Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region
Lane detection severs as one of the pivotal techniques to promote the development of local navigation and HD Map building of autonomous driving. However, lane detection remains an unresolved problem for the challenge of detection accuracy in diverse driving scenarios and computational limitation in...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8854974/ |
| _version_ | 1818642812872163328 |
|---|---|
| author | Weiwei Zhang Zeyang Mi Yaocheng Zheng Qiaoming Gao Wenjing Li |
| author_facet | Weiwei Zhang Zeyang Mi Yaocheng Zheng Qiaoming Gao Wenjing Li |
| author_sort | Weiwei Zhang |
| collection | DOAJ |
| description | Lane detection severs as one of the pivotal techniques to promote the development of local navigation and HD Map building of autonomous driving. However, lane detection remains an unresolved problem for the challenge of detection accuracy in diverse driving scenarios and computational limitation in on-board devices, let alone other road guidance markings. In this paper, we go beyond aforementioned limitations and propose a segmentation-by-detection method for road marking extraction. The architecture of this method consists of three modules: pre-processing, road marking detection and segmentation. In the pre-processing stage, image enhancement operation is used to highlight the contrast especially between road markings and road background. To reduce the computational complexity, the road region will be cropped by vanishing point detection algorithm in this module. Then, a lightweight network is dedicated designed for road marking detection. In order to enhance the network sensitivity to road markings and improve the detection accuracy, we further incorporate a Siamese attention module by integrating with the channel and spatial maps into the network. In the segmentation module, different from the method of semantic segmentation by neural network, our segmentation method is mainly based on conventional image morphological algorithms, which is less computational and also can achieve pixel-level accuracy. Additionally, the sliding search box and maximum stable external region (MSER) algorithms are utilized to compensate for missed detection and position error of bounding boxes. In the experiments, our proposed method delivers outstanding performances on cross datasets and achieves the real-time speed on the embedded devices. |
| first_indexed | 2024-12-16T23:49:01Z |
| format | Article |
| id | doaj.art-0ea24eeb597045b5a2aec68f3c7a3bb9 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-16T23:49:01Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-0ea24eeb597045b5a2aec68f3c7a3bb92022-12-21T22:11:23ZengIEEEIEEE Access2169-35362019-01-01714371014372010.1109/ACCESS.2019.29449938854974Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External RegionWeiwei Zhang0Zeyang Mi1https://orcid.org/0000-0003-1530-9708Yaocheng Zheng2Qiaoming Gao3Wenjing Li4School of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Mechanical and Transportation Engineering, Guangxi University of Science and Technology, Liuzhou, ChinaSchool of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, Shanghai, ChinaLane detection severs as one of the pivotal techniques to promote the development of local navigation and HD Map building of autonomous driving. However, lane detection remains an unresolved problem for the challenge of detection accuracy in diverse driving scenarios and computational limitation in on-board devices, let alone other road guidance markings. In this paper, we go beyond aforementioned limitations and propose a segmentation-by-detection method for road marking extraction. The architecture of this method consists of three modules: pre-processing, road marking detection and segmentation. In the pre-processing stage, image enhancement operation is used to highlight the contrast especially between road markings and road background. To reduce the computational complexity, the road region will be cropped by vanishing point detection algorithm in this module. Then, a lightweight network is dedicated designed for road marking detection. In order to enhance the network sensitivity to road markings and improve the detection accuracy, we further incorporate a Siamese attention module by integrating with the channel and spatial maps into the network. In the segmentation module, different from the method of semantic segmentation by neural network, our segmentation method is mainly based on conventional image morphological algorithms, which is less computational and also can achieve pixel-level accuracy. Additionally, the sliding search box and maximum stable external region (MSER) algorithms are utilized to compensate for missed detection and position error of bounding boxes. In the experiments, our proposed method delivers outstanding performances on cross datasets and achieves the real-time speed on the embedded devices.https://ieeexplore.ieee.org/document/8854974/Lane detectionsegmentation-by-detectionlightweight networkSiamese attention module |
| spellingShingle | Weiwei Zhang Zeyang Mi Yaocheng Zheng Qiaoming Gao Wenjing Li Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region IEEE Access Lane detection segmentation-by-detection lightweight network Siamese attention module |
| title | Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region |
| title_full | Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region |
| title_fullStr | Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region |
| title_full_unstemmed | Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region |
| title_short | Road Marking Segmentation Based on Siamese Attention Module and Maximum Stable External Region |
| title_sort | road marking segmentation based on siamese attention module and maximum stable external region |
| topic | Lane detection segmentation-by-detection lightweight network Siamese attention module |
| url | https://ieeexplore.ieee.org/document/8854974/ |
| work_keys_str_mv | AT weiweizhang roadmarkingsegmentationbasedonsiameseattentionmoduleandmaximumstableexternalregion AT zeyangmi roadmarkingsegmentationbasedonsiameseattentionmoduleandmaximumstableexternalregion AT yaochengzheng roadmarkingsegmentationbasedonsiameseattentionmoduleandmaximumstableexternalregion AT qiaominggao roadmarkingsegmentationbasedonsiameseattentionmoduleandmaximumstableexternalregion AT wenjingli roadmarkingsegmentationbasedonsiameseattentionmoduleandmaximumstableexternalregion |