Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds
The ability to segment moving objects from three-dimensional (3D) LiDAR scans is critical to advancing autonomous driving technology, facilitating core tasks like localization, collision avoidance, and path planning. In this paper, we introduce a novel deep neural network designed to enhance the per...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/2/195 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832587599658614784 |
|---|---|
| author | Fangzhou Tang Bocheng Zhu Junren Sun |
| author_facet | Fangzhou Tang Bocheng Zhu Junren Sun |
| author_sort | Fangzhou Tang |
| collection | DOAJ |
| description | The ability to segment moving objects from three-dimensional (3D) LiDAR scans is critical to advancing autonomous driving technology, facilitating core tasks like localization, collision avoidance, and path planning. In this paper, we introduce a novel deep neural network designed to enhance the performance of 3D LiDAR point cloud moving object segmentation (MOS) through the integration of image gradient information and the principle of motion consistency. Our method processes sequential range images, employing depth pixel difference convolution (DPDC) to improve the efficacy of dilated convolutions, thus boosting spatial information extraction from range images. Additionally, we incorporate Bayesian filtering to impose posterior constraints on predictions, enhancing the accuracy of motion segmentation. To handle the issue of uneven object scales in range images, we develop a novel edge-aware loss function and use a progressive training strategy to further boost performance. Our method is validated on the SemanticKITTI-based LiDAR MOS benchmark, where it significantly outperforms current state-of-the-art (SOTA) methods, all while working directly on two-dimensional (2D) range images without requiring mapping. |
| format | Article |
| id | doaj-art-cf6310f1fa3e4cde87046cae307f5543 |
| institution | Kabale University |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-cf6310f1fa3e4cde87046cae307f55432025-01-24T13:47:41ZengMDPI AGRemote Sensing2072-42922025-01-0117219510.3390/rs17020195Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point CloudsFangzhou Tang0Bocheng Zhu1Junren Sun2School of Electronics, Peking University, Beijing 100871, ChinaSchool of Electronics, Peking University, Beijing 100871, ChinaSchool of Electronics, Peking University, Beijing 100871, ChinaThe ability to segment moving objects from three-dimensional (3D) LiDAR scans is critical to advancing autonomous driving technology, facilitating core tasks like localization, collision avoidance, and path planning. In this paper, we introduce a novel deep neural network designed to enhance the performance of 3D LiDAR point cloud moving object segmentation (MOS) through the integration of image gradient information and the principle of motion consistency. Our method processes sequential range images, employing depth pixel difference convolution (DPDC) to improve the efficacy of dilated convolutions, thus boosting spatial information extraction from range images. Additionally, we incorporate Bayesian filtering to impose posterior constraints on predictions, enhancing the accuracy of motion segmentation. To handle the issue of uneven object scales in range images, we develop a novel edge-aware loss function and use a progressive training strategy to further boost performance. Our method is validated on the SemanticKITTI-based LiDAR MOS benchmark, where it significantly outperforms current state-of-the-art (SOTA) methods, all while working directly on two-dimensional (2D) range images without requiring mapping.https://www.mdpi.com/2072-4292/17/2/195LiDAR point cloudmoving object segmentationrange imagegradient enhancementmotion consistency |
| spellingShingle | Fangzhou Tang Bocheng Zhu Junren Sun Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds Remote Sensing LiDAR point cloud moving object segmentation range image gradient enhancement motion consistency |
| title | Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds |
| title_full | Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds |
| title_fullStr | Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds |
| title_full_unstemmed | Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds |
| title_short | Gradient Enhancement Techniques and Motion Consistency Constraints for Moving Object Segmentation in 3D LiDAR Point Clouds |
| title_sort | gradient enhancement techniques and motion consistency constraints for moving object segmentation in 3d lidar point clouds |
| topic | LiDAR point cloud moving object segmentation range image gradient enhancement motion consistency |
| url | https://www.mdpi.com/2072-4292/17/2/195 |
| work_keys_str_mv | AT fangzhoutang gradientenhancementtechniquesandmotionconsistencyconstraintsformovingobjectsegmentationin3dlidarpointclouds AT bochengzhu gradientenhancementtechniquesandmotionconsistencyconstraintsformovingobjectsegmentationin3dlidarpointclouds AT junrensun gradientenhancementtechniquesandmotionconsistencyconstraintsformovingobjectsegmentationin3dlidarpointclouds |