Camera-Based Online Vectorized HD Map Construction With Incomplete Observation
- 30 May 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 9 (7) , 6456-6463
Abstract
Camera-based online map construction focuses on learning map elements from surround-view images. Distinguished with previous methods that rely on complete observations, we explore a new map construction problem under incomplete observations where one or more perspectives of the surround-view are missing due to camera damage or occlusion. Incomplete observations lead to inferior performance and may even result in failure. Map construction based on incomplete observations faces two challenges: supplementing missing perspective features and reducing the complexity of high-dimensional feature learning. To address these issues, we propose a novel Panoramic Observation Prior Network (POP-Net). Firstly, based on the observation switch training mechanism, we propose a Panoramic Learning Module (PL-Module). It establishes a learnable panoramic feature space, facilitating the extraction of panoramic features from incomplete observations, thus supplementing missing perspective features. Secondly, based on the feature decomposition mechanism, we design a Panoramic Decomposition-Aggregation Operation (PDA-Operation), which decomposes high-dimensional panoramic features into low-dimensional local scene features. This allows limited local scene features to represent diverse panoramic features, alleviating computational and memory burdens of high-dimensional feature learning. Experimental results demonstrate that our method surpasses existing approaches under incomplete observation scenarios.Keywords
Funding Information
- National Natural Science Foundation of China (U22A2058, 62176138, 62176136)
- Shandong Outstanding Youth Funding (ZR2023YQ054)
- Shandong Provincial Key research and development (2019JZZY010130, 2020CXGC010207)
This publication has 27 references indexed in Scilit:
- End-to-End Vectorized HD-map Construction with Piecewise Bézier Curve2023
- Neural Map Prior for Autonomous Driving2023
- BEVFormer: Learning Bird’s-Eye-View Representation from Multi-camera Images via Spatiotemporal Transformers2022
- TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformers2022
- Video Prediction Recalling Long-term Motion Context via Memory Alignment Learning2021
- nuScenes: A Multimodal Dataset for Autonomous Driving2020
- VectorNet: Encoding HD Maps and Agent Dynamics From Vectorized Representation2020
- Deep Residual Learning for Image Recognition2016
- Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network2016
- Aggregating local descriptors into a compact image representation2010