자율 주행을 위한 딥러닝 기반 라이다 객체 인식 신경망 연구 분석 > Domestic Journals

자율 주행을 위한 딥러닝 기반 라이다 객체 인식 신경망 연구 분석
Year	2022
Month
Journal	2022 / 한국자동차공학회논문집 vol. 30, no. 9, pp. 635-647
Author	선민혁, 백동희, 공승현
Object detection is one of the most crucial functions for autonomous driving since path planning, obstacle avoidance, and numerous other functions rely on the acquired information regarding the positions of objects on the road. To enable accurate object detections, numerous works utilize lidar as the primary sensor since it can accurately acquire 3D measurements and are robust to adverse environmental conditions such as poor illumination. In this work, we aim to comprehensively review deep learning-based object detection using lidar which has shown remarkable detection performance on various datasets. First, we explain the general concepts of deep learning-based lidar object detection along with the datasets and benchmarks that are commonly used in existing works. We then thoroughly discuss the latest state-of-the-art neural networks for lidar object detection. Finally, we provide suggestions on how to employ these networks into an autonomous driving system.