LiDAR Sequence

LiDAR sequence processing focuses on efficiently and accurately extracting information from temporally ordered point cloud data acquired by LiDAR sensors, primarily for autonomous driving and robotics applications. Current research emphasizes developing novel architectures, such as transformers and autoencoders, to effectively model spatio-temporal relationships within these sequences, often incorporating self-supervised learning to address data scarcity. This work aims to improve tasks like object segmentation, prediction, and place recognition, leading to more robust and reliable perception systems for autonomous vehicles and other robotic applications.