Analysis of the experimental flow shadowgraph images by computer vision methods


  • Igor A. Doroshchenko


computer vision
digital image processing
object detection
convolutional neural networks
flow visualization
shock wave
bow shock


In this study, two examples of physical experiment automation using computer vision and deep learning techniques are considered. The first of them involves the use of classical computer vision techniques to detect and track the oblique shock wave on the experimental shadowgraph images. This was achieved using Canny edge detection and Hough transform, which allowed to obtain the line equation corresponding to the oblique shock wave. By automatically calculating the angle of this wave for each frame in the video, the process of extracting quantitative information from flow visualizations was significantly accelerated. In the second example, a convolutional neural network was trained to identify four classes of objects on the shadowgraph images, namely vertical shock waves, bow shocks, plumes, and opaque particles in the flow. The custom object detection model is based on the up-todate YOLOv8 architecture. To realize this task, a dataset of 1493 labeled shadowgraph images was collected. The model showed excellent performance during the learning process, with model precision and mAP50 scores exceeding 0.9. It was successfully applied to detect objects on the shadowgraph images, demonstrating the potential of deep learning techniques for automating the processing of flow visualizations. Overall, this study highlights the significant benefits of combining classical computer vision algorithms with deep learning techniques in the automation of physical experiments. However, classical algorithms demand the writing additional code to extract the required information. The deep neural networks can perform this task automatically, provided that a well-annotated dataset is available. This approach offers a promising avenue for accelerating the analysis of flow visualizations and the extraction of quantitative information in physical experiments.






Methods and algorithms of computational mathematics and their applications

Author Biography

Igor A. Doroshchenko


  1. A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” Advances in Neural Information Processing Systems 25 (NIPS 2012). . Cited May 11, 2023.
  2. J. Rienitz, “Schlieren Experiment 300 Years Ago,” Nature 254 (5498), 293-295 (1975).
    doi 10.1038/254293a0.
  3. G. S. Settles and M. J. Hargather, “A Review of Recent Developments in Schlieren and Shadowgraph Techniques,” Meas. Sci. Technol. 28 (4), Article Number 042001 (2017).
    doi 10.1088/1361-6501/aa5748.
  4. J. Wolfram and J. Martinez Schramm, “Pattern Recognition in High Speed Schlieren Visualization at the High Enthalpy Shock Tunnel Göttingen (HEG),” in Notes on Numerical Fluid Mechanics and Multidisciplinary Design (Springer, Berlin, 2010), Vol. 112, pp. 399-406.
    doi 10.1007/978-3-642-14243-7_49.
  5. N. T. Smith, M. J. Lewis, and R. Chellappa, “Extraction of Oblique Structures in Noisy Schlieren Sequences Using Computer Vision Techniques,” AIAA J. 50 (5), 1145-1155 (2012).
    doi 10.2514/1.J051335.
  6. C. Liu, R. Jiang, D. Wei, et al., “Deep Learning Approaches in Flow Visualization,” Adv. Aerodyn. 4, Article Number 17 (2022).
    doi 10.1186/s42774-022-00113-1.
  7. Y. Liu, Y. Lu, Y. Wang, et al., “A CNN-Based Shock Detection Method in Flow Visualization,” Comput. Fluids 184, 1-9 (2019).
    doi 10.1016/j.compfluid.2019.03.022.
  8. S. Cui, Y. Wang, X. Qian, and Z. Deng, “Image Processing Techniques in Shockwave Detection and Modeling,” J. Signal Inform. Process. 4 (3B), 109-113 (2013).
    doi 10.4236/jsip.2013.43B019.
  9. G. Li, M. Burak Agir, K. Kontis, et al., “Image Processing Techniques for Shock Wave Detection and Tracking in High Speed Schlieren and Shadowgraph Systems,” J. Phys.: Conf. Ser. 1215, Article Number 012021 (2019).
    doi 10.1088/1742-6596/1215/1/012021.
  10. N. T. Smith, M. J. Lewis, and R. Chellappa, “Detection, Localization, and Tracking of Shock Contour Salient Points in Schlieren Sequences,” AIAA J. 52 (6), 1249-1264 (2014).
    doi 10.2514/1.J052367.
  11. T. R. Fujimoto, T. Kawasaki, and K. Kitamura, “Canny-Edge-Detection/Rankine-Hugoniot-Conditions Unified Shock Sensor for Inviscid and Viscous Flows,” J. Comput. Phys. 396, 264-279 (2019).
    doi 10.1016/
  12. M. V. Srisha Rao and G. Jagadeesh, “Visualization and Image Processing of Compressible Flow in a Supersonic Gaseous Ejector,” J. Indian Inst. Sci. 93 (1), 57-66 (2013).
  13. P. V. Bulat, K. N. Volkov, and M. S. Yakovchuk, “Flow Visualization with Strong and Weak Gas Dynamic Discontinuities in Computational Fluid Dynamics,” Numerical Methods and Programming (Vychislitel’nye Metody i Programmirovanie). 17 (3), 245-257 (2016).
    doi 10.26089/NumMet.v17r323.
  14. P. V. Bulat and K. N. Volkov, “Visualization of Gas Dynamics Discontinuities in Supersonic Flows Using Digital Image Processing Methods,” Numerical Methods and Programming (Vychislitel’nye Metody i Programmirovanie). 20 (3), 237-253 (2019).
    doi 10.26089/NumMet.v20r322.
  15. C. Brossard, J. C. Monnier, P. Barricau, et al., “Principles and Applications of Particle Image Velocimetry,” Aerospace Lab. No. 1, 1-11 (2009). . Cited May 6, 2023.
  16. E. K. Akhmetbekov, A. V. Bilsky, Yu. A. Lozhkin, et al., “Software for Experiment Management and Processing of Data Obtained by Digital Flow Visualization Techniques (ActualFlow),” Numerical Methods and Programming. (Vychislitel’nye Metody i Programmirovanie). 7 (3), 79-85 (2006).
  17. E. Arnaud, E. Mémin, R. Sosa, and G. Artana, “A Fluid Motion Estimator for Schlieren Image Velocimetry,” in Lecture Notes in Computer Science (Springer, Berlin, 2006), Vol. 3951, pp. 198-210.
    doi 10.1007/11744023_16.
  18. M. Lawson, M. Hargather, G. Settles, et al., “Focusing-Schlieren PIV Measurements of a Supersonic Turbulent Boundary Layers,” in Proc. 47th AIAA Aerospace Sciences Meeting including The New Horizons Forum and Aerospace Exposition, Orlando, USA, January 5-8, 2009. . Cited May 5, 2023.
  19. A. W. Gena, C. Voelker, and G. S. Settles, “Qualitative and Quantitative Schlieren Optical Measurement of the Human Thermal Plume,” Indoor Air 30 (4), 757-766 (2020).
    doi 10.1111/ina.12674.
  20. M. J. Hargather, M. J. Lawson, G. S. Settles, and L. M. Weinstein, “Seedless Velocimetry Measurements by Schlieren Image Velocimetry,” AIAA J. 49 (3), 611-620 (2011).
    doi 10.2514/1.J050753.
  21. M. Debella-Gilo and A. Kääb, “Sub-Pixel Precision Image Matching for Measuring Surface Displacements on Mass Movements Using Normalized Cross-Correlation,” Remote Sens. Environ. 115 (1), 130-142 (2011).
    doi 10.1016/j.rse.2010.08.012.
  22. M. Berenjkoub, G. Chen, and T. Günther, “Vortex Boundary Identification Using Convolutional Neural Network,” in Proc. 2020 IEEE Visualization Conference (VIS), Salt Lake City, USA, October 25-30, 2020 (IEEE Press, New York, 2020), pp. 261-265.
    doi 10.1109/VIS47514.2020.00059.
  23. A. D. Beck, J. Zeifang, A. Schwarz, and D. G. Flad, “A Neural Network Based Shock Detection and Localization Approach for Discontinuous Galerkin Methods,” J. Comput. Phys. 423, Article Number 109824 (2020).
    doi 10.1016/
  24. M. Morimoto, K. Fukami, and K. Fukagata, “Experimental Velocity Data Estimation for Imperfect Particle Images Using Machine Learning,” Phys. Fluids. 33 (8), Article Number 087121 (2021).
    doi 10.1063/5.0060760.
  25. B. N. Ubald, P. Seshadri, and A. Duncan, “Density Reconstruction from Schlieren Images through Bayesian Nonparametric Models,” arXiv preprint: 2201.05233v3 [physics.flu-dyn] (Cornell Univ. Library, Ithaca, 2022).
    doi 10.48550/arXiv.2201.05233.
  26. M. Monfort, T. Luciani, J. Komperda, et al., “A Deep Learning Approach to Identifying Shock Locations in Turbulent Combustion Tensor Fields,” in Modeling, Analysis, and Visualization of Anisotropy (Springer, Cham, 2017), pp. 375-392.
    doi 10.1007/978-3-319-61358-1_16.
  27. G. B’iró, M. Pocsai, I. F. Barna, et al., “Machine Learning Methods for Schlieren Imaging of a Plasma Channel in Tenuous Atomic Vapor,” Opt. Laser Technol. 159, Article Number 108948 (2023).
    doi 10.1016/j.optlastec.2022.108948.
  28. B. Colvert, M. Alsalman, and E. Kanso, “Classifying Vortex Wakes Using Neural Networks,” Bioinspir. Biomim. 13 (2), Article Number 025003 (2018).
    doi 10.1088/1748-3190/aaa787.
  29. M. D. Manshadi, H. Vahdat-Nejad, M. Kazemi-Esfeh, and M. Alavi, “Speed Detection in Wind-Tunnels by Processing Schlieren Images,” Int. J. Eng. 29 (7), 962-967 (2016).
    doi 10.5829/idosi.ije.2016.29.07a.11.
  30. Shadowgraph Images. Datasets at Hugging Face. . Cited May 6, 2023.
  31. Ultralytics YOLOv8. . Cited May 6, 2023.
  32. I. A. Znamenskaya and I. A. Doroshchenko, “Edge Detection and Machine Learning for Automatic Flow Structures Detection and Tracking on Schlieren and Shadowgraph Images,” J. Flow Vis. Image Process. 28 (4), 1-26, (2021).
    doi 10.1615/JFlowVisImageProc.2021037690.
  33. I. A. Znamenskaya, I. A. Doroshchenko, N. N. Sysoev, and D. I. Tatarenkova, “Results of Quantitative Analysis of High-Speed Shadowgraphy of Shock Tube Flows Using Machine Vision and Machine Learning,” Dokl. Akad. Nauk 497 (1), 16-20 (2021) [Dokl. Phys. 66 (4), 93-96, (2021)].
    doi 10.1134/S1028335821040066.