SafeMap: Robust HD Map Construction from Incomplete Observations

Xiaoshuai Hao, Lingdong Kong, Rong Yin, Pengwei Wang, Jing Zhang, Yunfeng Diao, Shu Zhao

TL;DR: Robust HD Map Construction from Incomplete Observations

Abstract: Robust high-definition (HD) map construction is vital for autonomous driving, yet existing methods often struggle with incomplete multi-view camera data. This paper presents SafeMap, a novel framework specifically designed to ensure accuracy even when certain camera views are missing. SafeMap integrates two key components: the Gaussian-based Perspective View Reconstruction (G-PVR) module and the Distillation-based Bird’s-Eye-View (BEV) Correction (D-BEVC) module. G-PVR leverages prior knowledge of view importance to dynamically prioritize the most informative regions based on the relationships among available camera views. Furthermore, D-BEVC utilizes panoramic BEV features to correct the BEV representations derived from incomplete observations. Together, these components facilitate comprehensive data reconstruction and robust HD map generation. SafeMap is easy to implement and integrates seamlessly into existing systems, offering a plug-and-play solution for enhanced robustness. Experimental results demonstrate that SafeMap significantly outperforms previous methods in both complete and incomplete scenarios, highlighting its superior performance and resilience.

Lay Summary:

SafeMap is an innovative framework designed to create accurate high-definition maps for autonomous driving, even when some camera views are missing. Traditional methods often struggle with incomplete data from multiple cameras, which can lead to errors in map construction. SafeMap addresses this challenge by using two main features:

1. G-PVR identifies the most important areas to focus on based on the available camera views, ensuring that the most relevant information is prioritized.
2. D-BEVC uses advanced panoramic features to enhance the map representations, correcting any inaccuracies caused by missing views.

Together, these features allow for thorough data reconstruction and reliable map generation. SafeMap is user-friendly and can easily fit into existing systems, making it a practical solution for improving the robustness of autonomous driving technologies. Our experiments show that SafeMap performs significantly better than previous methods, proving its effectiveness in both complete and incomplete data scenarios.

Primary Area: Applications->Computer Vision

Keywords: Autonomous Driving；HD Map Construction；Incomplete Observations；

Submission Number: 8037