基于改进的MSRCR-CLAHE融合的水下图像增强算法

马敏慧; 王红茹; 王佳

基于改进的MSRCR-CLAHE融合的水下图像增强算法

马敏慧^1,,
王红茹^{1, 2, ,},
王佳^{1, 2}

1.
江苏科技大学机械工程学院，江苏镇江 212003
2.
江苏科技大学江苏省船海机械先进制造及工艺重点实验室，江苏镇江 212003

基金项目:

国家重点研发计划资助项目 2018YFC0309100

详细信息

作者简介:
马敏慧（1997-），男，硕士研究生，主要研究方向为机器视觉与图像处理。E-mail：1723164582@qq.com

通讯作者:
王红茹（1979-），女，工学博士，副教授，硕士生导师，研究方向为智能机器人技术。E-mail：wanghr@126.com

中图分类号: TP391
计量
- 文章访问数: 254
- HTML全文浏览量: 46
- PDF下载量: 57
- 被引次数: 0
出版历程
- 收稿日期: 2022-02-20
- 修回日期: 2022-04-06
- 刊出日期: 2023-01-20

An Underwater Image Enhancement Algorithm Based on Improved MSRCR-CLAHE Fusion

MA Minhui^1
,,
WANG Hongru^{1, 2
, ,},
WANG Jia^{1, 2}

1.
School of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, China
2.
Key Laboratory of Advanced Manufacture and Process for Marine Mechanical Equipment Research Institute, Jiangsu University of Science and Technology, Zhenjiang 212003, China

摘要

摘要: 针对海洋复杂成像环境导致的水下图像出现颜色衰退、对比度低等问题，提出一种改进的带色彩恢复的多尺度视网膜（Multi-Scale Retinex with Color Restore，MSRCR）与限制对比度自适应直方图均衡化（Contrast Limited Adaptive Histogram Equalization，CLAHE）多尺度融合的水下图像增强算法。首先，采用带有导向滤波的MSRCR算法解决水下图像颜色衰退的问题；其次，采用带有Gamma校正的CLAHE算法以提高水下图像的对比度；最后，对经过改进的MSRCR和CLAHE处理后的图像进行多尺度融合以获得细节增强后的水下图像。实验结果表明，和其他算法相比，文中算法的峰值信噪比（Peak Signal to Noise Ratio，PSNR）平均提高了9.3914、结构相似性（Structural Similarity Index Measure，SSIM）平均提高了0.3013、水下图像评价指标（Underwater Image Quality Evaluation，UIQE）平均提高了4.7047，能实现水下图像的有效增强。
- 水下图像增强 /
- CLAHE增强 /
- 带颜色恢复的多尺度Retinex /
- 导向滤波 /
- Gamma校正 /
- 多尺度融合
Abstract: To address the problems of color fading and low contrast in underwater images caused by the complex imaging environment in the ocean, improved Multi-Scale Retinex with Color Restore (MSRCR) and Contrast Limited Adaptive Histogram Equalization (CLAHE) multi-scale fusion algorithms for underwater image enhancement are proposed. First, the MSRCR algorithm with guided filtering was used to solve the problem of underwater image color fading. Second, the CLAHE algorithm with Gamma correction was used to improve the contrast of underwater images. Finally, the improved MSRCR and CLAHE images were fused at multi-scale to obtain an underwater image with enhanced detail. The experimental results show that, compared with other algorithms, the Peak Signal-To-Noise Ratio (PSNR) of the proposed algorithm is improved by 9.3914 on average, and the Structural Similarity Index Measure (SSIM) and Underwater Image Quality Evaluation (UIQE) increased by 0.3013 and 4.7047 on average, respectively, which can realize the effective enhancement of underwater images.
- underwater image enhancement /
- CLAHE enhancement /
- MSRCR /
- guided filtering /
- Gamma correction /
- multi-scale fusion

HTML全文

图 1 不同MSR的处理结果

Figure 1. Processing results of different MSR

下载: 全尺寸图片幻灯片

图 2 局部细节图

Figure 2. Local diagrams in details

下载: 全尺寸图片幻灯片

图 3 CLAHE算法直方图变换过程

Figure 3. Histogram transformation process of CLAHE algorithm

下载: 全尺寸图片幻灯片

图 4 不同直方图均衡算法的处理结果

Figure 4. Processing results of different histogram equalization algorithms

下载: 全尺寸图片幻灯片

图 5 颜色增强图像的权重图

Figure 5. Weight diagram of color enhanced images

下载: 全尺寸图片幻灯片

图 6 对比度增强图像的权重图

Figure 6. Weight diagram of contrast enhanced image

下载: 全尺寸图片幻灯片

图 7 图像多尺度融合

Figure 7. Image multi-scale fusion and reconstruction

下载: 全尺寸图片幻灯片

图 8 本文算法流程

Figure 8. Algorithm flow chart of this paper

下载: 全尺寸图片幻灯片

图 9 10种不同图像增强算法处理结果

Figure 9. Results of ten different image enhancement algorithms

下载: 全尺寸图片幻灯片

表 1 不同算法PSNR性能比较

Table 1. PSNR performance comparison of different algorithms

PNSR	Original	Reference[3]	Reference[7]	Reference[13]	Reference[16]	Ours
Picture 1	-	13.8014	14.2261	16.3929	21.0436	24.2896
Picture 2	-	6.2876	6.7461	6.2519	14.0142	18.9873
Picture 3	-	15.9585	12.5442	13.0328	20.9045	22.3212
Picture 4	-	7.3254	7.9521	8.1265	13.2158	19.9914
Picture 5	-	12.9871	14.8561	15.8516	17.3258	20.5563
Picture 6	-	15.6243	15.9985	16.2546	18.2319	24.7963
Picture 7	-	11.8274	13.2873	14.7931	19.2291	19.9639
Picture 8	-	10.2034	11.2544	15.2698	16.3245	19.3312
Picture 9	-	14.5758	15.9152	18.3223	20.5513	24.3698
Picture10	-	11.4522	12.3756	13.4851	16.6334	19.3497

下载: 导出CSV

表 2 不同算法SSIM性能比较

Table 2. Performance comparison of different SSIM algorithms

SSIM	Original	Reference[3]	Reference[7]	Reference[13]	Reference[16]	Ours
Picture 1	-	0.5609+	0.5943	0.8321	0.8384	0.9611
Picture 2	-	0.5223	0.5081	0.6869	0.8612	0.8874
Picture 3	-	0.6186	0.7926	0.8031	0.8299	0.9212
Picture 4	-	0.5743	0.5178	0.8163	0.8752	0.9649
Picture 5	-	0.7121	0.7963	0.8263	0.8998	0.9088
Picture 6	-	0.6933	0.7432	0.7966	0.8364	0.8997
Picture 7	-	0.6074	0.5927	0.6988	0.7411	0.8796
Picture 8	-	0.5871	0.5988	0.6355	0.7843	0.8894
Picture 9	-	0.6121	0.6028	0.7123	0.7652	0.9126
Picture 10	-	0.5386	0.6103	0.7521	0.8419	0.8696

下载: 导出CSV

表 3 不同算法UIQE性能比较

Table 3. UIQE performance comparison of different algorithms

UIQE	Original	Reference[3]	Reference[7]	Reference[13]	Reference[16]	Ours
Picture 1	2.6449	4.9242	4.6436	5.0271	3.6167	5.2238
Picture 2	1.7252	4.0814	0.2696	3.9252	2.1831	4.9121
Picture 3	1.9542	3.0251	1.3447	3.5738	2.2406	4.4633
Picture 4	1.5241	4.5296	1.0328	4.2153	3.6574	5.9685
Picture 5	1.7551	3.1221	3.0217	3.9746	4.9962	6.2312
Picture 6	1.9978	2.2173	2.1179	4.5023	4.8785	6.0178
Picture 7	1.2212	1.3258	2.3647	4.2589	4.5565	6.9872
Picture 8	2.0121	2.2365	3.4562	4.2199	4.8456	5.5463
Picture 9	2.7853	2.8742	3.9893	4.5631	4.7987	6.2971
Picture 10	0.6721	2.9255	3.2372	3.4801	1.5899	3.6943

下载: 导出CSV

参考文献(18)

[1]	Mishra A, Gupta M, Sharma P. Enhancement of underwater images using improved CLAHE[C]//2018 International Conference on Advanced Computation and Telecommunication (ICACAT), 2018: 1-6.
[2]	Sun J. Underwater Image Enhancement Algorithm Based on MSRCR[J]. Radio Engineering, 2019, 49(9): 783-787. doi: 10.3969/j.issn.1003-3106.2019.09.006
[3]	WANG F, ZHANG B, ZHANG C, et al. Low-light image joint enhancement optimization algorithm based on frame accumulation and multi-scale Retinex[J]. Ad Hoc Networks, 2020, 113(4): 102398.
[4]	Dhanya P R, Anilkumar S, AA Balakrishnan, et al. L-CLAHE intensification filter (L-CIF) algorithm for underwater image enhancement and color restoration[C]//2019 International Symposium on Ocean Technology (SYMPOL). IEEE, 2020: 117-128.
[5]	朱佳琦, 周丽丽, 闫晶晶, 等. 结合改进红通道先验与幂律校正CLAHE的水下图像复原方法[J]. 红外技术, 2021, 43(7): 696-701. http://hwjs.nvir.cn/article/id/f67e3336-e395-449d-88a7-3752f030808f ZHU J Q, ZHOU L L, YAN J J, et al. Underwater image restoration method combining improved red channel prior and power law correction CLAHE[J]. Infrared Technology, 2021, 43(7): 696-701. http://hwjs.nvir.cn/article/id/f67e3336-e395-449d-88a7-3752f030808f
[6]	范新南, 鹿亮, 史朋飞, 等. 结合MSRCR与多尺度融合的水下图像增强算法[J]. 国外电子测量技术, 2021, 40(2): 6-10. https://www.cnki.com.cn/Article/CJFDTOTAL-GWCL202102004.htm FAN X L, LU L, SHI P F, et al. Underwater image enhancement algorithm by MSRCR and multi-scale fusion[J]. Foreign Electronic Measurement Technology, 2021, 40(2): 6-10. https://www.cnki.com.cn/Article/CJFDTOTAL-GWCL202102004.htm
[7]	张薇, 郭继昌. 基于白平衡和相对全变分的低照度水下图像增强[J]. 激光与光电子学进展, 2020, 57(12): 213-220. https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202012022.htm ZHANG W, GUO J C. Low illumination underwater image enhancement based on white balance and relative total variation[J]. Advances in Laser and Optoelectronics, 2020, 57(12): 213-220. https://www.cnki.com.cn/Article/CJFDTOTAL-JGDJ202012022.htm
[8]	TENG L, XUE F. Remote sensing image enhancement via edge-preserving multiscale retinex[J]. IEEE Photonics Journal, 2019, 11(2): 1-10.
[9]	Rahman Z, Jobson D J, Woodell G. Retinex processing for automatic image enhancement[J]. J Electronic Imaging, 2004, 13(1): 100-110. doi: 10.1117/1.1636183
[10]	HE K, SUN J. Guided image filtering[C]//IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6): 1397-1409.
[11]	Kaur H, Rani J. MRI brain image enhancement using histogram equalization techniques[C]//2016 International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET), 2016: 770-773.
[12]	吕侃徽, 张大兴. 基于自适应直方图均衡化耦合拉普拉斯变换的红外图像增强算法[J]. 光学技术, 2021, 47(6): 747-753. https://www.cnki.com.cn/Article/CJFDTOTAL-GXJS202106018.htm LYU K H, ZHANG D X. Infrared image enhancement algorithm based on adaptive histogram equalization coupled with Laplace transform[J]. Optical Technology, 2021, 47(6): 747-753. https://www.cnki.com.cn/Article/CJFDTOTAL-GXJS202106018.htm
[13]	王晓俊, 李莉莉. 基于CLAHE-DCP融合的水下图像增强[J]. 中国水运(下半月), 2021, 21(8): 60-62. https://www.cnki.com.cn/Article/CJFDTOTAL-ZSUX202108027.htm WANG X J, LI L L. Underwater image enhancement based on CLAHE-DCP fusion[J]. China Water Transport (the second half of the month), 2021, 21(8): 60-62. https://www.cnki.com.cn/Article/CJFDTOTAL-ZSUX202108027.htm
[14]	王梓丞, 尹勇. 基于色彩平衡及校正的水下图像增强算法[J]. 舰船科学技术, 2021, 43(21): 154-159. https://www.cnki.com.cn/Article/CJFDTOTAL-JCKX202121030.htm WANG Z C, YI Y. Underwater image enhancement algorithm based on color balance and correction[J]. Ship Science and Technology, 2021, 43(21): 154-159. https://www.cnki.com.cn/Article/CJFDTOTAL-JCKX202121030.htm
[15]	MA Jinxiang, FAN Xinnan, ZHU Jianjun, et al. Multi-scale retinex with color restoration image enhancement based on Gaussian filtering and guided filtering[J]. International Journal of Modern Physics B, 2017, 31: 16-19.
[16]	杨亚绒, 李恒, 赵磊等. 改进的同态滤波与多尺度融合的水下图像增强[J/OL]. 机械科学与技术, 2021, 12(5): 1-9. YANG Y R, Li H, ZHAO L, et al. Improved homomorphic filtering and multi-scale fusion underwater image enhancement[J/OL]. Mechanical Science and Technology, 2021, 12(5): 1-9.
[17]	Horé A, Ziou D. Image quality metrics: PSNR vs. SSIM[C]//20th International Conference on Pattern Recognition, ICPR, 2010, 4(3): 28-29.
[18]	YANG Miao, Arcot Sowmya. An underwater color image quality evaluation metric[J]. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2015, 24(12): 24-29.