3D Object Detection for Autonomous Driving: A Comprehensive Survey

Jiageng Mao,Shaoshuai Shi,Xiaogang Wang,Hongsheng Li
2023-04-04
Abstract:Autonomous driving, in recent years, has been receiving increasing attention for its potential to relieve drivers' burdens and improve the safety of driving. In modern autonomous driving pipelines, the perception system is an indispensable component, aiming to accurately estimate the status of surrounding environments and provide reliable observations for prediction and planning. 3D object detection, which intelligently predicts the locations, sizes, and categories of the critical 3D objects near an autonomous vehicle, is an important part of a perception system. This paper reviews the advances in 3D object detection for autonomous driving. First, we introduce the background of 3D object detection and discuss the challenges in this task. Second, we conduct a comprehensive survey of the progress in 3D object detection from the aspects of models and sensory inputs, including LiDAR-based, camera-based, and multi-modal detection approaches. We also provide an in-depth analysis of the potentials and challenges in each category of methods. Additionally, we systematically investigate the applications of 3D object detection in driving systems. Finally, we conduct a performance analysis of the 3D object detection approaches, and we further summarize the research trends over the years and prospect the future directions of this area.
Computer Vision and Pattern Recognition,Artificial Intelligence,Robotics
What problem does this paper attempt to address?
### Problems the Paper Attempts to Solve This paper aims to comprehensively review and analyze 3D object detection technology in the field of autonomous driving. Specifically, the paper attempts to address the following core issues: 1. **Background Introduction and Challenges**: - **Background**: Autonomous driving has received increasing attention in recent years because it can reduce the burden on drivers and improve driving safety. In modern autonomous driving systems, the perception system is an indispensable component, aiming to accurately estimate the state of the surrounding environment and provide reliable observational data for prediction and planning. - **Challenges**: 3D object detection is an important part of the perception system, aiming to predict the position, size, and category of 3D objects near the autonomous vehicle. However, this task faces many challenges, including the heterogeneity of data representation, the need to process different views, and the requirement for high-precision localization. 2. **Method Review**: - **Models and Sensor Inputs**: The paper conducts a comprehensive survey of the progress in 3D object detection from the perspective of models and sensor inputs, including LiDAR-based, camera-based, and multi-modal detection methods. - **Classification Methods**: An in-depth analysis of the potential and challenges of each method is provided, including methods under different data representations such as point clouds, meshes, point-voxel, and range images. 3. **Application Analysis**: - **Applications in Driving Systems**: A systematic study of the applications of 3D object detection in driving systems, including object tracking, trajectory prediction, and path planning. 4. **Performance Evaluation and Trend Analysis**: - **Performance Analysis**: A comprehensive analysis of the performance of 3D object detection methods, including the evaluation of speed and accuracy. - **Research Trends and Future Directions**: A summary of recent research trends and an outlook on future development directions. ### Summary By comprehensively reviewing and analyzing 3D object detection technology, this paper aims to provide researchers with a systematic perspective to understand the advantages and challenges of different methods, thereby promoting further development in the field of autonomous driving.