Gas-liquid flow image segementation method based on deep convolutional neural network

doi:10.11772/j.issn.1001-9081.2022030376

Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (S1): 217-223.DOI: 10.11772/j.issn.1001-9081.2022030376

• Multimedia computing and computer simulation • Previous Articles

Gas-liquid flow image segementation method based on deep convolutional neural network

Ziliang CUI¹, Yuanyuan JU¹^,²^,³(), Dongdong LIU¹, Lin DAI¹, Qingtai XIAO³

^1.Faculty of Science，Kunming University of Science and Technology，Kunming Yunnan 650500，China
^2.Key Laboratory of Applied Statistics and Data Analysis of Department of Education of Yunnan Province （Kunming University of Science and Technology），Kunming Yunnan 650500，China
^3.State Key Laboratory of Complex Nonferrous Metal Resources Clean Utilization （Kunming University of Science and Technology），Kunming Yunnan 650093，China

Received:2022-03-23 Revised:2022-07-26 Accepted:2022-08-15 Online:2023-07-04 Published:2023-06-30
Contact: Yuanyuan JU

基于深度卷积神经网络的气液两相流图像分割方法

崔子良¹, 句媛媛¹^,²^,³(), 刘冬冬¹, 戴琳¹, 肖清泰³

^1.昆明理工大学理学院，昆明 650500
^2.云南省教育厅应用统计与数据分析重点实验室（昆明理工大学），昆明 650500
^3.省部共建复杂有色金属资源清洁利用国家重点实验室（昆明理工大学），昆明 650093

通讯作者: 句媛媛
作者简介:崔子良（1998—），男，山东滨州人，硕士研究生，主要研究方向：工业图像处理
句媛媛（1984—），女，云南昆明人，副教授，博士，主要研究方向：工业图像处理。jundeyy@126.com
刘冬冬（1996—），男，河南周口人，硕士研究生，主要研究方向：工业图像处理
戴琳（1968—），女，云南昆明人，教授，硕士，主要研究方向：应用统计分析
肖清泰（1989—），男，山东菏泽人，副教授，博士，主要研究方向：工业图像处理。
基金资助:
云南省博士后定向培养项目(109820210027);云南省科技厅科技计划项目(202101AU070031);云南省教育厅科学研究基金资助项目(2022J0059)

Abstract

Abstract:

To solve the problem of accurate identification of gas and liquid phases in gas-liquid two-phase flow， an image segmentation method based on deep convolutional neural network was proporsed. Firstly， the advantages and disadvantages of four image denoising methods and five image segmentation methods were compared. Secondly， the synthetic images were processed by image denoising and image segmentation， and the segmentation results were quantified by image segmentation evaluation indexes. Finally， image denoising and image segmentation were carried out on the public dataset images and the real gas-liquid two-phase flow images. The experimental results show that anisotropic diffusion filter， median filter， total variance filter and non-local mean filter have slightly different noise reduction performance on the bubble images， and non-local mean filter has the best effect. When the convolutional neural network method was used to segment bubble images， the values of Pixel Accuracy （PA）， Mean Pixel Accuracy （MPA）， Mean Intersection over Union （MIoU） and Frequency Weighted Intersection over Union （FWIoU） are more than 0.84. The proposed method has high accurary and excellent segmentation effect， but its calculation cost is higher than the traditional model. By comparing and analyzing the image processing technologies of gas-liquid two-phase flow， it can be found that deep learning method is an important research direction for gas-liquid two-phase flow images in the future.

Key words: gas-liquid flow, image preprocessing, image noise reduction, image segmentation, convolutional neural network

摘要：

为解决气液两相流中气相与液相的精准识别问题，提出了一种基于深度卷积神经网络的图像分割方法。首先，对比研究了4种图像去噪和5种图像分割方法的优缺点；其次，采用图像去噪和图像分割的方法研究了人工合成图像，并采用图像分割评价指标量化分割结果；最后，采用图像去噪和图像分割方法对公开数据集图像和真实气液两相流图像进行实验。实验结果表明，各向异性扩散滤波器、中值滤波器、全变差滤波器和非局部均值滤波器对气泡图像的降噪性能略有差异，非局部均值滤波器的效果最优；采用卷积神经网络方法分割气泡图像时，像素精确度（PA）、平均像素精确度（MPA）、平均交并比（MIoU）、频率加权交并比（FWIoU）四个评估指标的值均超过0.84，其精度较高、分割效果较为优异，但其计算成本高于传统方法。通过对比研究气液两相流图像的处理技术，可以发现深度学习方法是未来气液两相流图像的一个重要研究方向。

关键词: 气液两相流, 图像预处理, 图像去噪, 图像分割, 卷积神经网络

CLC Number:

TP391.41

Ziliang CUI, Yuanyuan JU, Dongdong LIU, Lin DAI, Qingtai XIAO. Gas-liquid flow image segementation method based on deep convolutional neural network[J]. Journal of Computer Applications, 2023, 43(S1): 217-223.

崔子良, 句媛媛, 刘冬冬, 戴琳, 肖清泰. 基于深度卷积神经网络的气液两相流图像分割方法[J]. 《计算机应用》唯一官方网站, 2023, 43(S1): 217-223.

Figures/Tables 10

References 30

1	曹玲玲. 转炉熔池气-渣-金多相流行为的模拟研究［D］. 北京：北京科技大学， 2018：1-32.
2	LIOTTI E， ARTETA C， ZISSERMAN A， et al. Crystal nucleation in metallic alloys using x-ray radiography and machine learning［J］. Science Advances， 2018， 4（4）： eaar4004. 10.1126/sciadv.aar4004
3	YOON S， KIM M， JANG M， et al. Deep optical imaging within complex scattering media［J］. Nature Reviews Physics， 2020， 2（3）： 141-158. 10.1038/s42254-019-0143-2
4	ROCHA B， RUIZ-ROMERO C， BLANCO F J. Mass spectrometry imaging： a novel technology in rheumatology［J］. Nature Reviews Rheumatology， 2017， 13（1）： 52-63. 10.1038/nrrheum.2016.184
5	MIKKO M， PAUL G. Hyperspectral near infrared imaging quantifies the heterogeneity of carbon materials［J］. Scientific Reports， 2018，8（1）： 1-7. 10.1038/s41598-018-28889-7
6	WILBURN B， JOSHI N， VAISH V， et al. High-speed videography using a dense camera array［C］// Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2004： 294-301.
7	刘珍丹，张莉娜，丁丰，等. 沉浸式环境下数字图像的两相流纹理特征提取［J］. 计算机仿真， 2021， 38（6）： 133-137. 10.3969/j.issn.1006-9348.2021.06.028
8	NIU H， HUANG Y. Application of the phase-averaged pre-processing method in identification of oscillatory two-phase flow patterns based on textural features［J］. Flow Measurement and Instrumentation， 2021， 77： 101872. 10.1016/j.flowmeasinst.2020.101872
9	张蕊，李锦涛.基于深度学习的场景分割算法研究综述［J］. 计算机研究与发展， 2020， 57（4）： 859-875. 10.7544/issn1000-1239.2020.20190513
10	翁润滢，孙斌，赵玉晓，等. 基于自适应最优核和卷积神经网络的气液两相流流型识别方法［J］. 化工学报， 2018， 69（12）： 5065-5072. 10.11949/j.issn.0438-1157.20180198
11	YANG Y， WANG D， NIU P， et al. Measurement of vertical gas-liquid two-phase flow by electromagnetic flowmeter and image processing based on the phase-isolation［J］. Experimental Thermal and Fluid Science， 2019， 101： 87-100. 10.1016/j.expthermflusci.2018.10.010
12	LIU L， BAI B. Flow regime identification of swirling gas-liquid flow with image processing technique and neural networks［J］. Chemical Engineering Science， 2019， 199： 588-601. 10.1016/j.ces.2019.01.037
13	JU Y， WU L， LI M， et al. A novel hybrid model for flow image segmentation and bubble pattern extraction［J］. Measurement， 2022，192： 110861. 10.1016/j.measurement.2022.110861
14	仝卫国，庞雪纯，朱赓宏. 基于卷积神经网络的气液两相流流型识别方法［J］. 系统仿真学报， 2021， 33（4）： 883-891. 10.16182/j.issn1004731x.joss.19-0619
15	GONZALEZ R C， WOODS R E. Digital image processing［J］. IEEE Transactions on Acoustics Speech and Signal Processing， 1980， 28（4）：484-486. 10.1109/tassp.1980.1163437
16	PERONA P， MALIK J. Scale-space and edge detection using anisotropic diffusion［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2002， 12（7）：629-639.
17	RUDIN L I， OSHER S， FATEMI E. Nonlinear total variation based noise removal algorithms［J］. Physica D： Nonlinear Phenomena， 1992， 60（1/2/3/4）： 259-268. 10.1016/0167-2789(92)90242-f
18	BUADES A， COLL B， MOREL J M. A non-local algorithm for image denoising［C］// Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2005， 2： 60-65.
19	LONG J， SHELHAMER E， DARRELL T. Fully convolutional networks for semantic segmentation［C］// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2015： 3431-3440. 10.1109/cvpr.2015.7298965
20	RONNEBERGER O， FISCHER P， BROX T. U-Net： Convolutional networks for biomedical image segmentation［C］// Proceedings of the 2015 International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham： Springer， 2015： 234-241. 10.1007/978-3-319-24574-4_28
21	BADRINARAYANAN V， KENDALL A， CIPOLLA R. SegNet： a deep convolutional encoder-decoder architecture for image segmentation［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2017， 39（12）： 2481-2495. 10.1109/tpami.2016.2644615
22	CHEN L C， PAPANDREOU G， KOKKINOS I， et al. Semantic image segmentation with deep convolutional nets and fully connected CRFs ［EB/OL］.（2014-12-22）［2021-06-20］. . 10.1109/tpami.2017.2699184
23	CHEN L C， PAPANDREOU G， KOKKINOS I， et al. DeepLab： semantic image segmentation with deep convolutional nets， atrous convolution， and fully connected CRFs［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2016， 40（4）：834-848.
24	CHEN L C， PAPANDREOU G， SCHROFF F， et al. Rethinking atrous convolution for semantic image segmentation［EB/OL］.（2017-07-17）［2021-06-20］. . 10.1007/978-3-030-01234-2_49
25	CHEN L C， ZHU Y， PAPANDREOU G， et al. Encoder-decoder with atrous separable convolution for semantic image segmentation［C］// Proceedings of the 2018 European Conference on Computer Vision. Cham： Springer， 2018： 801-818. 10.1007/978-3-030-01234-2_49
26	DIAKOGIANNIS F I， WALDNER F， CACCETTA P， et al. ResUNet-a： A deep learning framework for semantic segmentation of remotely sensed data［J］. ISPRS Journal of Photogrammetry and Remote Sensing， 2020， 162： 94-114. 10.1016/j.isprsjprs.2020.01.013
27	JAHEDSARAVANI A， MARHABAN M H， MASSINAEI M. Prediction of the metallurgical performances of a batch flotation system by image analysis and neural networks［J］. Minerals Engineering， 2014， 69： 137-145. 10.1016/j.mineng.2014.08.003
28	WANG W， BERGHOLM F， YANG B. Froth delineation based on image classification［J］. Minerals Engineering， 2003， 16（11）： 1183-1192. 10.1016/j.mineng.2003.07.014
29	JAHEDSARAVANI A， MASSINAEI M， MARHABAN M H. An image segmentation algorithm for measurement of flotation froth bubble size distributions［J］. Measurement， 2017， 111： 29-37. 10.1016/j.measurement.2017.07.023
30	ZHOU K， YANG C， GUI W， et al. Clustering-driven watershed adaptive segmentation of bubble image［J］. Journal of Central South University of Technology， 2010， 17（5）： 1049-1057. 10.1007/s11771-010-0597-y

类型	方法	PA	MPA	MIoU	FWIoU
传统图像分割方法	最大方差法	96.7	91.0	88.9	93.6
	最小误差法	88.5	69.0	62.7	79.0
	最大熵法	89.5	75.6	70.0	81.2
	K均值聚类	49.3	49.2	44.4	46.5
	模糊C均值聚类	51.3	51.0	46.2	48.3
	贝叶斯马尔可夫随机场	24.3	50.1	13.7	11.2
	分水岭算法	80.5	49.4	49.4	80.5
	活动轮廓算法	55.9	63.3	48.1	50.2
	边缘检测法	81.7	51.5	42.5	68.0
基于深度学习方法	FCN	98.5	96.9	95.1	97.1
	U-Net	99.7	99.2	99.0	99.5
	SegNet	99.8	99.6	99.2	99.6
	DeepLabv3+	98.3	95.4	93.9	96.7
	ResUNet-a	90.8	91.8	84.7	86.2

类型	方法	PA	MPA	MIoU	FWIoU
传统图像分割方法	最大方差法	96.7	91.0	88.9	93.6
	最小误差法	88.5	69.0	62.7	79.0
	最大熵法	89.5	75.6	70.0	81.2
	K均值聚类	49.3	49.2	44.4	46.5
	模糊C均值聚类	51.3	51.0	46.2	48.3
	贝叶斯马尔可夫随机场	24.3	50.1	13.7	11.2
	分水岭算法	80.5	49.4	49.4	80.5
	活动轮廓算法	55.9	63.3	48.1	50.2
	边缘检测法	81.7	51.5	42.5	68.0
基于深度学习方法	FCN	98.5	96.9	95.1	97.1
	U-Net	99.7	99.2	99.0	99.5
	SegNet	99.8	99.6	99.2	99.6
	DeepLabv3+	98.3	95.4	93.9	96.7
	ResUNet-a	90.8	91.8	84.7	86.2

[1]	Xiwei LIU, Xiaoyan GONG, Hongxia ZHAO, Siyu BIAN, Shuai SHAO, Yaping DAI, Wenxin DAI. Dynamic facial expression recognition based on hybrid attention mechanism [J]. Journal of Computer Applications, 2023, 43(S1): 1-7.
[2]	Kui JIANG, Zhihang YU, Xiaolei CHEN, Yuhao LI. Design and implementation of Webshell traffic detection system based on BERT-CNN [J]. Journal of Computer Applications, 2023, 43(S1): 126-132.
[3]	Pengliu TAN, Guangyong XU, Luyu ZHANG, Runshu WANG. Heart disease prediction model based on convolutional neural network and Adaboost [J]. Journal of Computer Applications, 2023, 43(S1): 19-25.
[4]	Jun LIANG, Zehong HONG, Songsen YU. Image segmentation model based on improved particle swarm optimization algorithm and genetic mutation [J]. Journal of Computer Applications, 2023, 43(6): 1743-1749.
[5]	Huibin ZHANG, Liping FENG, Yaojun HAO, Yining WANG. Ancient mural dynasty identification based on attention mechanism and transfer learning [J]. Journal of Computer Applications, 2023, 43(6): 1826-1832.
[6]	Liyao FU, Mengxiao YIN, Feng YANG. Transformer based U-shaped medical image segmentation network： a survey [J]. Journal of Computer Applications, 2023, 43(5): 1584-1595.
[7]	Bin WANG, Tian XIANG, Yidong LYU, Xiaofan WANG. Adaptive multi-scale feature channel grouping optimization algorithm based on NSGA‑Ⅱ [J]. Journal of Computer Applications, 2023, 43(5): 1401-1408.
[8]	Rui XU, Shuang LIANG, Hang WAN, Yimin WEN, Shiming SHEN, Jian LI. Extraction of PM_2.5 diffusion characteristics based on candlestick pattern matching [J]. Journal of Computer Applications, 2023, 43(5): 1394-1400.
[9]	Jiahong SUI, Yingchi MAO, Huimin YU, Zicheng WANG, Ping PING. Global image captioning method based on graph attention network [J]. Journal of Computer Applications, 2023, 43(5): 1409-1415.
[10]	Jianhui HE, Chunlong HU, Xin SHU. Multi-task age estimation method based on multi-peak label distribution learning [J]. Journal of Computer Applications, 2023, 43(5): 1578-1583.
[11]	Senqi YANG, Xuliang DUAN, Zhan XIAO, Songsong LANG, Zhiyong LI. Text classification of agricultural news based on ERNIE+DPCNN+BiGRU [J]. Journal of Computer Applications, 2023, 43(5): 1461-1466.
[12]	Xiaoyu FAN, Suzhen LIN, Yanbo WANG, Feng LIU, Dawei LI. Reconstruction algorithm for highly undersampled magnetic resonance images based on residual graph convolutional neural network [J]. Journal of Computer Applications, 2023, 43(4): 1261-1268.
[13]	Qingyu JIA, Liang CHANG, Xianyi YANG, Baohua QIANG, Shihao ZHANG, Wu XIE, Minghao YANG. Real-time reconstruction method of visual information for manipulator operation [J]. Journal of Computer Applications, 2023, 43(4): 1255-1260.
[14]	Haiyu YANG, Wenpu GUO, Kai KANG. Signal modulation recognition method based on convolutional long short-term deep neural network [J]. Journal of Computer Applications, 2023, 43(4): 1318-1322.
[15]	Haifeng LI, Fan ZHANG, Minnan PIAO, Huaichao WANG, Nansha LI, Zhongcheng GUI. Automatic detection of targets under airport pavement based on channel and spatial attention [J]. Journal of Computer Applications, 2023, 43(3): 930-935.

Gas-liquid flow image segementation method based on deep convolutional neural network

基于深度卷积神经网络的气液两相流图像分割方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 10

References 30

Related Articles 15

Recommended Articles

Metrics