Meter image enhancement method in high light substation based on improved CycleGAN

Xi Chen; Shaosheng Fan

doi:10.1117/12.2689816

31 July 2023 Meter image enhancement method in high light substation based on improved CycleGAN

Xi Chen, Shaosheng Fan

Proceedings Volume 12747, Third International Conference on Optics and Image Processing (ICOIP 2023); 127471Y (2023) https://doi.org/10.1117/12.2689816
Event: Third International Conference on Optics and Image Processing (ICOIP 2023), 2023, Hangzhou, China

Abstract

When the patrol robot recognizes the instrument under strong light, the collected instrument image has serious exposure, details loss and other phenomena. Aiming at the above problems, an improved CycleGAN method for enhancing the instrument image in the substation under strong light is proposed. First of all, for the generator, the CSD network and the self-attention mechanism ACmix module are improved to improve the lighting processing effect. Secondly, the idea of Patch-GAN is adopted in the discriminator, and the feature graph is mapped to an N×N matrix at the end to improve the processing ability of the details. The experimental results show that the peak signal-to-noise ratio of the image is increased from 15dB to 20dB after the above improvement; The index of structural similarity increased from 0.5 to 0.8; The recognition accuracy after enhancement has also been improved by 6%. In conclusion, the method in this paper can eliminate the influence of illumination on instrument recognition, and has strong robustness.

Citation Download Citation

Xi Chen and Shaosheng Fan "Meter image enhancement method in high light substation based on improved CycleGAN", Proc. SPIE 12747, Third International Conference on Optics and Image Processing (ICOIP 2023), 127471Y (31 July 2023); https://doi.org/10.1117/12.2689816

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