Real-time defect detection method and system for selective laser melting 3D printing

Abstract

The invention provides a real-time defect detection method and system for selective laser melting 3D printing. The real-time defect detection method comprises the steps that a corresponding defect database is called based on material attributes of a 3D printing finished product to construct a material defect feature set, and an image collection moment is determined for the real-time printing progress of a current printing layer based on a laser scanning array; the method comprises the following steps: acquiring a complete plane image of a current printing layer, identifying a plurality of current material defects contained in the complete plane image, generating defect information of the current printing layer, establishing a 3D image of a 3D printing finished product, and identifying a defect attribute corresponding to each printing part in the corresponding 3D image by utilizing the defect information. A defect report is generated and displayed, the type and size of the defect are automatically recorded in the printing process, labeling is conducted in combination with three-dimensional slicing software, and technical reference is provided for a user.

Description

Technical Field

[0001] The present invention relates to the field of defect recognition technology, and in particular to a real-time defect detection method and system for selective laser melting 3D printing. Background Art

[0002] SLM technology is developed on the basis of SLS. SLM technology needs to completely melt the metal powder and directly form metal parts. Therefore, a high-power density laser is required. Before the laser beam starts scanning, the horizontal powder roller first spreads the metal powder on the substrate of the processing chamber. Then the laser beam will selectively melt the powder on the substrate according to the contour information of the current layer, process the contour of the current layer, and then the lifting system will drop a layer thickness. The rolling powder roller will spread the metal powder on the processed current layer, and the equipment will be transferred to the next layer for processing. This process is repeated layer by layer until the entire part is processed. However, SLM technology is the same as all 3D printing technologies. Selective laser melting is also processed layer by layer until the entire part is processed. When all the processing work is completed, it is extremely difficult to find out whether there are defects inside the part. For example, since the process of laser melting powder is similar to welding, just like welding will have "false welding", in the process of 3D printing, some powders will not be completely melted. How to detect defects in the printed product during the printing process has become an urgent problem to be solved.

[0003] Therefore, the present invention provides a real-time defect detection method and system for selective laser melting 3D printing. Summary of the invention

[0004] The present invention is a real-time defect detection method and system for selective laser melting 3D printing. By adding a machine vision system to the laser scanning array, a plane image of each layer is quickly captured before the next layer of powder is laid after each layer is printed, and then the image is analyzed to determine the printing defects of the layer. After printing is completed, an internal defect map of each part is generated, and whether the part is usable is determined according to the requirements of the demander.

[0005] The present invention provides a real-time defect detection method for selective laser melting 3D printing, comprising: Step 1: Retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; Step 2: determining the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtaining a complete plane image of the current printing layer; Step 3: using the material defect feature set to identify a number of current material defects contained in the complete plane image, performing depth identification on each of the current material defects, and generating defect information of the current printing layer; Step 4: Create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

[0006] In one practicable manner, The step 1 comprises: Step 11: Determine the printing raw material of the 3D printed product based on the current printing task, and search for a defect database corresponding to the material property in big data based on the material property of the printing raw material; Step 12: Draw a defect image of the printing raw material using the defect data, and determine defect features corresponding to each defect data; Step 13: Perform cluster analysis on the defect features to generate a material defect feature set corresponding to this printing task and display it.

[0007] In one practicable manner, The step 12 comprises: Step 121: generating a flawless image of the printing raw material based on the material properties, and obtaining a data difference between each defect data and the flawless data corresponding to the flawless image; Step 122: performing corresponding defect adjustment training on the flawless image using the data difference to obtain a defect image corresponding to each defect data; Step 123: Identify the image presentation features corresponding to each of the defective images respectively, and generate defect features corresponding to the defect data.

[0008] In one practicable manner, The step 13 comprises: Step 131: performing cluster analysis on the defect features to obtain a plurality of defect clusters, identifying the defect weight of each defect feature in the corresponding defect cluster, and reorganizing the corresponding defect features based on the defect weight to obtain a cluster defect feature corresponding to each defect cluster; Step 132: Draw a virtual print product map according to the current print task, map each cluster defect feature in the virtual print product map to obtain several material defect features of the current print task, generate a material defect feature set corresponding to the current print task and display it.

[0009] In one practicable manner, Also includes: Obtain the indoor environment information of the 3D printing processing room; When the indoor environment information meets the specified printing requirements, the laser scanning array is controlled to perform printing.

[0010] In one practicable manner, The step 2 comprises: Step 21: determining a number of layers to be printed of the 3D printed product according to the current printing task, and determining the contour to be printed corresponding to each of the layers to be printed, generating a number of printing guide instructions to control the laser scanning array to perform real-time printing, and determining the current printing layer according to the real-time printing data of the laser scanning array; Step 22: determining the complete printing state of the current printing layer based on the current contour to be printed corresponding to the current printing layer, constructing the real-time printing state of the current printing layer using the real-time printing data, and determining the real-time printing progress of the current printing layer; Step 23: When the real-time printing progress is 100%, and the displayed printing status of the current printing layer is consistent with the complete printing status, it is determined that the current printing layer has completed printing, and the preset liquid lens is controlled to collect and display the complete plane image of the current printing layer.

[0011] In one practicable manner, The step 3 comprises: Step 31: grayscale enhancement is performed on the print contour contained in the complete plane image to obtain enhanced contour information of the current print layer, straight line fitting is performed on the enhanced contour information in the complete plane image, and a plurality of contour key points are located in the print contour according to the fitting result; Step 32: Divide the printed contour into a plurality of contour regions, establish a regional contour feature corresponding to each contour region based on the distribution of contour key points corresponding to each contour region, and search for relevant material defect features corresponding to each regional contour feature in the material defect feature set; Step 33: Map each of the relevant material defect features in the corresponding contour area, obtain the point overlap ratio between each contour key point and the corresponding relevant material defect feature, and select the target relevant material defects with a point overlap ratio higher than a specified ratio as the current material defects corresponding to the contour area; Step 34: Based on the point overlap ratio corresponding to each current material defect, feature recognition is performed on the complete plane image to obtain defect interference information of each current material defect on the printing contour, obtain several defect appearances of the complete plane image and defect specifications corresponding to each defect appearance, and generate defect information of the current printing layer.

[0012] In one practicable manner, The step 4 comprises: Step 41: establishing a 3D stereoscopic image according to a plurality of complete plane images corresponding to the 3D printed product, and matching a corresponding stereoscopic image area for each defect information based on the corresponding relationship between the defect information and the complete plane image; Step 42: identifying a plurality of printed parts contained in the 3D stereoscopic image, determining a stereoscopic image region corresponding to each of the printed parts, and marking each of the defect information in the corresponding stereoscopic image region to obtain a part defect feature corresponding to each of the printed parts; Step 43: Mark each of the part defect features in the 3D stereoscopic image to obtain complete part information corresponding to each of the printed parts, analyze defect interference information between different printed parts in the 3D stereoscopic image, and establish defect attributes corresponding to each of the printed parts in combination with the part position corresponding to each of the printed parts; Step 44: Perform an overall defect evaluation on the 3D printed product according to the defect attributes corresponding to each of the printed parts, perform an overall function evaluation on the 3D printed product according to the part functions corresponding to each of the printed parts, generate a defect report for the 3D printed product, and display it.

[0013] In one practicable manner, Also includes: Screening defective 3D printed products that fail the overall defect evaluation or the overall function evaluation; Several defect attributes corresponding to the defective 3D printed product are respectively marked in the corresponding defective 3D stereoscopic image, and a multi-dimensional defect map of the defective 3D printed product is generated and displayed.

[0014] The present invention provides a real-time defect detection system for selective laser melting 3D printing, comprising: A defect presetting module is used to retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; A synchronous monitoring module, used to determine the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtain a complete plane image of the current printing layer; A defect recognition module, used to use the material defect feature set to identify a plurality of current material defects contained in the complete plane image, perform depth recognition on each of the current material defects, and generate defect information of the current printing layer; The report generation module is used to create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

[0015] The achievable beneficial effects of the above technical solution are: In order to solve the defects in the prior art, a method for real-time defect detection is constructed. Before printing, the corresponding defect database is screened according to the material properties of the raw materials used this time, so as to construct the material defect feature set for this printing. Then, after completing one layer of printing, its complete plane image is collected, and the defect information of the current printing layer is determined by identifying the current material defects in the complete plane image and performing depth recognition on it. Finally, the defect properties of each printed part are analyzed by constructing a 3D stereo image, so as to generate a defect report for the 3D printed product. In this way, real-time detection can be performed during the printing process, the defects in the finished product can be determined, and a defect report can be generated for reference by the demander. The demander can perform different processing on the printed product according to actual needs, and various printing defects can be identified during the printing process, which is convenient for manufacturers to discover generation problems in time and deal with them, thereby promoting the development of productivity.

[0016] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0017] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 Schematic diagram of the workflow of the real-time defect detection method for selective laser melting 3D printing in an embodiment of the present invention; Figure 2Schematic diagram of the composition of a real-time defect detection system for selective laser melting 3D printing in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0020] Example 1 This embodiment provides a real-time defect detection method for selective laser melting 3D printing, such as Figure 1 As shown, including: Step 1: Retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; Step 2: determining the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtaining a complete plane image of the current printing layer; Step 3: using the material defect feature set to identify a number of current material defects contained in the complete plane image, performing depth identification on each of the current material defects, and generating defect information of the current printing layer; Step 4: Create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

[0021] In this example, the material attribute indicates the attribute of the raw material used in printing, which is generally a powdered raw material; In this example, the defect database represents data information about defects that may occur after the finished product is constructed using the raw material; In this example, the material defect feature set represents the features of defects that may be generated after the 3D printed product is constructed from the raw material; In this example, the current printing layer indicates the layer currently being 3D printed; In this example, the image acquisition time is: the time corresponding to when the real-time printing progress of a current printing layer is 100%; In this example, the complete plane image represents the result of collecting all image information of the current printing layer; In this example, the current material defect represents a defect present in the complete planar image; In this example, printed parts represent individual parts used to make up the finished 3D printed product; In this example, the defect attribute represents the attribute of the defect presented by the printed part.

[0022] The working principle and beneficial effects of the above technical solution: In order to solve the defects in the prior art, a method for real-time defect detection is constructed. Before printing, the corresponding defect database is screened according to the material properties of the raw materials used this time, so as to construct a material defect feature set for this printing. Then, after completing a layer of printing, its complete plane image is collected, and the defect information of the current printing layer is determined by identifying the current material defects in the complete plane image and performing depth identification on it. Finally, the defect properties of each printed part are analyzed by constructing a 3D stereoscopic image, so as to generate a defect report for the 3D printed product. In this way, real-time detection can be performed during the printing process, the defects in the finished product can be determined, and a defect report can be generated for reference by the demander. The demander can perform different processing on the printed product according to actual needs, and various printing defects can be identified during the printing process, which is convenient for manufacturers to promptly discover generation problems and deal with them, thereby promoting the development of productivity.

[0023] Example 2 On the basis of Example 1, the real-time defect detection method for selective laser melting 3D printing, the step 1, comprises: Step 11: Determine the printing raw material of the 3D printed product based on the current printing task, and search for a defect database corresponding to the material property in big data based on the material property of the printing raw material; Step 12: Draw a defect image of the printing raw material using the defect data, and determine defect features corresponding to each defect data; Step 13: Perform cluster analysis on the defect features to generate a material defect feature set corresponding to this printing task and display it.

[0024] The working principle and beneficial effects of the above technical solution are as follows: by searching the defect database of printing raw materials in big data and then drawing the corresponding defect images to analyze the defect characteristics corresponding to each defect data, the material defect feature set of this printing task is constructed. In this way, a highly referenced, high-quality defect comparison set can be provided for this printing work, which is convenient for subsequent defect identification.

[0025] Example 3 On the basis of Example 2, the real-time defect detection method for selective laser melting 3D printing, the step 12 comprises: Step 121: generating a flawless image of the printing raw material based on the material properties, and obtaining a data difference between each defect data and the flawless data corresponding to the flawless image; Step 122: performing corresponding defect adjustment training on the flawless image using the data difference to obtain a defect image corresponding to each defect data; Step 123: Identify the image presentation features corresponding to each of the defective images respectively, and generate defect features corresponding to the defect data.

[0026] In this example, the flawless image refers to an image of an object without defects constructed using printing raw materials, and the flawless image can be generated using AI technology; In this example, the data difference represents the different data segments between the defective data and the flawless data; In this example, defect adjustment training refers to the process of adjusting the flawless data to be consistent with the defective data by using the data difference; In this example, the image presentation feature refers to a feature related to the defect presented in the defect image.

[0027] The working principle and beneficial effects of the above technical solution: in order to better identify the defect features corresponding to the defect data, a flawless image is constructed using material properties, and defect adjustment training is performed on the flawless image based on the data difference between the defect data and the flawless data, so that the defect image corresponding to each defect data is obtained, and the defect features of each defect number are determined by identifying the image presentation features corresponding to each defect image. In this way, the defect features of multiple defect data can be analyzed at the same time, which improves the generation efficiency and quality of defect features and improves the accuracy of subsequent defect detection.

[0028] Example 4 On the basis of Example 2, the real-time defect detection method for selective laser melting 3D printing, the step 13 comprises: Step 131: performing cluster analysis on the defect features to obtain a plurality of defect clusters, identifying the defect weight of each defect feature in the corresponding defect cluster, and reorganizing the corresponding defect features based on the defect weight to obtain a cluster defect feature corresponding to each defect cluster; Step 132: Draw a virtual print product map according to the current print task, map each cluster defect feature in the virtual print product map to obtain several material defect features of the current print task, generate a material defect feature set corresponding to the current print task and display it.

[0029] In this example, the defect cluster represents the result of clustering defect features, and a defect cluster contains several similar defect features; In this example, the defect weight represents the proportion of a defect feature in the defect cluster; In this example, the cluster defect feature represents a feature commonly contained among several defect features contained in a defect cluster; In this example, the virtual printed product image represents an image of the 3D printed product drawn in a virtual space.

[0030] The working principle and beneficial effects of the above technical solution: In order to improve the efficiency of constructing the material defect feature set and reduce the impact of a single defect on the feature set, the defect features are first divided into several defect clusters by cluster analysis, and then the cluster defect features of each defect cluster are analyzed according to the weight of each defect feature in its corresponding defect cluster. The virtual printed product image is processed using the cluster defect features, and several material defect features are determined, thereby generating a material defect feature set for this printing task. In this way, not only can the defect features be classified, but also the preliminary identification of defects can be achieved based on the common features between different defect features, thereby improving the quality of subsequent defect detection.

[0031] Example 5 Based on Example 1, the real-time defect detection method for selective laser melting 3D printing further includes: Obtain the indoor environment information of the 3D printing processing room; When the indoor environment information meets the specified printing requirements, the laser scanning array is controlled to perform printing.

[0032] In this example, the printing requirement is specified as one of the following environments: 1. vacuum environment, 2. with protective gas.

[0033] The working principle and beneficial effects of the above technical solution are as follows: the entire processing process is carried out in a vacuum or gas-protected processing chamber to prevent the metal from reacting with other gases at high temperatures.

[0034] Example 6 Based on Example 1, the real-time defect detection method for selective laser melting 3D printing, step 2, comprises: Step 21: determining a number of layers to be printed of the 3D printed product according to the current printing task, and determining the contour to be printed corresponding to each of the layers to be printed, generating a number of printing guide instructions to control the laser scanning array to perform real-time printing, and determining the current printing layer according to the real-time printing data of the laser scanning array; Step 22: determining the complete printing state of the current printing layer based on the current contour to be printed corresponding to the current printing layer, constructing the real-time printing state of the current printing layer using the real-time printing data, and determining the real-time printing progress of the current printing layer; Step 23: When the real-time printing progress is 100%, and the displayed printing status of the current printing layer is consistent with the complete printing status, it is determined that the current printing layer has completed printing, and the preset liquid lens is controlled to collect and display the complete plane image of the current printing layer.

[0035] In this example, the layer to be printed means that multiple prints are required to complete a 3D printed product, and the printing surface corresponding to each print is a printing layer; In this example, the contour to be printed represents the contour generated when printing on a layer to be printed when a 3D printed product is completed; In this example, the printing guide instruction refers to an instruction for guiding the laser scanning array to print, and one layer to be printed corresponds to one printing guide instruction; In this example, the complete printing state means the state presented when all the contours to be printed in the current printing layer are printed completely; In this example, the real-time printing state indicates the state of the current printing layer at the current moment; In this example, a preset liquid lens is set in the printing processing room to capture images.

[0036] The working principle and beneficial effects of the above technical solution: In order to perform effective defect detection and ensure that the detection results are real-time, the number of layers to be printed and the contours to be printed corresponding to each layer to be printed are first determined according to the current printing task, and then a printing guidance instruction is generated to control the laser scanning array to perform the corresponding printing work, and then the real-time printing progress of the printing work is determined according to the contour printing completion status of the current printing layer. After the current printing layer is printed, the current printing layer is photographed by a preset liquid lens to obtain a complete plane image. In this way, image acquisition can be completed and image acquisition can be performed in time after printing is completed, avoiding unnecessary damage to the printed product due to long-term video monitoring, and corresponding detection is performed based on the quality of the printed product.

[0037] Example 7 On the basis of Example 1, the real-time defect detection method for selective laser melting 3D printing, the step 3, comprises: Step 31: grayscale enhancement is performed on the print contour contained in the complete plane image to obtain enhanced contour information of the current print layer, straight line fitting is performed on the enhanced contour information in the complete plane image, and a plurality of contour key points are located in the print contour according to the fitting result; Step 32: Divide the printed contour into a plurality of contour regions, establish a regional contour feature corresponding to each contour region based on the distribution of contour key points corresponding to each contour region, and search for relevant material defect features corresponding to each regional contour feature in the material defect feature set; Step 33: Map each of the relevant material defect features in the corresponding contour area, obtain the point overlap ratio between each contour key point and the corresponding relevant material defect feature, and select the target relevant material defects with a point overlap ratio higher than a specified ratio as the current material defects corresponding to the contour area; Step 34: Based on the point overlap ratio corresponding to each current material defect, feature recognition is performed on the complete plane image to obtain defect interference information of each current material defect on the printing contour, obtain several defect appearances of the complete plane image and defect specifications corresponding to each defect appearance, and generate defect information of the current printing layer.

[0038] In this example, the enhanced contour information indicates the result of enhancing the contour contained in the current printing layer. The purpose of the enhancement is to increase the contrast and obtain the detail information in the contour. In this example, the contour keypoint represents the intersection between two or two straight lines; In this example, the related material characteristic defect represents a material defect characteristic related to the regional profile characteristic; In this example, the coincidence ratio represents the ratio between the number of coincidences and the number of non-coincidences between the contour key points and the related material defect features; In this example, the prescribed proportion is 90%; In this example, the defect interference information indicates the interference caused by the current material defect on the printing profile; In this example, the defect specification indicates the specification size of the defect appearance; In this example, the defect appearance refers to the appearance of the defect contained in the complete plane image.

[0039] The working principle and beneficial effects of the above technical solution are as follows: by grayscale enhancement of the printed contour on the complete plane image and straight line fitting to determine several contour key points contained in the printed contour, then dividing the printed contour into several contour areas, performing feature recognition on each contour area, and finding the relevant material defect features corresponding to each area contour, further determining the current material defect of the area contour by analyzing the overlap ratio between the relevant material defect features and the contour key points, and finally analyzing the defect interference information of each current material defect on the printed contour to determine the defect appearance contained in the complete plane image and its defect specifications, and finally determining the defect information of the current printed layer. In this way, the defect appearance in the current printed layer can be uniformly identified and processed, the position of each defect appearance and its specifications can be determined, and accurate defect information can be obtained to provide a powerful data reference for the demander.

[0040] Example 8 On the basis of Example 1, the real-time defect detection method for selective laser melting 3D printing, the step 4, comprises: Step 41: establishing a 3D stereoscopic image according to a plurality of complete plane images corresponding to the 3D printed product, and matching a corresponding stereoscopic image area for each defect information based on the corresponding relationship between the defect information and the complete plane image; Step 42: identifying a plurality of printed parts contained in the 3D stereoscopic image, determining a stereoscopic image region corresponding to each of the printed parts, and marking each of the defect information in the corresponding stereoscopic image region to obtain a part defect feature corresponding to each of the printed parts; Step 43: Mark each of the part defect features in the 3D stereoscopic image to obtain complete part information corresponding to each of the printed parts, analyze defect interference information between different printed parts in the 3D stereoscopic image, and establish defect attributes corresponding to each of the printed parts in combination with the part position corresponding to each of the printed parts; Step 44: Perform an overall defect evaluation on the 3D printed product according to the defect attributes corresponding to each of the printed parts, perform an overall function evaluation on the 3D printed product according to the part functions corresponding to each of the printed parts, generate a defect report for the 3D printed product, and display it.

[0041] In this example, the 3D stereoscopic image refers to a three-dimensional image generated by combining several complete plane images corresponding to a 3D printed product in a layer order; In this example, the stereoscopic image region represents a region of the 3D printed product corresponding to a defect information; In this example, the part defect feature represents a feature of a defect presented on a printed part; In this example, the part completeness information indicates the completeness of each position of the printed part; In this example, the defect attribute represents the attribute of a defect presented by a printed part, such as a cold solder joint attribute, a leaking solder joint attribute, etc.

[0042] The working principle and beneficial effects of the above technical solution are as follows: by utilizing a complete plane image to construct a 3D stereoscopic image of a 3D printed product to process defect information, the defect characteristics of each printed part are determined, and then the various information presented in the 3D stereoscopic image is analyzed to determine the defect attributes of each printed part. Finally, the 3D printed product is evaluated multiple times to generate a defect report. In this way, the defects of the 3D printed product can be analyzed from all angles and angles, thereby improving the accuracy of defect capture and providing the demander with a well-reasoned defect report.

[0043] Example 9 On the basis of Example 8, the real-time defect detection method for selective laser melting 3D printing further includes: Screening defective 3D printed products that fail the overall defect evaluation or the overall function evaluation; Several defect attributes corresponding to the defective 3D printed product are respectively marked in the corresponding defective 3D stereoscopic image, and a multi-dimensional defect map of the defective 3D printed product is generated and displayed.

[0044] The working principle and beneficial effects of the above technical solution are as follows: by generating an exploded view of a defective 3D printed product to show all its defects, it is convenient for the demander to understand the work errors of this printing and help it make technical improvements.

[0045] Example 10 This example provides a real-time defect detection system for selective laser melting 3D printing, such as Figure 2 As shown, including: A defect presetting module is used to retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; A synchronous monitoring module, used to determine the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtain a complete plane image of the current printing layer; A defect recognition module, used to use the material defect feature set to identify a plurality of current material defects contained in the complete plane image, perform depth recognition on each of the current material defects, and generate defect information of the current printing layer; The report generation module is used to create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

[0046] In this example, the material attribute indicates the attribute of the raw material used in printing, which is generally a powdered raw material; In this example, the defect database represents data information about defects that may occur after the finished product is constructed using the raw material; In this example, the material defect feature set represents the features of defects that may be generated after the 3D printed product is constructed from the raw material; In this example, the current printing layer indicates the layer currently being 3D printed; In this example, the image acquisition time is: the time corresponding to when the real-time printing progress of a current printing layer is 100%; In this example, the complete plane image represents the result of collecting all image information of the current printing layer; In this example, the current material defect represents a defect present in the complete planar image; In this example, printed parts represent individual parts used to make up the finished 3D printed product; In this example, the defect attribute represents the attribute of the defect presented by the printed part.

[0047] The working principle and beneficial effects of the above technical solution: In order to solve the defects in the prior art, a method for real-time defect detection is constructed. Before printing, the corresponding defect database is screened according to the material properties of the raw materials used this time, so as to construct a material defect feature set for this printing. Then, after completing a layer of printing, its complete plane image is collected, and the defect information of the current printing layer is determined by identifying the current material defects in the complete plane image and performing depth identification on it. Finally, the defect properties of each printed part are analyzed by constructing a 3D stereoscopic image, so as to generate a defect report for the 3D printed product. In this way, real-time detection can be performed during the printing process, the defects in the finished product can be determined, and a defect report can be generated for reference by the demander. The demander can perform different processing on the printed product according to actual needs, and various printing defects can be identified during the printing process, which is convenient for manufacturers to promptly discover generation problems and deal with them, thereby promoting the development of productivity.

[0048] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A real-time defect detection method for selective laser melting 3D printing, characterized in that: include: Step 1: Retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; Step 2: determining the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtaining a complete plane image of the current printing layer; Step 3: using the material defect feature set to identify a number of current material defects contained in the complete plane image, performing depth identification on each of the current material defects, and generating defect information of the current printing layer; Step 4: Create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

2. The real-time defect detection method for selective laser melting 3D printing according to claim 1, characterized in that: The step 1 comprises: Step 11: Determine the printing raw material of the 3D printed product based on the current printing task, and search for a defect database corresponding to the material property in big data based on the material property of the printing raw material; Step 12: Draw a defect image of the printing raw material using the defect data, and determine defect features corresponding to each defect data; Step 13: Perform cluster analysis on the defect features to generate a material defect feature set corresponding to this printing task and display it.

3. The real-time defect detection method for selective laser melting 3D printing according to claim 2, characterized in that: The step 12 comprises: Step 121: generating a flawless image of the printing raw material based on the material properties, and respectively obtaining a data difference between each defect data and the flawless data corresponding to the flawless image; Step 122: performing corresponding defect adjustment training on the flawless image using the data difference to obtain a defect image corresponding to each defect data; Step 123: Identify the image presentation features corresponding to each of the defective images respectively, and generate defect features corresponding to the defect data.

4. The real-time defect detection method for selective laser melting 3D printing according to claim 2, characterized in that: The step 13 comprises: Step 131: performing cluster analysis on the defect features to obtain a plurality of defect clusters, identifying the defect weight of each defect feature in the corresponding defect cluster, and reorganizing the corresponding defect features based on the defect weight to obtain a cluster defect feature corresponding to each defect cluster; Step 132: Draw a virtual print product map according to the current print task, map each cluster defect feature in the virtual print product map to obtain several material defect features of the current print task, generate a material defect feature set corresponding to the current print task and display it.

5. The real-time defect detection method for selective laser melting 3D printing according to claim 1, characterized in that: Also includes: Obtain the indoor environment information of the 3D printing processing room; When the indoor environment information meets the specified printing requirements, the laser scanning array is controlled to perform printing.

6. The real-time defect detection method for selective laser melting 3D printing according to claim 1, characterized in that: The step 2 comprises: Step 21: determining a number of layers to be printed of the 3D printed product according to the current printing task, and determining the contour to be printed corresponding to each of the layers to be printed, generating a number of printing guidance instructions to control the laser scanning array to perform real-time printing, and determining the current printing layer according to the real-time printing data of the laser scanning array; Step 22: determining the complete printing state of the current printing layer based on the current contour to be printed corresponding to the current printing layer, constructing the real-time printing state of the current printing layer using the real-time printing data, and determining the real-time printing progress of the current printing layer; Step 23: When the real-time printing progress is 100%, and the displayed printing status of the current printing layer is consistent with the complete printing status, it is determined that the current printing layer has completed printing, and the preset liquid lens is controlled to collect and display the complete plane image of the current printing layer.

7. The real-time defect detection method for selective laser melting 3D printing according to claim 1, characterized in that: The step 3 comprises: Step 31: grayscale enhancement is performed on the print contour contained in the complete plane image to obtain enhanced contour information of the current print layer, straight line fitting is performed on the enhanced contour information in the complete plane image, and a plurality of contour key points are located in the print contour according to the fitting result; Step 32: Divide the printed contour into a plurality of contour regions, establish a regional contour feature corresponding to each contour region based on the distribution of contour key points corresponding to each contour region, and search for relevant material defect features corresponding to each regional contour feature in the material defect feature set; Step 33: Map each of the relevant material defect features in the corresponding contour area, obtain the point overlap ratio between each contour key point and the corresponding relevant material defect feature, and select the target relevant material defects with a point overlap ratio higher than a specified ratio as the current material defects corresponding to the contour area; Step 34: Based on the point overlap ratio corresponding to each current material defect, feature recognition is performed on the complete plane image to obtain defect interference information of each current material defect on the printing contour, obtain several defect appearances of the complete plane image and defect specifications corresponding to each defect appearance, and generate defect information of the current printing layer.

8. The real-time defect detection method for selective laser melting 3D printing according to claim 1, characterized in that: The step 4 comprises: Step 41: establishing a 3D stereoscopic image according to a plurality of complete plane images corresponding to the 3D printed product, and matching a corresponding stereoscopic image area for each defect information based on the corresponding relationship between the defect information and the complete plane image; Step 42: identifying a plurality of printed parts contained in the 3D stereoscopic image, determining a stereoscopic image region corresponding to each of the printed parts, and marking each of the defect information in the corresponding stereoscopic image region to obtain a part defect feature corresponding to each of the printed parts; Step 43: Mark each of the part defect features in the 3D stereoscopic image to obtain complete part information corresponding to each of the printed parts, analyze defect interference information between different printed parts in the 3D stereoscopic image, and establish defect attributes corresponding to each of the printed parts in combination with the part position corresponding to each of the printed parts; Step 44: Perform an overall defect evaluation on the 3D printed product according to the defect attributes corresponding to each of the printed parts, perform an overall function evaluation on the 3D printed product according to the part functions corresponding to each of the printed parts, generate a defect report for the 3D printed product, and display it.

9. The real-time defect detection method for selective laser melting 3D printing according to claim 8, characterized in that: Also includes: Screening defective 3D printed products that fail the overall defect evaluation or the overall function evaluation; Several defect attributes corresponding to the defective 3D printed product are respectively marked in the corresponding defective 3D stereoscopic image, and a multi-dimensional defect map of the defective 3D printed product is generated and displayed.

10. A real-time defect detection system for selective laser melting 3D printing, characterized in that: include: A defect presetting module is used to retrieve a corresponding defect database based on the material properties of the 3D printed product, identify the defect features corresponding to each defect data in the defect database, and obtain a material defect feature set; A synchronous monitoring module, used to determine the image acquisition time based on the real-time printing progress of the current printing layer by the laser scanning array, and obtain a complete plane image of the current printing layer; A defect recognition module, used to use the material defect feature set to identify a plurality of current material defects contained in the complete plane image, perform depth recognition on each of the current material defects, and generate defect information of the current printing layer; The report generation module is used to create a 3D stereoscopic image of the 3D printed product, use the defect information to identify the defect attributes corresponding to each printed part in the corresponding 3D stereoscopic image, generate a defect report of the 3D printed product and display it.

Images