Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds
Powder bed defects are critical factors affecting the print quality and stability in Laser Powder Bed Fusion (LPBF). However, traditional 2D image-based powder bed defect monitoring methods are limited by sensitivity to lighting conditions and insufficient data capture. This study proposes a real-ti...
| Main Authors: | , , , , , , , |
|---|---|
| 格式: | Article |
| 語言: | English |
| 出版: |
Taylor & Francis Group
2025-12-01
|
| 叢編: | Virtual and Physical Prototyping |
| 主題: | |
| 在線閱讀: | https://www.tandfonline.com/doi/10.1080/17452759.2024.2449171 |
| _version_ | 1826866540535873536 |
|---|---|
| author | Junlai Zhao Zihan Yang Qingpeng Chen Chen Zhang Jianhui Zhao Guoqing Zhang Fang Dong Sheng Liu |
| author_facet | Junlai Zhao Zihan Yang Qingpeng Chen Chen Zhang Jianhui Zhao Guoqing Zhang Fang Dong Sheng Liu |
| author_sort | Junlai Zhao |
| collection | DOAJ |
| description | Powder bed defects are critical factors affecting the print quality and stability in Laser Powder Bed Fusion (LPBF). However, traditional 2D image-based powder bed defect monitoring methods are limited by sensitivity to lighting conditions and insufficient data capture. This study proposes a real-time defect monitoring system based on 3D point cloud data and deep learning approach. The system uses binocular vision to capture point cloud data in real time, enabling high-precision defect segmentation with advanced deep learning models. However, direct deep learning on point clouds can result in the loss of small defect features during downsampling. To address this, an indirect point cloud deep learning method based on 2D projection is introduced, which improves segmentation accuracy for small defects while reducing inference time. By deploying the trained model, this study establishes a closed-loop control system for powder bed defect detection and conducts real-world printing tests, demonstrating effective defect remediation capabilities. Although larger-scale industrial testing is still required, this research illustrates the significant potential of 3D point cloud-based deep learning in enhancing defect detection and quality control in additive manufacturing. |
| first_indexed | 2025-02-16T22:03:20Z |
| format | Article |
| id | doaj.art-acf1ef3567734b51a917c3c34908d6e7 |
| institution | Directory Open Access Journal |
| issn | 1745-2759 1745-2767 |
| language | English |
| last_indexed | 2025-02-16T22:03:20Z |
| publishDate | 2025-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Virtual and Physical Prototyping |
| spelling | doaj.art-acf1ef3567734b51a917c3c34908d6e72025-01-15T18:44:35ZengTaylor & Francis GroupVirtual and Physical Prototyping1745-27591745-27672025-12-0120110.1080/17452759.2024.2449171Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point cloudsJunlai Zhao0Zihan Yang1Qingpeng Chen2Chen Zhang3Jianhui Zhao4Guoqing Zhang5Fang Dong6Sheng Liu7School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, People’s Republic of ChinaThe Institute of Technological Sciences, Wuhan University, Wuhan, People’s Republic of ChinaThe Institute of Technological Sciences, Wuhan University, Wuhan, People’s Republic of ChinaThe Institute of Technological Sciences, Wuhan University, Wuhan, People’s Republic of ChinaSchool of Computer Science, Wuhan University, Wuhan, People’s Republic of ChinaThe Institute of Technological Sciences, Wuhan University, Wuhan, People’s Republic of ChinaThe Institute of Technological Sciences, Wuhan University, Wuhan, People’s Republic of ChinaSchool of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, People’s Republic of ChinaPowder bed defects are critical factors affecting the print quality and stability in Laser Powder Bed Fusion (LPBF). However, traditional 2D image-based powder bed defect monitoring methods are limited by sensitivity to lighting conditions and insufficient data capture. This study proposes a real-time defect monitoring system based on 3D point cloud data and deep learning approach. The system uses binocular vision to capture point cloud data in real time, enabling high-precision defect segmentation with advanced deep learning models. However, direct deep learning on point clouds can result in the loss of small defect features during downsampling. To address this, an indirect point cloud deep learning method based on 2D projection is introduced, which improves segmentation accuracy for small defects while reducing inference time. By deploying the trained model, this study establishes a closed-loop control system for powder bed defect detection and conducts real-world printing tests, demonstrating effective defect remediation capabilities. Although larger-scale industrial testing is still required, this research illustrates the significant potential of 3D point cloud-based deep learning in enhancing defect detection and quality control in additive manufacturing.https://www.tandfonline.com/doi/10.1080/17452759.2024.2449171Laser powder bed fusion (LPBF)powder bed defect detection3D point clouddeep learningreal-time monitoring |
| spellingShingle | Junlai Zhao Zihan Yang Qingpeng Chen Chen Zhang Jianhui Zhao Guoqing Zhang Fang Dong Sheng Liu Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds Virtual and Physical Prototyping Laser powder bed fusion (LPBF) powder bed defect detection 3D point cloud deep learning real-time monitoring |
| title | Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds |
| title_full | Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds |
| title_fullStr | Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds |
| title_full_unstemmed | Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds |
| title_short | Real-time detection of powder bed defects in laser powder bed fusion using deep learning on 3D point clouds |
| title_sort | real time detection of powder bed defects in laser powder bed fusion using deep learning on 3d point clouds |
| topic | Laser powder bed fusion (LPBF) powder bed defect detection 3D point cloud deep learning real-time monitoring |
| url | https://www.tandfonline.com/doi/10.1080/17452759.2024.2449171 |
| work_keys_str_mv | AT junlaizhao realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT zihanyang realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT qingpengchen realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT chenzhang realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT jianhuizhao realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT guoqingzhang realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT fangdong realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds AT shengliu realtimedetectionofpowderbeddefectsinlaserpowderbedfusionusingdeeplearningon3dpointclouds |