Bundle Adjustment in the Eager Mode

TL;DR

This paper introduces a GPU-accelerated, eager-mode bundle adjustment library integrated with PyTorch, significantly improving runtime efficiency for robotic applications like SLAM and AR.

Contribution

It presents a novel eager-mode BA library with native PyTorch integration, sparsity-aware auto-differentiation, and GPU acceleration for enhanced performance.

Findings

Achieves an average speedup of 18.5x over GTSAM

Achieves an average speedup of 22x over g2o

Achieves an average speedup of 23x over Ceres

Abstract

Bundle adjustment (BA) is a critical technique in various robotic applications such as simultaneous localization and mapping (SLAM), augmented reality (AR), and photogrammetry. BA optimizes parameters such as camera poses and 3D landmarks to align them with observations. With the growing importance of deep learning in perception systems, there is an increasing need to integrate BA with deep learning frameworks for enhanced reliability and performance. However, widely-used C++-based BA libraries, such as GTSAM, g$^2$o, and Ceres Solver, lack native integration with modern deep learning libraries like PyTorch. This limitation affects their flexibility, ease of debugging, and overall implementation efficiency. To address this gap, we introduce an eager-mode BA library seamlessly integrated with PyTorch with high efficiency. Our approach includes a sparsity-aware auto-differentiation design and GPU-accelerated sparse operations designed for 2nd-order optimization. Our eager-mode BA on GPU demonstrates substantial runtime efficiency, achieving an average speedup of 18.5$\times$, 22$\times$, and 23$\times$ across all benchmarks compared to GTSAM, g$^2$o, and Ceres, respectively.