LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network
Accurate and robust detection of lane and curb in urban areas is essential to many real-world intelligent vehicle applications. Existing vision-based researches treat lane detection and curb detection separately due to the nature of curb detection problem relying on 3D features, which is not efficie...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/146087 |
| _version_ | 1824454112958218240 |
|---|---|
| author | Zhou, Hui Wang, Han Zhang, Handuo Hasith, Karunasekera |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Zhou, Hui Wang, Han Zhang, Handuo Hasith, Karunasekera |
| author_sort | Zhou, Hui |
| collection | NTU |
| description | Accurate and robust detection of lane and curb in urban areas is essential to many real-world intelligent vehicle applications. Existing vision-based researches treat lane detection and curb detection separately due to the nature of curb detection problem relying on 3D features, which is not efficient when driving in a real world. This paper presents a unified network to incorporate these two tasks together by taking advantage of the powerful feature learning ability brought by deep convolutional neural networks. The resulting unified network provides valuable road boundary information by curb detection even when lane markings are not visible during vehicle navigation. Another significant capability coming with the proposed method is able to accurately differentiate various lane and curb instances with tiny gaps or complex spatial relationships which are thought as the biggest challenge in real driving situations. To achieve this, the network is specially designed to guide lane and curb instances to learnable kernel through pixel grouping. This learnable kernel is capable to handle any number of arbitrary-shaped lanes and curbs no matter which angle the vehicle is heading at. In the end, the presented approach is evaluated on two datasets (BDD100K and self-collected dataset) scoring 32 FPS processing speed. Results are very encouraging with more than 98% F 1 measure for both lane and curb detection on the self-collected dataset. |
| first_indexed | 2025-02-19T03:17:08Z |
| format | Conference Paper |
| id | ntu-10356/146087 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:17:08Z |
| publishDate | 2021 |
| record_format | dspace |
| spelling | ntu-10356/1460872021-01-26T07:13:09Z LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network Zhou, Hui Wang, Han Zhang, Handuo Hasith, Karunasekera School of Electrical and Electronic Engineering 2020 IEEE 16th International Conference on Control, Automation, Robotics and Vision (ICARCV) Engineering::Electrical and electronic engineering Instance Segmentation Road Detection Accurate and robust detection of lane and curb in urban areas is essential to many real-world intelligent vehicle applications. Existing vision-based researches treat lane detection and curb detection separately due to the nature of curb detection problem relying on 3D features, which is not efficient when driving in a real world. This paper presents a unified network to incorporate these two tasks together by taking advantage of the powerful feature learning ability brought by deep convolutional neural networks. The resulting unified network provides valuable road boundary information by curb detection even when lane markings are not visible during vehicle navigation. Another significant capability coming with the proposed method is able to accurately differentiate various lane and curb instances with tiny gaps or complex spatial relationships which are thought as the biggest challenge in real driving situations. To achieve this, the network is specially designed to guide lane and curb instances to learnable kernel through pixel grouping. This learnable kernel is capable to handle any number of arbitrary-shaped lanes and curbs no matter which angle the vehicle is heading at. In the end, the presented approach is evaluated on two datasets (BDD100K and self-collected dataset) scoring 32 FPS processing speed. Results are very encouraging with more than 98% F 1 measure for both lane and curb detection on the self-collected dataset. National Research Foundation (NRF) Accepted version 2021-01-26T07:13:09Z 2021-01-26T07:13:09Z 2020 Conference Paper Zhou, H., Wang, H., Zhang, H., & Hasith, K. (2020). LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network. Proceedings of International Conference on Control, Automation, Robotics and Vision (ICARCV), 184-189. doi:10.1109/ICARCV50220.2020.9305341 978-1-7281-7709-0 978-1-7281-7710-6 https://hdl.handle.net/10356/146087 10.1109/ICARCV50220.2020.9305341 184 189 en MRP1A © 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/ICARCV50220.2020.9305341 application/pdf |
| spellingShingle | Engineering::Electrical and electronic engineering Instance Segmentation Road Detection Zhou, Hui Wang, Han Zhang, Handuo Hasith, Karunasekera LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title | LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title_full | LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title_fullStr | LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title_full_unstemmed | LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title_short | LaCNet : real-time end-to-end arbitrary-shaped lane and curb detection with instance segmentation network |
| title_sort | lacnet real time end to end arbitrary shaped lane and curb detection with instance segmentation network |
| topic | Engineering::Electrical and electronic engineering Instance Segmentation Road Detection |
| url | https://hdl.handle.net/10356/146087 |
| work_keys_str_mv | AT zhouhui lacnetrealtimeendtoendarbitraryshapedlaneandcurbdetectionwithinstancesegmentationnetwork AT wanghan lacnetrealtimeendtoendarbitraryshapedlaneandcurbdetectionwithinstancesegmentationnetwork AT zhanghanduo lacnetrealtimeendtoendarbitraryshapedlaneandcurbdetectionwithinstancesegmentationnetwork AT hasithkarunasekera lacnetrealtimeendtoendarbitraryshapedlaneandcurbdetectionwithinstancesegmentationnetwork |