In autonomous driving, different sensors such as cameras, dynamic vision sensors(DVS), and light detection and ranging(LiDAR) products have been wide. Perceptions of outdoor environments are subjected to complex scenarios, among which raindrops and snowflakes are the most common factors. Raindrops or snowflakes appear inside the perceived data, resulting in degradation of data quality. Therefore, this report introduces a series of perception enhancement methods with different sensors in complex outdoor scenarios. Using the information extracted from continuous temporal space, the outliers such as raindrops or snowflakes are removed while the non-noise is preserved. Experiments in CARLA simulators and real-world environments are reported as well.
Director, the Institute of Artificial Intelligence and Unmanned Systems of the School of Computer Science
Sun Yat-Sen University
Kai Huang joined Sun Yat-Sen University as a Professor in 2015. He was appointed as the director of the Institute of Artificial Intelligence and Unmanned Systems of the School of Computer Science in 2020. He was a senior researcher in the Computer Science Department, the Technical University of Munich, Germany from 2012 to 2015. He obtained his Ph.D. in ETH Zurich in 2010, MSc from University Leiden in 2005, and BSc from Fudan University in 1999. His research interests include techniques for the analysis, design, and optimization of embedded/CPS systems, particularly in the automotive, medical, and robotic domains. He was the recipient of the best paper awards/candidates for a number of conferences.
