High precision polarization image reconstruction algorithm based on depth residual convolutional neural network- in situ replacement optimization

Fan Bu; Dalei Yao; Weicheng Cao; Yongqing Yang; Xueli Wang

doi:10.1117/12.2690430

6 November 2023 High precision polarization image reconstruction algorithm based on depth residual convolutional neural network- in situ replacement optimization

Fan Bu, Dalei Yao, Weicheng Cao, Yongqing Yang, Xueli Wang

Author Affiliations +

Proceedings Volume 12921, Third International Computing Imaging Conference (CITA 2023); 129212G (2023) https://doi.org/10.1117/12.2690430
Event: Third International Computing Imaging Conference (CITA 2023), 2023, Sydney, Australia

Abstract

Compared to traditional polarization imaging methods such as time sharing, amplitude sharing, and aperture sharing, the Division of Focal Plane Polarimeter (DOFP) polarization imaging method has obvious advantages such as simultaneous imaging, compact optical and mechanical structure, small size, low power consumption, and high reliability. Therefore, this technology is currently a research hotspot in the field of polarization imaging. However, the focal plane polarization imaging technology uses a Micro Polarization Array (MPA) detector, which only captures a single polarization direction for a pixel, resulting in reduced spatial resolution of polarization images. In order to improve spatial resolution and the impact of factors such as unit mismatch in traditional interpolation methods on the detection system, this paper proposes a polarization image interpolation method based on depth residual convolution neural network. This algorithm closely combines the periodic phase characteristics of focal plane polarized images, designs a matched phase convolution kernel for feature extraction based on the pattern periodicity of four channel polarized image blocks, and designs a demosaic image interpolation method based on generating confrontation networks and phase convolution. Experimental results show that this algorithm can effectively reconstruct full resolution polarized images and is superior to traditional methods in terms of vision and Grayscale Mean Gradient (G).

Citation Download Citation

Fan Bu, Dalei Yao, Weicheng Cao, Yongqing Yang, and Xueli Wang "High precision polarization image reconstruction algorithm based on depth residual convolutional neural network- in situ replacement optimization", Proc. SPIE 12921, Third International Computing Imaging Conference (CITA 2023), 129212G (6 November 2023); https://doi.org/10.1117/12.2690430

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
10 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Polarization

Reconstruction algorithms

Polarized light

Image restoration

Interpolation

Convolution

Polarization imaging

Show All Keywords

Keywords/Phrases

Search In:

Publication Years