NeuralBlox: Real-Time Neural Representation Fusion for Robust Volumetric Mapping

TL;DR

NeuralBlox introduces a real-time neural implicit mapping method that incrementally fuses scene data into a robust, large-scale 3D map, outperforming traditional methods especially under noisy conditions.

Contribution

It proposes a novel fusion strategy and training pipeline for incremental neural implicit scene reconstruction, enabling large-scale, real-time mapping from partial observations.

Findings

Real-time incremental scene reconstruction on CPU.

More robust and complete maps compared to TSDFs.

Effective performance on real-world noisy datasets.

Abstract

We present a novel 3D mapping method leveraging the recent progress in neural implicit representation for 3D reconstruction. Most existing state-of-the-art neural implicit representation methods are limited to object-level reconstructions and can not incrementally perform updates given new data. In this work, we propose a fusion strategy and training pipeline to incrementally build and update neural implicit representations that enable the reconstruction of large scenes from sequential partial observations. By representing an arbitrarily sized scene as a grid of latent codes and performing updates directly in latent space, we show that incrementally built occupancy maps can be obtained in real-time even on a CPU. Compared to traditional approaches such as Truncated Signed Distance Fields (TSDFs), our map representation is significantly more robust in yielding a better scene completeness given noisy inputs. We demonstrate the performance of our approach in thorough experimental validation on real-world datasets with varying degrees of added pose noise.