A Survey of LiDAR-Based Monomodal 3D Object Detection

3D object detection is a key technology for identifying objects by analyzing the spatial information contained in sensor data. Its goal is to obtain the position, category, and boundaries of objects, and it is widely applied in fields such as autonomous driving and robotic navigation. This paper first introduces the main characteristics of point cloud data, including sparsity, irregularity, discreteness, and rotational invariance. Then, according to the different data representation methods, this paper comprehensively sorts out and analyzes the current research status and progress of monomodal 3D object detection from four methods: voxel-based, point-based, point-voxel based, and graph-based methods. Furthermore, the paper discusses commonly used datasets and evaluation metrics in the field of 3D object detection, providing a foundation for algorithm performance evaluation and comparative analysis. Finally, the paper looks forward to future development trends in 3D object detection, offering insights and directions for future research.