InfiniTAM v3: A Framework for Large-Scale 3D Reconstruction with Loop Closure

Abstract

Volumetric models have become a popular representation for 3D scenes in recent years. One breakthrough leading to their popularity was KinectFusion, which focuses on 3D reconstruction using RGB-D sensors. However, monocular SLAM has since also been tackled with very similar approaches. Representing the reconstruction volumetrically as a TSDF leads to most of the simplicity and efficiency that can be achieved with GPU implementations of these systems. However, this representation is memory-intensive and limits applicability to small-scale reconstructions. Several avenues have been explored to overcome this. With the aim of summarizing them and providing for a fast, flexible 3D reconstruction pipeline, we propose a new, unifying framework called InfiniTAM. The idea is that steps like camera tracking, scene representation and integration of new data can easily be replaced and adapted to the user's needs. This report describes the technical implementation details of InfiniTAM v3, the third version of our InfiniTAM system. We have added various new features, as well as making numerous enhancements to the low-level code that significantly improve our camera tracking performance. The new features that we expect to be of most interest are (i) a robust camera tracking module; (ii) an implementation of Glocker et al.'s keyframe-based random ferns camera relocaliser; (iii) a novel approach to globally-consistent TSDF-based reconstruction, based on dividing the scene into rigid submaps and optimising the relative poses between them; and (iv) an implementation of Keller et al.'s surfel-based reconstruction approach.

This article largely supersedes arxiv:1410.0925 (it describes version 3 of the InfiniTAM framework)

References in corpus (1)

• On-the-Fly Adaptation of Regression Forests for Online Camera Relocalisation

Cited by in corpus (16)

• Unsupervised Scale-consistent Depth Learning from Video
• Real-Time RGB-D Camera Pose Estimation in Novel Scenes using a Relocalisation Cascade
• Collaborative Large-Scale Dense 3D Reconstruction with Online Inter-Agent Pose Optimisation
• R${}^3$SGM: Real-time Raster-Respecting Semi-Global Matching for Power-Constrained Systems
• A Survey of Simultaneous Localization and Mapping with an Envision in 6G Wireless Networks
• NeuralBlox: Real-Time Neural Representation Fusion for Robust Volumetric Mapping
• ASH: A Modern Framework for Parallel Spatial Hashing in 3D Perception
• Structure Aware SLAM using Quadrics and Planes
• Efficient 3D Reconstruction, Streaming and Visualization of Static and Dynamic Scene Parts for Multi-client Live-telepresence in Large-scale Environments
• Rendering the Directional TSDF for Tracking and Multi-Sensor Registration with Point-To-Plane Scale ICP
• Self-supervised Depth Denoising Using Lower- and Higher-quality RGB-D sensors
• DI-Fusion: Online Implicit 3D Reconstruction with Deep Priors
• A Survey of FPGA-Based Robotic Computing
• Semantic Dense Reconstruction with Consistent Scene Segments
• PSDF Fusion: Probabilistic Signed Distance Function for On-the-fly 3D Data Fusion and Scene Reconstruction
• RGB-D SLAM with Structural Regularities