Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces

Mingjun Ren, G. Xiao, L. Zhu, Wenhan Zeng, D. Whitehouse

Research output: Contribution to journalArticle

Abstract

This paper presents an in-process inspection approach for quality control of non-diffuse curved surfaces based upon an evolution of conventional photometric stereo. An inverse reflectance model is proposed to reveal the non-linear reflectance behaviour of general non-diffuse surfaces from images based on a neural network. The model can directly be used to drive a high accurate photometric stereo which only requires a single RGB image with a pre-captured collocated image. This allows the technique to realize in-process inspection of moving surfaces with micro-second level capturing rate. Experimental study confirms the excellent texture recovery and defect detection capabilities in mass production.

Original languageEnglish
Pages (from-to)563-566
Number of pages4
JournalCIRP Annals
Volume68
Issue number1
Early online date24 Apr 2019
DOIs
Publication statusPublished - 2019

Fingerprint

Inspection
Quality control
Textures
Neural networks
Recovery
Defect detection

Cite this

Ren, Mingjun ; Xiao, G. ; Zhu, L. ; Zeng, Wenhan ; Whitehouse, D. / Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces. In: CIRP Annals. 2019 ; Vol. 68, No. 1. pp. 563-566.
@article{830a8f171a114cfa9a3a0b8f02c0ccc6,
title = "Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces",
abstract = "This paper presents an in-process inspection approach for quality control of non-diffuse curved surfaces based upon an evolution of conventional photometric stereo. An inverse reflectance model is proposed to reveal the non-linear reflectance behaviour of general non-diffuse surfaces from images based on a neural network. The model can directly be used to drive a high accurate photometric stereo which only requires a single RGB image with a pre-captured collocated image. This allows the technique to realize in-process inspection of moving surfaces with micro-second level capturing rate. Experimental study confirms the excellent texture recovery and defect detection capabilities in mass production.",
keywords = "Quality control, Surface, Visual inspection",
author = "Mingjun Ren and G. Xiao and L. Zhu and Wenhan Zeng and D. Whitehouse",
year = "2019",
doi = "10.1016/j.cirp.2019.04.013",
language = "English",
volume = "68",
pages = "563--566",
journal = "CIRP Annals - Manufacturing Technology",
issn = "0007-8506",
publisher = "Elsevier USA",
number = "1",

}

Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces. / Ren, Mingjun; Xiao, G.; Zhu, L.; Zeng, Wenhan; Whitehouse, D.

In: CIRP Annals, Vol. 68, No. 1, 2019, p. 563-566.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces

AU - Ren, Mingjun

AU - Xiao, G.

AU - Zhu, L.

AU - Zeng, Wenhan

AU - Whitehouse, D.

PY - 2019

Y1 - 2019

N2 - This paper presents an in-process inspection approach for quality control of non-diffuse curved surfaces based upon an evolution of conventional photometric stereo. An inverse reflectance model is proposed to reveal the non-linear reflectance behaviour of general non-diffuse surfaces from images based on a neural network. The model can directly be used to drive a high accurate photometric stereo which only requires a single RGB image with a pre-captured collocated image. This allows the technique to realize in-process inspection of moving surfaces with micro-second level capturing rate. Experimental study confirms the excellent texture recovery and defect detection capabilities in mass production.

AB - This paper presents an in-process inspection approach for quality control of non-diffuse curved surfaces based upon an evolution of conventional photometric stereo. An inverse reflectance model is proposed to reveal the non-linear reflectance behaviour of general non-diffuse surfaces from images based on a neural network. The model can directly be used to drive a high accurate photometric stereo which only requires a single RGB image with a pre-captured collocated image. This allows the technique to realize in-process inspection of moving surfaces with micro-second level capturing rate. Experimental study confirms the excellent texture recovery and defect detection capabilities in mass production.

KW - Quality control

KW - Surface

KW - Visual inspection

UR - http://www.scopus.com/inward/record.url?scp=85064593425&partnerID=8YFLogxK

U2 - 10.1016/j.cirp.2019.04.013

DO - 10.1016/j.cirp.2019.04.013

M3 - Article

VL - 68

SP - 563

EP - 566

JO - CIRP Annals - Manufacturing Technology

JF - CIRP Annals - Manufacturing Technology

SN - 0007-8506

IS - 1

ER -