留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于多尺度域红外目标分割及特征点匹配的泡沫流速检测

施雯玲 廖一鹏 许志猛 严欣 朱坤华

施雯玲, 廖一鹏, 许志猛, 严欣, 朱坤华. 基于多尺度域红外目标分割及特征点匹配的泡沫流速检测[J]. 红外技术, 2023, 45(5): 463-473.
引用本文: 施雯玲, 廖一鹏, 许志猛, 严欣, 朱坤华. 基于多尺度域红外目标分割及特征点匹配的泡沫流速检测[J]. 红外技术, 2023, 45(5): 463-473.
SHI Wenling, LIAO Yipeng, XU Zhimeng, YAN Xin, ZHU Kunhua. Foam Flow Rate Detection Based on Infrared Target Segmentation and SURF Matching in NSST Domain[J]. Infrared Technology , 2023, 45(5): 463-473.
Citation: SHI Wenling, LIAO Yipeng, XU Zhimeng, YAN Xin, ZHU Kunhua. Foam Flow Rate Detection Based on Infrared Target Segmentation and SURF Matching in NSST Domain[J]. Infrared Technology , 2023, 45(5): 463-473.

基于多尺度域红外目标分割及特征点匹配的泡沫流速检测

基金项目: 

国家自然科学基金项目 61601126

国家自然科学基金项目 61904031

福建省自然科学基金项目 2019J01224

详细信息
    作者简介:

    施雯玲(1996-),女,硕士研究生,主要从事机器视觉方面的研究。E-mail:382935094@qq.com

    通讯作者:

    廖一鹏(1982-),男,博士,讲师,主要从事图像处理与机器视觉方面的研究。E-mail:fzu_lyp@163.com

  • 中图分类号: TP391

Foam Flow Rate Detection Based on Infrared Target Segmentation and SURF Matching in NSST Domain

  • 摘要: 为减少浮选气泡合并、破碎等变化对泡沫表面流动特征提取的影响,提出了一种非下采样剪切波变换(Nonsubsampled Shearlet Transform,NSST)域红外目标分割及改进加速鲁棒特征(Speeded Up Robust Features,SURF)匹配的泡沫表面流速检测方法。首先,对相邻两帧泡沫红外图像NSST分解,在多尺度域构建图割能量函数的边界、亮度、显著性约束项实现对合并、破碎气泡的分割;然后,对分割后的背景区域进行SURF特征点检测,通过统计扇形区域内的尺度相关系数确定特征点主方向,采用特征点邻域的多方向高频系数构造特征描述符;最后,对相邻两帧泡沫红外图像进行特征点匹配,根据匹配结果计算泡沫流速的大小、方向、加速度、无序度。实验结果表明,本文方法能有效分割出合并、破碎的气泡,具有较高的分割精度,提升了SURF算法的匹配精度,流速检测受气泡合并、破碎的影响小,检测精度和效率较现有方法有一定提升,能准确地表征不同工况下泡沫表面的流动特性, 为后续的工况识别奠定基础。
  • 图  1  气泡图像多尺度变换

    Figure  1.  Multiscale transform of bubble images

    图  2  浮选泡沫的红外热像

    Figure  2.  Infrared images of flotation foam

    图  3  特征主方向确定

    Figure  3.  Feature main direction determination

    图  4  构建特征描述符

    Figure  4.  Constructing feature descriptors

    图  5  泡沫流动速度检测流程

    Figure  5.  Foam flow rate detection flow chart

    图  6  目标分割结果及对比

    Figure  6.  Target segmentation results and comparison

    图  7  改进SURF的匹配结果及比较

    Figure  7.  Improved SURF algorithm matching results and comparison

    图  8  流速检测效果及比较

    Figure  8.  Flow rate detection effect and comparison

    图  9  不同加药状态下的流速检测结果

    Figure  9.  Flow rate detection results under different dosing conditions

    表  1  四种方法分割性能比较

    Table  1.   Comparison of the performance of four methods of segmentation

    Algorithm Under-flotation Normal flotation Over-flotation Running time/s
    IOU Error IOU Error IOU Error
    Ref.[14] 0.7957 0.2799 0.8316 0.2239 0.8682 0.1727 0.4017
    Ref.[18] 0.8251 0.2087 0.8174 0.2185 0.8876 0.1624 0.2240
    Ref.[19] 0.8188 0.2745 0.8541 0.2523 0.8871 0.1361 0.3119
    Ours 0.8357 0.2054 0.8547 0.1847 0.9014 0.1523 0.7301
    下载: 导出CSV

    表  2  匹配精度及运行时间比较

    Table  2.   Comparison of matching accuracy and running time

    Algorithm Noise variance 10% Scale 1:2
    Matching accuracy/% Time/s Matching accuracy/% Time/s
    Improved SIFT 95.91 0.9625 98.72 0.9041
    Improved SURF 93.76 0.2604 96.24 0.1943
    Ours 96.12 0.3128 98.64 0.2612
    Algorithm Noise variance 30% Scale 1:8
    Matching accuracy/% Time/s Matching accuracy/% Time/s
    Improved SIFT 82.23 1.1254 81.25 0.8547
    Improved SURF 78.54 0.2321 75.68 0.1269
    Ours 87.12 0.3513 85.03 0.2098
    下载: 导出CSV

    表  3  流速检测结果及比较

    Table  3.   Flow rate detection effect and comparison

    Algorithm vx/(pixel⋅s-1) vy/(pixel⋅s-1) V/(pixel⋅s-1) θv Ev/% Eθ/% Time/s
    Ref. [6] 15.5132 14.9353 21.5342 43.9127 1.5132 2.4162 0.3413
    Ref. [7] 15.0139 14.8926 21.1473 44.7676 0.3107 0.5164 0.5451
    Ref. [8] 14.9860 14.9577 21.1734 44.9458 0.1876 0.1204 1.5220
    Ref. [9] 14.9877 15.0659 21.2512 45.1491 0.1791 0.3313 0.6011
    Ours 14.9943 15.0064 21.2137 45.0231 0.0023 0.0513 1.0743
    下载: 导出CSV
  • [1] Fonseca R R, Franco I C, Thompson J P, et al. Turbidity control on dissolved air flotation process using fuzzy logic[J]. Water Science and Technology, 2018, 78(12): 2586-2596. doi:  10.2166/wst.2019.015
    [2] XIE Y, WU J, XU D, et al. Reagent addition control for stibium rougher flotation based on sensitive froth image features[J]. IEEE Transactions on Industrial Electronics, 2016, 64(5): 4199-4206. http://www.onacademic.com/detail/journal_1000039773149010_f4aa.html
    [3] TAN J, LIANG L, PENG Y, et al. The concentrate ash content analysis of coal flotation based on froth images[J]. Minerals Engineering, 2016, 92: 9-20. doi:  10.1016/j.mineng.2016.02.006
    [4] ZHANG J, TANG Z, LIU J, et al. Recognition of flotation working conditions through froth image statistical modeling for performance monitoring[J]. Minerals Engineering, 2016, 86: 116-129. doi:  10.1016/j.mineng.2015.12.008
    [5] 陈良琴, 王卫星. 基于气泡跟踪与相位相关的浮选表面气泡平移运动估计[J]. 四川大学学报: 工程科学版, 2016, 48(5): 143-152. https://www.cnki.com.cn/Article/CJFDTOTAL-SCLH201605020.htm

    CHEN L Q, WANG W X. Flotation surface bubble displacement motion estimation based on bubble tracking and phase correlation[J]. Journal of Sichuan University: Engineering Science Edition, 2016, 48(5): 143-152. https://www.cnki.com.cn/Article/CJFDTOTAL-SCLH201605020.htm
    [6] Kaartinen J, Hätönen J, Hyötyniemi H, et al. Machine-vision-based control of zinc flotation—a case study[J]. Control Engineering Practice, 2006, 14(12): 1455-1466. doi:  10.1016/j.conengprac.2005.12.004
    [7] Nakhaei F, Irannajad M, Mohammadnejad S. Column flotation performance prediction: PCA, ANN and image analysis-based approaches[J]. Physicochemical Problems of Mineral Processing, 2019, 55(5): 1298-1310. http://www.xueshufan.com/publication/2974543226
    [8] WANG Y, SUN B, ZHANG R, et al. Sulfur flotation performance recognition based on hierarchical classification of local dynamic and static froth features[J]. IEEE Access, 2018, 6: 14019-14029. doi:  10.1109/ACCESS.2018.2805265
    [9] 廖一鹏, 陈诗媛, 杨洁洁. NSST域改进ORB的泡沫流动特征提取及加药状态识别. [J]. 光学精密工程, 2020, 28(12): 2684-2699. doi:  10.37188/OPE.20202812.2684

    LIAO Y P, CHEN S Y, YANG J J, et al. Dosing status identification and froth flow feature extraction based on improved ORB in NSST domain[J]. Optics and Precision Engineering, 2020, 28(12): 2684-2699. doi:  10.37188/OPE.20202812.2684
    [10] BAY H, TUYTELAARS T, VAN GOOL L. Surf: Speeded up robust features[C]//European Conference on Computer Vision, 2006: 404-417.
    [11] SINGH S, ANAND R S, GUPTA D. CT and MR image information fusion scheme using a cascaded framework in ripplet and NSST domain[J]. IET Image Processing, 2018, 12(5): 696-707. doi:  10.1049/iet-ipr.2017.0214
    [12] 廖一鹏, 王卫星, 付华栋, 等. 结合分数阶微分的浮选泡沫图像NSCT多尺度增强[J]. 华南理工大学学报: 自然科学版, 2018, 46(3): 92-102. doi:  10.3969/j.issn.1000-565X.2018.03.014

    LIAO Y P, WANG W X, FU H D, et al. Flotation foam image NSCT multi-scale enhancement with fractional differential[J]. Journal of South China University of Technology: Natural Science Edition, 2018, 46(3): 92-102. doi:  10.3969/j.issn.1000-565X.2018.03.014
    [13] PENG Z, QU S, LI Q. Interactive image segmentation using geodesic appearance overlap graph cut[J]. Signal Processing: Image Communication, 2019, 78: 159-170. doi:  10.1016/j.image.2019.06.012
    [14] 廖苗, 刘毅志, 欧阳军林, 等. 基于非线性增强和图割的CT序列肝脏肿瘤自动分割[J]. 计算机辅助设计与图形学学报, 2019, 31(6): 1030-1038. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJF201906019.htm

    LIAO M, LIU Y Z, OUYANG J L, et al. Automatic segmentation of liver tumor in CT volumes using nonlinear enhancement and graph cuts [J]. Journal of Computer-Aided Design & Computer Graphics, 2019, 31(6): 1030-1038. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJF201906019.htm
    [15] 林剑萍, 廖一鹏. 结合分数阶显著性检测及量子烟花算法的NSST域图像融合[J]. 光学精密工程, 2021, 29(6): 1406-1419. https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202106021.htm

    LIN J P, LIAO Y P. A novel image fusion method with fractional saliency detection and QFWA in NSST[J]. Optics and Precision Engineering, 2021, 29(6): 1406-1419. https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM202106021.htm
    [16] 袁丽英, 刘佳, 王飞越. 基于SURF的图像配准改进算法[J]. 探测与控制学报, 2020, 42(2): 65-70. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYX202002014.htm

    YUAN L Y, LIU J, WANG F Y. An improved algorithm of SURF image registration[J]. Journal of Detection & Control, 2020, 42(2): 65-70. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYX202002014.htm
    [17] SONG J L, YANG J H, LIU F J, et al. High temperature strain measurement method by combining digital image correlation of laser speckle and improved RANSAC smoothing algorithm[J]. Optics and Lasers in Engineering, 2018, 111: 8-18. doi:  10.1016/j.optlaseng.2018.07.012
    [18] 邵磊, 张一鸣, 李季, 等. 基于改进的两维Otsu管道红外图像高温区域分割研究[J]. 光谱学与光谱分析, 2019, 39(5): 1637-1642. https://www.cnki.com.cn/Article/CJFDTOTAL-GUAN201905055.htm

    SHAO L, ZHANG Y M, LI J, et al. Research on high temperature region segmentation of infrared pipeline image based on improved two-dimensional-Otsu[J]. Spectroscopy and Spectral Analysis, 2019, 39(5): 1637-1642. https://www.cnki.com.cn/Article/CJFDTOTAL-GUAN201905055.htm
    [19] BI H, TANG H, YANG G Y, et al. Accurate image segmentation using Gaussian mixture model with saliency map[J]. Pattern Analysis and Applications, 2018, 21(3): 869-878. doi:  10.1007/s10044-017-0672-1
    [20] 牛燕雄, 陈梦琪, 张贺. 基于尺度不变特征变换的快速景象匹配方法[J]. 电子与信息学报, 2019, 41(3): 626-631. https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX201903017.htm

    NIU Y X, CHEN M Q, ZHANG H. Fast scene matching method based on scale invariant feature transform[J]. Journal of Electronics & Information Technology, 2019, 41(3): 626-631. https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX201903017.htm
    [21] 罗久飞, 邱广, 张毅, 等. 基于自适应双阈值的SURF双目视觉匹配算法研究[J]. 仪器仪表学报, 2020, 41(3): 240-247. https://www.cnki.com.cn/Article/CJFDTOTAL-YQXB202003026.htm

    LUO J F, QIU G, ZHANG Y, et al. Research on speeded up robust feature binocular vision matching algorithm based on adaptive double threshold[J]. Chinese Journal of Scientific Instrument, 2020, 41(3): 240-247. https://www.cnki.com.cn/Article/CJFDTOTAL-YQXB202003026.htm
  • 加载中
图(9) / 表(3)
计量
  • 文章访问数:  93
  • HTML全文浏览量:  15
  • PDF下载量:  24
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-05-27
  • 修回日期:  2022-06-24
  • 刊出日期:  2023-05-20

目录

    /

    返回文章
    返回