Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation
The intelligent extraction of highway assets is pivotal for advancing transportation infrastructure and autonomous systems, yet traditional methods relying on manual inspection or 2D imaging struggle with sparse, occluded environments, and class imbalance. This study proposes an enhanced MinkUNet-ba...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/13/2147 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849427510792552448 |
|---|---|
| author | Dalong Zhou Yuanyang Yi Yu Wang Zhenfeng Shao Yanjun Hao Yuyan Yan Xiaojin Zhao Junkai Guo |
| author_facet | Dalong Zhou Yuanyang Yi Yu Wang Zhenfeng Shao Yanjun Hao Yuyan Yan Xiaojin Zhao Junkai Guo |
| author_sort | Dalong Zhou |
| collection | DOAJ |
| description | The intelligent extraction of highway assets is pivotal for advancing transportation infrastructure and autonomous systems, yet traditional methods relying on manual inspection or 2D imaging struggle with sparse, occluded environments, and class imbalance. This study proposes an enhanced MinkUNet-based framework to address data scarcity, occlusion, and imbalance in highway point cloud segmentation. A large-scale dataset (PEA-PC Dataset) was constructed, covering six key asset categories, addressing the lack of specialized highway datasets. A hybrid conical masking augmentation strategy was designed to simulate natural occlusions and enhance local feature retention, while semi-supervised learning prioritized foreground differentiation. The experimental results showed that the overall mIoU reached 73.8%, with the IoU of bridge railings and emergency obstacles exceeding 95%. The IoU of columnar assets increased from 2.6% to 29.4% through occlusion perception enhancement, demonstrating the effectiveness of this method in improving object recognition accuracy. The framework balances computational efficiency and robustness, offering a scalable solution for sparse highway scenes. However, challenges remain in segmenting vegetation-occluded pole-like assets due to partial data loss. This work highlights the efficacy of tailored augmentation and semi-supervised strategies in refining 3D segmentation, advancing applications in intelligent transportation and digital infrastructure. |
| format | Article |
| id | doaj-art-0fce671f20bd482895e2111758cfbcf0 |
| institution | Kabale University |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-0fce671f20bd482895e2111758cfbcf02025-08-20T03:29:00ZengMDPI AGRemote Sensing2072-42922025-06-011713214710.3390/rs17132147Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud SegmentationDalong Zhou0Yuanyang Yi1Yu Wang2Zhenfeng Shao3Yanjun Hao4Yuyan Yan5Xiaojin Zhao6Junkai Guo7State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, ChinaShanxi Intelligent Transportation Laboratory Co., Ltd., Shanxi Transportation Research Institute Group Co., Ltd., Taiyuan 030032, ChinaSchool of Remote Sensing Information Engineering, Wuhan University, Wuhan 430072, ChinaShanxi Intelligent Transportation Laboratory Co., Ltd., Shanxi Transportation Research Institute Group Co., Ltd., Taiyuan 030032, ChinaShanxi Intelligent Transportation Laboratory Co., Ltd., Shanxi Transportation Research Institute Group Co., Ltd., Taiyuan 030032, ChinaThe intelligent extraction of highway assets is pivotal for advancing transportation infrastructure and autonomous systems, yet traditional methods relying on manual inspection or 2D imaging struggle with sparse, occluded environments, and class imbalance. This study proposes an enhanced MinkUNet-based framework to address data scarcity, occlusion, and imbalance in highway point cloud segmentation. A large-scale dataset (PEA-PC Dataset) was constructed, covering six key asset categories, addressing the lack of specialized highway datasets. A hybrid conical masking augmentation strategy was designed to simulate natural occlusions and enhance local feature retention, while semi-supervised learning prioritized foreground differentiation. The experimental results showed that the overall mIoU reached 73.8%, with the IoU of bridge railings and emergency obstacles exceeding 95%. The IoU of columnar assets increased from 2.6% to 29.4% through occlusion perception enhancement, demonstrating the effectiveness of this method in improving object recognition accuracy. The framework balances computational efficiency and robustness, offering a scalable solution for sparse highway scenes. However, challenges remain in segmenting vegetation-occluded pole-like assets due to partial data loss. This work highlights the efficacy of tailored augmentation and semi-supervised strategies in refining 3D segmentation, advancing applications in intelligent transportation and digital infrastructure.https://www.mdpi.com/2072-4292/17/13/2147point cloud segmentationhighway asset extractiondata augmentationMinkUNetsemi-supervised learning |
| spellingShingle | Dalong Zhou Yuanyang Yi Yu Wang Zhenfeng Shao Yanjun Hao Yuyan Yan Xiaojin Zhao Junkai Guo Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation Remote Sensing point cloud segmentation highway asset extraction data augmentation MinkUNet semi-supervised learning |
| title | Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation |
| title_full | Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation |
| title_fullStr | Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation |
| title_full_unstemmed | Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation |
| title_short | Enhancing Highway Scene Understanding: A Novel Data Augmentation Approach for Vehicle-Mounted LiDAR Point Cloud Segmentation |
| title_sort | enhancing highway scene understanding a novel data augmentation approach for vehicle mounted lidar point cloud segmentation |
| topic | point cloud segmentation highway asset extraction data augmentation MinkUNet semi-supervised learning |
| url | https://www.mdpi.com/2072-4292/17/13/2147 |
| work_keys_str_mv | AT dalongzhou enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT yuanyangyi enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT yuwang enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT zhenfengshao enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT yanjunhao enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT yuyanyan enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT xiaojinzhao enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation AT junkaiguo enhancinghighwaysceneunderstandinganoveldataaugmentationapproachforvehiclemountedlidarpointcloudsegmentation |