Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar
Object detection is a critical task in autonomous driving. Currently, 3D object detection methods for autonomous driving primarily rely on stereo cameras and LiDAR, which are susceptible to adverse weather conditions and low lighting, resulting in limited robustness. In contrast, automotive mmWave r...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/8/1465 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850180884607533056 |
|---|---|
| author | Weigang Shi Panpan Tong Xin Bi |
| author_facet | Weigang Shi Panpan Tong Xin Bi |
| author_sort | Weigang Shi |
| collection | DOAJ |
| description | Object detection is a critical task in autonomous driving. Currently, 3D object detection methods for autonomous driving primarily rely on stereo cameras and LiDAR, which are susceptible to adverse weather conditions and low lighting, resulting in limited robustness. In contrast, automotive mmWave radar offers advantages such as resilience to complex weather, independence from lighting conditions, and a low cost, making it a widely studied sensor type. Modern 4D millimeter-wave (mmWave) radar can provide spatial dimensions (x, y, z) as well as Doppler information, meeting the requirements for 3D object detection. However, the point cloud density of 4D mmWave radar is significantly lower than that of LiDAR in the case of short distances, and existing point cloud object detection methods struggle to adapt to such sparse data. To address this challenge, we propose a novel 4D mmWave radar point cloud object detection framework. First, we employ moving least squares (MLS) to densify multi-frame fused point clouds, effectively increasing the point cloud density. Next, we construct a 3D object detection network based on point pillar encoding and utilize an SSD detection head for detection on feature maps. Finally, we validate our method on the VoD dataset. Experimental results demonstrate that our proposed framework outperforms comparative methods, and the MLS-based point cloud densification method significantly enhances the object detection performance. |
| format | Article |
| id | doaj-art-4df82e597c7d47e09ebbb7f7643e0066 |
| institution | OA Journals |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-4df82e597c7d47e09ebbb7f7643e00662025-08-20T02:18:01ZengMDPI AGRemote Sensing2072-42922025-04-01178146510.3390/rs17081465Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave RadarWeigang Shi0Panpan Tong1Xin Bi2School of Automotive Studies, Tongji University, Shanghai 201804, ChinaSchool of Automotive Studies, Tongji University, Shanghai 201804, ChinaSchool of Automotive Studies, Tongji University, Shanghai 201804, ChinaObject detection is a critical task in autonomous driving. Currently, 3D object detection methods for autonomous driving primarily rely on stereo cameras and LiDAR, which are susceptible to adverse weather conditions and low lighting, resulting in limited robustness. In contrast, automotive mmWave radar offers advantages such as resilience to complex weather, independence from lighting conditions, and a low cost, making it a widely studied sensor type. Modern 4D millimeter-wave (mmWave) radar can provide spatial dimensions (x, y, z) as well as Doppler information, meeting the requirements for 3D object detection. However, the point cloud density of 4D mmWave radar is significantly lower than that of LiDAR in the case of short distances, and existing point cloud object detection methods struggle to adapt to such sparse data. To address this challenge, we propose a novel 4D mmWave radar point cloud object detection framework. First, we employ moving least squares (MLS) to densify multi-frame fused point clouds, effectively increasing the point cloud density. Next, we construct a 3D object detection network based on point pillar encoding and utilize an SSD detection head for detection on feature maps. Finally, we validate our method on the VoD dataset. Experimental results demonstrate that our proposed framework outperforms comparative methods, and the MLS-based point cloud densification method significantly enhances the object detection performance.https://www.mdpi.com/2072-4292/17/8/1465autonomous drivingmmWave radarobject detectionpoint cloud |
| spellingShingle | Weigang Shi Panpan Tong Xin Bi Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar Remote Sensing autonomous driving mmWave radar object detection point cloud |
| title | Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar |
| title_full | Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar |
| title_fullStr | Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar |
| title_full_unstemmed | Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar |
| title_short | Moving-Least-Squares-Enhanced 3D Object Detection for 4D Millimeter-Wave Radar |
| title_sort | moving least squares enhanced 3d object detection for 4d millimeter wave radar |
| topic | autonomous driving mmWave radar object detection point cloud |
| url | https://www.mdpi.com/2072-4292/17/8/1465 |
| work_keys_str_mv | AT weigangshi movingleastsquaresenhanced3dobjectdetectionfor4dmillimeterwaveradar AT panpantong movingleastsquaresenhanced3dobjectdetectionfor4dmillimeterwaveradar AT xinbi movingleastsquaresenhanced3dobjectdetectionfor4dmillimeterwaveradar |