Low-Light Image Enhancement Algorithm for Underground Coal Mines Based on Retinex Theory

DOI: https://doi.org/10.62517/jike.202504216

Author(s)

Guangyao Yang1, Qingqing Ran2, Lihong Dong2,*, Naining Wen2

Affiliation(s)

1SHCCIG Yubei Coal Industry Co., Ltd, Yulin, China 2College of Computer Science and Technology, Xi’an University of Science and Technology, Xi’an, China *Corresponding Author

Abstract

The underground environment of coal mines is complex, causing issues such as insufficient lighting, low contrast, and high noise in the images collected by underground monitoring equipment. These problems severely affect subsequent image analysis and intelligent decision-making. This paper introduces a low-illumination image enhancement algorithm for underground coal mines, grounded in Retinex theory. Firstly, the projection module is utilized to process the original image, reducing the interference of noise on Retinex decomposition. Secondly, a decomposition network integrated with the U-Net structure is employed to accurately separate the illumination and reflection components. Finally, a self-calibration illumination network is introduced. Through multi-stage residual learning and self-calibration mapping, it can automatically adjust the illumination component. Experiments demonstrate that on the self-built underground image dataset, the proposed algorithm outperforms mainstream methods in terms of PSNR and SSIM metrics. The algorithm presented in this paper outperforms comparative algorithms in subjective human vision analysis, demonstrating its effectiveness in enhancing the visual quality of low-illumination images in underground coal mines.

Keywords

Low Illumination Imaget; Image Enhancement; Retinex Theory; Underground Coalmine.

References

[1]Deqiang Chen, Jiansheng Qiu, Xingge Guo. Review on key technologies of AI recognition for videos in coal mine Coal Science and Technology, 2023, 51(2):349-365. [2]A. Paul. Adaptive tri-plateau limit tri-histogram equalization algorithm for digital image enhancement The Visual Computer, 2023, 39(1):297-318. [3]Xiaojie Guo, Yutao Li, Haibin Ling. LIME: Low-light image enhancement via illumination map estimation IEEE Transactions on Image Processing, 2016, 26(2):982-993. [4]Yifan Jiang, Xiaowei Gong, Dong Liu, et al. EnlightenGAN: Deep light enhancement without paired supervision IEEE Transactions on Image Processing, 2021, 30: 2340-2349. [5]Ying Zhang, Jing Zhang, and Xiaojie Guo. Kindling the darkness: A practical low-light image enhancer Proceedings of the 27th ACM International Conference on Multimedia, Nice, France, 2019. [6]Chunle Guo, Chongyi Li, Jinjian Guo, Chen Change Loy, Junhui Hou, and Sam Kwong. Zero-reference deep curve estimation for low-light image enhancement Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, USA, 2020. [7]Chen Wei, Wenjing Wang, Wenhan Yang, and Jiaying Liu. Deep Retinex decomposition for low-light enhancement Proceedings of the 2018 British Machine Vision Conference, Newcastle, UK, 2018. [8]Ziteng Cui, Kunchang Li, Lin Gu, Shenglong Su, Pan Gao, and Zheng Jiang. You only need 90k parameters to adapt light: A lightweight transformer for image enhancement Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, Montreal, Canada, 2022. [9]Long Ma, Tengyu Ma, Risheng Liu, Xin Fan, and Zhongxuan Luo. Toward fast, flexible, and robust low-light image enhancement Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, USA, 2022. [10]Zheng Fu, Yang Yang, Xiaogang Tu, Yuhui Huang, Xiaoming Ding, and Kaiming Ma. Learning a simple low-light image enhancer from paired low-light instances Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, Canada, 2023.