Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LiDAR-CS Dataset: LiDAR Point Cloud Dataset with Cross-Sensors for 3D Object Detection

Jin Fang, Dingfu Zhou, Jingjing Zhao, Chenming Wu, Chulin Tang, Cheng-Zhong Xu, Liangjun Zhang

TL;DR

The paper tackles the sensor-related domain gaps in 3D LiDAR-based object detection by introducing LiDAR-CS Dataset, a large-scale cross-sensor, annotated dataset generated with a pattern-aware LiDAR simulator. The framework comprises a Pattern Generation module that recovers LiDAR ray patterns from real scans and a Data Generation module that renders multi-sensor point clouds with consistent ground-truth annotations. Empirical results on five detectors reveal clear cross-sensor domain gaps and demonstrate that simple domain-alignment strategies (NNDS) can substantially reduce transfer losses. The dataset enables robust cross-sensor evaluation, sensor-parameter selection, and development of distribution-insensitive features, with public release to spur further research.

Abstract

Over the past few years, there has been remarkable progress in research on 3D point clouds and their use in autonomous driving scenarios has become widespread. However, deep learning methods heavily rely on annotated data and often face domain generalization issues. Unlike 2D images whose domains usually pertain to the texture information present in them, the features derived from a 3D point cloud are affected by the distribution of the points. The lack of a 3D domain adaptation benchmark leads to the common practice of training a model on one benchmark (e.g. Waymo) and then assessing it on another dataset (e.g. KITTI). This setting results in two distinct domain gaps: scenarios and sensors, making it difficult to analyze and evaluate the method accurately. To tackle this problem, this paper presents LiDAR Dataset with Cross Sensors (LiDAR-CS Dataset), which contains large-scale annotated LiDAR point cloud under six groups of different sensors but with the same corresponding scenarios, captured from hybrid realistic LiDAR simulator. To our knowledge, LiDAR-CS Dataset is the first dataset that addresses the sensor-related gaps in the domain of 3D object detection in real traffic. Furthermore, we evaluate and analyze the performance using various baseline detectors and demonstrated its potential applications. Project page: https://opendriving.github.io/lidar-cs.

LiDAR-CS Dataset: LiDAR Point Cloud Dataset with Cross-Sensors for 3D Object Detection

TL;DR

The paper tackles the sensor-related domain gaps in 3D LiDAR-based object detection by introducing LiDAR-CS Dataset, a large-scale cross-sensor, annotated dataset generated with a pattern-aware LiDAR simulator. The framework comprises a Pattern Generation module that recovers LiDAR ray patterns from real scans and a Data Generation module that renders multi-sensor point clouds with consistent ground-truth annotations. Empirical results on five detectors reveal clear cross-sensor domain gaps and demonstrate that simple domain-alignment strategies (NNDS) can substantially reduce transfer losses. The dataset enables robust cross-sensor evaluation, sensor-parameter selection, and development of distribution-insensitive features, with public release to spur further research.

Abstract

Over the past few years, there has been remarkable progress in research on 3D point clouds and their use in autonomous driving scenarios has become widespread. However, deep learning methods heavily rely on annotated data and often face domain generalization issues. Unlike 2D images whose domains usually pertain to the texture information present in them, the features derived from a 3D point cloud are affected by the distribution of the points. The lack of a 3D domain adaptation benchmark leads to the common practice of training a model on one benchmark (e.g. Waymo) and then assessing it on another dataset (e.g. KITTI). This setting results in two distinct domain gaps: scenarios and sensors, making it difficult to analyze and evaluate the method accurately. To tackle this problem, this paper presents LiDAR Dataset with Cross Sensors (LiDAR-CS Dataset), which contains large-scale annotated LiDAR point cloud under six groups of different sensors but with the same corresponding scenarios, captured from hybrid realistic LiDAR simulator. To our knowledge, LiDAR-CS Dataset is the first dataset that addresses the sensor-related gaps in the domain of 3D object detection in real traffic. Furthermore, we evaluate and analyze the performance using various baseline detectors and demonstrated its potential applications. Project page: https://opendriving.github.io/lidar-cs.
Paper Structure (16 sections, 2 equations, 3 figures, 3 tables)

This paper contains 16 sections, 2 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. LiDAR-CS Dataset
  4. Cross-Sensor Data Generation
  5. Pattern Generation Module
  6. Data Generation Module
  7. LiDAR-CS Dataset
  8. Dataset Details
  9. Evaluation Metric
  10. Empirical Studies
  11. 3D Object Detection Baselines
  12. Cross-Sensors Evaluation
  13. Domain Alignment Evaluation
  14. Potential Applications and Discussion
  15. Conclusion
  16. ...and 1 more sections

Figures (3)

  • Figure 1: (a) and (b) are LiDAR point cloud examples collected from different types of sensors which are from 64-beams and 16-beams LiDAR sensors respectively. The vehicle has been cropped and zoomed in for detailed visualization purposes. Sub-figure (c) gives a cross-sensors evaluation of experimental results where four baseline detectors are trained on VLD 64 LiDAR data and evaluated on five different sensors in the same scenario. The results show that the domain gaps are obvious across different sensors.
  • Figure 2: The overview of the proposed pattern-aware LiDAR Simulator framework. First of all, the real LiDAR points are normalized to a spherical surface. Due to points missing, statistics information from multiple scans is required to build the LiDAR Ray Pattern. Then, the Ray Pattern vectors are simultaneously projected and query the depth value from the depth map to generate the simulation point cloud.
  • Figure 3: An example of the LiDAR-CS dataset. All the point clouds are generated from the same scenario under different sensor patterns. The points in the cycle are zoomed in and shown in the white boxes for a better view. The point clouds are colorized by the height of the points. Better viewed in color.