Fusion of infrared and visible light images of power equipment based on Schatten-p LatLRR
-
摘要:
为了解决潜在低秩表示(LatLRR)方法中使用的核函数可能导致的对秩函数逼近出现偏差问题，采用基于Schatten-p范数与潜在低秩分解的方法，进行了理论分析和实验验证。通过中值滤波方法对图像去噪，利用基于Schatten-p范数和LatLRR的图像分解方法，将图像分解为低秩部分与显著部分；采用算数平均策略融合红外与可见光的低秩部分，采用求和策略融合红外与可见光图像的显著部分；最终采用求和策略融合已融合好的低秩部分与显著部分，得到兼具清晰的纹理信息和显著的热故障信息的红外与可见光融合图像。结果表明, 最佳融合效果的p值为0.6, 在7种算法中有最好的融合性能。该方法能够有效地捕捉电力系统红外与可见光源图像中丰富的整体结构和局部结构信息。
-
关键词:
- 图像处理 /
- 潜在低秩表示 /
- Schatten-p范数 /
- 中值滤波
Abstract:In order to address the potential deviation in rank function approximation caused by the kernel function used in the latent low-rank representation(LatLRR) method, an approach based on Schatten-p norm and latent low-rank decomposition was proposed. Theoretical analysis and experimental validation were conducted using this method. The images were first denoised using a median filtering method. The images were decomposed into low-rank and salient parts using the Schatten-p norm-based latent low-rank decomposition method combined with LatLRR. Then, an arithmetic mean strategy was employed to fuse the low-rank parts of the infrared and visible light images, while a summation strategy was used to fuse their salient parts. Finally, a summation strategy was applied to fuse the already merged low-rank and salient parts, resulting in fused infrared and visible light images with clear texture information and prominent thermal fault information. Through qualitative and quantitative experimental analysis, a p-value of 0.6 was determined to achieve the optimal fusion effect, and the proposed method outperformed seven other algorithms in fusion performance comparison. Through this approach, rich structural information at both global and local levels in infrared and visible light source images of power systems can be effectively captured.
-
-
![]()
图 2 图像分解
a—红外图像分解 b—可见光图像分解
Figure 2. Image decomposition
a—infrared image decomposition b—visible light image decomposition
![]()
图 6 源图像对
a—红外图像 b—可见光图像
Figure 6. Source image pair
a—infrared image b—visible light image
![]()
图 10 8种算法在PSNR、CC、MSE、Nabf上的评估结果
Figure 10. Evaluation results of 8 algorithms on PSNR, CC, MSE, NABF
表 1 p值效果比较
Table 1 Comparison of p-value effects
p-value PSNR CC MSE NABF 0.1 62.97063 0.63547 0.03485 0.09971 0.2 62.96614 0.63543 0.03482 0.09971 0.3 62.96836 0.63538 0.03479 0.09974 0.4 62.97593 0.63544 0.03476 0.09857 0.5 62.98138 0.63531 0.03474 0.09912 0.6 62.98348 0.63558 0.03461 0.09738 0.7 62.9802 0.63526 0.03468 0.09813 0.8 62.97511 0.63524 0.03467 0.09903 0.9 62.97606 0.63526 0.03473 0.09854 1.0 62.98052 0.63527 0.03470 0.10683 
下载: 导出CSV
-
[1] TANG W, HE F, LIU Y, et al. DATFuse: Infrared and visible image fusion via dual attention transformer[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023, 33(7): 3159-3172. DOI: 10.1109/TCSVT.2023.3234340
[2] 陈龙. 电力设备故障检测中图像融合算法的研究[D]. 吉林: 东北电力大学, 2021. CHEN L. Research on image fusion algorithm in fault detection of power equipment [D]. Jilin: Northeast Electric Power University, 2021(in Chinese).
[3] SHREYAMSHA KUMAR B K. Image fusion based on pixel significance using cross bilateral filter[J]. Signal, Image and Video Processing, 2015, 9: 1193-1204. DOI: 10.1007/s11760-013-0556-9
[4] ZHOU Z, WANG B, LI S, et al. Perceptual fusion of infrared and visible images through a hybrid multi-scale decomposition with Gaussian and bilateral filters[J]. Information Fusion, 2016, 30: 15-26. DOI: 10.1016/j.inffus.2015.11.003
[5] LIU Y, CHEN X, CHENG J, et al. Infrared and visible image fusion with convolutional neural networks[J]. International Journal of Wavelets, Multiresolution and Information Processing, 2018, 16(3): 1850018. DOI: 10.1142/S0219691318500182
[6] MA J, ZHOU Z, WANG B, et al. Infrared and visible image fusion based on visual saliency map and weighted least square optimization[J]. Infrared Physics & Technology, 2017, 82: 8-17.
[7] LI H, WU X J. Infrared and visible image fusion using latent low-rank representation [EB/OL]. (2022-01-29)[ 2024-04-02]. https://doi.org/10.48550/arXiv.1804.08992.
[8] LI H, WU X J, KITTLER J. MDLatLRR: A novel decomposition method for infrared and visible image fusion[J]. IEEE Transactions on Image Processing, 2020, 29: 4733-4746. DOI: 10.1109/TIP.2020.2975984
[9] 张亚加, 邱啟蒙, 刘恒, 等. 结合潜在低秩分解和稀疏表示的脑部图像融合[J]. 光电子·激光, 2023, 34(11): 1225-1232. ZHANG Y J, QIU Q M, LIU H, et al. Fusion of brain images combining latent low-rank decomposition and sparse representation[J]. Journal of Optoelectronics · Laser, 2023, 34(11): 1225-1232(in Chinese).
[10] 杨亚东, 黄胜一, 谭毅华. 基于低秩和重加权稀疏表示的红外弱小目标检测算法[J]. 应用科学学报, 2023, 41(5): 753-765. YANG Y D, HUANG Sh Y, TAN Y H. Infrared weak target detection algorithm based on low-rank and reweighted sparse representation[J]. Journal of Applied Sciences, 2023, 41(5): 753-765(in Chinese).
[11] 潘巧英. 基于多级潜在低秩表示的红外与可见光图像融合研究[D]. 杭州: 杭州电子科技大学, 2023. PAN Q Y. Research on infrared and visible image fusion based on multi-level latent low-rank representation[D]. Hangzhou: Hangzhou University of Electronic Science and Technology, 2023(in Chin-ese).
[12] 陈家益, 战荫伟, 曹会英, 等. 消除椒盐噪声的基于纹理特征的决策滤波[J]. 电子测量与仪器学报, 2019, 33(3): 126-135. CHEN J Y, ZHAN Y W, CAO H Y, et al. Decision filtering based on texture features for eliminating salt-and-pepper noise[J]. Journal of Electronic Measurement and Instrumentation, 2019, 33(3): 126-135(in Chinese).
[13] 马飞, 王梓璇, 刘思雨. 基于深度图像先验的高光谱图像去噪方法[J]. 激光技术, 2024, 48(3): 379-386. DOI: 10.7510/jgjs.issn.1001-3806.2024.03.013 MA F, WANG Z X, LIU S Y. Hyperspectral image denoising method based on depth image prior[J]. Laser Technology, 2024, 48(3): 379-386(in Chinese). DOI: 10.7510/jgjs.issn.1001-3806.2024.03.013
[14] 赖烨辉, 黄慧英, 彭绍婷, 等. 利用加权对数范数分解的矩阵填充算法[J]. 赣南师范大学学报, 2023, 44(6): 112-119. [15] 孙彬, 诸葛吴为, 高云翔, 等. 基于潜在低秩表示的红外和可见光图像融合[J]. 红外技术, 2022, 44(8): 853-86. SUN B, ZHUGE W W, GAO Y X, et al. Infrared and visible image fusion based on latent low-rank representation[J]. Infrared Techno-logy, 2022, 44(8): 853-86(in Chinese).
[16] 张晓婷. 基于稀疏梯度与结构化矩阵分解的显著性目标检测[D]. 深圳: 深圳大学, 2020. ZHANG X T. Saliency object detection based on sparse gradient and structured matrix decomposition[D]. Shenzhen: Shenzhen University, 2020(in Chinese).
[17] 潘伟, 胡春安. 基于加权Schatten-p范数的矩阵填充及其应用[J]. 计算机应用与软件, 2023, 40(4): 230-235. PAN W, HU Ch A. Matrix completion based on weighted Schatten-p norm and its application[J]. Computer Applications and Software, 2023, 40(4): 230-235(in Chinese).
[18] 赵辽英, 潘巧英, 厉小润. 基于WSN-LatLRR的红外和可见光图像融合方法: CN113362281A[P]. 2021-09-07. ZHAO L Y, PAN Q Y, LI X R. Infrared and visible image fusion method based on WSN-LatLRR: CN113362281A[P]. 2021-09-07(in Chinese).
[19] 徐慧娴, 田洋川, 陈明举, 等. 基于潜在低秩表示的多聚焦图像融合方法[J]. 传感器与微系统, 2023, 42(5): 156-160. XU H X, TIAN Y Ch, CHEN M J, et al. Multi-focus image fusion method based on latent low-rank representation[J]. Sensors and Microsystems, 2023, 42(5): 156-160(in Chinese).
[20] 袁代玉, 袁丽华, 习腾彦, 等. 潜在低秩表示下的双判别器生成对抗网络的图像融合[J]. 光学精密工程, 2023, 31(7): 1085-1095. YUAN D Y, YUAN L H, XI T Y, et al. Image fusion based on dual discriminator generative adversarial networks under latent low-rank representation[J]. Optics and Precision Engineering, 2023, 31(7): 1085-1095(in Chinese).
计量
- 文章访问数: 41
- HTML全文浏览量: 14
- PDF下载量: 8