RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations
Lane detection plays an essential part in advanced driver-assistance systems and autonomous driving systems. However, lane detection is affected by many factors such as some challenging traffic situations. Multilane detection is also very important. To solve these problems, we proposed a lane detect...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2021-01-01
|
| Series: | Journal of Advanced Transportation |
| Online Access: | http://dx.doi.org/10.1155/2021/7544355 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832550944495108096 |
|---|---|
| author | Ronghui Zhang Yueying Wu Wanting Gou Junzhou Chen |
| author_facet | Ronghui Zhang Yueying Wu Wanting Gou Junzhou Chen |
| author_sort | Ronghui Zhang |
| collection | DOAJ |
| description | Lane detection plays an essential part in advanced driver-assistance systems and autonomous driving systems. However, lane detection is affected by many factors such as some challenging traffic situations. Multilane detection is also very important. To solve these problems, we proposed a lane detection method based on instance segmentation, named RS-Lane. This method is based on LaneNet and uses Split Attention proposed by ResNeSt to improve the feature representation on slender and sparse annotations like lane markings. We also use Self-Attention Distillation to enhance the feature representation capabilities of the network without adding inference time. RS-Lane can detect lanes without number limits. The tests on TuSimple and CULane datasets show that RS-Lane has achieved comparable results with SOTA and has improved in challenging traffic situations such as no line, dazzle light, and shadow. This research provides a reference for the application of lane detection in autonomous driving and advanced driver-assistance systems. |
| format | Article |
| id | doaj-art-e897afdf25f04e4cab8cabc94e12dd2e |
| institution | Kabale University |
| issn | 0197-6729 2042-3195 |
| language | English |
| publishDate | 2021-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Journal of Advanced Transportation |
| spelling | doaj-art-e897afdf25f04e4cab8cabc94e12dd2e2025-02-03T06:05:34ZengWileyJournal of Advanced Transportation0197-67292042-31952021-01-01202110.1155/2021/75443557544355RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic SituationsRonghui Zhang0Yueying Wu1Wanting Gou2Junzhou Chen3Guangdong Provincial Key Laboratory of Intelligent Transport System, School of Intelligent Systems Engineering, Sun Yat-Sen University, Guangzhou 510275, ChinaGuangdong Provincial Key Laboratory of Intelligent Transport System, School of Intelligent Systems Engineering, Sun Yat-Sen University, Guangzhou 510275, ChinaGuangdong Provincial Key Laboratory of Intelligent Transport System, School of Intelligent Systems Engineering, Sun Yat-Sen University, Guangzhou 510275, ChinaGuangdong Provincial Key Laboratory of Intelligent Transport System, School of Intelligent Systems Engineering, Sun Yat-Sen University, Guangzhou 510275, ChinaLane detection plays an essential part in advanced driver-assistance systems and autonomous driving systems. However, lane detection is affected by many factors such as some challenging traffic situations. Multilane detection is also very important. To solve these problems, we proposed a lane detection method based on instance segmentation, named RS-Lane. This method is based on LaneNet and uses Split Attention proposed by ResNeSt to improve the feature representation on slender and sparse annotations like lane markings. We also use Self-Attention Distillation to enhance the feature representation capabilities of the network without adding inference time. RS-Lane can detect lanes without number limits. The tests on TuSimple and CULane datasets show that RS-Lane has achieved comparable results with SOTA and has improved in challenging traffic situations such as no line, dazzle light, and shadow. This research provides a reference for the application of lane detection in autonomous driving and advanced driver-assistance systems.http://dx.doi.org/10.1155/2021/7544355 |
| spellingShingle | Ronghui Zhang Yueying Wu Wanting Gou Junzhou Chen RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations Journal of Advanced Transportation |
| title | RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations |
| title_full | RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations |
| title_fullStr | RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations |
| title_full_unstemmed | RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations |
| title_short | RS-Lane: A Robust Lane Detection Method Based on ResNeSt and Self-Attention Distillation for Challenging Traffic Situations |
| title_sort | rs lane a robust lane detection method based on resnest and self attention distillation for challenging traffic situations |
| url | http://dx.doi.org/10.1155/2021/7544355 |
| work_keys_str_mv | AT ronghuizhang rslanearobustlanedetectionmethodbasedonresnestandselfattentiondistillationforchallengingtrafficsituations AT yueyingwu rslanearobustlanedetectionmethodbasedonresnestandselfattentiondistillationforchallengingtrafficsituations AT wantinggou rslanearobustlanedetectionmethodbasedonresnestandselfattentiondistillationforchallengingtrafficsituations AT junzhouchen rslanearobustlanedetectionmethodbasedonresnestandselfattentiondistillationforchallengingtrafficsituations |