Application of Partial Differential Segmentation Model with Adaptive Weight in Infrared Image of Substation Equipment
-
摘要: 针对变电设备运维中的红外检测图像无法被准确分割的问题,本文应用了一种改进的自适应权重偏微分图像分割方法。通过分析红外图像信噪比低、边缘模糊、对比度低以及最常出现的灰度不均匀问题,在明确传统分割方法不足的基础上,对基于偏微分方程的分割模型开展改进。本文所提出的自适应权重的LGIF分割模型利用目标设备和背景灰度不均匀程度不同的特点,将其与区域内的平均灰度值联系起来,针对性调整模型中全局能量项和局部能量项权重,以弥补现有算法不足。在多种场景下经实验验证,本文模型相较阈值法、CV分割模型和固定权重LGIF模型均更为有效准确,表现稳定,方便了后续的特征提取和识别。Abstract: To address the problem whereby the equipment area cannot be accurately segmented in the infrared image while maintaining substation equipment, this study applied an improved adaptive weight partial differential image segmentation method to segment the equipment area. By analyzing problems such as low signal-to-noise ratio, blurred edges, low contrast, and uneven grayscale of images, we investigated the disadvantages of traditional image segmentation methods, and the segmentation model based on partial differential equations was improved. The proposed adaptive weight LGIF model utilizes the different gray scale inhomogeneity of the target equipment and the background, associated it with the respective average gray scale, and adjusted the weights of the model's global and local energy items. Experiments in a variety of scenarios have verified that the model in this study is more effective and accurate than the OTSU method, CV model, and fixed-weight LGIF model, which is convenient for follow-up feature extraction and recognition.
-
Keywords:
- substation equipment /
- infrared image /
- image segmentation /
- active contour
-
-
表 1 三种情况设备分割情况统计
Table 1 Statistics of equipment segmentation in three situations
Background type Total Correct outcome Correct percentage/% OTSU CV LGIF Ours OTSU CV LGIF Ours Pure 69 42 50 30 58 62 74 44 85 Mess 41 30 26 14 28 73 65 35 70 Intensity inhomogeneity 110 64 66 52 69 58 60 47 63 表 2 本文方法与阈值法对比评价
Table 2 Comparison between the method in this paper and the Otsu method
Fig Method IOU Dice VOE RVD Precision Recall Eg.1 Ours 0.68 0.81 0.17 0.19 0.75 0.89 Eg.1 OTSU 0.55 0.71 0.27 0.31 0.63 0.83 Eg.2 Ours 0.54 0.70 0.25 0.29 0.62 0.80 Eg.2 OTSU 0.00 0.00 −1.05 −0.69 0.00 0.00 Eg.3 Ours 0.57 0.73 0.35 0.42 0.62 0.88 Eg.3 OTSU 0.26 0.41 1.12 2.52 0.26 0.93 Eg.4 Ours 0.47 0.64 0.31 0.37 0.55 0.76 Eg.4 OTSU 0.35 0.52 0.81 1.36 0.37 0.88 Eg.5 Ours 0.61 0.76 −0.08 −0.08 0.79 0.73 Eg.5 OTSU 0.44 0.61 0.48 0.64 0.49 0.81 Eg.6 Ours 0.48 0.65 0.62 0.89 0.50 0.94 Eg.6 OTSU 0.43 0.60 0.76 1.23 0.43 0.97 Eg.7 Ours 0.69 0.82 0.01 0.01 0.82 0.83 Eg.7 OTSU 0.63 0.77 0.26 0.29 0.68 0.88 Eg.8 Ours 0.67 0.80 0.09 0.09 0.76 0.83 Eg.8 OTSU 0.63 0.77 0.29 0.34 0.67 0.89 Eg.9 Ours 0.38 0.55 0.73 1.15 0.40 0.86 Eg.9 OTSU 0.33 0.50 0.91 1.68 0.34 0.93 -
[1] 罗军川. 电气设备红外诊断技术及在四川电网的应用研究[D]. 重庆: 重庆大学, 2003. LUO Junchuan. Researches on Infrared Diagnostic Technology of Electric Equipments and Application in Electric Power Grid of Sichuan[D]. Chongqing: Chongqing University, 2003.
[2] 李德刚. 红外诊断技术在电气设备状态检测中的研究与应用[D]. 济南: 山东大学, 2010. LI Degang. Research and Application on the Infrared Diagnostics of Electrical Equipment Testing[D]. Jinan: Shandong University, 2010.
[3] 周立辉, 张永生, 孙勇, 等. 智能变电站巡检机器人研制及应用[J]. 电力系统自动化, 2011, 35(19): 85-88. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201119018.htm ZHOU Lihui, ZHANG Yongsheng, SUN Yong, et al. Development and application of equipment inspection robot for smart substations[J]. Automation of Electric Power Systems, 2011, 35(10): 85-88. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201119018.htm
[4] OTSU N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62-66. DOI: 10.1109/TSMC.1979.4310076
[5] 陈跃伟, 彭道刚, 夏飞, 等. 基于区域生长法和BP神经网络的红外图像识别[J]. 激光与红外, 2018, 48(3): 401-408. DOI: 10.3969/j.issn.1001-5078.2018.03.024 CHEN Yuewei, PENG Daogang, XIA Fei, et al. Infrared image recognition based on region growing method and BP neural network[J]. Laser & Infrared, 2018, 48(3): 401-408. DOI: 10.3969/j.issn.1001-5078.2018.03.024
[6] 施兢业. 基于红外图像处理的变电设备识别与热故障诊断[D]. 上海: 上海电机学院, 2017. SHI Jingye. Substation Equipment Recognition and Thermal Fault Diagnosis Based on Infrared Image Processing[D]. Shanghai: Shanghai Dianji University, 2017.
[7] 袁建军. 基于偏微分方程图像分割技术的研究[D]. 重庆: 重庆大学, 2012. YUAN Jianjun. Image Segmentation Technology Based on Partical Differential Equation[D]. Chongqing: Chongqing University, 2012.
[8] 汤茂飞. 基于主动轮廓模型的红外图像分割方法研究[D]. 南京: 南京理工大学, 2015. TANG Maofei. Research on Infrared Image Segmentation Method Based on Active Contour Model[D]. Nanjing: Nanjing University of Science & Technology, 2015.
[9] CHAN T, VESE L. Active contours without edges[J]. IEEE Transcactions on Image Processing, 2001, 2(10): 266-277.
[10] LI C, KAO C Y, Gore J C, et al. Implicit active contours driven by local binary fitting energy[C]//2007 IEEE Conference on Computer Vision and Pattern Recognition., 2007: (DOI: 10.1109/CVPR.2007.383014).
[11] WANG Li, LI Chunming, SUN Quansen, et al. Active contours driven by local and global intensity fitting energy with application to brain MR image segmentation[J]. Computerized Medical Imaging and Graphics, 2009, 33(7): 520-531. DOI: 10.1016/j.compmedimag.2009.04.010
-
期刊类型引用(1)
1. 朱强,周维虎,陈晓梅,石俊凯,李冠楠. 高速实时近红外弱信号检测系统. 光学精密工程. 2022(24): 3116-3127 . 百度学术
其他类型引用(2)