Pure Vision and Multi-modal Perception in Autonomous Driving: Performance, Challenges and Architectural Insights

DOI

10.2991/978-94-6239-648-7_79How to use a DOI?

Keywords

Autonomous driving; Pure visual perception; Multimodal fusion; Generalization ability; Robustness

Abstract

Environmental perception is a core technical aspect of autonomous driving systems, and its architecture design directly determines the vehicle’s understanding ability of the surrounding environment and driving safety. With the development of artificial intelligence and sensor technology, pure visual perception and multimodal fusion perception have become two mainstream technical routes. The trade-off in performance, cost, and reliability between the two is of great significance for the practical deployment of autonomous driving systems. This article focuses on the design of a new perception architecture and explores how to enhance the overall performance and safety of autonomous driving systems. The article systematically analyzes and compares the performance differences between pure visual perception systems and multimodal perception systems (fusing cameras, millimeter-wave radars, laser radars, etc.) in terms of perception accuracy, generalization ability, and robustness. It summarizes the challenges faced by each path (such as generalization, stability, and reliability) and their impacts on downstream tasks, and reviews several representative research works. In addition, this article also analyzes the advantages of pure visual perception in system construction from multiple dimensions, while also pointing out its shortcomings and deficiencies in dealing with complex environments, in order to provide theoretical references and practical inspirations for the design of future perception architectures.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yaode Han
AU  - Kaiyang Li
PY  - 2026
DA  - 2026/04/24
TI  - Pure Vision and Multi-modal Perception in Autonomous Driving: Performance, Challenges and Architectural Insights
BT  - Proceedings of the International Workshop on Advances in Deep Learning for Image Analysis and Computer Vision (IWADIC 2025)
PB  - Atlantis Press
SP  - 731
EP  - 738
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6239-648-7_79
DO  - 10.2991/978-94-6239-648-7_79
ID  - Han2026
ER  -