Automatic Generation of Vector Maps from Aerial Images for Autonomous Driving

PhD Thesis Proposal Defence


Title: "Automatic Generation of Vector Maps from Aerial Images for Autonomous 
Driving"

by

Mr. Zhenhua XU


Abstract:

The vector map of road elements is critical for autonomous driving. It contains 
the connectivity and topology information of the road for multiple downstream 
tasks of the autonomous vehicles, such as prediction, planning and motion 
control. To assure the normal and safe operations of autonomous vehicles, 
vector maps, including the SD (standard-definition) map and the HD 
(high-definition) map need to be pre-built before the deployment of autonomous 
vehicles in the target region. However, manually annotating the vector map is 
labor-intensive, expensive and time-consuming, which severely restricts the 
fast application and deployment of the autonomous vehicle. Therefore, automatic 
approaches to effectively and efficiently generate the vector map of the target 
region are needed. With the rapid development of aerial imaging, 
high-resolution aerial images can be easily accessed all over the globe, which 
can provide high-quality BEV (bird-eye-view) images around the vehicle. Thus, 
aerial images are leveraged in this proposal for vector map generation. 
Previous works on our task could be classified into three categories: 
segmentation-based approaches, two-stage graph-based approaches and 
decision-making graph-based approaches. Compared with segmentation-based 
approaches that only focus on pixel-level prediction, graph-based approaches 
can better optimize the graph structure and present superior topology 
correctness. Therefore, this proposal mainly focuses on graph-based approaches.

In this proposal, we will discuss three graph-based systems for vector map 
generation. First, we propose to detect the graph of road curbs from small 
aerial image patches with a decision-making network trained by imitation 
learning. Road curbs possess simple topology without complicated intersections 
or overlapping. Then, we design a two-stage graph-based approach to generate 
the vector map of road boundaries in city-scale aerial images. After obtaining 
key points of road boundaries as graph vertices, we predict the adjacency 
matrix for graph edges. Finally, we manage to detect the graph structure of the 
road network, whose topology is much more complicated than road curbs and road 
boundaries, such as intersections of arbitrary numbers of roads and overpasses. 
A decision-making system based on transformer is proposed to generate the 
vector map vertex by vertex. The proposed system can handle road networks with 
any topological changes, which presents state-of-the-art performance compared 
with previous works. In the end, we also analyze remaining problems that cannot 
be well solved by existing approaches as potential future works, such as vector 
map generation from vehicle-mounted sensors, map update problems and how to 
adapt advanced decision-making algorithms to our task.


Date:			Friday, 4 November 2022

Time:                  	4:30pm - 6:30pm

Venue: 			Room 5501
 			Lifts 25/26

Committee Members:	Dr. Ming Liu (Supervisor)
 			Prof. Huamin Qu (Supervisor)
 			Dr. Dan Xu (Chairperson)
 			Prof Ling SHI (ECE)


**** ALL are Welcome ****