LiDAR Point Cloud
LiDAR point clouds are collections of 3D points representing a scene, crucial for autonomous systems needing precise environmental understanding. Current research emphasizes efficient processing of these large datasets, focusing on learned feature extraction to reduce computational load and improve accuracy in tasks like simultaneous localization and mapping (SLAM), place recognition, and object detection. This involves developing novel neural network architectures, such as transformers and graph convolutional networks, often combined with multimodal fusion (e.g., integrating LiDAR with camera data) to enhance robustness and accuracy. The resulting advancements have significant implications for autonomous driving, robotics, and 3D mapping applications.
Papers
Joint Intrinsic and Extrinsic LiDAR-Camera Calibration in Targetless Environments Using Plane-Constrained Bundle Adjustment
Liang Li, Haotian Li, Xiyuan Liu, Dongjiao He, Ziliang Miao, Fanze Kong, Rundong Li, Zheng Liu, Fu Zhang
SCP: Spherical-Coordinate-based Learned Point Cloud Compression
Ao Luo, Linxin Song, Keisuke Nonaka, Kyohei Unno, Heming Sun, Masayuki Goto, Jiro Katto
LiDAR-BEVMTN: Real-Time LiDAR Bird's-Eye View Multi-Task Perception Network for Autonomous Driving
Sambit Mohapatra, Senthil Yogamani, Varun Ravi Kumar, Stefan Milz, Heinrich Gotzig, Patrick Mäder
Ada3D : Exploiting the Spatial Redundancy with Adaptive Inference for Efficient 3D Object Detection
Tianchen Zhao, Xuefei Ning, Ke Hong, Zhongyuan Qiu, Pu Lu, Yali Zhao, Linfeng Zhang, Lipu Zhou, Guohao Dai, Huazhong Yang, Yu Wang