基于图像的道路语义分割检测方法

doi:10.6040/j.issn.1672-3961.0.2022.365

摘要/Abstract

摘要： 针对现有语义分割网络模型在道路语义分割方面检测精度低、计算量大等问题,基于BiSeNet V2网络模型进行优化改进,引入一种高效的通道注意力(efficient channel attention, ECA)模块,在BiSeNet V2的语义分支和细节分支的每个阶段末端分别加入ECA,得到ECA-Semantic-BiSeNet V2网络。使用实车采集道路图像数据进行标注并构建自采数据集,在Cityscapes数据集、KITTI数据集及自采数据集上分别对改进前后的网络模型进行试验验证。试验结果表明,与BiSeNet V2模型方法相比,本研究方法在Cityscapes数据集上M_IoU提高14.01%,在KITTI数据集上M_IoU提高1.86%,同时在BiSeNet V2的语义分支加入ECA后运算量增加0.02 GFlops的条件下,模型推理速度提高了7.82帧/s。

关键词: 语义分割, 深度学习, BiSeNet V2, 车道线检测, 注意力机制

中图分类号:

TP183

王碧瑶,韩毅,崔航滨,刘毅超,任铭然,高维勇,陈姝廷,刘嘉巍,崔洋. 基于图像的道路语义分割检测方法[J]. 山东大学学报 (工学版), 2023, 53(5): 37-47.

WANG Biyao, HAN Yi, CUI Hangbin, LIU Yichao, REN Mingran, GAO Weiyong, CHEN Shuting, LIU Jiawei, CUI Yang. Road semantic segmentation detection method based on image[J]. Journal of Shandong University(Engineering Science), 2023, 53(5): 37-47.

参考文献

[1] 陈毅, 张帅, 汪贵平. 基于激光雷达和摄像头信息融合的车辆检测算法[J]. 机械与电子, 2020, 38(1): 52-56. CHEN Yi, ZHANG Shuai, WANG Guiping. Vehicle detection algorithm based on information fusion of lidar and camera[J]. Machinery & Electronics, 2020, 38(1): 52-56.
[2] CSURKA G, PERRONNIN F. An efficient approach to semantic segmentation[J]. International Journal of Computer Vision, 2011, 95(2): 198-212.
[3] 王龙飞, 严春满. 道路场景语义分割综述[J]. 激光与光电子学进展, 2021, 58(12): 44-66. WANG Longfei, YAN Chunman. Overview of semantic segmentation of road scenes[J]. Laser & Optoelectronics Progress, 2021, 58(12): 44-66.
[4] 田萱, 王亮, 丁琪. 基于深度学习的图像语义分割方法综述[J]. 软件学报, 2019, 30(2): 440-468. TIAN Xuan, WANG Liang, DING Qi. Review of image semantic segmentation based on deep learning[J]. Journal of Software, 2019, 30(2): 440-468.
[5] 范跃文. 基于复杂环境下的车道线识别[J]. 中国新技术新产品, 2022(7): 28-30. FAN Yuewen. Based on lane line identification in complex environments[J]. China New Technologies and Products, 2022(7): 28-30.
[6] 黄鹤, 孟维明. 基于视觉的大半径圆曲线车道线识别[J]. 测绘通报, 2022(4): 134-137. HUANG He, MENG Weiming. Visual-based lane line identification of large radius circular curves[J]. Bulletin of Surveying and Mapping, 2022(4): 134-137.
[7] 朱鸿宇, 杨帆, 高晓倩, 等. 基于级联霍夫变换的车道线快速检测算法[J]. 计算机技术与发展, 2021, 31(1): 88-93. ZHU Hongyu, YANG Fan, GAO Xiaoqian, et al. A fast lane detection algorithm based on cascade Hough transform[J]. Computer Technology and Development, 2021, 31(1): 88-93.
[8] 李茂月, 吕虹毓, 王飞, 等. 基于机器视觉的智能车辆鲁棒车道线识别方法[J]. 中国机械工程, 2021, 32(2): 242-251. LI Maoyue, LÜ Hongyu, WANG Fei, et al. Robust lane line identification method of intelligent vehicles based on machine vision[J]. China Mechanical Engineering, 2021, 32(2): 242-251.
[9] 王源隆, 井志强, 张峻玮, 等. 智能车辆车道线识别方法研究[J]. 重庆理工大学学报(自然科学), 2022, 36(2): 1-10. WANG Yuanlong, JING Zhiqiang, ZHANG Junwei, et al. Research on intelligent vehicle lane line recognition methods[J]. Journal of Chongqing University of Technology(Natural Science), 2022, 36(2): 1-10.
[10] 田晟, 张剑锋, 张裕天, 等. 基于扩张卷积金字塔网络的车道线检测算法[J]. 西南交通大学学报, 2020, 55(2): 386-392. TIAN Sheng, ZHANG Jianfeng, ZHANG Yutian, et al. Lane detection algorithm based on extended convolutional pyramid network[J]. Journal Center of Southwest Jiaotong University, 2020, 55(2): 386-392.
[11] 周苏, 吴迪, 金杰. 基于卷积神经网络的车道线实例分割算法[J]. 激光与光电子学进展, 2021, 58(8): 381-388. ZHOU Su, WU Di, JIN Jie. Lane segmentation algorithm based on convolutional neural network[J]. Laser & Optoelectronics Progress, 2021, 58(8): 381-388.
[12] LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C] //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE Computer Society, 2015: 431-440.
[13] NEVEN D, DE BRABANDERE B, GEORGOULIS S, et al. Towards end-to-end lane detection: an instance segmentation approach[C] //2018 IEEE Intelligent Vehicles Symposium(Ⅳ). Suzhou, China: IEEE, 2018: 286-291.
[14] 陈正斌, 叶东毅. 带语义分割的轻量化车道线检测算法[J]. 小型微型计算机系统, 2021, 42(9): 1877-1883. CHEN Zhengbin, YE Dongyi. Lightweight lane detection algorithm with semantic segmentation[J]. Journal of Chinese Computer Systems, 2021, 42(9): 1877-1883.
[15] 时小虎, 吴佳琦, 吴春国, 等. 基于残差网络的弯道增强车道线检测方法[J/OL]. 吉林大学学报(工学版).(2022-02-15)[2022-11-26]. https://kns.cnki.net/kcms/detail/22.1341.T.20220214.1736.010.html. SHI Xiaohu, WU Jiaqi, WU Chunguo, et al. Curve-enhanced lane line detection method based on residual network[J/OL]. Journal of Jilin University(Engineering and Technology Edition).(2022-02-15)[2022-10-26]. https://kns.cnki.net/kcms/detail/22.1341.T.20220214.1736.010. html.
[16] 李守彪, 武志斐. 基于多尺寸分解卷积的车道线检测[J]. 汽车技术, 2022(8): 32-37. LI Shoubiao, WU Zhifei. Lane detection based on multi-size decomposition convolution[J]. Automobile Technology, 2022(8): 32-37.
[17] 杨鹏强, 张艳伟, 胡钊政. 基于改进RepVGG网络的车道线检测算法[J]. 交通信息与安全, 2022, 40(2): 73-81. YANG Pengqiang, ZHANG Yanwei, HU Zhaozheng. Lane detection algorithm based on improved RepVGG network[J]. Journal of Transport Information and Safety, 2022, 40(2): 73-81.
[18] 邓天民, 王琳, 杨其芝, 等. 基于改进SegNet算法的车道线检测方法[J]. 科学技术与工程, 2020, 20(36): 14988-14993. DENG Tianmin, WANG Lin, YANG Qizhi, et al. Lane detection method based on improved SegNet algorithm[J]. Science Technology and Engineering, 2020, 20(36): 14988-14993.
[19] 胡杰, 熊宗权, 徐文才, 等. 基于轻量化卷积神经网络的车道线检测[J]. 激光与光电子学进展, 2022, 59(10): 352-360. HU Jie, XIONG Zongquan, XU Wencai, et al. Lane detection based on lightweight convolutional neural network[J]. Laser & Optoelectronics Progress, 2022, 59(10): 352-360.
[20] 张冲, 黄影平, 郭志阳, 等. 基于语义分割的实时车道线检测方法[J]. 光电工程, 2022, 49(5): 26-37. ZHANG Chong, HUANG Yingping, GUO Zhiyang, et al. Real-time lane detection method based on semantic segmentation[J]. Opto-Electronic Engineering, 2022, 49(5): 26-37.
[21] YU C Q, GAO C X, WANG J B, et al. BiSeNet V2: bilateral network with guided aggregation for real-time semantic segmentation[J]. International Journal of Computer Vision, 2021, 129(11): 3051-3068.
[22] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C] //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA: IEEE Computer Society, 2018: 7132-7141.
[23] WANG Q, WU B, ZHU P, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C] //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Virtual, Online, USA: IEEE, 2020: 11531-11539.
[24] GEIGER A, LENZ P, URTASUN R. Are we ready for autonomous driving? The KITTI vision benchmark suite[C] //2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence, USA: IEEE Computer Society, 2012: 3354-3361.
[25] FRITSCH J, KUEHNL T, GEIGER A. A new performance measure and evaluation benchmark for road detection algorithms[C] //16th International IEEE Conference on Intelligent Transportation Systems(ITSC 2013). Hague, Netherlands: IEEE, 2013: 1693-1700.
[26] CORDTS M, OMRAN M, RAMOS S, et al. The cityscapes dataset for semantic urban scene understanding[C] //Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Las Vegas, USA: IEEE Computer Society, 2016: 3213-3223.
[27] 全国交通工程设施(公路)标准化技术委员会. 道路交通标志和标线:第3部分道路交通标线:GB 5768.3—2009[S]. 北京:中国标准出版社, 2009.
[28] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C] //Proceedings of the IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017: 2980-2988.
[29] LI X, SUN X, MENG Y, et al. Dice loss for data-imbalanced NLP tasks[EB/OL].(2020-08-29)[2022-10-26]. https://arxiv.org/abs/1911.02855.
[30] GOYAL P, DOLLÁR P, GIRSHICK R, et al. Accurate, large minibatch SGD:training imagenet in 1 hour[EB/OL].(2018-04-30)[2022-10-26]. https://arxiv.org/abs/1706.02677.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed