Top Image 1
Top Image 2

Water-DSLAM: An Elastic Multi-Modal Dense SLAM Architecture for Sustained Underwater In-Situ Observation

Exploring the fine-grained underwater world
Welcome to exchange ideas and collaborate

Project Overall Introduction

Abstract

Existing underwater SLAM systems often struggle to sustain reliable operation in underwater environments with weak texture and degraded geometry, resulting in intermittent tracking and sparse mapping. Therefore, we present Water-DSLAM, an elastic multi-modal dense SLAM architecture to sustain reliable in-situ observation in diverse complex underwater scenarios. For evaluation, we use Water-Scanner, a multi-sensor robotic platform equipped with a self-scanning Underwater Binocular Structured Light (UBSL) module for accurate 3D data acquisition. A flexible and extensible multi-subsystem front-end design is adopted, where the motion estimation module provides high-rate position and orientation estimates and is tightly coupled with the Water-Stereo subsystem to enhance performance in environments with weak visual features, as well as with the Water-UBSL subsystem to improve robustness in structurally degraded scenarios. Furthermore, a multi-modal factor graph back-end dynamically fuses available heterogeneous sensor constraints. The proposed factor graph maintenance strategy associates constraints with their observation times to accommodate asynchronous updates and partial data loss. Experimental results demonstrate that Water-DSLAM is capable of sustained localization and dense mapping in multiple challenging underwater environments, including an artificial pool, a 16-meter-deep sinkhole, a natural river, and an offshore zone of the South China Sea. This architecture supports continuous underwater robotic dense in-situ observation through elastic use of external constraints under degraded sensing conditions. To support the underwater robotics community, we will open-source our hardware designs and code after the paper is accepted.

Robot: Water-Scanner

Framework: Water-DSLAM

Selected Key Experiment

Quick Highlights: Water-DSLAM

Ablation Experiment: Multi-Subsystem Ablation Verification

Comparative Experiment: Free-Motion Dense Mapping in Pool

Application Experiment 1: Underwater Dark Scene In-situ Observation

Application Experiment 2: Underwater Sinkhole In-situ Observation

Application Experiment 3: Field River In-situ Observation

Application Experiment 4: Offsore Seabed In-situ Observation

BibTeX


        Will come later..