面向自动驾驶的轻量级道路场景语义分割

doi:10.3778/j.issn.1002-8331.2206-0433

摘要/Abstract

摘要： 自动驾驶领域中，现有的道路场景语义分割算法开销巨大，无法满足自动驾驶的实时性。基于DeepLabV3+的整体结构，提出了一种并行特征处理的轻量级图像语义分割模型，兼顾了高精度和实时性。采用MobileNetV2作为主干网络，精简上采样过程，提升分割速度，并减少网络参数量，以便于网络迁移和训练；引入双注意力机制，与空洞卷积空间金字塔模块结合组成并行特征处理结构，提高分割精度；最后，将MobileNetV2与该并行特征处理结构相结合，以完成对图像特征的提取。实验结果表明，相比于传统模型，所提出模型能以少量的系统开销和网络参数量保证高效且精准的图像分割。模型在Cityscapes数据集mIoU达到73.61%，处理一张512×512的图片仅需25 ms。

关键词: 语义分割, 自动驾驶, 轻量级, MobileNetV2, 注意力机制

Abstract: In the field of autonomous driving, existing semantic segmentation algorithms for road scenes have huge overhead and cannot meet the real-time performance of autonomous driving. Based on the overall structure of DeepLabV3+, this paper proposes a lightweight image semantic segmentation model with parallel feature processing, which takes into account both high accuracy and real-time performance. Firstly, MobileNetV2 is used as the backbone network to streamline the upsampling process and reduce the number of network parameters for network migration and training. Then, a dual-attention mechanism is introduced to combine with the null convolutional space pyramid module to form a parallel feature processing structure to improve segmentation accuracy. Finally, the parallel feature processing structure is then combined with MobileNetV2 to complete the extraction of image features. The experimental results show that the proposed model can guarantee efficient and accurate image segmentation with less system overhead and number of network parameters than the traditional model. The model achieves 73.61% mIoU in the Cityscapes dataset and processes a 512×512 image in 25 ms.

Key words: semantic segmentation, autopilot, lightweight, MobileNetV2, attention mechanism

李顺新, 吴桐. 面向自动驾驶的轻量级道路场景语义分割[J]. 计算机工程与应用, 2023, 59(19): 177-183.

LI Shunxin, WU Tong. Lightweight Semantic Segmentation of Road Scenes for Autonomous Driving[J]. Computer Engineering and Applications, 2023, 59(19): 177-183.

参考文献

[1] BAI Y Q，ZHENG Y F，TIAN H.Semantic segmentation method of road scene based on Deeplabv3+ and attention mechanism[J].Journal of Measurement Science & Instrumentation，2021，12（4）：412-422.
[2] 胡云卿，潘文波，侯志超，等.面向街景自动驾驶的DSC-MB-PSPNet语义分割技术研究[J].控制与信息技术，2020（4）：1-9.
HU Y Q，PAN W B，HOU Z C，et al.Research on semantic segmentation technology of DSC-MB-PSPNet for street view autonomous driving[J].Control and Information Technology，2020（4）：1-9.
[3] 李彦辰，张小俊，张明路，等.基于改进Efficientdet的自动驾驶场景目标检测[J].计算机工程与应用，2022，58（6）：183-191.
LI Y C，ZHANG X J，ZHANG M L，et al.Object detection in autonomous driving scene based on improved efficientdet[J].Computer Engineering and Applications，2022，58（6）：183-191.
[4] 王雨.面向自动驾驶场景的高效实时语义分割方法研究[D].南京：南京邮电大学，2020.
WANG Y.Research on efficient real-time semantic segmentation method for autonomous driving scenarios[D].Nanjing：Nanjing University of Posts and Telecommunications，2020.
[5] 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 X，WANG L，DING Q.A Review of image semantic segmentation methods based on deep learning[J].Journal of Software，2019，30（2）：440-468.
[6] 何淼楹，崔宇超.面向自动驾驶的交通景语义分割[J].计算机应用，2021，41（S1）：25-30.
HE M Y，CUI Y C.Semantic segmentation of traffic scenarios for autonomous driving[J].Journal of Computer Applications，2021，41（S1）：25-30.
[7] 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 X，WANG L，DING Q.Review of image semantic segmentation based on deep learning[J].Journal of software，2019，30（2）：440-468.
[8] BADRINARAYANAN V，KENDALL A，CIPOLLA R.SegNet：a deep convolutional encoder-decoder architecture for image segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（12）：2481-2495.
[9] RONNBERGER O，FISCHER P，BROX T.U-Net：convolutional networks for biomedical image segmentation[C]//Proceedings of International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham：Springer，2015：234-241.
[10] CHEN L C，PAPANDREOU G，KOKKINOS I，et al.DeepLab：semantic image segmentation with deep convolutional nets，atrous convolution，and fully connected CRFS[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，40（4）：834-848.
[11] LIU C，CHEN L C，SCHROFF F，et al.Auto-deepLab：hierarchical neural architecture search for semantic image segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：82-92.
[12] WANG Y，GAO L，HONG D，et al.Mask DeepLab：end-to-end image segmentation for change detection in high-resolution remote sensing images[J].International Journal of Applied Earth Observation and Geoinformation，2021，104：102582.
[13] CHEN J，WANG C，TONG Y.AtICNet：semantic segmentation with atrous spatial pyramid pooling in image cascade network[J].EURASIP Journal on Wireless Communications and Networking，2019（1）：1-7.
[14] WANG Y，ZHANG J，KAN M，et al.Self-supervised equivariant attention mechanism for weakly supervised semantic segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：12275-12284.
[15] FU J，LIU J，TIAN H，et al.Dual attention network for scene segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：3146-3154.
[16] 张汉，张德祥，陈鹏，等.并行注意力机制在图像语义分割中的应用[J].计算机工程与应用，2022，58（9）：151-160.
ZHANG H，ZHANG D X，CHEN P，et al.Application of parallel attention mechanism in image semantic segmentation[J].Computer Engineering and Applications，2022，58（9）：151-160.
[17] 孟琭，徐磊，郭嘉阳.一种基于改进的 MobileNetV2 网络语义分割算法[J].电子学报，2020，48（9）：1769.
MENG L，XU L，GUO J Y.A MobileNetV2 network semantic segmentation algorithm based on improvement[J].Chinese Journal of Electronics，2020，48（9）：1769.
[18] SANDLER M，HOWARD A，ZHU M，et al.Mobilenetv2：inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：4510-4520.
[19] 马书浩，安居白，于博.改进DeepLabv2的实时图像语义分割算法[J].计算机工程与应用，2020，56（18）：157-164.
MA S H，AN J B，YU B.Improved DeepLabv2 real-time image semantic segmentation algorithm[J].Computer Engineering and Applications，2020，56（18）：157-164.
[20] 王欣，王美丽，边党伟.融合MobileNetv2和注意力机制的轻量级人像分割算法[J].计算机工程与应用，2022，58（7）：220-228.
WANG X，WANG M L，BIAN D W.Algorithm for Portrait segmentation combined with MobileNetv2 and attention mechanism[J].Computer Engineering and Applications，2022，58（7）：220-228
[21] ZHAO H，SHI J，QI X，et al.Pyramid scene parsing network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2017：6230-6239.