Scale Estimation with Dual Quadrics for Monocular Object SLAM

TL;DR

This paper introduces a novel scale estimation method for monocular SLAM using dual quadrics to accurately represent objects and optimize their dimensions, enabling absolute scale recovery without prior gravity information.

Contribution

The paper proposes a new nonlinear optimization approach with dual quadrics for object-level SLAM to estimate absolute scale without gravity priors.

Findings

Accurate scale estimation achieved without gravity prior

Dual quadrics enable compact and precise object representation

System effectively integrates object dimensions for scale recovery

Abstract

The scale ambiguity problem is inherently unsolvable to monocular SLAM without the metric baseline between moving cameras. In this paper, we present a novel scale estimation approach based on an object-level SLAM system. To obtain the absolute scale of the reconstructed map, we derive a nonlinear optimization method to make the scaled dimensions of objects conforming to the distribution of their sizes in the physical world, without relying on any prior information of gravity direction. We adopt the dual quadric to represent objects for its ability to fit objects compactly and accurately. In the proposed monocular object-level SLAM system, dual quadrics are fastly initialized based on constraints of 2-D detections and fitted oriented bounding box and are further optimized to provide reliable dimensions for scale estimation.