$\nu$-DBA: Neural Implicit Dense Bundle Adjustment Enables Image-Only Driving Scene Reconstruction

TL;DR

This paper introduces $ u$-DBA, a neural implicit dense bundle adjustment framework that improves autonomous driving scene reconstruction by jointly optimizing trajectories and dense maps using optical flow and neural surfaces.

Contribution

It presents a novel neural implicit dense bundle adjustment method that enhances trajectory and dense map optimization in autonomous driving scenes.

Findings

Achieves superior trajectory optimization accuracy.

Provides more detailed and accurate dense scene reconstructions.

Improves performance with scene-specific self-supervised optical flow fine-tuning.

Abstract

The joint optimization of the sensor trajectory and 3D map is a crucial characteristic of bundle adjustment (BA), essential for autonomous driving. This paper presents $\nu$-DBA, a novel framework implementing geometric dense bundle adjustment (DBA) using 3D neural implicit surfaces for map parametrization, which optimizes both the map surface and trajectory poses using geometric error guided by dense optical flow prediction. Additionally, we fine-tune the optical flow model with per-scene self-supervision to further improve the quality of the dense mapping. Our experimental results on multiple driving scene datasets demonstrate that our method achieves superior trajectory optimization and dense reconstruction accuracy. We also investigate the influences of photometric error and different neural geometric priors on the performance of surface reconstruction and novel view synthesis. Our method stands as a significant step towards leveraging neural implicit representations in dense bundle adjustment for more accurate trajectories and detailed environmental mapping.