A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces
High-quality reconstruction of polished surfaces is a promising yet challenging task in the industrial field. Due to its extreme reflective properties, state-of-the-art methods have not achieved a satisfying trade-off between retaining texture and removing the effects of specular outliers. In this p...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-02-01
|
| Series: | Machines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-1702/10/2/120 |
| _version_ | 1797478570569760768 |
|---|---|
| author | Yanlong Cao Xiaoyao Wei Wenyuan Liu Binjie Ding Jiangxin Yang Yanpeng Cao |
| author_facet | Yanlong Cao Xiaoyao Wei Wenyuan Liu Binjie Ding Jiangxin Yang Yanpeng Cao |
| author_sort | Yanlong Cao |
| collection | DOAJ |
| description | High-quality reconstruction of polished surfaces is a promising yet challenging task in the industrial field. Due to its extreme reflective properties, state-of-the-art methods have not achieved a satisfying trade-off between retaining texture and removing the effects of specular outliers. In this paper, we propose a learning based pixel-level photometric stereo method to estimate the surface normal. A feature fusion convolutional neural network is used to extract the features from the normal map solved by the least square method and from the original images respectively, and combine them to regress the normal map. The proposed network outperforms the state-of-the-art methods on the DiLiGenT benchmark dataset. Meanwhile, we use the polished rail welding surface to verify the generalization of our method. To fit the complex geometry of the rails, we design a flexible photometric stereo information collection hardware with multi-angle lights and multi-view cameras, which can collect the light and shade information of the rail surface for photometric stereo. The experimental results indicate that the proposed method is able to reconstruct the normal of the polished surface at the pixel level with abundant texture information. |
| first_indexed | 2024-03-09T21:34:39Z |
| format | Article |
| id | doaj.art-c5c8e3db8ae24df9801f90fddbbb4394 |
| institution | Directory Open Access Journal |
| issn | 2075-1702 |
| language | English |
| last_indexed | 2024-03-09T21:34:39Z |
| publishDate | 2022-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Machines |
| spelling | doaj.art-c5c8e3db8ae24df9801f90fddbbb43942023-11-23T20:48:34ZengMDPI AGMachines2075-17022022-02-0110212010.3390/machines10020120A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished SurfacesYanlong Cao0Xiaoyao Wei1Wenyuan Liu2Binjie Ding3Jiangxin Yang4Yanpeng Cao5State Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou 310027, ChinaHigh-quality reconstruction of polished surfaces is a promising yet challenging task in the industrial field. Due to its extreme reflective properties, state-of-the-art methods have not achieved a satisfying trade-off between retaining texture and removing the effects of specular outliers. In this paper, we propose a learning based pixel-level photometric stereo method to estimate the surface normal. A feature fusion convolutional neural network is used to extract the features from the normal map solved by the least square method and from the original images respectively, and combine them to regress the normal map. The proposed network outperforms the state-of-the-art methods on the DiLiGenT benchmark dataset. Meanwhile, we use the polished rail welding surface to verify the generalization of our method. To fit the complex geometry of the rails, we design a flexible photometric stereo information collection hardware with multi-angle lights and multi-view cameras, which can collect the light and shade information of the rail surface for photometric stereo. The experimental results indicate that the proposed method is able to reconstruct the normal of the polished surface at the pixel level with abundant texture information.https://www.mdpi.com/2075-1702/10/2/120polished surfacespecular reflectionphotometric stereofeature fusion3D reconstruction/modeling |
| spellingShingle | Yanlong Cao Xiaoyao Wei Wenyuan Liu Binjie Ding Jiangxin Yang Yanpeng Cao A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces Machines polished surface specular reflection photometric stereo feature fusion 3D reconstruction/modeling |
| title | A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces |
| title_full | A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces |
| title_fullStr | A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces |
| title_full_unstemmed | A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces |
| title_short | A Novel Learning Based Non-Lambertian Photometric Stereo Method for Pixel-Level Normal Reconstruction of Polished Surfaces |
| title_sort | novel learning based non lambertian photometric stereo method for pixel level normal reconstruction of polished surfaces |
| topic | polished surface specular reflection photometric stereo feature fusion 3D reconstruction/modeling |
| url | https://www.mdpi.com/2075-1702/10/2/120 |
| work_keys_str_mv | AT yanlongcao anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT xiaoyaowei anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT wenyuanliu anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT binjieding anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT jiangxinyang anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT yanpengcao anovellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT yanlongcao novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT xiaoyaowei novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT wenyuanliu novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT binjieding novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT jiangxinyang novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces AT yanpengcao novellearningbasednonlambertianphotometricstereomethodforpixellevelnormalreconstructionofpolishedsurfaces |