Lane Detection Algorithm Based on Instance Segmentation

被引：0
|
作者：
Wu Z. [1 ]
Li S. [1 ]
机构：
[1] College of Mechanical and Vehicle Engineering, Taiyuan University of Technology, Taiyuan
来源：
Qiche Gongcheng/Automotive Engineering | 2023年 / 45卷 / 02期
关键词：
clustering; instance segmentation; lane detection; vehicle active safety;
D O I：
10.19562/j.chinasae.qcgc.2023.02.011
中图分类号：
学科分类号：
摘要：
To realize the detection of lane lines with changing numbers in complex scenarios of autonomous driving，a lane detection algorithm based on instance segmentation is proposed. Firstly，the ResNet18 network is used as the backbone network to extract image features，and the feature pyramid network is used for feature fusion，at the same time a dilated convolutional residual module is designed to improve the detection accuracy. Then instance segmentation is carried out based on the location of lane，and the corresponding clustering point locations is predicted using the lane point locations from the semantic segmentation. Finally，the DBSCAN clustering algorithm is used for the cluster points to realize the lane instance distinction. The results show that the algorithm can effectively detect multi-lane lines in complex and changeable autonomous driving scenarios，and the F1 indicators on the CULane dataset and TuSimple dataset reach 75.2% and 97.0%，respectively. © 2023 SAE-China. All rights reserved.
引用
收藏
页码：263 / 272
页数：9
相关论文
共 22 条
  • [1] WU Y Q, LIU L., Research and development of the vision-based lane detection methods［J］, Chinese Journal of Scientific Instrument, 40, 12, pp. 92-109, (2019)
  • [2] LIANG D, ZHANG S K，, Et al., Lane detection：a survey with new results［J］, Journal of Computer Science and Technology, 35, 3, pp. 493-505, (2020)
  • [3] SUN T Y, CHAN V., HSI color model based lane-marking detection［C］, 2006 IEEE Intelligent Transportation Systems Conference, pp. 1168-1172, (2006)
  • [4] CHEN T, ZHANG H D，, CHEN D，, Et al., Lane detection based on high priority pixels and tracking by Kalman filter［J］, Automotive Engineering, 38, 2, pp. 200-205, (2016)
  • [5] JIN Z L，, HE L X, ZHAO W Z., Detection and tracking method of lane line for intelligent vehicles under complex illumination condition［J］, Journal of Automotive Safety and Energy, 10, 4, pp. 459-466, (2019)
  • [6] WU H Y, ZHAO X M., Multi-interference lane recognition based on IPM and edge image filtering［J］, China Journal of Highway and Transport, 33, 5, pp. 153-164, (2020)
  • [7] ZHANG J M, QIAN L J, Et al., A multiple lane detection algorithm［J］, Journal of Hefei University of Technology（Natural Science）, 43, 4, pp. 536-542, (2020)
  • [8] CHEN W W, JIANG Y T, TAN D K., A fast lane marking recognition algorithm based on edge projection［J］, Automotive Engineering, 39, 3, pp. 357-363, (2017)
  • [9] Spatial as deep：spatial CNN for traffic scene understanding［C］, Proceedings of the AAAI Conference on Artificial Intelligence, pp. 7276-7283, (2018)
  • [10] Learning lightweight lane detection CNNs by self attention distillation［C］, Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1013-1021, (2019)
123