Process detection system and method for monitoring 3D printing quality in real time

Abstract

The invention discloses a process detection system and method for monitoring 3D printing quality in real time, and the system is characterized in that at least four industrial cameras are fixedly arranged around a target 3D printing object, and are used for collecting a plurality of angle images of the target 3D printing object; the industrial host is used for acquiring the data acquired by the industrial camera, preprocessing the data to form a multi-view image, performing stereo matching and calculation by using the multi-view image to obtain three-dimensional point cloud data, and performing point cloud splicing according to the three-dimensional point cloud data to obtain a point cloud image of the target 3D printing object; a three-dimensional reconstruction model of the target 3D printing object is obtained through three-dimensional reconstruction of the point cloud picture; performing size detection on a reconstruction model obtained by three-dimensional reconstruction to judge whether the size of the object is in an allowable range, and performing defect detection to judge whether the object has local defects; 3D printing parameters are adjusted in real time according to the detected defect types and positions, and the subsequent printing process is optimized. According to the method, the defect identification precision and the real-time feedback effect in the 3D printing process can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the fields of additive manufacturing and machine vision inspection, and in particular to a process inspection system and method for real-time monitoring of 3D printing quality. Background Art

[0002] With the rapid development of additive manufacturing (3D printing) technology, an increasing number of industries are beginning to apply 3D printing to create complex structures and high-precision components. However, during the printing process, defects such as delamination, material accumulation, warping, and bubbles are prone to occur due to factors such as equipment performance, material properties, and process parameters. These defects not only affect the appearance of printed parts but can also lead to a decrease in mechanical properties or even failure.

[0003] Traditional 3D printing quality inspection methods typically involve checking printed parts after printing through manual visual inspection, non-destructive testing, or scanning measurements. However, these methods are time-consuming and limited, and offline inspections are inefficient, making it difficult to promptly adjust the printing process to prevent or correct defects. Therefore, there is an urgent need for an inspection method that can monitor 3D printing quality in real time, efficiently and accurately detecting defects during the printing process, thereby improving print quality and production efficiency. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems of insufficient image acquisition range and defect detection accuracy in the prior art, and to provide a process detection system and method for real-time monitoring of 3D printing quality. It integrates multi-visual sensor information fusion technology, point cloud stitching and 3D reconstruction technology, and real-time parameter feedback control technology, and is suitable for 3D printing detection in scenarios such as industrial manufacturing, medical equipment, and personalized customization.

[0005] The technical solution adopted in the present invention is: The process detection system for real-time monitoring of 3D printing quality includes an industrial host and an LED ring light source, an FA industrial lens and at least four industrial cameras connected to the industrial host; the LED ring light source provides uniform lighting, and at least four industrial cameras are fixedly configured around the target 3D printed object to collect images of the target 3D printed object from multiple angles. The FA industrial lens is used in conjunction with the industrial camera to control imaging distortion; the industrial host is used to obtain data collected by the industrial camera and pre-process it to form multi-view images, and use the multi-view images for stereo matching and calculation to obtain three-dimensional point cloud data, and then perform point cloud splicing based on the three-dimensional point cloud data to obtain a point cloud map of the target 3D printed object, and use the point cloud Figure 3 The 3D reconstruction model of the target 3D printed object is obtained by 3D reconstruction, and the 3D reconstruction model is compared with the ideal model (CAD model) or the preset reference geometry to obtain the defect type and location, and the subsequent printing process is optimized by adjusting the 3D printing parameters in real time.

[0006] The LED ring light source provides uniform illumination, enhances image contrast, and improves imaging stability. The FA industrial lens ensures image clarity and controls image distortion. Each industrial lens, paired with a corresponding industrial camera, ensures a consistent working distance and field of view, covering all critical areas of the target object.

[0007] Furthermore, the calibrated field of view range is used to fix at least four industrial cameras around the target 3D printed object.

[0008] Furthermore, the preprocessing of the data collected by the industrial camera includes bilateral filtering denoising and lens distortion correction operations.

[0009] Furthermore, 3D printing parameters include print head movement speed, nozzle temperature, and nozzle material extrusion volume.

[0010] A process detection method for real-time monitoring of 3D printing quality includes the following steps: Step 1: Build a visual inspection system and calibrate the inspection system: Build a visual inspection system that includes at least four industrial cameras, an LED ring light source, and a FA industrial lens. Configure and calibrate each component of the visual inspection system, perform imaging parameter analysis, and optimize the system to obtain a high-precision calibration inspection system. Step 2, Image Acquisition and Preprocessing: Based on the calibrated field of view, multiple industrial cameras are used to capture multi-view images of the target 3D printed object from multiple angles, ensuring coverage of all areas of the printed object. The multi-view images captured by the cameras are acquired in real time for transmission and preprocessed simultaneously. Step 3: Stereo matching and point cloud generation: The SGBM algorithm is used to match multi-view images to generate a high-quality disparity map and convert it into a depth map containing depth information. The depth map is combined with camera parameters to calculate 3D point cloud data, which faithfully restores the surface morphology of the printed object. These point clouds provide accurate 3D data for subsequent stitching.

[0011] Step 4, point cloud stitching: Use the calibration results to convert the point clouds from all viewpoints into the same global coordinate system for preliminary alignment; optimize the image after preliminary alignment and fine-tune the relative positions of adjacent point clouds to minimize the matching error between adjacent point clouds; then remove duplicates and fuse the overlapping areas to obtain the point cloud image of the object to be measured; Specifically, since point clouds acquired by multiple cameras may have overlapping areas, these overlapping areas need to be deduplicated and fused to obtain a point cloud map of the object to be measured.

[0012] Step 5, 3D reconstruction: Scan the spliced ​​point cloud data point by point, remove outliers and invalid data in concave areas, and generate a continuous 3D mesh model. Finally, reconstruct a high-precision and complete 3D model. Step 6, visual inspection: The reconstructed model obtained by 3D reconstruction is subjected to size inspection to determine whether the size of the object is within the allowable range, and defect inspection is also performed to determine whether the object has local defects; Step 7: Dynamically correct defects: Based on the defect type and location detected in real time, the subsequent printing process is optimized by adjusting the 3D printing parameters in real time to correct the defects.

[0013] Furthermore, in step 2, the data captured by the camera is transmitted in real time via a data transmission channel. Preprocessing, including bilateral filtering for denoising and lens distortion correction, optimizes image quality. This process reduces the latency of storing intermediate data, improves processing efficiency, and provides high-precision, real-time image data for subsequent analysis.

[0014] Furthermore, in step 4, the ICP algorithm is applied to the preliminarily aligned image.

[0015] Furthermore, in step 5, a rolling ball method is used to perform three-dimensional reconstruction.

[0016] Specifically, after point cloud data generation and stitching are complete, a rolling ball method is used for 3D reconstruction to further enhance the depiction of the printed object's surface morphology. By scanning the stitched point cloud data point by point, invalid data from outliers and concave areas is removed, and a continuous 3D mesh model is generated. Ultimately, a highly accurate and complete 3D model is reconstructed, providing a reliable data foundation for subsequent surface quality inspection, defect analysis, and dimensional measurement.

[0017] Furthermore, during the size detection in step 6, the reconstructed model obtained by 3D reconstruction is compared with the ideal model (CAD model) or the preset reference geometric shape to detect whether the object size is greater than a set threshold to determine whether the size error of the object is within the allowable range.

[0018] Furthermore, in step 6, the defect detection detects convex and concave defects by calculating the local curvature of the point cloud, and determines whether there is a defect in the corresponding area by judging whether the curvature exceeds a set curvature threshold.

[0019] Furthermore, the 3D printing parameters in step 7 include the print head movement speed, nozzle temperature, and nozzle material extrusion volume.

[0020] Specifically, based on the defect type and location monitored in real time, the subsequent printing process is optimized by adjusting key parameters such as the print head movement speed, nozzle temperature, and nozzle material extrusion volume in real time to ensure immediate correction of the printing process. The detection, feedback, and parameter integration system transforms defect repair from manual processing to a fully automated and intelligent process, reducing human intervention and improving efficiency.

[0021] The present invention adopts the above technical solution, deploying multiple industrial cameras within the printing area, combined with an LED ring light source and FA lens, to capture multi-view images of the printed object in real time. The system optimizes image quality through image processing in an online detection mode, generating a complete 3D reconstructed model. By comparing the image with the ideal CAD model, it detects defects that may occur during the printing process. Based on the real-time detection results, the system dynamically adjusts printing parameters to automatically repair defects and optimize print quality. This invention significantly improves the accuracy of defect recognition and provides real-time feedback during 3D printing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments; Figure 1 Schematic diagram of the three-dimensional reconstruction process of the present invention; Figure 2 This is a schematic diagram of the visual inspection process of the present invention; Figure 3 Schematic diagram of the process of real-time monitoring of 3D printing quality according to the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0024] like Figures 1 to 3 As shown in one, the present invention discloses a process detection system for real-time monitoring of 3D printing quality, including an industrial host and an LED ring light source, an FA industrial lens and at least four industrial cameras connected to the industrial host; the LED ring light source provides uniform lighting, and the at least four industrial cameras are fixedly arranged around the target 3D printed object to collect multiple angle images of the target 3D printed object, and the FA industrial lens is used in conjunction with the industrial camera to control imaging distortion; the industrial host is used to obtain data collected by the industrial camera and pre-process it to form a multi-view image, and the multi-view image is used to perform stereo matching and calculation to obtain three-dimensional point cloud data, and the point cloud is spliced ​​according to the three-dimensional point cloud data to obtain a point cloud map of the target 3D printed object, and the point cloud is used to obtain a point cloud map of the target 3D printed object. Figure 3The 3D reconstruction model of the target 3D printed object is obtained by 3D reconstruction, and the 3D reconstruction model is compared with the ideal model (CAD model) or the preset reference geometry to obtain the defect type and location, and the subsequent printing process is optimized by adjusting the 3D printing parameters in real time.

[0025] The LED ring light source provides uniform illumination, enhances image contrast, and improves imaging stability. The FA industrial lens ensures image clarity and controls image distortion. Each industrial lens, paired with a corresponding industrial camera, ensures a consistent working distance and field of view, covering all critical areas of the target object.

[0026] Furthermore, the calibrated field of view range is used to fix at least four industrial cameras around the target 3D printed object.

[0027] Furthermore, the preprocessing of the data collected by the industrial camera includes bilateral filtering denoising and lens distortion correction operations.

[0028] Furthermore, 3D printing parameters include print head movement speed, nozzle temperature, and nozzle material extrusion volume.

[0029] A process detection method for real-time monitoring of 3D printing quality includes the following steps: Step 1: Build a visual inspection system and calibrate the inspection system: Build a visual inspection system that includes at least four industrial cameras, an LED ring light source, and a FA industrial lens. Configure and calibrate each component of the visual inspection system, perform imaging parameter analysis, and optimize the system to obtain a high-precision calibration inspection system. Step 2, Image Acquisition and Preprocessing: Based on the calibrated field of view, multiple industrial cameras are used to capture multi-view images of the target 3D printed object from multiple angles, ensuring coverage of all areas of the printed object. The multi-view images captured by the cameras are acquired in real time for transmission and preprocessed simultaneously. Step 3: Stereo matching and point cloud generation: The SGBM algorithm is used to match multi-view images to generate a high-quality disparity map and convert it into a depth map containing depth information. The depth map is combined with camera parameters to calculate 3D point cloud data, which faithfully restores the surface morphology of the printed object. These point clouds provide accurate 3D data for subsequent stitching.

[0030] Step 4, point cloud stitching: Use the calibration results to convert the point clouds from all viewpoints into the same global coordinate system for preliminary alignment; optimize the image after preliminary alignment and fine-tune the relative positions of adjacent point clouds to minimize the matching error between adjacent point clouds; then remove duplicates and fuse the overlapping areas to obtain the point cloud image of the object to be measured; Specifically, since point clouds acquired by multiple cameras may have overlapping areas, these overlapping areas need to be deduplicated and fused to obtain a point cloud map of the object to be measured.

[0031] Step 5, 3D reconstruction: Scan the spliced ​​point cloud data point by point, remove outliers and invalid data in concave areas, and generate a continuous 3D mesh model. Finally, reconstruct a high-precision and complete 3D model. Step 6, visual inspection: The reconstructed model obtained by 3D reconstruction is subjected to size inspection to determine whether the size of the object is within the allowable range, and defect inspection is also performed to determine whether the object has local defects; Step 7: Dynamically correct defects: Based on the defect type and location detected in real time, the subsequent printing process is optimized by adjusting the 3D printing parameters in real time to correct the defects.

[0032] Furthermore, in step 2, real-time camera data is transmitted through different process data transmission channels (pipelines) within Windows. Preprocessing, including bilateral filtering for denoising and lens distortion correction, optimizes image quality. This process reduces latency in storing intermediate data, improves processing efficiency, and provides high-precision, real-time image data for subsequent analysis.

[0033] Furthermore, in step 4, the ICP algorithm is applied to the preliminarily aligned image.

[0034] Furthermore, in step 5, a rolling ball method is used to perform three-dimensional reconstruction.

[0035] Specifically, after point cloud data generation and stitching are complete, a rolling ball method is used for 3D reconstruction to further enhance the depiction of the printed object's surface morphology. By scanning the stitched point cloud data point by point, invalid data from outliers and concave areas is removed, and a continuous 3D mesh model is generated. Ultimately, a highly accurate and complete 3D model is reconstructed, providing a reliable data foundation for subsequent surface quality inspection, defect analysis, and dimensional measurement.

[0036] Furthermore, in step 6, the size detection room compares the reconstructed model obtained by 3D reconstruction with the ideal model (CAD model) or the preset reference geometry to detect whether the object size is greater than a set threshold to determine whether the size error of the object is within the allowable range.

[0037] Furthermore, in step 6, the defect detection detects convex and concave defects by calculating the local curvature of the point cloud, and determines whether there is a defect in the corresponding area by judging whether the curvature exceeds a set curvature threshold.

[0038] Furthermore, the 3D printing parameters in step 7 include the print head movement speed, nozzle temperature, and nozzle material extrusion amount.

[0039] Specifically, based on the defect type and location monitored in real time, the subsequent printing process is optimized by adjusting key parameters such as the print head movement speed, nozzle temperature, and nozzle material extrusion volume in real time to ensure immediate correction of the printing process. The detection, feedback, and parameter integration system transforms defect repair from manual processing to a fully automated and intelligent process, reducing human intervention and improving efficiency.

[0040] The present invention adopts the above technical solution, deploying multiple industrial cameras within the printing area, combined with an LED ring light source and FA lens, to capture multi-view images of the printed object in real time. The system optimizes image quality through image processing in an online detection mode, generating a complete 3D reconstructed model. By comparing the image with the ideal CAD model, it detects defects that may occur during the printing process. Based on the real-time detection results, the system dynamically adjusts printing parameters to automatically repair defects and optimize print quality. This invention significantly improves the accuracy of defect recognition and provides real-time feedback during 3D printing.

[0041] Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

Claims

1. A process detection system for real-time monitoring of 3D printing quality, characterized by: It includes an industrial host and an LED ring light source, an FA industrial lens and at least 4 industrial cameras connected to the industrial host; the LED ring light source provides uniform lighting, and at least 4 industrial cameras are fixedly arranged around the target 3D printed object to collect multiple angle images of the target 3D printed object. The FA industrial lens is used in conjunction with the industrial camera to control imaging distortion; the industrial host is used to obtain the data collected by the industrial camera and pre-process it to form a multi-view image, and use the multi-view image to perform stereo matching and calculation to obtain three-dimensional point cloud data, and use the three-dimensional point cloud data to perform point cloud splicing to obtain a point cloud map of the target 3D printed object, and use the point cloud map to three-dimensionally reconstruct a three-dimensional reconstructed model of the target 3D printed object; the reconstructed model obtained by three-dimensional reconstruction is subjected to size detection to determine whether the size of the object is within the allowable range, and defect detection is performed to determine whether the object has local defects; the 3D printing parameters are adjusted in real time according to the type and location of the detected defects to optimize the subsequent printing process.

2. The process detection system for real-time monitoring of 3D printing quality according to claim 1, characterized in that: The calibrated field of view range will be determined by fixing at least 4 industrial cameras around the target 3D printing object.

3. The process detection system for real-time monitoring of 3D printing quality according to claim 1, characterized in that: The preprocessing of data collected by industrial cameras includes bilateral filtering denoising and lens distortion correction operations.

4. The process detection system for real-time monitoring of 3D printing quality according to claim 1, characterized in that: Size detection compares the reconstructed model obtained by 3D reconstruction with the ideal model or preset reference geometry to detect whether the object size is greater than the set threshold to determine whether the object's size error is within the allowable range; Defect detection detects convex and concave defects by calculating the local curvature of the point cloud, and determines whether there is a defect in the corresponding area by judging whether the curvature exceeds the set curvature threshold.

5. A process detection method for real-time monitoring of 3D printing quality, according to the process detection system for real-time monitoring of 3D printing quality according to any one of claims 1 to 4, characterized in that: The method comprises the following steps: Step 1: Build a visual inspection system and calibrate the inspection system: Build a visual inspection system that includes at least four industrial cameras, an LED ring light source, and a FA industrial lens; Configure and calibrate each component in the visual inspection system, perform imaging parameter analysis and system optimization adjustments to obtain a high-precision calibration inspection system; Step 2, Image Acquisition and Preprocessing: Based on the calibrated field of view, multiple industrial cameras are used to capture multi-view images of the target 3D printed object from multiple angles, ensuring coverage of all areas of the printed object. The multi-view images captured by the cameras are acquired in real time for transmission and preprocessed simultaneously. Step 3: Stereo matching and point cloud generation: Use the SGBM algorithm to match multi-view images to generate a high-quality disparity map and convert it into a depth map containing depth information. Combine the depth map with camera parameters to calculate 3D point cloud data, truly restoring the surface morphology of the printed object. Step 4, point cloud stitching: Use the calibration results to transform the point clouds from all perspectives into the same global coordinate system for preliminary alignment; Optimize the image after preliminary alignment and fine-tune the relative positions of adjacent point clouds to minimize the matching error between adjacent point clouds; Then, the overlapping areas are deduplicated and fused to obtain the point cloud image of the object to be measured; Step 5, 3D reconstruction: Scan the spliced ​​point cloud data point by point, remove outliers and invalid data in concave areas, and generate a continuous 3D mesh model. Finally, reconstruct a high-precision and complete 3D model. Step 6, visual inspection: The reconstructed model obtained by 3D reconstruction is subjected to size inspection to determine whether the size of the object is within the allowable range, and defect inspection is also performed to determine whether the object has local defects; Step 7: Dynamically correct defects: Based on the defect type and location detected in real time, the subsequent printing process is optimized by adjusting the 3D printing parameters in real time to correct the defects.

6. The process detection method for real-time monitoring of 3D printing quality according to claim 5, characterized in that: In step 2, the data collected by the camera is transmitted in real time through the data transmission channel. Preprocessing includes bilateral filtering denoising and lens distortion correction to optimize image quality.

7. The process detection method for real-time monitoring of 3D printing quality according to claim 5, characterized in that: In step 4, the ICP algorithm is applied to the image after preliminary alignment.

8. The process detection method for real-time monitoring of 3D printing quality according to claim 5, characterized in that: In step 5, the rolling ball method is used for three-dimensional reconstruction.

9. The process detection method for real-time monitoring of 3D printing quality according to claim 5, characterized in that: In step 6, the size detection is performed by comparing the reconstructed model obtained by 3D reconstruction with the ideal model or the preset reference geometry to detect whether the object size is greater than a set threshold to determine whether the size error of the object is within the allowable range; Defect detection detects convex and concave defects by calculating the local curvature of the point cloud, and determines whether there is a defect in the corresponding area by judging whether the curvature exceeds the set curvature threshold.

10. The process detection method for real-time monitoring of 3D printing quality according to claim 5, characterized in that: Step 7: 3D printing parameters include print head movement speed, nozzle temperature, and nozzle material extrusion volume.

Images