多尺度加权Retinex变压器油下图像增强

强虎; 钟羽中; 佃松宜

doi:10.11999/JEIT240645

多尺度加权Retinex变压器油下图像增强

doi: 10.11999/JEIT240645

强虎¹,
钟羽中¹,
佃松宜^{1, 2, ,}

1.
四川大学电气工程学院成都 610065
2.
四川大学深地工程智能建造与健康运维全国重点实验室成都 610065

基金项目: 国家自然科学基金(62203314)

详细信息

作者简介:
强虎：男，博士生，研究方向为人工智能、计算机视觉

钟羽中：女，副教授，研究方向为计算机视觉、图像处理

佃松宜：男，教授，研究方向为先进控制、感知与人工智能

通讯作者:
佃松宜　scudiansy@scu.edu.cn

中图分类号: TN911.73; TP391.41
计量
- 文章访问数: 69
- HTML全文浏览量: 23
- PDF下载量: 7
- 被引次数: 0
出版历程
- 收稿日期: 2024-07-23
- 修回日期: 2024-11-08
- 网络出版日期: 2024-11-13

Image Enhancement under Transformer Oil Based on Multi-Scale Weighted Retinex

1.
College of Electrical Engineering, Sichuan University, Chengdu 610065, China
2.
State Key Laboratory of Intelligent Construction and Healthy Operation and Maintenance of Deep Underground Engineering, Sichuan University, Chengdu 610065, China

Funds: The National Natural Science Foundation of China (62203314)

摘要

摘要: 针对变压器油下图像存在颜色失真、亮度低和细节失真问题，该文提出一种多尺度加权Retinex变压器油下图像增强算法。首先，为了缓解变压器油下图像颜色失真问题，提出一种混合动态颜色通道补偿算法，根据拍摄图像各通道的衰减状态对衰减通道进行动态补偿。然后，为了解决细节失真问题，提出一种锐化权重加权策略。最后，该文创新性采用金字塔多尺度融合策略对不同尺度Retinex反射分量和相应权重图进行加权融合得到变压器油下清晰图像。实验结果表明所提算法可以有效解决变压器油下图像复杂退化问题。
- 变压器油下图像增强 /
- Retinex /
- 通道补偿 /
- 多尺度加权
Abstract: To solve the degradation problems such as color distortion, low brightness, and detail loss in images under transformer oil, a multi-scale weighted Retinex algorithm for image enhancement is proposed in this paper. Firstly, in order to alleviate the color distortion problem of image under transformer oil, a hybrid dynamic color channel compensation algorithm is proposed, which dynamically compensates according to the attenuation state of each channel of the captured image. Then, in order to solve the problem of detail loss, a sharpening weight strategy is proposed. Finally, pyramid multi-scale fusion strategy is used to weighted fuse different-scale Retinex reflection components and corresponding weight maps to obtain clear images under transformer oil. Experimental results demonstrate that the algorithm proposed in this paper can effectively solve the complex degradation problem of image under transformer oil.
- Image enhancement under transformer oil /
- Retinex /
- Channel compensation /
- Multi-scale weighted

HTML全文

图 1 算法流程图

下载: 全尺寸图片幻灯片

图 2 不同尺度Retinex反射分量图

下载: 全尺寸图片幻灯片

图 3 不同实验场景

下载: 全尺寸图片幻灯片

图 4 不同场景下采集的变压器油下图像

下载: 全尺寸图片幻灯片

图 5 不同算法对图4(a)增强结果

下载: 全尺寸图片幻灯片

图 6 不同算法对图4(b)增强结果

下载: 全尺寸图片幻灯片

图 7 不同算法对图4(c)增强结果

下载: 全尺寸图片幻灯片

图 8 不同子模块对变压器油下图像增强效果

下载: 全尺寸图片幻灯片

图 9 所提算法对水下退化图像增强效果

下载: 全尺寸图片幻灯片

1 混合动态颜色通道补偿

输入：相机拍摄图像${I_{{\text{in}}}}$，增益系数$\omega $
输出：补偿后的图像${I_{{\text{out}}}}$
(1) $B,G,R \leftarrow {\text{split}}({I_{{\text{in}}}})$
(2) ${I_{{\text{Max}}}} \leftarrow \max (R,G,B)$
(3) ${I_{{\text{Min}}}} \leftarrow \min (R,G,B)$
(4) if ${I_{{\text{Max}}}} = \bar R$
(5) 　if ${I_{{\text{Min}}}} = \bar G$ then
(6) 　　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据$ {V}_{\mathrm{min}}=G $, 　　　　${V_{{\text{med}}}} = B,{V_{\max }} = R$
(7) 　end if
(8) 　if ${I_{{\text{Min}}}} = \bar B$ then
(9) 　　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据${V_{\min }} = B$, ${V_{{\text{med}}}} = G $, 　　　　${V_{\max }} = R$
(10) end if
(11) end if
(12) if ${I_{{\text{Max}}}} = \bar B$
(13) if ${I_{{\text{Min}}}} = \bar R$ then
(14) 　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据${V_{\min }} = R$, ${V_{{\text{med}}}} = G $, 　　　　 ${V_{\max }} = B$
(15) end if
(16) if ${I_{{\text{Min}}}} = \bar G$ then
(17) 　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据${V_{\min }} = G$, ${V_{{\text{med}}}} = R $, 　　　　${V_{\max }} = B$
(18) end if
(19) end if
(20) if ${I_{{\text{Max}}}} = \bar G$
(21) if ${I_{{\text{Min}}}} = \bar R$ then
(22) 　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据${V_{\min }} = R$, ${V_{{\text{med}}}} = B $, 　　　　${V_{\max }} = G$
(23) end if
(24) if ${I_{{\text{Min}}}} = \bar B$ then
(25) 　计算${V_{{\text{com\_min}}}},{V_{{\text{com\_med}}}}$根据${V_{\min }} = B$, ${V_{{\text{med}}}} = R $, 　　　　${V_{\max }} = G$
(26) end if
(27) end if
(28) ${I_{{\text{out}}}} \leftarrow {\text{merge}}(\bar B,\bar G,\bar R)$
(29) return ${I_{{\text{out}}}}$

下载: 导出CSV

表 1 UIQM, FUDM和NIQE无参考图像质量评估结果

指标	方法
指标	原图	UCM	UDCP	IBLA	ULAP	Water-Net	Shallow-UWnet	UDnet	本文
UIQM	1.476	1.943	1.417	2.144	1.379	2.272	1.273	1.880	3.265
FDUM	0.184	0.224	0.229	0.298	0.294	0.249	0.187	0.183	0.379
NIQE	5.021	4.864	5.315	5.714	4.815	4.859	5.240	4.754	4.681

下载: 导出CSV

表 2 不同模块消融实验

HCC	DW	PF	UIQM	FDUM	NIQE
			1.476	0.184	5.021
√			1.508	0.206	4.982
		√	2.965	0.283	4.630
	√	√	3.112	0.343	4.501
√	√	√	3.265	0.379	4.681

下载: 导出CSV

参考文献(26)

[1]	JHA M and BHANDARI A K. CBLA: Color balanced locally adjustable underwater image enhancement[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 5020911. doi: 10.1109/TIM.2024.3396850.
[2]	ZHANG Dehuan, WU Chenyu, ZHOU Jingchun, et al. Robust underwater image enhancement with cascaded multi-level sub-networks and triple attention mechanism[J]. Neural Networks, 2024, 169: 685–697. doi: 10.1016/j.neunet.2023.11.008.
[3]	YANG H Y, CHEN Peiyin, HUANG C C, et al. Low complexity underwater image enhancement based on dark channel prior[C]. 2011 Second International Conference on Innovations in Bio-inspired Computing and Applications, Shenzhen, China, 2011: 17–20. doi: 10.1109/IBICA.2011.9.
[4]	QIANG Hu, ZHONG Yuzhong, ZHU Yuqi, et al. Underwater image enhancement based on multichannel adaptive compensation[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 5014810. doi: 10.1109/TIM.2024.3378290.
[5]	DREWS JR P, DO NASCIMENTO E, MORAES F, et al. Transmission estimation in underwater single images[C]. 2013 IEEE International Conference on Computer Vision Workshops, Sydney, Australia, 2013: 825–830. doi: 10.1109/ICCVW.2013.113.
[6]	SONG Wei, WANG Yan, HUANG Dongmei, et al. A rapid scene depth estimation model based on underwater light attenuation prior for underwater image restoration[C]. Proceedings of the 19th Pacific-Rim Conference on Multimedia on Advances in Multimedia Information Processing – PCM 2018, Hefei, China, 2018: 678–688. doi: 10.1007/978-3-030-00776-8_62.
[7]	ZHANG Song, ZHAO Shili, AN Dong, et al. LiteEnhanceNet: A lightweight network for real-time single underwater image enhancement[J]. Expert Systems with Applications, 2024, 240: 122546. doi: 10.1016/j.eswa.2023.122546.
[8]	WANG Zhengyong, SHEN Liquan, XU Mai, et al. Domain adaptation for underwater image enhancement[J]. IEEE Transactions on Image Processing, 2023, 32: 1442–1457. doi: 10.1109/TIP.2023.3244647.
[9]	米泽田, 晋洁, 李圆圆, 等. 基于多尺度级联网络的水下图像增强方法[J]. 电子与信息学报, 2022, 44(10): 3353–3362. doi: 10.11999/JEIT220375. MI Zetian, JIN Jie, LI Yuanyuan, et al. Underwater image enhancement method based on multi-scale cascade network[J]. Journal of Electronics & Information Technology, 2022, 44(10): 3353–3362. doi: 10.11999/JEIT220375.
[10]	LI Chongyi, ANWAR S, and PORIKLI F. Underwater scene prior inspired deep underwater image and video enhancement[J]. Pattern Recognition, 2020, 98: 107038. doi: 10.1016/j.patcog.2019.107038.
[11]	RAO Yuan, LIU Wenjie, LI Kunqian, et al. Deep color compensation for generalized underwater image enhancement[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2024, 34(4): 2577–2590. doi: 10.1109/TCSVT.2023.3305777.
[12]	WANG Keyan, HU Yan, CHEN Jun, et al. Underwater image restoration based on a parallel convolutional neural network[J]. Remote Sensing, 2019, 11(13): 1591. doi: 10.3390/rs11131591.
[13]	WU Shengcong, LUO Ting, JIANG Gangyi, et al. A two-stage underwater enhancement network based on structure decomposition and characteristics of underwater imaging[J]. IEEE Journal of Oceanic Engineering, 2021, 46(4): 1213–1227. doi: 10.1109/JOE.2021.3064093.
[14]	BUCHSBAUM G. A spatial processor model for object colour perception[J]. Journal of the Franklin Institute, 1980, 310(1): 1–26. doi: 10.1016/0016-0032(80)90058-7.
[15]	LAND E H and MCCANN J J. Lightness and retinex theory[J]. Journal of the Optical Society of America, 1971, 61(1): 1–11. doi: 10.1364/JOSA.61.000001.
[16]	JOBSON D J, RAHMAN Z, and WOODELL G A. Properties and performance of a center/surround retinex[J]. IEEE Transactions on Image Processing, 1997, 6(3): 451–462. doi: 10.1109/83.557356.
[17]	RAHMAN Z, JOBSON D J, and WOODELL G A. Multi-scale retinex for color image enhancement[C]. The 3rd IEEE International Conference on Image Processing, Lausanne, Switzerland, 1996: 1003–1006. doi: 10.1109/ICIP.1996.560995.
[18]	PANETTA K, GAO Chen, and AGAIAN S. Human-visual-system-inspired underwater image quality measures[J]. IEEE Journal of Oceanic Engineering, 2016, 41(3): 541–551. doi: 10.1109/JOE.2015.2469915.
[19]	YANG Ning, ZHONG Qihang, LI Kun, et al. A reference-free underwater image quality assessment metric in frequency domain[J]. Signal Processing: Image Communication, 2021, 94: 116218. doi: 10.1016/j.image.2021.116218.
[20]	MITTAL A, SOUNDARARAJAN R, and BOVIK A C. Making a “completely blind” image quality analyzer[J]. IEEE Signal Processing Letters, 2013, 20(3): 209–212. doi: 10.1109/LSP.2012.2227726.
[21]	IQBAL K, ODETAYO M, JAMES A, et al. Enhancing the low quality images using unsupervised colour correction method[C]. 2010 IEEE International Conference on Systems, Man and Cybernetics, Istanbul, Turkey, 2010: 1703–1709. doi: 10.1109/ICSMC.2010.5642311.
[22]	PENG Y T and COSMAN P C. Underwater image restoration based on image blurriness and light absorption[J]. IEEE Transactions on Image Processing, 2017, 26(4): 1579–1594. doi: 10.1109/TIP.2017.2663846.
[23]	LI Chongyi, GUO Chunle, REN Wenqi, et al. An underwater image enhancement benchmark dataset and beyond[J]. IEEE Transactions on Image Processing, 2020, 29: 4376–4389. doi: 10.1109/TIP.2019.2955241.
[24]	NAIK A, SWARNAKAR A, and MITTAL K. Shallow-UWnet: Compressed model for underwater image enhancement (student abstract)[C]. The Thirty-Fifth AAAI Conference on Artificial Intelligence, Palo Alto, USA, 2021: 15853–15854. doi: 10.1609/aaai.v35i18.17923.
[25]	SALEH A, SHEAVES M, JERRY D, et al. Adaptive uncertainty distribution in deep learning for unsupervised underwater image enhancement[J]. arXiv preprint arXiv: 2212.08983, 2022.
[26]	LIU Risheng, FAN Xin, ZHU Ming, et al. Real-world underwater enhancement: Challenges, benchmarks, and solutions under natural light[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2020, 30(12): 4861–4875. doi: 10.1109/TCSVT.2019.2963772.