Research on Underwater Non-Contact Displacement Measurement Based on Vision

To address the problem that the contact measurement instrument exerted external force on the experimental target and affected the experimental accuracy during the underwater dynamic experiment, a non-contact underwater displacement measurement method based on video tracking was proposed, the traditional Meanshift algorithm and Kalman filter were combined to track fast-moving targets underwater. In addition, this article ran and compared four algorithms: Lucas-Kanade(LK) optical flow algorithm, traditional Meanshift algorithm, MOSSE algorithm and Circle Hough Transform (CHT). First, a free fall experiment was carried out. The relative errors of the gravitational acceleration calculated by the five algorithms were all less than 5%, which verified the feasibility of the five algorithms. In the displacement measurement experiment, the displacement-time curves and velocity-time curves of the five algorithms were fitted respectively, and the effects of the five algorithms in the underwater environment were studied. The research results show that although the improved Meanshift algorithm takes more time without the influence of environmental factors, the accuracy of tracking fast moving targets is improved.