Asso.Prof. Xuesong Li
Project Description and Objectives
Automotive engineering and related technologies have gained substantial attention and investment in this information era, surprising or not. With the technology of hybrid vehicle, electric vehicle, vehicle-to-everything (V2X), assisted driving and autonomous driving, etc., new insights and rapid growth are seen in this industry, with the embrace of both traditional technologies and new methodologies. This summer research program will focus on some of the hot topics in this field, such as computer vision for the application of assisted-driving and autonomous driving. This program will aim at both theoretical studies including a literature review and report drafting, as well as gaining hands-on experience from programming and image processing to recognize vehicles/pedestrian/signal lights from real images captured. The objective of the project is to help the students build up ideas about how research projects and engineering projects are performed, and also help the students to understand the fundamentals for the chosen topics so that they would be better equipped in continuing or starting education/employment in the field of automotive engineering.
Basic understanding of programming (Python preferred) and image processing.
(1) Traditional computer technology study such as camera calibration, color space conversion, edge detection, etc. for lane line detection.
(2) Image classification using machine technology like SVM and deep neural network such as LeNet, VGGNet, etc. These architectures are used in the up-to-date object detection algorithms like region-proposed CNN.
(3) Semantic segmentation of the image for drivable area detection using a fully connected neural network.