Abstract: Compared with traditional 2D object detection techniques, 3D object detection techniques can effectively solve the occlusion problems and enhance the spatial perception ability by utilizing the depth information, thus significantly improving the accuracy of object detection algorithms. This paper analyzes the background of the development of 3D object detection techniques and dissects its significant advantages compared with 2D detection. Subsequently, it systematically classifies and reviews the 3D object detection algorithms, focusing on methods based on points, voxels, point-voxel fusion, and multimodal and image fusion, and analyzes the technical characteristics of each type of method. Meanwhile, it conducts a comparative analysis of commonly used large-scale open-source datasets such as KITTI, Waymo, and NuScenes, revealing their crucial role in promoting the development of 3D object detection techniques. Finally, it summarizes the current research status of 3D object detection techniques, points out the challenges it faces in practical applications, and prospects the future development directions of this field.