Implementation of visual positioning system for underwater vehicles


Kropotov A. N.*, Makashov A. A.**, Plyasunov V. M.***

Bauman Moscow State Technical University, MSTU, 5, bldg. 1, 2-nd Baumanskaya str., Moscow, 105005, Russia



Underwater vehicle's visual positioning system (VPS) is designed to determine its local motion parameters near object site. This article describes implementation of such system and its adaptation for small-size ROVs.

Using specially developed algorithms for video images processing the VSP processes video information obtained from underwater vehicle (UV) video camera. It allows to measure UV’s both cruising and lag coordinates and velocities, as well as its distance from the seabed which are used for closing feedback loops in motion control system.

Comparison of correlation-extreme algorithm, optical flow method and methods based on descriptors collating was made. As a result we choose modified Kanade-Lucas tracker (KLT), representing one of the methods for optical flow evaluation. We also consider it worthwhile to further study of binary descriptor-based tracker for VPS.

We adapted KLT for use with small-sized remote controlled ROV. To achieve this goal we increased frequency and conducted laboratory and semi-natural testing. Algorithm efficiency with trim and list angles up to 30° was experimentally proved, under condition of obtaining these angles values from external sensors. The absence of drift error of coordinates measured by VPS in length of time was also shown. Based on Bode diagram we came to a conclusion that visual positioning system can be represented as pure time delay link within the control system.

A full-scale tests of VPS on ROV in the special natatorium showed that developed VPS can be used for small-sized highly dynamic ROVs local positioning within object site. The represented video positioning algorithm can be also used for automated drone hovering mode.


local navigation, video processing, underwater vehicle, control system, computer vision, image fitting, dynamic positioning


    1.  Aulinas J., Carreras M., Llado, X. et al. Feature extraction for underwater visual SLAM, Ocean, IEEE - Spain. IEEE, 2011, pp. 1–7.

    2.  Negahdaripour S., Firoozfam P. An ROV Stereovision System for Ship-Hull Inspection , Ocean, 2006, vol. 31, issue 3, pp. 551–564.

    3.  Kostenko V.V., Pavin A.M. Podvodnye issledovaniya i robototekhnika, 2014, no. 1(17), pp. 39–47.

    4.  Lipatov A.N., Lyash A.N., Makarov S.V. Podvodnye issledovaniya i robototekhnika, 2014, no. 1(17), pp. 18–32.

    5.   Shi J., Tomasi C. Good features to track, Proc. IEEE Conference of Computer Vision and Pattern Recognition, Seatle, June 1994, pp. 593-600.

    6.  Lowe D. Object recognition from local scale-invariant features, Proceedings of the International Conference on Computer Vision 2, pp. 1150–1157.

    7.   Bay H. Speeded-up robust features (SURF), IEEE Computer Vision and image understanding, 2008, vol. 110, issue 3, pp. 346–359.

    8.   Leutenegger S., Chli M., Siegwart R.Y. BRISK: Binary Robust invariant scalable keypoints, Int. Conf. Comput. Vis. IEEE, 2011, pp. 2548–2555.

    9.   Artyukhov M.Yu., Kropotov A.N., Makashov A. A., Sakharova E.I. Trudy nauchno-tekhnicheskoy konferentsii-seminara “Tekhnicheskoe zrenie v sistemakh upravleniya mobil'nymi ob"ektami”, (Tarusa – 2010), IKI RAS 2010, p. 230.

    10.  Kropotov A.N., Makashov A.A., Artyukhov M.Yu. Materialy chetvertoi vserossiiskoi nauchno-tekhnicheskoi konferentsii “Tekhnicheskie problemy osvoeniya mirovogo okeana” (Vladivostok-2011), Vladivostok, IPMT DVO RAS, 2011,  pp. 316-321.

    11.  Vishnyakov B.V., Gorbatsevich V.S. Materialy konferentsii “Tekhnicheskoe zrenie v sistemakh upravleniya” (Moscow - 2013), Moscow, IKI RAS, 2011, pp. 119-121.

    12.   Kropotov A.N., Makashov A.A., Sakharova E.I., Plyasunov V.M. Materialy konferentsii “Sovremennye metody i sredstva okeanologicheskikh issledovanii”, (Moscow - 2013), Moscow, IO RAS, 2013, pp. 73-76.

      Download — informational site MAI

      Copyright © 2000-2021 by MAI