PAPL-SLAM: Principal Axis-Anchored Monocular Point-Line SLAM

TL;DR

This work models the line-axis probabilistic data association using the Expectation Maximization (EM) algorithm and provides the pipelines for axis creation, updating, and optimization, enhancing the system's robustness and avoiding mismatch.

Abstract

In point-line SLAM systems, the utilization of line structural information and the optimization of lines are two significant problems. The former is usually addressed through structural regularities, while the latter typically involves using minimal parameter representations of lines in optimization. However, separating these two steps leads to the loss of constraint information to each other. We anchor lines with similar directions to a principal axis and optimize them with $n+2$ parameters for $n$ lines, solving both problems together. Our method considers scene structural information, which can be easily extended to different world hypotheses while significantly reducing the number of line parameters to be optimized, enabling rapid and accurate mapping and tracking. To further enhance the system's robustness and avoid mismatch, we have modeled the line-axis probabilistic data association and provided the algorithm for axis creation, updating, and optimization. Additionally, considering that most real-world scenes conform to the Atlanta World hypothesis, we provide a structural line detection strategy based on vertical priors and vanishing points. Experimental results and ablation studies on various indoor and outdoor datasets demonstrate the effectiveness of our system.

Figures (4)

• Figure 1: Overall of our system. Our system is based on VPL-SLAM chen2024vpl and ORB-SLAM2 mur2017orb, consisting of tracking, local mapping, and loop closure threads. (a) shows the tracking thread workflow, where we detect structural lines in the scene and perform pose optimization. (b) shows the local mapping thread, where we perform principal axis management and local optimization. (c) show the loop closing process, using structural lines to optimize the essential graph. (d) shows our factor graph for BA, where we use a three-parameter optimization method based on principal axes in addition to traditional point and line feature optimization.
• Figure 2: Principal axis and line handling algorithms. (a) illustrates the definition of our three-parameter line representation. (b) defines our principal axis probabilistic association model. (c) shows the structural line and vanishing point detection strategy based on the vertical prior. (d) demonstrates the triangulation process of 3D lines and the reprojection error during optimization.
• Figure 3: Qualitative comparison between our representation and the orthogonal representation before and after optimization, under given pose and observation perturbations.
• Figure 4: The qualitative mapping results on the garage dataset show lines in different colors representing different principal axes. Using line features and various colors, our map visually and intuitively illustrates the corresponding scene.