Feature analysis method and system for additive manufacturing process based on deep learning

Abstract

The invention provides a feature analysis method and system for an additive manufacturing process based on deep learning, and can be applied to the technical field of additive manufacturing. According to the method, after the to-be-analyzed image including the macroscopic molten pool image and the macroscopic metallographic image or the microstructure image of the cross section of the deposition sample is obtained, the to-be-analyzed image is segmented through the preset segmentation model to obtain the molten pool area image, the deposition layer contour image or the secondary phase image of the microstructure; then, the deposition layer dilution rate, the additive manufacturing effective deposition rate or the microstructure secondary phase content in the additive manufacturing process are obtained through calculation according to the target area image, and therefore the problems that when the dilution rate, the effective deposition rate and the microstructure secondary phase content are calculated, parameters are sensitive, manual work is depended, precision is low, and efficiency is poor can be effectively solved; and the evaluation accuracy of the component precision, the forming quality and the microstructure evolution in the functionally graded material additive manufacturing process is improved.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing technology, and in particular to a feature analysis method and system for additive manufacturing processes based on deep learning. Background Technology

[0002] Functionally graded material additive manufacturing (FJD) achieves gradient changes in material properties by continuously altering the composition ratios of two or more materials, showing broad application prospects in aerospace, biomedicine, and other fields. However, the core of quality control in this process lies in accurately assessing the compositional precision, forming quality, and microstructure evolution during deposition.

[0003] The dilution rate of the deposited layer is a key indicator for evaluating the accuracy of the composition of interlayer deposition of heterogeneous materials. During interlayer deposition, the preceding deposited layer remelts, causing a shift in the composition of the new deposited layer. To control this compositional shift, the degree of remelting, i.e., the dilution rate, must be precisely quantified. In existing technologies, the dilution rate is calculated by analyzing metallographic images of the deposited layer cross-section and is defined as the ratio of the molten area of ​​the substrate to the area of ​​the entire molten pool. However, traditional calculation methods heavily rely on manual measurement or automated measurement based on traditional image processing algorithms. These traditional algorithms require multiple preprocessing steps, including grayscale conversion, filtering and noise reduction, and threshold binarization, exhibiting significant parameter sensitivity and dependence on human intervention. Operators need to repeatedly adjust parameters for different imaging conditions (such as illumination and contrast), a cumbersome and inefficient process, and the contour extraction accuracy decreases significantly under complex backgrounds or noise interference.

[0004] Similarly, when evaluating the effective deposition rate (the ratio of effective deposited material to total deposited material) in additive manufacturing, existing techniques typically require calculating the total area of ​​the deposited layer and the area of ​​the effective region within it from a cross-sectional image. This method struggles to automatically and accurately identify the largest effective region within irregular contours.

[0005] In microstructure analysis, the content, size, and distribution of secondary phases (such as intermetallic compounds and carbides) in functionally graded materials directly determine the final mechanical properties of the material. Existing quantitative metallographic analysis methods often employ manually set fixed thresholds for image binarization. This method is highly susceptible to subjective influence, poorly adaptable to variations in lighting and contrast across different viewing fields, easily introduces measurement bias, and makes it difficult to achieve large-scale, standardized, and reproducible statistical analysis. Summary of the Invention

[0006] The main objective of this invention is to provide a feature analysis method and system for additive manufacturing processes based on deep learning, which can effectively improve the accuracy of evaluation of composition precision, forming quality, and microstructure evolution during the additive manufacturing of functionally graded materials.

[0007] To achieve the above objectives, the present invention provides a feature analysis method for additive manufacturing processes based on deep learning, the method comprising the following steps:

[0008] Acquire images to be analyzed during the additive manufacturing process, including macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images;

[0009] The image to be analyzed is segmented by a preset segmentation model to obtain a target region image; the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure;

[0010] The deposition layer dilution rate, effective deposition rate, or secondary phase content of the microstructure during the additive manufacturing process are calculated based on the target area image.

[0011] In some embodiments, calculating the deposition layer dilution rate during the additive manufacturing process based on the target region image includes:

[0012] Obtain the positive and negative marker points input from the interactive interface;

[0013] The cladding area is identified in the molten pool area image based on the positive marker points;

[0014] The substrate melting region is identified in the cladding region based on the reverse point;

[0015] The deposition layer dilution rate during the additive manufacturing process is calculated based on the cladding area and the substrate melting area.

[0016] In some embodiments, calculating the deposition layer dilution rate during the additive manufacturing process based on the cladding region and the substrate melting region includes:

[0017] Obtain the first pixel area of ​​the cladding region and the second pixel area of ​​the substrate melting region;

[0018] The first physical area corresponding to the cladding region is calculated based on the preset pixel size calibration coefficient and the area of ​​the first pixel.

[0019] The second physical area of ​​the substrate melting region is calculated based on the preset pixel size calibration coefficient and the second pixel area.

[0020] The deposition layer dilution rate during the additive manufacturing process is calculated based on the second physical area and the first physical area.

[0021] In some embodiments, calculating the effective deposition rate of additive manufacturing in the additive manufacturing process based on the target region image includes:

[0022] Obtain the positive marker points input from the interactive interface;

[0023] The sedimentary layer contour is determined in the sedimentary layer contour image based on the positive marker points;

[0024] The total area of ​​the sedimentary layer is calculated based on the sedimentary layer profile.

[0025] The maximum incorporation matrix within the deposition layer profile is determined based on the maximum matrix incorporation algorithm.

[0026] Calculate the area of ​​the target rectangle corresponding to the maximum inscribed matrix;

[0027] The effective deposition rate of additive manufacturing is calculated based on the area of ​​the target rectangle and the total area of ​​the deposition layer.

[0028] In some embodiments, determining the maximum inscribed matrix within the deposition layer profile based on the maximum matrix inscribed algorithm includes:

[0029] Obtain the deposition contour polygon within the deposition layer contour;

[0030] Traverse all pixels within the deposited contour polygon to obtain the area of ​​the candidate matrix for all polygons;

[0031] The largest inscribed matrix is ​​determined from all polygons based on the area of ​​the candidate matrix.

[0032] In some embodiments, calculating the secondary phase content of the microstructure during the additive manufacturing process based on the target region image includes:

[0033] Convert the secondary phase image of the microstructure into a grayscale image;

[0034] The grayscale image is converted into a binary image based on an adaptive segmentation algorithm;

[0035] Perform pixel statistics on the binary image to obtain the total number of white pixels;

[0036] The content of secondary phase in the microstructure during the additive manufacturing process is calculated based on the total number of white pixels.

[0037] In some embodiments, converting the secondary phase image of the microstructure into a grayscale image includes:

[0038] The secondary phase image of the microstructure is preprocessed to obtain a preprocessed image. The preprocessing process includes contrast enhancement and brightness adjustment.

[0039] The preprocessed image is converted into a grayscale image.

[0040] In some embodiments, calculating the secondary phase content of the microstructure during the additive manufacturing process based on the total number of white pixels includes:

[0041] Calculate the percentage of the total number of white pixels in the secondary phase image of the microstructure;

[0042] The pixel percentage is used as the content of secondary phase in the microstructure during the additive manufacturing process.

[0043] To achieve the above objectives, another aspect of the present invention provides a feature analysis system for additive manufacturing processes based on deep learning, the system comprising:

[0044] The image acquisition module is used to acquire images to be analyzed during the additive manufacturing process. The images to be analyzed include macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images.

[0045] The image processing module is used to segment the image to be analyzed using a preset segmentation model to obtain a target region image; the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure;

[0046] The feature calculation module is used to calculate the deposition layer dilution rate, effective deposition rate, or secondary phase content of microstructure during the additive manufacturing process based on the target area image.

[0047] To achieve the above objectives, another aspect of the present invention provides a feature analysis system for additive manufacturing processes based on deep learning, the system comprising:

[0048] At least one processor;

[0049] At least one memory for storing at least one program;

[0050] When the at least one program is executed by the at least one processor, the at least one processor performs the method described above.

[0051] The present invention provides the following beneficial effects: This embodiment acquires an image to be analyzed, including a macroscopic molten pool image, a macroscopic metallographic image of the cross-section of the deposited sample, or a microstructure image. Then, it segments the image using a preset segmentation model integrated with an image segmentation unit based on a deep convolutional neural network, obtaining images of the molten pool region, the deposited layer contour, or the secondary phase of the microstructure. Based on the target region image, the deposition layer dilution rate, effective deposition rate, or secondary phase content of the microstructure during the additive manufacturing process is calculated. This effectively solves the problems of parameter sensitivity, reliance on manual labor, low accuracy, and poor efficiency in calculating the dilution rate, effective deposition rate, and secondary phase content. It achieves automated, high-precision, and robust analysis of key quality indicators in additive manufacturing, thereby improving the accuracy of compositional precision, forming quality, and microstructure evolution assessment during the additive manufacturing of functionally graded materials. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0053] Figure 1 This is a flowchart of a feature analysis method for additive manufacturing processes based on deep learning, provided in this application.

[0054] Figure 2 This is a login diagram of the visual interface provided in this application;

[0055] Figure 3 This is a schematic diagram of the calculation interface for the sediment layer dilution rate provided in this application;

[0056] Figure 4 This is a flowchart illustrating the calculation of the sediment layer dilution rate provided in this application;

[0057] Figure 5 This is a schematic diagram of the calculation interface for the effective deposition rate provided in this application;

[0058] Figure 6 This is a flowchart illustrating the calculation of the effective deposition rate provided in this application;

[0059] Figure 7 This is a schematic diagram of the calculation interface for the secondary phase content of microstructure provided in this application;

[0060] Figure 8 This is a flowchart of the calculation of the secondary phase content of the microstructure provided in this application;

[0061] Figure 9 This is a schematic diagram of a feature analysis system for additive manufacturing processes based on deep learning, provided in this application. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0063] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The terms "upper," "lower," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are used only for the convenience of describing this application and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. Unless otherwise expressly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly, for example, as a fixed connection, a detachable connection, or an integral connection; a mechanical connection or an electrical connection; a direct connection or an indirect connection through an intermediate medium; or a connection within two elements. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0064] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0065] Reference Figure 1This application provides a feature analysis method for additive manufacturing processes based on deep learning, the method comprising the following steps:

[0066] Step S110: Obtain images to be analyzed during the additive manufacturing process, including macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images.

[0067] Step S120: Segment the image to be analyzed using a preset segmentation model to obtain the target region image; wherein, the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure.

[0068] Step S130: Calculate the deposition layer dilution rate, effective deposition rate, or secondary phase content of microstructure during the additive manufacturing process based on the target area image.

[0069] It is understood that the macroscopic molten pool image in this embodiment can be a macroscopic additive manufacturing molten pool image, which is used for dilution rate analysis; the macroscopic image of the cross-section of the deposited sample in this embodiment is used for effective deposition rate analysis; the microstructure image in this embodiment can be a high-magnification microstructure SEM / optical microscope image, which is used for secondary phase analysis.

[0070] It is understood that the preset segmentation model in this embodiment integrates an image segmentation unit based on a deep convolutional neural network. This unit employs a SAM model pre-trained on massive datasets (such as the SA-1B dataset). This SAM model possesses excellent zero-shot transfer capability and can be directly applied to the analysis scenario of this embodiment without requiring additional training for a specific task. Specifically, when segmenting the image to be analyzed, this embodiment can... Figure 2 After logging in to the visualization interface shown, users can specify the target area to be segmented by adding positive markers (such as green dots) and negative markers (such as red dots) to exclude interference or background areas. The SAM model can generate a high-precision segmentation mask after processing the image to be analyzed based on the added markers, accurately identifying specific phases in the cladding area, sedimentary layer outline, or microstructure of the image.

[0071] In the embodiments of this application, Figure 2 The system shown in the visualization interface can consist of a high-performance computer and image acquisition equipment (including a high-definition camera, stereo microscope, scanning electron microscope, etc.). The system software is developed based on the Python language, integrating the PYQT5 toolkit to build the graphical user interface (GUI), and the core algorithms rely on deep learning frameworks such as PyTorch and the SAM model library.

[0072] It is understood that, in this embodiment, after obtaining the target area image, the deposition layer dilution rate in the additive manufacturing process is obtained by analyzing the molten pool area image within the target area image. Specifically, this embodiment can obtain the forward and reverse marker points input in the interactive interface, identify the cladding area in the molten pool area image based on the forward marker points, identify the substrate melting area in the cladding area based on the reverse marker points, and then calculate the deposition layer dilution rate in the additive manufacturing process based on the cladding area and the substrate melting area. In this embodiment, this embodiment can obtain the first pixel area of ​​the cladding area and the second pixel area of ​​the substrate melting area, calculate the first physical area corresponding to the cladding area based on a preset pixel size calibration coefficient and the first pixel area, and calculate the second physical area of ​​the substrate melting area based on the preset pixel size calibration coefficient and the second pixel area, and then calculate the deposition layer dilution rate in the additive manufacturing process based on the second physical area and the first physical area.

[0073] For example, based on such Figure 3 When calculating the sediment layer dilution rate using the visualization interface shown, as follows: Figure 4 As shown, the processing procedure in this embodiment is as follows:

[0074] Step S201: The user imports a cross-sectional metallographic image of a single-pass, single-layer deposition of 316L / IN625 functionally graded material through the system interface.

[0075] Step S202: The user selects the "Dilution Calculation" function. The system interface displays an image.

[0076] Step S203: By adding dividing lines to the image, the entire molten pool is divided into the cladding area and the substrate melting area;

[0077] Step S204: Select and exclude areas (add green dots to selected areas and red dots to excluded areas), click Start Detection to obtain the cladding area;

[0078] Step S205: Click to switch to the second volume, select and exclude areas (add green dots to the selected areas and red dots to the excluded areas), click Start Detection to obtain the substrate melting area.

[0079] Step S206: After the model is segmented, the area of ​​the first pixel corresponding to the cladding area and the area of ​​the second pixel corresponding to the substrate melting area are automatically calculated, and converted into the actual first physical area (A) according to the preset pixel-actual size calibration coefficient. clad ) and second physical area (A melt ).

[0080] Step S207: Feature calculation module calls formula It automatically calculates the dilution rate of the deposited layer during the additive manufacturing process and displays it in the results box on the interface.

[0081] Step S208: Users can save the calculation results and segmented images for recording and reporting.

[0082] It is understood that, after obtaining the target area image, this embodiment calculates the effective deposition rate of additive manufacturing in the additive manufacturing process based on the sedimentary layer contour image in the target area image. Specifically, this embodiment can obtain the positive marker points input in the interactive interface, determine the sedimentary layer contour in the sedimentary layer contour image based on the positive marker points, calculate the total area of ​​the sedimentary layer based on the sedimentary layer contour, and determine the maximum inline matrix within the sedimentary layer contour based on the maximum matrix inline algorithm; then calculate the area of ​​the target rectangle corresponding to the maximum inline matrix, and calculate the effective deposition rate of additive manufacturing in the additive manufacturing process based on the area of ​​the target rectangle and the total area of ​​the sedimentary layer.

[0083] For example, based on such Figure 5 When performing effective deposition rate calculations using the visualization interface shown, as follows: Figure 6 As shown, the processing procedure in this embodiment is as follows:

[0084] Step S301: The user imports a macroscopic metallographic image of the cross-section of an FGM sample after multi-channel and multi-layer deposition.

[0085] Step S302: The user selects the "Effective Deposition Rate Calculation" function.

[0086] Step S303: The user roughly outlines the sedimentary layer by adding positive marker points (green). Based on this, the SAM model accurately segments the cross-sectional outline of the entire sedimentary layer and automatically calculates the total area S of the sedimentary layer. total .

[0087] Step S304: Call the built-in optimized maximum inscribed rectangle algorithm.

[0088] Step S305: After determining the maximum inscribed rectangle by optimizing the maximum inscribed rectangle algorithm, highlight the area S of the target rectangle corresponding to the maximum inscribed rectangle on the interface image. effective .

[0089] Step S306: According to the formula Automatically calculate and display the effective deposition rate in the additive manufacturing process.

[0090] Step S307: The user saves the analysis results.

[0091] It is understood that the processing procedure of the maximum matrix incorporation algorithm in this embodiment may involve obtaining the deposition contour polygons within the deposition layer contour, traversing all pixels within the deposition contour polygons to obtain the candidate matrix areas of all polygons, and then determining the maximum incorporation matrix from all polygons based on the candidate matrix areas. For example, the execution process of the maximum matrix incorporation algorithm in this embodiment is as follows:

[0092] a) Obtain the polygonal outline of the sedimentary layer.

[0093] b) Traverse all pixels (or simplified polygon vertices) on the contour of the sedimentation layer polygon.

[0094] c) For each pixel as a reference point, calculate the distance of other pixels relative to that reference point. .

[0095] d) During the traversal, dynamically update and record the leftmost and rightmost x-coordinates corresponding to each distance h. min and x max .

[0096] e) For each distance h, calculate the width of the rectangle. .

[0097] f) Calculate the area of ​​the rectangle at the current distance. .

[0098] g) Iteratively compare all possible distances of all pixels. Find the rectangle with the largest area; the area of ​​this rectangle is S. effective .

[0099] It is understood that, in this embodiment, after obtaining the target region image, the secondary phase content of the microstructure in the additive manufacturing process is calculated based on the secondary phase image of the microstructure in the target region image. Specifically, this embodiment can convert the secondary phase image of the microstructure into a grayscale image, then convert the grayscale image into a binary image based on an adaptive segmentation algorithm, then perform pixel statistics on the binary image to obtain the total number of white pixels, and then calculate the secondary phase content of the microstructure in the additive manufacturing process based on the total number of white pixels. In this embodiment, a preprocessed image can be obtained by performing contrast enhancement and brightness adjustment preprocessing on the secondary phase image of the microstructure, and then the preprocessed image can be converted into a grayscale image. Furthermore, when calculating the secondary phase content of the microstructure, the pixel percentage of the total number of white pixels in the secondary phase image of the microstructure can be calculated, and this pixel percentage can be used as the secondary phase content of the microstructure in the additive manufacturing process.

[0100] For example, based on such Figure 7When performing secondary phase content calculations in the visualization interface shown, such as... Figure 8 As shown, the processing procedure in this embodiment is as follows:

[0101] Step S401: The user imports a SEM or optical microscope image (RGB format) of the microstructure of a 316L / IN625 functionally graded material.

[0102] Step S402: After the user selects the "Secondary Phase Analysis" function, the contrast and brightness of the image are adjusted appropriately to enhance the clarity of the phase boundary.

[0103] Step S403: Automatically convert the enhanced RGB color image to a grayscale image to simplify the data.

[0104] Step S404: Apply an adaptive thresholding algorithm (such as the cv2.adaptiveThreshold function in OpenCV) to the grayscale image to obtain a binary image. In this embodiment, the adaptive thresholding algorithm does not use a global threshold, but instead calculates local thresholds in different small regions of the image. This allows for better handling of images with uneven lighting, accurately converting the grayscale image into a binary image. The secondary phase is typically segmented into white (pixel value 255) and the matrix into black (pixel value 0).

[0105] Step S405: Perform pixel statistics on the binary image and calculate the total number of all white pixels.

[0106] Step S406: Calculate the percentage of the total number of white pixels to the total number of pixels. This percentage is approximately the area fraction of the quadratic phase, and the result is displayed on the interface.

[0107] Step S407: Display the quadratic phase portion represented in the binary image as a pseudo-color (such as red) superimposed on the original grayscale image for the user to visually verify the segmentation effect.

[0108] Step S408: The user saves the quantitative results and effect diagram.

[0109] As can be seen from the above, the method of this embodiment can effectively solve the problems of parameter sensitivity, reliance on manual labor, low accuracy, and poor efficiency in calculating dilution rate, effective deposition rate, and secondary phase content of microstructure. It realizes automated, high-precision, and robust analysis of key quality indicators of additive manufacturing, thereby improving the accuracy of compositional accuracy, forming quality, and microstructure evolution assessment in the additive manufacturing process of functional graded materials.

[0110] Reference Figure 9 This application provides a feature analysis system for additive manufacturing processes based on deep learning. The system includes:

[0111] The image acquisition module is used to acquire images to be analyzed during the additive manufacturing process. These images include macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images.

[0112] The image processing module is used to segment the image to be analyzed using a preset segmentation model to obtain a target region image; wherein, the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure;

[0113] The feature calculation module is used to calculate the deposition layer dilution rate, effective deposition rate, or secondary phase content of microstructure in the additive manufacturing process based on the target area image.

[0114] It is understood that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0115] This application provides a feature analysis system for a deep learning-based additive manufacturing process, the system comprising:

[0116] At least one processor;

[0117] At least one memory for storing at least one program;

[0118] When the at least one program is executed by the at least one processor, the at least one processor implements Figure 1 The method shown.

[0119] It is understood that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0120] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical coding feature maps; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A feature analysis method for additive manufacturing processes based on deep learning, characterized in that, The method includes the following steps: Acquire images to be analyzed during the additive manufacturing process, including macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images; The image to be analyzed is segmented by a preset segmentation model to obtain a target region image; the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure; The deposition layer dilution rate, effective deposition rate, or secondary phase content of the microstructure during the additive manufacturing process are calculated based on the target area image.

2. The feature analysis method according to claim 1, characterized in that, The step of calculating the deposition layer dilution rate during the additive manufacturing process based on the target region image includes: Obtain the positive and negative marker points input from the interactive interface; The cladding area is identified in the molten pool area image based on the positive marker points; The substrate melting region is identified in the cladding region based on the reverse point; The deposition layer dilution rate during the additive manufacturing process is calculated based on the cladding area and the substrate melting area.

3. The feature analysis method according to claim 2, characterized in that, The calculation of the deposition layer dilution rate during the additive manufacturing process based on the cladding area and the substrate melting area includes: Obtain the first pixel area of ​​the cladding region and the second pixel area of ​​the substrate melting region; The first physical area corresponding to the cladding region is calculated based on the preset pixel size calibration coefficient and the area of ​​the first pixel. The second physical area of ​​the substrate melting region is calculated based on the preset pixel size calibration coefficient and the second pixel area. The deposition layer dilution rate during the additive manufacturing process is calculated based on the second physical area and the first physical area.

4. The feature analysis method according to claim 1, characterized in that, The step of calculating the effective deposition rate of additive manufacturing in the additive manufacturing process based on the target area image includes: Obtain the positive marker points input from the interactive interface; The sedimentary layer contour is determined in the sedimentary layer contour image based on the positive marker points; The total area of ​​the sedimentary layer is calculated based on the sedimentary layer profile. The maximum incorporation matrix within the deposition layer profile is determined based on the maximum matrix incorporation algorithm. Calculate the area of ​​the target rectangle corresponding to the maximum inscribed matrix; The effective deposition rate of additive manufacturing is calculated based on the area of ​​the target rectangle and the total area of ​​the deposition layer.

5. The feature analysis method according to claim 4, characterized in that, The algorithm based on maximum matrix incorporation determines the maximum incorporation matrix within the deposition layer profile, including: Obtain the deposition contour polygon within the deposition layer contour; Traverse all pixels within the deposited contour polygon to obtain the area of ​​the candidate matrix for all polygons; The largest inscribed matrix is ​​determined from all polygons based on the area of ​​the candidate matrix.

6. The feature analysis method according to claim 1, characterized in that, The step of calculating the secondary phase content of the microstructure during the additive manufacturing process based on the target region image includes: Convert the secondary phase image of the microstructure into a grayscale image; The grayscale image is converted into a binary image based on an adaptive segmentation algorithm; Perform pixel statistics on the binary image to obtain the total number of white pixels; The content of secondary phase in the microstructure during the additive manufacturing process is calculated based on the total number of white pixels.

7. The feature analysis method according to claim 6, characterized in that, The step of converting the secondary phase image of the microstructure into a grayscale image includes: The secondary phase image of the microstructure is preprocessed to obtain a preprocessed image. The preprocessing process includes contrast enhancement and brightness adjustment. The preprocessed image is converted into a grayscale image.

8. The feature analysis method according to claim 6, characterized in that, The calculation of the secondary phase content of the microstructure in the additive manufacturing process based on the total number of white pixels includes: Calculate the percentage of the total number of white pixels in the secondary phase image of the microstructure; The pixel percentage is used as the content of secondary phase in the microstructure during the additive manufacturing process.

9. A feature analysis system for additive manufacturing processes based on deep learning, characterized in that, The system includes: The image acquisition module is used to acquire images to be analyzed during the additive manufacturing process. The images to be analyzed include macroscopic molten pool images, macroscopic metallographic images of the cross-section of the deposited sample, or microstructure images. The image processing module is used to segment the image to be analyzed using a preset segmentation model to obtain a target region image; the preset segmentation model integrates an image segmentation unit based on a deep convolutional neural network, and the target region image includes a melt pool region image, a sedimentary layer contour image, or a secondary phase image of microstructure; The feature calculation module is used to calculate the deposition layer dilution rate, effective deposition rate, or secondary phase content of microstructure during the additive manufacturing process based on the target area image.

10. A feature analysis system for additive manufacturing processes based on deep learning, characterized in that, The system includes: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method as described in any one of claims 1-8.

Images