Search papers, labs, and topics across Lattice.
This paper introduces ISOPoT, a novel imaging sonar odometry method that leverages modern point tracking techniques to enhance navigation in challenging underwater environments. By utilizing multi-frame point tracks as the primary correspondence representation and incorporating lightweight optimizations, ISOPoT significantly improves robustness against the inherent noise and artifacts of sonar images. Evaluations on both the Aracati 2017 dataset and a proprietary underwater dataset demonstrate that ISOPoT consistently outperforms existing state-of-the-art methods in both sonar-only and multi-sensor scenarios.
ISOPoT achieves superior underwater navigation by effectively transforming noisy sonar data into reliable point tracks, outpacing existing methods.
Reliable navigation in underwater environments remains a key challenge in marine robotics. In such scenarios, forward-looking sonars are a natural choice for long-range perception, offering wide coverage even in turbid, low-visibility conditions. However, sonar images are inherently noisy, contain artifacts, and lack rich semantic structure, causing standard computer vision methods for keypoint detection and matching to perform poorly. In this paper, we introduce ISOPoT, an imaging sonar odometry method based on modern point tracking techniques. We propose a sonar odometry pipeline that uses multi-frame point tracks as its primary correspondence representation, augmented with lightweight optimizations to improve robustness. We evaluated the proposed method on the Aracati 2017 dataset, as well as on an internal sonar dataset collected in real-world underwater environments. Our results show that ISOPoT outperforms previous state-of-the-art methods consistently in both sonar-only scenarios and in multi-sensor settings.
