Method, device and equipment for detecting broken rod defect of additive manufacturing lattice component

Abstract

The invention discloses an additive manufacturing dot matrix component broken rod defect detection method, device and equipment, and the method comprises the steps: carrying out the CT detection of a dot matrix component, and obtaining a dot matrix component CT reconstruction result; converting a dot matrix component CT reconstruction result into a two-dimensional slice image, wherein the method comprises the following steps: reconstructing to generate a three-dimensional CT object; cutting the three-dimensional CT object to obtain an interested dot matrix area of the detected object; a dot matrix feature extraction algorithm is adopted, and dot matrix rods in the interested dot matrix area are converted into line segment features with the width of a single pixel; on the basis of the extracted dot matrix features, generating two cutting planes which are perpendicular to each other; correcting the three-dimensional CT object by adopting a section cutting plane according to a posture adjustment algorithm; on the basis of the three-dimensional CT object subjected to straightening processing, sectioning parameters are obtained; according to sectioning parameters, sectioning the three-dimensional CT object to generate a plurality of two-dimensional slice images; and performing defect detection and target positioning on the two-dimensional slice image by adopting an intelligent recognition algorithm.

Description

Technical Field

[0001] This application belongs to the technical field of additive manufacturing, and particularly relates to a method, device, and equipment for detecting broken rod defects in additive manufactured lattice components. Background Art

[0002] Additive manufacturing is an advanced manufacturing technology that has developed rapidly in recent years. It has the capabilities of high efficiency, low cost, and near-net shaping of complex parts. When preparing complex three-dimensional lattice components, additive manufacturing technology has obvious technical advantages.

[0003] A lattice component is a component composed of orderly arranged units. The units in the lattice component have various forms, such as BCC (body-centered cubic). Lattice components made of metal materials usually have excellent mechanical properties and can achieve the integration of structure and function. Metal lattice components have structural characteristics such as ultra-light weight, high specific strength, and high specific stiffness, as well as functional characteristics such as heat insulation, fire resistance, and electromagnetic wave absorption, and have broad application prospects in many fields such as aerospace, electronics, and military industry.

[0004] During the additive manufacturing process of lattice components, some defects (including surface defects, internal holes, inaccurate external dimensions, etc.) will occur, which will cause the manufactured lattice components to deviate from the ideal geometric model, thus causing the actual performance of the lattice components to deviate from the expected situation. Broken rod defect is a macroscopic defect in lattice components, and the occurrence of broken rods is related to factors such as additive manufacturing process parameters and internal stress of workpieces. To ensure the service performance of lattice components, strict control of broken rod defects in lattice components is required.

[0005] Currently, the detection of broken rod defects in additive manufactured lattice components mainly relies on CT technology. In the quality evaluation of additive manufactured parts, CT technology is used for defect detection, dimensional measurement, pore analysis, etc. After obtaining the three-dimensional reconstruction result of the part using CT technology, the internal structure and internal defect state of the workpiece can be analyzed and evaluated in detail. However, the analysis of CT detection results currently mainly relies on manual methods: after obtaining the CT reconstruction result of the part to be inspected, professionals use three-dimensional analysis software to observe and judge. The difficulties in defect judgment for the CT detection results of additive manufactured lattice components are concentrated in: inconvenient viewing of three-dimensional data, various shapes and types of defects, and random defect positions. Due to the huge amount of data in CT detection results, the work intensity of manual defect judgment is high. When manually judging defects, the judgment accuracy is significantly affected by the inspectors, which is prone to misjudgment and missed detection, and the consistency of judgment results is poor.

[0006] To address the difficulty of analyzing CT results for lattice components, foreign researchers developed Matlab code to process and analyze CT results and identify broken rod defects in lattice components. The code aligns the CT results with the defined lattice coordinates, then acquires two-dimensional slice images and uses connectivity criteria to identify broken rod defects. However, this method requires a three-dimensional model of the component, making it difficult to align the 3D model with the CT results, especially for large-scale CT inspection data such as lattice models.

[0007] With the continuous development of artificial intelligence technology, people have tried to use neural network methods to perform defect analysis on the CT inspection results of additively manufactured lattice components. Domestic scholars used the YOLO v3 method to analyze the CT images of lattice components and detect and identify two forms of defects (i.e., local protrusions and local breakpoints). Domestic scholars also selected CT inspection images of metal lattice components containing defects, created a defect sample data set, and used the Faster R-CNN model to identify defects in CT results, achieving a high recognition rate. However, the CT results are three-dimensional data and can be sliced from multiple directions and angles, which will result in different grayscale features of the lattice components on the image, thereby affecting the defect recognition effect.

[0008] In summary, after the additive manufacturing lattice components are inspected by CT, there are great technical difficulties in the analysis of the CT inspection results. It relies on manual methods to check the three-dimensional structure one by one, which is time-consuming and labor-intensive. Using the image registration method, the cross-sectional image of the rod diameter can be obtained with the help of the model, but this method requires a three-dimensional model of the inspection object, which has a high computational cost and is not applicable to large-scale lattice components. The use of artificial intelligence to analyze and judge defects in lattice components requires a two-dimensional cross-sectional image that is compatible with the sample set, but the three-dimensional structure ( Figure 1 ) can be viewed from any angle ( Figure 2 ) for sectioning, the sectioned images show great differences due to different sectioning positions, resulting in low accuracy when artificial intelligence applications are used for analysis. Summary of the Invention

[0009] In order to solve the technical problems existing in the prior art, the present application proposes a method, device and equipment for detecting broken rod defects in additively manufactured lattice components.

[0010] This application is implemented through the following technical solutions:

[0011] A method for detecting broken rod defects in an additively manufactured lattice component, comprising:

[0012] Performing CT detection on the lattice components to obtain CT reconstruction results of the lattice components;

[0013] Converting the CT reconstruction result of the dot matrix component into a two-dimensional slice image includes: according to the CT reconstruction result of the dot matrix component, reconstructing and generating a three-dimensional CT object; cropping the three-dimensional CT object to obtain the interested dot matrix region of the detected object; using a dot matrix feature extraction algorithm to convert the dot matrix rods in the interested dot matrix region into line segment features with a single pixel width; generating two mutually perpendicular cutting planes based on the extracted dot matrix features; using the cutting planes to straighten the three-dimensional CT object according to the pose adjustment algorithm; based on the straightened three-dimensional CT object, obtaining cutting parameters; cutting the three-dimensional CT object according to the cutting parameters to generate a number of two-dimensional slice images;

[0014] Using an intelligent recognition algorithm to perform defect detection and target positioning on the two-dimensional slice image.

[0015] In some embodiments, the cropping of the three-dimensional CT object to obtain the interested dot matrix region of the detected object includes:

[0016] The length, width, and height of the interested dot matrix region are:

[0017] blk_x = (1.5 - 3.0) * gx

[0018] blk_y = (1.5 - 3.0) * gy

[0019] blk_z = (1.5 - 2.5) * gz

[0020] blk_x, blk_y, and blk_z are the length, width, and height of the interested dot matrix region respectively; gx is the horizontal side length of the lattice square in the dot matrix component model; gy is the vertical side length of the lattice square in the dot matrix component model; gz is the interval of the repeating unit in the dot matrix component model.

[0021] In some embodiments, the dot matrix feature extraction algorithm includes:

[0022] Binarizing the connecting rods in the interested dot matrix region according to a set threshold;

[0023] Performing morphological erosion processing on the binary image;

[0024] Numbering the binary image and performing feature analysis;

[0025] Screening the entities numbered in binary according to the features of the binary image;

[0026] Performing morphological skeletonization processing on the screened entities.

[0027] In some embodiments, the generating of two mutually perpendicular cutting planes based on the extracted dot matrix features includes:

[0028] Marking points are selected on the extracted lattice features, and multiple position coordinates are selected from two perpendicular directions to generate section planes A and B.

[0029] After the section plane is generated, the control parameters of the section plane are fine-tuned to make the position of the section plane coincide with the position of the lattice rod.

[0030] In some embodiments, the posture adjustment algorithm includes:

[0031] Obtain the characteristics of the section plane, including the origin coordinates, normal vector, u vector and v vector;

[0032] Generate a first transformation matrix according to the characteristics of the section plane A;

[0033] Calculate the angles θ_u and θ_v between the normal vector of section plane B and the u vector and v vector of section plane A respectively;

[0034] Calculate the second transformation matrix of the entity's rotation around the Z axis;

[0035] Calculating a composite transformation matrix based on the first transformation matrix and the second transformation matrix;

[0036] According to the composite transformation matrix, coordinate transformation processing is performed on the three-dimensional CT object.

[0037] In some embodiments, obtaining section parameters based on the straightened 3D CT object includes:

[0038] After aligning the 3D CT object, use the Figure X The X section and the Y section in the YZ coordinate system are used as the section plane A2 and the section plane B2;

[0039] By setting the starting position, ending position and moving interval of the sectioning plane, multiple position parameters that need to be sectioned are obtained.

[0040] In some embodiments, sectioning the three-dimensional CT object according to the sectioning parameters includes:

[0041] According to the sectioning parameters, the positions of the sectioning plane A2 and the sectioning plane B2 are set to obtain a sectioned volume of the three-dimensional CT object;

[0042] By specifying the magnification parameters, the cross-section body is rendered from both the front and back sides to generate the corresponding two-dimensional slice images;

[0043] According to the cutting algorithm, the 3D CT object is output as a series of slice images and saved.

[0044] In some embodiments, the segmentation algorithm comprises:

[0045] Switch to the XY view;

[0046] Moving step: Move the cutting plane A2 to the initial position of the specified slice and the cutting plane B2 to the termination position of the specified slice;

[0047] Cutting step: Cut the three-dimensional solid according to the cutting plane A2 and the cutting plane B2 to obtain a solid slice;

[0048] Display from both the front and the back directions respectively;

[0049] Save the displayed image according to the serial number;

[0050] Move the cutting plane A2 and the cutting plane B2 at intervals, and return to the above cutting step until the cutting process for all slice positions is completed;

[0051] Switch to the YZ view, return to the above moving step, and obtain two-dimensional slice images in the other direction.

[0052] In a second aspect, the present application proposes a defect detection device for a broken rod in an additive manufacturing lattice component, including:

[0053] A CT detection module for performing CT detection on the lattice component to obtain a CT reconstruction result of the lattice component;

[0054] A two-dimensional conversion module for converting the CT reconstruction result of the lattice component into a two-dimensional slice image;

[0055] And an intelligent recognition module that uses an intelligent recognition algorithm to perform defect detection and target positioning on the two-dimensional slice image;

[0056] Among them, the two-dimensional conversion module further includes:

[0057] A reconstruction unit for reconstructing and generating a three-dimensional CT object according to the CT reconstruction result of the lattice component;

[0058] A cutting unit for cutting the three-dimensional CT object to obtain an interested lattice region of the detected object;

[0059] A feature extraction unit that uses a lattice feature extraction algorithm to convert the lattice rods in the interested lattice region into line segment features with a single pixel width;

[0060] A cutting plane generation unit for generating two mutually perpendicular cutting planes based on the extracted lattice features;

[0061] An alignment processing unit that uses the cutting plane to perform alignment processing on the three-dimensional CT object according to the attitude adjustment algorithm;

[0062] A parameter acquisition unit for acquiring cutting parameters based on the three-dimensional CT object after alignment processing;

[0063] And, a sectioning unit sections the three-dimensional CT object according to the sectioning parameters to generate a plurality of two-dimensional slice images.

[0064] In a third aspect, the present application proposes an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements any one of the above-mentioned detection methods when executing the computer program.

[0065] This application proposes a method for detecting broken rod defects in additively manufactured lattice components. Using a posture adjustment algorithm, the lattice component's three-dimensional CT object is accurately and reliably aligned. The lattice component's three-dimensional CT object is then solid-sectioned at intervals and converted into a series of two-dimensional slice images, significantly reducing the amount of analysis data. Finally, an intelligent recognition algorithm is used for defect detection and positioning, effectively improving the efficiency and accuracy of detecting broken rod defects in the lattice component. When converting the lattice component's three-dimensional CT object into two-dimensional slice images, the application uses cut three-dimensional solids, observes and saves them in two directions, rather than using a single slice method to save the image. This ensures accurate setting of the sectioning position and the quality of the sectioned image, further improving the accuracy of defect detection.

[0066] Correspondingly, the device and equipment for detecting broken rod defects in additively manufactured lattice components proposed in this application also have the same technical effects as mentioned above. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The drawings described herein are used to provide a further understanding of the embodiments of the present application, constitute a part of the present application, and do not constitute a limitation of the embodiments of the present application. In the drawings:

[0068] Figure 1 is the reconstructed 3D CT object;

[0069] Figure 2 Cut images of three-dimensional entities at different angles;

[0070] Figure 3 A schematic diagram of the defect detection method flow chart proposed in an embodiment of the present application;

[0071] Figure 4 An example of the obtained region of interest.

[0072] Figure 5 is an example of the extracted lattice features;

[0073] Figure 6 is the original three-dimensional image of the point array area of interest;

[0074] Figure 7 For Figure 6 The three-dimensional image after the image is binarized;

[0075] Figure 8 It is the three-dimensional image after morphological corrosion processing;

[0076] Figure 9 is a numbered three-dimensional image;

[0077] Figure 10 is the three-dimensional image after screening;

[0078] Figure 11 Examples of generated section planes A and B are shown below;

[0079] Figure 12 This is an example of aligning a straightened 3D object with the coordinate system;

[0080] Figure 13 The following are examples of selected section planes A2 and B2;

[0081] Figure 14 An example of the position of the section plane of a sectioned body;

[0082] Figure 15 This is an example of a two-dimensional slice image;

[0083] Figure 16 Examples of defect detection and target location results;

[0084] Figure 17 This is a block diagram of the principle of the defect detection device proposed in the embodiment of the present application;

[0085] Figure 18 A schematic diagram of the principle of an electronic device proposed in an embodiment of the present application;

[0086] Figure 19 The two-dimensional slice image of the acquired lattice component;

[0087] Figure 20 Identification results of the broken rod defect in the lattice component;

[0088] Reference numerals and corresponding component names:

[0089] 200-Detection device, 201-CT detection module, 202-Two-dimensional conversion module, 203-Intelligent recognition module, 310-Reconstruction unit, 311-Cut unit, 312-Feature extraction unit, 313-Section plane generation unit, 314-Straightening processing unit, 315-Parameter acquisition unit, 316-Section unit, 400-Electronic device, 410-Memory, 420-Processor, 411-Computer program. DETAILED DESCRIPTION

[0090] Hereinafter, the term "comprising" or "may comprise" that may be used in various embodiments of the present application indicates the presence of an invented function, operation, or element, and does not limit the addition of one or more functions, operations, or elements. Further, as used in various embodiments of the present application, the terms "comprising", "having", and their cognates are only intended to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing items, and should not be construed as precluding the existence or addition of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing items first.

[0091] In various embodiments of the present application, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the recited words. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0092] Expressions (such as "first", "second", etc.) used in various embodiments of the present application may modify various constituent elements in the various embodiments, but do not limit the corresponding constituent elements. For example, the foregoing expressions do not limit the order and / or importance of the elements. The foregoing expressions are only for the purpose of distinguishing one element from other elements. For example, a first user device and a second user device indicate different user devices, although both are user devices. For example, without departing from the scope of various embodiments of the present application, the first element may be referred to as the second element, and similarly, the second element may also be referred to as the first element.

[0093] It should be noted that: if it is described that one constituent element is "connected" to another constituent element, the first constituent element may be directly connected to the second constituent element, and a third constituent element may be "connected" between the first constituent element and the second constituent element. Conversely, when one constituent element is "directly connected" to another constituent element, it can be understood that there is no third constituent element between the first constituent element and the second constituent element.

[0094] The terms used in various embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the various embodiments of the present application. As used herein, the singular form is also intended to include the plural form unless the context clearly indicates otherwise. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the present application belong. The terms (such as those defined in a commonly used dictionary) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning unless clearly defined in the various embodiments of the present application.

[0095] To make the objectives, technical solutions and advantages of this application more clear and understandable, the following further details this application in combination with embodiments and the accompanying drawings. The illustrative embodiments of this application and their descriptions are only used to explain this application and are not intended to limit this application.

[0096] Embodiment 1

[0097] This embodiment proposes a method for detecting broken rod defects in additively manufactured lattice components. As Figure 3 shown, the method proposed in this embodiment includes the following steps:

[0098] Step S1: Perform CT detection on the lattice component to obtain the CT reconstruction result of the lattice component.

[0099] It can be understood that performing CT detection on the lattice component belongs to the prior art. First, according to technical requirements (such as the detection area, detection resolution, etc.), the parameters for performing CT detection on the lattice component are determined. Then, by setting the parameters of ray detection (such as the tube voltage of the ray tube, tube current, exposure time of the detector), CT scan parameters (such as the number of projections, the distance from the ray source to the detector, object magnification), and CT reconstruction parameters (such as the size of the reconstruction matrix, filter parameters used), CT detection is performed on the lattice component, and the CT reconstruction result is obtained.

[0100] Step S2: Convert the CT reconstruction result of the lattice component into a two-dimensional slice image.

[0101] Optionally, the specific implementation process of this step S2 is as follows:

[0102] Step S2.1: Reconstruct the three-dimensional CT object: Read a series of CT reconstruction results (such as files in.raw format) into analysis and processing software according to specified parameters (image length and width, color depth, voxel size, etc.), and reconstruct and generate a three-dimensional CT object, as Figure 1 shown. It can be understood that the analysis and processing software can use existing commercial software Avizo, or can also use self-written software, such as software developed based on VTK, namely The Visualization Toolkit.

[0103] Step S2.2: Crop the three-dimensional CT object: In the analysis and processing software, specify a smaller three-dimensional cuboid area in an automatic or manual manner to crop the three-dimensional CT object to obtain the lattice area of interest of the detected object, as Figure 4 shown.

[0104] In the lattice component model, the horizontal side length of the lattice square is gx, the vertical side length of the lattice square is gy, and the interval of the repeating unit in the lattice component model is gz. Then, the length, width, and height of the lattice cutting region (blk_x, blk_y, and blk_z respectively) are as follows:

[0105] blk_x = (1.5 - 3.0) * gx

[0106] blk_y = (1.5 - 3.0) * gy

[0107] blk_z = (1.5 - 2.5) * gz

[0108] Step S2.3, extracting lattice features: Using the lattice feature extraction algorithm, convert the lattice rods in the lattice region of interest into line segment features with a single pixel width, as Figure 5 shown.

[0109] Among them, the lattice feature extraction algorithm is as follows:

[0110] Step a1, thresholding. Automatically set the threshold and perform binary processing on the connecting rods in the lattice region of interest (as Figure 6 shown) to obtain a binary three-dimensional image as Figure 7 shown.

[0111] Step b1, morphological erosion processing. According to the rod diameter value and the CT image resolution, set the neighborhood parameters (8, 18, or 26) for the erosion processing. It can be understood that the morphological erosion processing is a spatial operation based on the structural element, used to eliminate the boundary voxels of the binary object and shrink the object. The basic principle is to define a three-dimensional small cube as the structural element, and then traverse each voxel in the three-dimensional image and use the structural element for analysis and processing: if all the neighborhood points in the structural element are 1, then retain the voxel; otherwise, set it as the background. The three-dimensional image after the morphological erosion processing is as Figure 8 shown.

[0112] Step c1, binary image numbering processing. Analyze the features such as the three-dimensional volume and the central coordinates of the binary image. It can be understood that the numbering processing is to assign a unique identifier to different regions for distinguishing different structures. Since the lattice structure is composed of multiple connecting rods, after the numbering processing, the main body can be easily extracted with the help of the 3D volume index and further analyzed. Perform numbering processing on the Figure 8 shown image to obtain the result as Figure 9 shown. Figure 9 The volume of the connecting rod main body in is 36.9, which is much larger than the volumes of No. 2 - No. 7 (corresponding to the separated lattice rods, that is, the Figure 9 annotation boxes shown in).

[0113] Step d1: Filter the binary - numbered entities based on features such as three - dimensional volume. It can be understood that after numbering, the 3D volume of each region can be calculated. On this basis, a threshold (for example, 5) is set, and regions with a 3D volume greater than the threshold can be filtered out. At this time Figure 9 the regions of the upper bounding box are kicked out, that is, regions smaller than the threshold are removed, and the result is as shown in Figure 10 the figure shown.

[0114] Step e1: Perform morphological skeletonization on the filtered entities.

[0115] Step S2.4: Generate two mutually perpendicular cross - sectional planes based on the extracted dot - matrix features.

[0116] In this step S2.4, marker points are selected on the extracted dot - matrix features. More than three position coordinates are selected to determine the cross - sectional planes of the dot - matrix rods. For the dot - matrix of the BCC dot - matrix structure (BCCZ, regularBcc, dualDensityBcc, etc.), points for defining the cross - sectional planes are selected from two perpendicular directions respectively to generate cross - sectional plane A and cross - sectional plane B, as shown in Figure 11 the figure shown.

[0117] After generating the cross - sectional planes, the control parameters of the cross - sectional planes are fine - tuned to make the positions of the cross - sectional planes fit well with the positions of the dot - matrix rods.

[0118] Step S2.5: Use the cross - sectional planes to straighten the three - dimensional CT object according to the pose adjustment algorithm.

[0119] Among them, the pose adjustment algorithm is as follows:

[0120] Step a2: Obtain the features of the cross - sectional plane, including the origin (x0, y0, z0), the normal vector (nx, ny, nz), the u - vector (ux, uy, uz), and the v - vector (vx, vy, vz). Generate transformation matrix A from the features of cross - sectional plane A.

[0121]

[0122] Step b2: Calculate the angles θ_u and θ_v between the normal vector of cross - sectional plane B and the u - vector and v - vector of cross - sectional plane A.

[0123] Step c2: Calculate the transformation matrix B for the entity to rotate around the Z - axis.

[0124] Generate the transformation matrix B for the entity to rotate around the Z - axis at an angle of θ=(θ_u - 90).

[0125] If θ_u is greater than 90 degrees, then set θ=(90 - θ_v).

[0126]

[0127] Step d2, calculate the composite transformation matrix T.

[0128] T=A×B.

[0129] Step e2: performing coordinate transformation processing on the 3D CT object according to the composite transformation matrix T.

[0130] After the above alignment process, the dot matrix direction of the three-dimensional object will be consistent with the visual Figure X Align the YZ coordinate system, that is, align the dot matrix unit square with the viewing Figure X The XY axis in the YZ coordinate system is aligned as Figure 12 shown.

[0131] Step S2.6: Obtaining section parameters based on the straightened 3D CT object.

[0132] After aligning the 3D CT object, use the Figure X The X section (X-axis section) and the Y section (Y-axis section) in the YZ coordinate system are used as the section plane A2 and the section plane B2, as shown in FIG. Figure 13 By setting the starting position, ending position and moving interval of the sectioning plane, multiple position parameters that need to be sectioned are obtained.

[0133] Step S2.7, sectioning the 3D CT object (after alignment) according to the sectioning parameters.

[0134] By setting the positions of the section plane A2 and the section plane B2 using the section parameters determined in the previous step, a section volume of the 3D CT object is obtained, such as Figure 14 shown.

[0135] By specifying the magnification parameters, the section body is rendered from the front and back sides (i.e., two opposite directions) to generate the corresponding two-dimensional slice images, such as Figure 15 shown.

[0136] According to the sectioning algorithm, the three-dimensional CT object is output as a series of two-dimensional slice images and saved in a specific format (such as .png).

[0137] The specific cutting algorithm is:

[0138] Step a3, switch to XY view;

[0139] Step b3, moving the cutting plane A2 to the initial position of the designated slice, and the cutting plane B2 to the end position of the designated slice;

[0140] Step c3, cutting the three-dimensional solid along the cutting plane A2 and the cutting plane B2 to obtain solid slices;

[0141] Step d3, displaying from both the forward and reverse directions;

[0142] Step e3, save the displayed images in order of sequence number;

[0143] Step f3, moving the section plane A2 and the section plane B2 at intervals, repeating the operations from step c3 to step e3, and processing all the slice positions;

[0144] Step g3, switch to the YZ view, and follow steps b3 to f3 to obtain the lattice component slice image.

[0145] Step S3: Using intelligent recognition algorithms, perform defect detection and target positioning on the two-dimensional slice image.

[0146] It is understandable that the intelligent recognition algorithm of this embodiment can use existing artificial intelligence models to automatically identify and locate broken rod defects in two-dimensional slice images, such as Figure 16 As shown in the figure, artificial intelligence models can use deep learning models, such as two-stage object detection models: Fast-RCNN (Region-based Convolutional Neural Networks, RCNN) and Faster-RCNN; single-stage object detection models: Yolo (You Only Look Once) and SSD (Single Shot multi-box Detection).

[0147] The detection method proposed in this embodiment, by performing straightening operations and sectioning processing on the three-dimensional entity, not only avoids the complex calculation of model matching of the three-dimensional entity, but also avoids the random features in the sectioned image and the one-sidedness of the image features, thereby ensuring the accurate setting of the sectioning position and the quality of the sectioned image, while greatly reducing the amount of analyzed data and improving the detection efficiency; and combines with the intelligent recognition model to perform automated detection of broken rod defects, solving the problem of large workload when manually observing and judging the dot matrix rods one by one.

[0148] In another embodiment, this embodiment also provides a device 200 for detecting defects in broken rods of additively manufactured lattice components, such as Figure 17 As shown, the detection device 200 includes:

[0149] The CT detection module 201 is used to perform CT detection on the lattice components to obtain CT reconstruction results of the lattice components. The specific detection method is as described in the above method and will not be repeated here.

[0150] The two-dimensional conversion module 202 is used to convert the CT reconstruction results of the lattice components into two-dimensional slice images.

[0151] And, an intelligent recognition module 203, which uses an intelligent recognition algorithm to perform defect detection and target positioning on two-dimensional slice images. The specific recognition process is as described in the above method and will not be elaborated here.

[0152] Optionally, the two-dimensional conversion module 202 further includes:

[0153] A reconstruction unit 310, configured to reconstruct a three-dimensional CT object: read a series of CT reconstruction results (such as files in.raw format) into analysis and processing software according to specified parameters (image length and width, color depth, voxel size, etc.) to reconstruct and generate a three-dimensional CT object.

[0154] A cutting unit 311, configured to cut the three-dimensional CT object: in the analysis and processing software, specify a smaller three-dimensional cuboid area in an automatic or manual manner to cut the three-dimensional CT object to obtain an interested dot matrix area of the detected object. The specific cutting method is as described in the above method and will not be elaborated here.

[0155] A feature extraction unit 312, configured to extract dot matrix features: use a dot matrix feature extraction algorithm to convert the dot matrix rods in the interested dot matrix area into line segments with a single pixel width. The dot matrix feature extraction algorithm used is as described in the above method and will not be elaborated here.

[0156] A cross-section generation unit 313, which generates two mutually perpendicular cross-sections based on the extracted dot matrix features. The specific cross-section generation method is as described in the above method and will not be elaborated here.

[0157] An alignment processing unit 314, which uses the initial cross-section to perform alignment processing on the three-dimensional CT object according to an attitude adjustment algorithm. The attitude adjustment algorithm used is as described in the above method and will not be elaborated here.

[0158] A parameter acquisition unit 315, which acquires cutting parameters based on the three-dimensional CT object after alignment processing. The method for acquiring cutting parameters is as described in the above method and will not be elaborated here.

[0159] And, a cutting unit 316, which cuts the three-dimensional CT object according to the cutting parameters. The specific cutting algorithm is as described in the above method and will not be elaborated here.

[0160] In another embodiment, this embodiment also proposes an electronic device 400, as Figure 18 shown. The electronic device 400 includes: a memory 410, a processor 420, and a computer program 411 stored on the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented:

[0161] Perform CT detection on the dot matrix component to obtain the CT reconstruction result of the dot matrix component;

[0162] Converting CT reconstruction results of lattice components into two-dimensional slice images, specifically comprising: reconstructing a three-dimensional CT object based on the CT reconstruction results; cutting the three-dimensional CT object to obtain a lattice region of interest of the detected object; converting the lattice rods in the lattice region of interest into line segments with a single pixel width using a lattice feature extraction algorithm; generating two mutually perpendicular section planes based on the extracted lattice features; using the section planes and a posture adjustment algorithm to straighten the three-dimensional CT object; obtaining section parameters based on the straightened three-dimensional CT object; and sectioning the three-dimensional CT object according to the section parameters to generate a plurality of two-dimensional slice images;

[0163] Intelligent recognition algorithm is used to perform defect detection and target positioning on two-dimensional slice images.

[0164] Optionally, when the processor 420 executes the computer program 411 , any implementation method in the corresponding embodiment of the above detection method can be implemented.

[0165] It should be noted that the electronic device proposed in this embodiment is a device used to implement the above-mentioned detection method. Therefore, based on the above-mentioned detection method proposed in this embodiment, technical personnel in this field can understand the specific implementation method of the electronic device of this embodiment and its various variations. Therefore, how the electronic device specifically implements the above-mentioned detection method will not be introduced in detail here. As long as the electronic device used by technical personnel in this field to implement the above-mentioned detection method falls within the scope of protection to be protected by this application.

[0166] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0167] Example 2:

[0168] This embodiment Figure 1 The CT scan results for defects in an additively manufactured lattice component are 1027*1013*1038 matrix data. Manual analysis of broken rods within the lattice component requires manually positioning the object in the appropriate position within the 3D analysis software and then viewing it from two viewing angles. This requires reviewing 1027 + 1038 = 2065 images. Furthermore, the images contain a large number of rod diameters, requiring human eyes to individually examine and determine if any broken rods are present. This makes the analysis process time-consuming and labor-intensive, and accuracy is difficult to guarantee.

[0169] By using the detection method proposed in the above embodiment 1, it is possible to conveniently realize the alignment and sectioning of a three-dimensional entity, and obtain 42 two-dimensional slice images, such as Figure 19As shown, this greatly reduces the amount of data for detection and analysis (about 1 / 49). At the same time, an artificial intelligence model is adopted, and the automatic recognition of broken pole defects in the image can be completed within 1 minute, such as Figure 20 as shown, thus solving the problem that the analysis of the CT detection results of lattice components is time-consuming and laborious.

[0170] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0171] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0172] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0173] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0174] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present application. It should be understood that the above description is only the specific embodiments of the present application and is not used to limit the protection scope of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A method for detecting broken rod defects in additive manufacturing lattice components, characterized in that, Including: Performing CT detection on the dot matrix component to obtain the CT reconstruction result of the dot matrix component; Converting the CT reconstruction result of the dot matrix component into two-dimensional slice images, including: reconstructing and generating a three-dimensional CT object according to the CT reconstruction result of the dot matrix component; cropping the three-dimensional CT object to obtain the dot matrix region of interest of the detected object; using a dot matrix feature extraction algorithm to convert the dot matrix rods in the dot matrix region of interest into line segment features with a single pixel width; generating two mutually perpendicular cutting planes based on the extracted dot matrix features; using the cutting planes to straighten the three-dimensional CT object according to the pose adjustment algorithm; obtaining cutting parameters based on the straightened three-dimensional CT object; cutting the three-dimensional CT object according to the cutting parameters to generate a plurality of two-dimensional slice images; Using an intelligent recognition algorithm to perform defect detection and target positioning on the two-dimensional slice images.

2. The defect detection method for the broken rod of the additive manufacturing lattice component according to claim 1, characterized in that, The cropping of the three-dimensional CT object to obtain the dot matrix region of interest of the detected object includes: The length, width, and height of the dot matrix region of interest are: blk_x = (1.5 - 3.0) * gx blk_y = (1.5 - 3.0) * gy blk_z = (1.5 - 2.5) * gz blk_x, blk_y, and blk_z are the length, width, and height of the dot matrix region of interest respectively; gx is the horizontal side length of the lattice square in the dot matrix component model; gy is the vertical side length of the lattice square in the dot matrix component model; gz is the interval of the repeating unit in the dot matrix component model.

3. The method for detecting the broken rod defect of the additive manufacturing lattice component according to claim 1, wherein, The dot matrix feature extraction algorithm includes: Performing binarization processing on the connecting rods in the dot matrix region of interest according to a set threshold; Performing morphological erosion processing on the binary image; Performing numbering processing and feature analysis on the binary image; Screening the entities numbered by binarization according to the features of the binary image; Performing morphological skeletonization processing on the screened entities.

4. The defect detection method for the broken rod of the additive manufacturing lattice component according to claim 1, characterized in that, The generating of two mutually perpendicular cutting planes based on the extracted dot matrix features includes: Selecting marking points on the extracted dot matrix features, and respectively selecting multiple position coordinates from two perpendicular directions to generate cutting plane A and cutting plane B; After generating the cutting planes, finely adjusting the control parameters of the cutting planes to make the positions of the cutting planes coincide with the positions of the dot matrix rods.

5. The defect detection method for the broken rod of the additive manufacturing lattice component according to claim 1, characterized in that, The pose adjustment algorithm includes: Obtaining the features of the cutting plane, including the origin coordinates, normal vector, u vector, and v vector; Generating a first transformation matrix according to the features of cutting plane A; Calculating the angles θ_u and θ_v between the normal vector of cutting plane B and the u vector and v vector of cutting plane A respectively; Calculating a second transformation matrix for the entity to rotate around the Z axis; Calculating a composite transformation matrix according to the first transformation matrix and the second transformation matrix; Performing coordinate transformation processing on the three-dimensional CT object according to the composite transformation matrix.

6. A method for detecting broken rod defects in an additive manufacturing lattice component according to any one of claims 1-5, characterized in that The obtaining of the cutting parameters based on the straightened three-dimensional CT object includes: After straightening the three-dimensional CT object, using the X section and Y section in the view XYZ coordinate system as cutting plane A2 and cutting plane B2; Obtaining a plurality of position parameters to be subjected to cutting processing by setting the starting position, ending position, and moving interval of the cutting plane.

7. The method for detecting the broken rod defect of the additive manufacturing lattice component according to claim 6, characterized in that Slicing the three-dimensional CT object according to the slicing parameters includes: Setting the positions of cutting plane A2 and cutting plane B2 according to the slicing parameters to obtain a certain sliced body of the three-dimensional CT object; Rendering the sliced body from both the front and back by specifying the magnification parameter to generate corresponding two-dimensional slice images; Outputting the three-dimensional CT object as a series of slice images according to the slicing algorithm and saving them.

8. The method for detecting the defect of the broken rod in the additive manufacturing lattice component according to claim 7, characterized in that, The slicing algorithm includes: Switching to the XY view; Moving step: Moving cutting plane A2 to the initial position of the specified slice and cutting plane B2 to the termination position of the specified slice; Slicing step: Slicing the three-dimensional entity according to cutting plane A2 and cutting plane B2 to obtain entity slices; Displaying separately from both the front and back directions; Saving the displayed images according to the serial numbers; Moving cutting plane A2 and cutting plane B2 at intervals and returning to the above slicing step until the slicing process for all slice positions is completed; Switching to the YZ view and returning to the above moving step to obtain two-dimensional slice images in another direction.

9. An additive manufacturing lattice component broken rod defect detection device, characterized in that, Including: A CT detection module for performing CT detection on the lattice component to obtain the CT reconstruction result of the lattice component; A two-dimensional conversion module for converting the CT reconstruction result of the lattice component into two-dimensional slice images; And an intelligent recognition module that uses an intelligent recognition algorithm to perform defect detection and target positioning on the two-dimensional slice images; Wherein, the two-dimensional conversion module further includes: A reconstruction unit that reconstructs and generates a three-dimensional CT object according to the CT reconstruction result of the lattice component; A cutting unit for cutting the three-dimensional CT object to obtain the interested lattice region of the detected object; A feature extraction unit that uses a lattice feature extraction algorithm to convert the lattice rods in the interested lattice region into line segment features with a single pixel width; A cutting plane generation unit that generates two mutually perpendicular cutting planes based on the extracted lattice features; An attitude adjustment unit that uses the cutting planes to perform attitude adjustment on the three-dimensional CT object according to the attitude adjustment algorithm; A parameter acquisition unit that acquires slicing parameters based on the three-dimensional CT object after attitude adjustment; And a slicing unit that slices the three-dimensional CT object according to the slicing parameters to generate a plurality of two-dimensional slice images.

10. An electronic device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the detection method described in any one of claims 1-8.

Images