based on the first machining deviation, the second machining deviation, and the machining deviation trend, the overall machining part deviation at each stage is determined;

[0100] The second division is performed based on the defect data of each part at all stages of the 3D printing process, to obtain the full-stage defect data of each processed part.

[0101] Based on the full-stage defect data of the part, the cumulative defect characteristics of the machined part are determined, and each extreme defect is marked to obtain the cumulative defect marking characteristics.

[0102] Based on a neural network model, the accumulated features of the defect markers are learned to determine the influence features of the difference between the previous stage and the subsequent stage in the processing. Based on the influence features of all stages, the eliminateable deviations of the processed parts caused by the processing sequence are determined.

[0103] The printing defect data is supplemented based on the comprehensive machining part deviations at each stage and the eliminable deviations caused by the machining sequence of the parts.

[0104] In this embodiment, the deviation can be eliminated because the deviation of the previous process causes the deviation of the current process. Adjusting the deviation of the previous process can eliminate the deviation of the current process.

[0105] In this embodiment, the first processing deviation is a deviation related to the contour, such as size and area, and the second processing deviation is a deviation related to the texture, such as cracks and porosity.

[0106] In this embodiment, the defect accumulation feature is the defect feature accumulated from the defects of each stage.

[0107] In this embodiment, the processing deviation trend is the direction of the deviation of the equipment in processing the part as time increases, for example, the leftward offset increases as time increases.

[0108] In this embodiment, based on a neural network model, the accumulated features of the defect markers are learned to obtain the difference data before and after the learning of the neural network model, as well as the stage processing features, so as to obtain the pattern of the influence of the difference in the previous stage on the processing of the next stage.

[0109] The beneficial effects of the above design scheme are as follows: By obtaining single contour differences from stage batch defect data, a contour difference sequence is established; by obtaining single texture differences from stage batch defect data, a texture difference sequence is established; and by performing a second division based on the defect data of each part in all stages of the 3D printing process, full-stage defect data of each processed part is obtained. The defect features are processed and analyzed from two aspects: the horizontal stage processing situation and the vertical processing process of a single processed part. Finally, based on the comprehensive processing part deviation at each stage and the eliminable deviation caused by the processing sequence, the printing defect data is supplemented, enabling further analysis of the printing defect data. This improves the accuracy and intuitiveness of the final printing defect data, providing a foundation for feedback optimization of the 3D printing process based on visual inspection.

[0110] Example 5:

[0111] Based on Example 4, this embodiment of the invention provides a visual inspection-based 3D printing process feedback optimization method, which supplements printing defect data based on the comprehensive machining part deviation at each stage and the eliminable deviations caused by the machining sequence, including:

[0112] Incorporate both machining part deviations and eliminable deviations into the printing defect data;

[0113] The difference between the overall machining part deviation and the eliminable deviation is taken as the part deviation to be eliminated and added to the printing defect data.

[0114] The beneficial effects of the above design scheme are: by incorporating the comprehensive machining part deviation and the eliminable deviation into the printing defect data, taking the difference between the comprehensive machining part deviation and the eliminable deviation as the part deviation to be eliminated, and incorporating it into the printing defect data, the accuracy and intuitiveness of the final printing defect data are improved, providing a basis for feedback optimization of the 3D printing process based on vision inspection.

[0115] Example 6:

[0116] Based on Embodiment 1, this embodiment of the invention provides a 3D printing process feedback optimization method based on visual inspection. In step S3, based on the correlation between printing defect data and printing process data, adjustment parameters for the initial printing process data are determined, including:

[0117] From the initial contour features and initial texture features corresponding to the printed defect data, and based on the deviation of the part to be eliminated, the initial contour features and initial texture features are eliminated to obtain the target contour features and target texture features, and the defect type corresponding to the target contour features and target texture features is obtained.

[0118] Calculate the correlation value between each defect type and the printing process data of each 3D printing device, and sort the printing process data based on the correlation value to determine the correlation sequence between defect types and multiple printing process data.

[0119] Printing process data with a correlation value greater than a preset value are obtained from the relevant sequences and used as initial running data related to the defect type.

[0120] Historical operating data and historical defect data related to the initial operating data are obtained, and a linear relationship between the historical operating data and historical defect data is learned based on a machine model to construct a defect association identification model.

[0121] Based on the defect association identification model and combined with relevant sequences, a secondary association analysis is performed on the printing defect data and printing process data to obtain the target operating data related to the printing defect data, and to determine the degree of difference between the target operating data and the standard operating data.

[0122] Adjustment parameters for the target operating data are determined based on the degree of difference.

[0123] In this embodiment, the correlation is, for example, that when the printing temperature is too low, the porosity increases significantly.

[0124] In this embodiment, the association relationship is one-to-one or one-to-many. The initial running data is a potentially related running data type, while the target running data is a truly related running data type. Furthermore, the specific impact of the target running data on the defect type is also determined.

[0125] In this embodiment, the adjustment parameters are, for example, adjustments to temperature, speed, distance, etc.

[0126] The beneficial effects of the above design scheme are: by analyzing and learning the correlation from both statistical analysis and machine learning analysis, the accuracy of the obtained correlation is guaranteed, thereby ensuring the accuracy of the adjustment parameters for the initial printing process data, and providing a data foundation for real-time adjustment of the 3D printing process.

[0127] Example 7:

[0128] Based on Example 6, this embodiment of the invention provides a visual inspection-based 3D printing process feedback optimization method. Based on a defect association identification model and combined with relevant sequences, a secondary association analysis is performed on printing defect data and printing process data to obtain target operational data related to the printing defect data, including:

[0129] The data type of the initial running data related to the defect type is determined based on the relevant sequence.

[0130] Based on the defect association identification model, the correlation between defect data corresponding to the defect type and the initial running data is identified. Based on the correlation, the target running data related to the printed defect data is obtained.

[0131] The beneficial effects of the above design scheme are: by determining the type of initial running data related to the defect type based on the relevant sequence, identifying the correlation between the defect data corresponding to the defect type and the initial running data based on the defect association identification model, and obtaining the target running data related to the printing defect data based on the correlation, the accurate analysis of the correlation based on the machine model is realized, and the accuracy of the obtained correlation is guaranteed.

[0132] Example 8:

[0133] Based on Example 6, this embodiment of the invention provides a visual inspection-based 3D printing process feedback optimization method to obtain the defect types corresponding to the target contour features and target texture features, including:

[0134] The target contour features and target texture features are compared with the standard contour features and standard texture features of the standard part image data to obtain the difference contour features and difference texture features;

[0135] Obtain the defect types corresponding to the differential contour features and differential texture features from the defect type feature library.

[0136] The beneficial effects of the above design scheme are: by comparing the target contour features and target texture features with the standard contour features and standard texture features of the standard part image data, the difference contour features and difference texture features are obtained, and the defect types corresponding to the difference contour features and difference texture features are obtained from the defect type feature library, thus ensuring the accuracy of the obtained defect types.

[0137] Example 9:

[0138] Based on Embodiment 1, this embodiment of the invention provides a feedback optimization method for the 3D printing process based on visual inspection. In step S4, the adjustment parameters are input into a feedback control device to automatically control the adjustment of parameters during the 3D printing process, including:

[0139] The adjustment parameters are automatically input into the feedback control device;

[0140] The feedback control device automatically generates adjustment strategies based on the adjustment parameters to automatically control the adjustment of parameters during the 3D printing process.

[0141] In this embodiment, the automatically generated adjustment strategy is, for example, to adjust step by step according to units of two.

[0142] The beneficial effects of the above design scheme are: by automatically inputting the adjustment parameters into the feedback control device, the feedback control device automatically generates an adjustment strategy based on the adjustment parameters to automatically control the adjustment of parameters during the 3D printing process, thereby achieving dynamic optimization control. This scheme improves printing accuracy and quality, reduces manual intervention, and is suitable for the precision manufacturing of complex parts.

[0143] Example 10:

[0144] This invention provides a visual inspection-based 3D printing process feedback optimization system, such as... Figure 3 As shown, it includes:

[0145] The data acquisition module is used to collect part processing image data at each stage of the 3D printing process based on visual inspection technology;

[0146] The data analysis module is used to analyze part processing image data based on standard part image data and combined with image processing technology to obtain printing defect data;

[0147] The parameter determination module is used to determine the adjustment parameters for the initial printing process data based on the correlation between printing defect data and printing process data.

[0148] The parameter adjustment module is used to input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

[0149] In this embodiment, each stage of the 3D printing process is determined based on the structure of the complex part being processed, thereby determining the processing level and the manufacturing process for each layer.

[0150] In this embodiment, visual inspection techniques include, for example, camera inspection and laser scanning inspection.

[0151] In this embodiment, standard part image data and data such as the size and shape of the part when it is qualified are included. Each stage corresponds to a set of standard part image data.

[0152] In this embodiment, a stage includes multiple part processing image data, which are data from various orientations.

[0153] In this embodiment, image processing techniques include, for example, image preprocessing and image comparison.

[0154] In this embodiment, the printing defect data includes, for example, bumps, depressions, dimensional deviations, etc.

[0155] In this embodiment, the correlation between printing defect data and printing process data is, for example, that dimensional deviation is related to the printing distance in the printing process data, and that protrusions and depressions are related to the temperature and air pressure in the printing process data.

[0156] In this embodiment, a feedback control device is pre-installed in the 3D printing equipment to achieve automatic feedback adjustment of the 3D printing process.

[0157] The beneficial effects of the above design scheme are as follows: By collecting part processing image data at each stage of the 3D printing process based on visual inspection technology, image data is collected to provide a basis for further inspection of the parts. Based on standard processing data and combined with image processing technology, the part processing image data is analyzed to obtain printing defect data, realizing real-time monitoring and analysis of the processed parts, providing a basis for feedback optimization of the 3D printing process. Based on the correlation between printing defect data and printing process data, adjustment parameters for the initial printing process data are determined, realizing the analysis of printing defects and deviations based on monitoring data and providing adjustment data. The adjustment parameters are input into the feedback control device to automatically control the adjustment of parameters during the 3D printing process. Based on the automatic adjustment of printing parameters, dynamic optimization control is achieved. This scheme improves printing accuracy and quality, reduces manual intervention, and is suitable for the precision manufacturing of complex parts.

[0158] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this application and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A feedback optimization method for 3D printing process based on visual inspection, characterized in that, include: S1: Based on visual inspection technology, collect part processing image data at each stage of the 3D printing process; S2: Based on standard part image data, combined with image processing technology, the part processing image data is analyzed to obtain printing defect data; After obtaining the printing defect data, it also includes: The printing defect data is first divided according to the different stages of the 3D printing process to obtain stage defect data. From the stage defect data, the stage batch defect data of all processed parts in each stage is obtained. Single contour differences are obtained from stage batch defect data, and a contour difference sequence is established. Single texture differences are obtained from stage batch defect data, and a texture difference sequence is established. The first processing deviation is determined based on the average contour difference of all single contour differences, the second processing deviation is determined based on the average texture difference of all single texture differences, and the processing deviation trend is determined based on the sequence trends of the contour difference sequence and the texture difference sequence. Based on the first machining deviation, the second machining deviation, and the machining deviation trend, the overall machining part deviation at each stage is determined; The second division is performed based on the defect data of each part at all stages of the 3D printing process, to obtain the full-stage defect data of each processed part. Based on the full-stage defect data of the part, the cumulative defect characteristics of the machined part are determined, and each extreme defect is marked to obtain the cumulative defect marking characteristics. Based on a neural network model, the accumulated features of the defect markers are learned to determine the influence features of the difference between the previous stage and the subsequent stage in the processing. Based on the influence features of all stages, the eliminateable deviations of the processed parts caused by the processing sequence are determined. The printing defect data is supplemented based on the comprehensive machining part deviations at each stage and the eliminable deviations caused by the machining sequence, including: Incorporate both machining part deviations and eliminable deviations into the printing defect data; The difference between the overall machining part deviation and the eliminable deviation is taken as the part deviation to be eliminated and added to the printing defect data; S3: Based on the correlation between printing defect data and printing process data, determine the adjustment parameters for the initial printing process data; S4: Input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

2. The 3D printing process feedback optimization method based on vision detection according to claim 1, characterized in that, In step S1, based on visual inspection technology, image data of part processing at each stage of the 3D printing process are collected, including: Visual inspection devices are installed at each stage of the 3D printing process; Image data of part processing at each stage are collected based on a vision inspection device; The part processing image data is transmitted to the server.

3. The 3D printing process feedback optimization method based on vision detection according to claim 1, characterized in that, In step S2, based on standard part image data and combined with image processing technology, the part machining image data is analyzed to obtain printing defect data, including: The part processing image data is sequentially subjected to grayscale conversion, noise reduction, and contrast enhancement to obtain a preprocessed image; Based on the positional information of the preprocessed image, the preprocessed image is fused to obtain a comprehensive part processing image; Contour and texture features are extracted from the full part machining images to obtain contour and texture features; The standard contour features and standard texture features of the standard part image data are compared with the contour features and texture features respectively to obtain the contour differences and texture differences; Printing defect data is obtained based on the contour differences and texture differences.

4. The 3D printing process feedback optimization method based on vision detection according to claim 1, characterized in that, In step S3, based on the correlation between printing defect data and printing process data, adjustment parameters for the initial printing process data are determined, including: From the initial contour features and initial texture features corresponding to the printed defect data, and based on the deviation of the part to be eliminated, the initial contour features and initial texture features are eliminated to obtain the target contour features and target texture features, and the defect type corresponding to the target contour features and target texture features is obtained. Calculate the correlation value between each defect type and the printing process data of each 3D printing device, and sort the printing process data based on the correlation value to determine the correlation sequence between defect types and multiple printing process data. Printing process data with a correlation value greater than a preset value are obtained from the relevant sequences and used as initial running data related to the defect type. Historical operating data and historical defect data related to the initial operating data are obtained, and a linear relationship between the historical operating data and historical defect data is learned based on a machine model to construct a defect association identification model. Based on the defect association identification model and combined with relevant sequences, a secondary association analysis is performed on the printing defect data and printing process data to obtain the target operating data related to the printing defect data, and to determine the degree of difference between the target operating data and the standard operating data. Adjustment parameters for the target operating data are determined based on the degree of difference.

5. The 3D printing process feedback optimization method based on vision detection according to claim 4, characterized in that, Based on the defect association identification model and combined with relevant sequences, a secondary association analysis is performed on printing defect data and printing process data to obtain target operational data related to the printing defect data, including: The data type of the initial running data related to the defect type is determined based on the relevant sequence. Based on the defect association identification model, the correlation between defect data corresponding to the defect type and the initial running data is identified. Based on the correlation, the target running data related to the printed defect data is obtained.

6. The 3D printing process feedback optimization method based on vision detection according to claim 4, characterized in that, Obtain the defect types corresponding to the target contour features and target texture features, including: The target contour features and target texture features are compared with the standard contour features and standard texture features of the standard part image data to obtain the difference contour features and difference texture features; Obtain the defect types corresponding to the differential contour features and differential texture features from the defect type feature library.

7. The 3D printing process feedback optimization method based on vision detection according to claim 1, characterized in that, In step S4, inputting the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process includes: The adjustment parameters are automatically input into the feedback control device; The feedback control device automatically generates adjustment strategies based on the adjustment parameters to automatically control the adjustment of parameters during the 3D printing process.

8. A visual inspection-based 3D printing process feedback optimization system, used to implement the feedback optimization method as described in claim 1, characterized in that, include: The data acquisition module is used to collect part processing image data at each stage of the 3D printing process based on visual inspection technology; The data analysis module is used to analyze part processing image data based on standard part image data and combined with image processing technology to obtain printing defect data; The parameter determination module is used to determine the adjustment parameters for the initial printing process data based on the correlation between printing defect data and printing process data. The parameter adjustment module is used to input the adjustment parameters into the feedback control device to automatically control the adjustment of parameters during the 3D printing process.

Images