IMU-Guided Adaptive Video Stabilization for Unmanned Surface Vehicles
Abstract
This study addresses the critical challenge of video stabilization in Unmanned Surface Vehicles (USVs) operating under high-wave and harsh maritime conditions. Conventional stabilization methods often struggle to distinguish erratic vessel motions caused by waves from intentional camera maneuvers, leading to significant image distortion, horizon tilt, and loss of visual data. To overcome this, we propose a novel hybrid stabilization framework that fuses Inertial Measurement Unit (IMU) data with Optical Flow analysis. Unlike traditional fusion techniques, the core innovation of this approach is using IMU data strictly as a control signal to guide the stabilization process, rather than as a direct geometric warping tool. A key contribution of the proposed method is the implementation of an adaptive α (alpha) coefficient, which dynamically adjusts the weighting between inertial data and visual cues based on the instantaneous dynamics and motion intensity of the marine vehicle. This adaptive mechanism allows the algorithm to effectively suppress high-frequency jitter and wave-induced oscillations while preserving the integrity of intentional navigation maneuvers along the cruise path. Experimental results demonstrate that the proposed method significantly reduces unwanted vibrations and improves visual clarity compared to baseline stabilization models. The findings suggest that the IMU-guided adaptive approach provides a robust and computationally efficient solution for real-time perception and situational awareness in autonomous maritime systems. The source code and trained model generated in this study are publicly available at: https://github.com/KTUN-Institute-Works/acc_flow
Keywords
References
- Y. Wang, Q. Huang, C. Jiang, J. Liu, M. Shang, and Z. Miao, “Video stabilization: A comprehensive survey,” Neurocomputing, vol. 516, pp. 205–230, Jan. 2023, doi: 10.1016/j.neucom.2022.10.008.
- Z. Teed and J. Deng, “RAFT: Recurrent All-Pairs Field Transforms for Optical Flow,” in Computer Vision – ECCV 2020, vol. 12347, A. Vedaldi, H. Bischof, T. Brox, and J.-M. Frahm, Eds., in Lecture Notes in Computer Science, vol. 12347. , Cham: Springer International Publishing, 2020, pp. 402–419. doi: 10.1007/978-3-030-58536-5_24.
- J. Auysakul, H. Xu, and V. Pooneeth, “A Hybrid Motion Estimation for Video Stabilization Based on an IMU Sensor,” Sensors, vol. 18, no. 8, p. 2708, Aug. 2018, doi: 10.3390/s18082708.
- H. Li, K. Luo, and S. Liu, “GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning,” in 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada: IEEE, Oct. 2021, pp. 12849–12858. doi: 10.1109/ICCV48922.2021.01263.
- J. Yu, T. Zhang, F. Shi, L. He, and C.-K. Liang, “SensorFlow: Sensor and Image Fused Video Stabilization,” in 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Tucson, AZ, USA: IEEE, 2025.
- B. D. Lucas and T. Kanade, “An Iterative Image Registration Technique with an Application to Stereo Vision,” in Proceedings of the 7th International Joint Conference on Artificial Intelligence (IJCAI), Vancouver, BC, Canada, 1981, pp. 674–679.
- R. Hartley and A. Zisserman, Multiple view geometry in computer vision, 2. edition, 17.printing. Cambridge: Cambridge Univ. Press, 2018.
- M. Grundmann, V. Kwatra, and I. Essa, “Auto-directed video stabilization with robust L1 optimal camera paths,” in CVPR 2011, Colorado Springs, CO, USA: IEEE, Jun. 2011, pp. 225–232. doi: 10.1109/CVPR.2011.5995525.
Details
Primary Language
English
Subjects
Video Processing, Photonic and Electro-Optical Devices, Sensors and Systems (Excl. Communications)
Journal Section
Research Article
Authors
Early Pub Date
November 10, 2026
Publication Date
-
Submission Date
August 12, 2026
Acceptance Date
October 13, 2026
Published in Issue
Year 2026 Number: Advanced Online Publication