A 3D measurement method for specular surfaces based on polarization image sequences and machine learning

Abstract Image highlights caused by specular reflections always conceal or attenuate the feature information of the samples in vision measurement. This paper investigates countermeasures to highlighted specular surface measurements. A measurement system is developed to capture a series of images with different polarization angles, the highlighted regions of which are taken as the trained samples for a back-propagation neural network, while the initial weights of the neural network are set as Gaussian distributions. Experimental results show that the proposed method can efficiently increase the stereo matching accuracy and hence recover the information in highlighted regions in vision measurements of specular surfaces.

[1]  Qionghai Dai,et al.  Fast and High Quality Highlight Removal From a Single Image , 2015, IEEE Transactions on Image Processing.

[2]  Guy Godin,et al.  Separation of diffuse and specular components of surface reflection by use of polarization and statistical analysis of images , 2004, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Kazuo Yamazaki,et al.  Quick 3D Modeling of Machining Environment by Means of On-machine Stereo Vision with Digital Decomposition , 2007 .

[4]  Jun Sun,et al.  Highlight removal for camera captured documents based on image stitching , 2016, 2016 IEEE 13th International Conference on Signal Processing (ICSP).

[5]  David J. Whitehouse,et al.  Model-driven photometric stereo for in-process inspection of non-diffuse curved surfaces , 2019, CIRP Annals.

[6]  Shoji Tominaga,et al.  Surface Identification Using the Dichromatic Reflection Model , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  Yi Ping Tang,et al.  Research on Identification of Color Structured Light Based on Dichromatic Reflection Model , 2011 .

[8]  Quanxin Liu,et al.  Camera calibration method based on optimal polarization angle , 2019, Optics and Lasers in Engineering.

[9]  Basel Salahieh,et al.  Multi-polarization fringe projection imaging for high dynamic range objects. , 2014, Optics express.

[10]  Surya Ganguli,et al.  Deep Information Propagation , 2016, ICLR.

[11]  Peter Kinnell,et al.  Autonomous metrology for robot mounted 3D vision systems , 2017 .

[12]  Min Xu,et al.  Research and development of light field microscope for measuring 3D microstructures , 2019, International Symposium on Advanced Optical Manufacturing and Testing Technologies (AOMATT).

[13]  Zhengyou Zhang,et al.  A Flexible New Technique for Camera Calibration , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[14]  Shijie Feng,et al.  General solution for high dynamic range three-dimensional shape measurement using the fringe projection technique , 2014 .

[15]  Bernd Scholz-Reiter,et al.  Design of deep convolutional neural network architectures for automated feature extraction in industrial inspection , 2016 .

[16]  C. Mersmann,et al.  A method for the automated positioning and alignment of fibre-reinforced plastic structures based on machine vision , 2008 .