Keyframe-based Dense Mapping with the Graph of View-Dependent Local Maps

TL;DR

This paper introduces a keyframe-based dense mapping system that uses view-dependent local maps and a pose graph for improved accuracy and global consistency in RGB-D sensor environments.

Contribution

It presents a novel method for updating and merging local NDT maps with view-dependent structures and a pose graph for global map correction.

Findings

Outperforms Octomap and NDT-OM in accuracy

Enables effective global map correction after loop closure

Provides detailed example applications

Abstract

In this article, we propose a new keyframe-based mapping system. The proposed method updates local Normal Distribution Transform maps (NDT) using data from an RGB-D sensor. The cells of the NDT are stored in 2D view-dependent structures to better utilize the properties and uncertainty model of RGB-D cameras. This method naturally represents an object closer to the camera origin with higher precision. The local maps are stored in the pose graph which allows correcting global map after loop closure detection. We also propose a procedure that allows merging and filtering local maps to obtain a global map of the environment. Finally, we compare our method with Octomap and NDT-OM and provide example applications of the proposed mapping method.