A Dense Subframe-based SLAM Framework with Side-scan Sonar

TL;DR

A novel subframe-based dense SLAM framework utilizing SSS data is introduced, enabling effective dense matching in overlapping regions of paired side-scan images and presenting a feasible way of evaluating mapping quality against multi-beam echosounder data without the influence of pose.

Abstract

Side-scan sonar (SSS) is a lightweight acoustic sensor that is commonly deployed on autonomous underwater vehicles (AUVs) to provide high-resolution seafloor images. However, leveraging side-scan images for simultaneous localization and mapping (SLAM) presents a notable challenge, primarily due to the difficulty of establishing sufficient amount of accurate correspondences between these images. To address this, we introduce a novel subframe-based dense SLAM framework utilizing side-scan sonar data, enabling effective dense matching in overlapping regions of paired side-scan images. With each image being evenly divided into subframes, we propose a robust estimation pipeline to estimate the relative pose between each paired subframes, by using a good inlier set identified from dense correspondences. These relative poses are then integrated as edge constraints in a factor graph to optimize the AUV pose trajectory. The proposed framework is evaluated on three real datasets collected by a Hugin AUV. Among one of them includes manually-annotated keypoint correspondences as ground truth and is used for evaluation of pose trajectory. We also present a feasible way of evaluating mapping quality against multi-beam echosounder (MBES) data without the influence of pose. Experimental results demonstrate that our approach effectively mitigates drift from the dead-reckoning (DR) system and enables quasi-dense bathymetry reconstruction. An open-source implementation of this work is available.

Figures (15)

• Figure 1: Qualitative results of our proposed dense side-scan SLAM method on the three evaluated mission data named by their mission number. The green lines refer to the estimated trajectories, and the yellow points refer to the reconstructed quasi-dense map.
• Figure 2: A demonstration of overlapping side-scan images captured along two adjacent survey lines from our self-collected data (Mission58), where the sketch on top represents the AUV and its side-scan covering ranges. The non-transparent areas of the bottom images indicate the overlapped parts.
• Figure 3: Overview of our subframe-based side-scan sonar SLAM framework. The framework takes raw side-scan images with dead-reckoning poses as input and outputs an optimized pose trajectory together with a quasi-dense map. The grey color rectangles highlights the main components, while rectangles with no color-filled represents intermediate outputs.
• Figure 4: Illustration of the 3-step dense matching algorithm using two overlapping SSS images ($\mathbf{I}_{A}$ and $\mathbf{I}_{B}$). The figure demonstrates how the 4 sample patches in $\mathbf{I}_{A}$ (shown as rectangles with centre points in different colors) are aligned with areas in $\mathbf{I}_{B}$. After initialization, the approximate correspondences of the 4 patches (shown as dashed rectangles in corresponding colors) are found in $\mathbf{I}_{B}$ up to DR precision (top right). The positions of these correspondences are then adjusted through random search (bottom right). Finally, the propagation step further refines the results, leading to smooth and accurate correspondences (bottom left). Note that the last two steps are usually repeated a few times to obtain a satisfactory result.
• Figure 5: Factor graph representation of our proposed SSS SLAM framework. Left: an AUV surveying trajectory and a demo of an evenly-divided side-scan image aligned to the first survey line. Centre: factor graph of global pose graph optimization. Right: notations (top) and factor graph of relative pose estimation (bottom).