RUSSO: Robust Underwater SLAM with Sonar Optimization against Visual Degradation

TL;DR

This work tackles the challenge of underwater SLAM under visual degradation by introducing RUSSO, a tightly coupled system that fuses stereo camera, IMU, and imaging sonar for robust $6$-DoF localization. A key contribution is a sonar-based IMU propagation prior that improves pose estimates during visual dropouts, complemented by a sonar-enabled SLAM initialization to counteract feature-poor scenarios. The approach uniquely integrates imaging sonar with both stereo vision and inertial sensing, and it is validated through extensive experiments in simulation, a lab pool, and open water, showing improved robustness and accuracy over state-of-the-art VI-SLAM methods. The findings enable more reliable navigation and mapping in turbid or texture-poor underwater environments, with potential impact on underwater surveying and inspection tasks."

Abstract

Visual degradation in underwater environments poses unique and significant challenges, which distinguishes underwater SLAM from popular vision-based SLAM on the ground. In this paper, we propose RUSSO, a robust underwater SLAM system which fuses stereo camera, inertial measurement unit (IMU), and imaging sonar to achieve robust and accurate localization in challenging underwater environments for 6 degrees of freedom (DoF) estimation. During visual degradation, the system is reduced to a sonar-inertial system estimating 3-DoF poses. The sonar pose estimation serves as a strong prior for IMU propagation, thereby enhancing the reliability of pose estimation with IMU propagation. Additionally, we propose a SLAM initialization method that leverages the imaging sonar to counteract the lack of visual features during the initialization stage of SLAM. We extensively validate RUSSO through experiments in simulator, pool, and sea scenarios. The results demonstrate that RUSSO achieves better robustness and localization accuracy compared to the state-of-the-art visual-inertial SLAM systems, especially in visually challenging scenarios. To the best of our knowledge, this is the first time fusing stereo camera, IMU, and imaging sonar to realize robust underwater SLAM against visual degradation.

Figures (14)

• Figure 1: (a) The schematic of our method. The feature tracking of image sonar acts as a constraint to robot pose estimation when visual degradation occurs in $\Delta t$ frame, thus diminishing the pose drift. (b) The samples of visual available and visual degradation scenarios in the underwater simulator, pool, and sea.
• Figure 2: Overview of the RUSSO system, where imaging sonar fusion is integrated with the VIO system. The purple box and lines denote new add-on to the VIO system.
• Figure 3: (a) The imaging model of imaging sonar which has FOV in both horizontal and vertical directions. (b) The real sonar image is a 2D image which is a projection of the whole vertical sonar data onto a plane.
• Figure 4: The green circles represent the extracted feature points of sonar image using A-KAZE. The initial matching contains many incorrect matches. Outliers are rejected after using RANSAC algorithm, and consequently feature points are correctly matched between the keyframe and the current frame.
• Figure 5: (a) The experimental environment with varying light and haze in UUV simulator. (b) The stereo camera image and corresponding imaging sonar image of rexrov robot.