Robust Road Surface Normal and Pitch Estimation via IMU-Camera Fusion

1Machine Perception Research Laboratory, HUN-REN SZTAKI, Budapest, Hungary 2Vehicle Industry Research Center, Széchenyi István University, Győr, Hungary
Pitch GIF

Abstract

Predicting road surface normal and pitch with image-based algorithms remains a significant challenge, especially when steep inclines, declines, and sudden changes in road inclination are involved.

To solve this problem, we propose a novel image-based algorithm that leverages homography decomposition to achieve accurate ground plane normal and pitch estimation.By integrating a Kalman Filter, our method en- hances prediction stability. Further, we incorporate robust sensor fu- sion by integrating IMU-based odometry, ensuring that the estimates are accurately aligned with real-world motion. Overall, our approach outperforms existing techniques in both accuracy and responsiveness for dynamic driving environments.

Experimental results show that our approach achieves superior performance, reducing average pitch and normal errors by 0.493° and 0.483°, respectively, compared to the current state-of-the-art, and exhibits a shorter transient response in case of a sudden road inclination change. Code will be available at https://norbertmarko.github.io/ground-normal-prediction/

Normal GIF

Result Figures

Sequence 020 comparison with SOTA
Sequence 020 comparison with SOTA.
Sequence 029 comparison with SOTA
Sequence 029 comparison with SOTA.
Sequence 034 comparison with SOTA
Sequence 034 comparison with SOTA.
Sequence 039 comparison with SOTA
Sequence 039 comparison with SOTA.
Sequence 040 comparison with SOTA
Sequence 040 comparison with SOTA.
Sequence 043 comparison with SOTA
Sequence 043 comparison with SOTA.
Sequence 044 comparison with SOTA
Sequence 044 comparison with SOTA.

Result Tables

Result Table 1
Result Table 2

BibTeX

@article{marko2025roboust,
  author    = {Norbert Markó and Zoltán Rózsa and Áron Ballagi and Tamás Szirányi},
  title     = {Robust Road Surface Normal and Pitch Estimation via IMU-Camera Fusion},
  journal   = {},
  year      = {2025},
}