面向非结构化道路的轻量化轴向上下文分割网络

doi:10.19678/j.issn.1000-3428.0253402

计算机工程 ›› 2026, Vol. 52 ›› Issue (8): 33-45. doi: 10.19678/j.issn.1000-3428.0253402

所属专题：人工智能赋能建材建工、汽车、交通及相关产业创新专题；

• 人工智能赋能建材建工、汽车、交通及相关产业创新 • 上一篇下一篇

面向非结构化道路的轻量化轴向上下文分割网络

曹福¹, 邢雯彬¹, 左勇², 张荣辉¹, 陈俊周¹^,*()

1. 中山大学智能工程学院, 广东广州 510275
2. 武汉理工大学汽车学院, 湖北武汉 430070

收稿日期:2025-12-08 修回日期:2026-03-17 出版日期:2026-08-15 发布日期:2026-04-14
通讯作者: 陈俊周
作者简介:
曹福, 男, 硕士研究生, 主研方向为计算机视觉
邢雯彬, 硕士研究生
左勇, 博士研究生
张荣辉, 副教授
陈俊周(通信作者), 副教授
基金资助:
国家自然科学基金(52172350); 国家自然科学基金(W2421069); 广州市重点研发计划(2024B01W0079); 深圳市科技攻关面上项目(JSGG20210802153412036)

Lightweight Axial Context Segmentation Network for Unstructured Road

CAO Fu¹, XING Wenbin¹, ZUO Yong², ZHANG Ronghui¹, CHEN Junzhou¹^,*()

1. School of Intelligent Systems Engineering, Sun Yat-sen University, Guangzhou 510275, Guangdong, China
2. School of Automotive Engineering, Wuhan University of Technology, Wuhan 430070, Hubei, China

Received:2025-12-08 Revised:2026-03-17 Online:2026-08-15 Published:2026-04-14
Contact: CHEN Junzhou

摘要/Abstract

摘要：

非结构化道路分割是自动驾驶技术环境感知的重要组成部分, 面临全局拓扑建模不完整、边界细节难以保持, 及模型效率与精度的权衡等挑战。针对这些挑战, 设计了一种轻量化轴向上下文网络AXON-Net。该网络采用编码器-解码器架构, 在编码器中引入通道-空间注意力模块(CASAB), 通过聚合多维统计信息自适应重标定特征权重, 有效抑制环境噪声, 以增强复杂背景下的特征区分度; 在瓶颈层设计轻量化部分上下文(LightPCT)模块, 利用部分通道交互策略降低计算冗余, 高效捕获长程依赖关系以修复道路拓扑连通性。在解码器中集成双路径通道融合(DPCF)与轴向细结构增强(TSE)模块, 旨在缩小特征语义鸿沟并显式强化轴向几何特征, 改善模糊道路边缘的精细化恢复效果。在基于印度驾驶数据集(IDD)与越野空间检测(ORFD) 数据集二次构建的非结构化道路数据集上的实验结果表明, AXON-Net在道路交并比(IoU_Road)分别达到95.3%、88.1%, 参数量仅为8.49×10³, 实现了分割精度与模型效率的较优平衡。消融实验验证了各模块协同作用的有效性, 展示了该网络在非结构化道路感知任务中的应用潜力。

关键词: 非结构化道路, 轴向上下文, 通道-空间注意力, 细结构增强, 长程依赖

Abstract:

Unstructured road segmentation is a crucial component of environmental perception for autonomous driving systems, and it suffers from challenges such as the integrity of global topological modeling, preservation of boundary details, and trade-off between model efficiency and accuracy. To address these challenges, this paper proposes a lightweight axial context network AXON-Net. The network employs an encoder—decoder architecture and introduces a Channel And Spatial Attention Block (CASAB) in the encoder, which adaptively recalibrates feature weights by aggregating multidimensional statistical information to effectively suppress environmental noise, thereby enhancing feature discriminability in complex backgrounds. A Lightweight Partial Context (LightPCT) model is integrated at the bottleneck, utilizing a partial channel interaction strategy to reduce computational redundancy and efficiently capture long-range dependencies for restoring road topological connectivity. In addition, the decoder integrates Dual-Path Channel Fusion (DPCF) and Thin Structure Enhancer (TSE) modules for bridging the feature semantic gap and explicitly enhancing axial geometric features for the refined recovery of blurred road edges. Experimental results on unstructured road datasets constructed from the India Driving Dataset (IDD) and Off-Road Freespace Detection (ORFD) dataset show that AXON-Net achieves road Intersection over Union (IoU_Road) scores of 95.3% and 88.1%, respectively. With only 8.49×10³ parameters, it achieves a superior balance between segmentation accuracy and model efficiency. Ablation studies further validate the synergistic effectiveness of the proposed modules, demonstrating the potential application of the network for unstructured road perception tasks.

Key words: unstructured road, axial context, channel and spatial attention, Thin Structure Enhancer (TSE), long-range dependencies

曹福, 邢雯彬, 左勇, 张荣辉, 陈俊周. 面向非结构化道路的轻量化轴向上下文分割网络[J]. 计算机工程, 2026, 52(8): 33-45.

CAO Fu, XING Wenbin, ZUO Yong, ZHANG Ronghui, CHEN Junzhou. Lightweight Axial Context Segmentation Network for Unstructured Road[J]. Computer Engineering, 2026, 52(8): 33-45.

https://www.ecice06.com/CN/Y2026/V52/I8/33

图/表 14

图1 AXON-Net模型总架构

Fig.1 Overall architecture of AXON-Net model

图2 特征编码器及CASAB模块示意图

Fig.2 Schematic diagram of the feature encoder and CASAB module

图3 上下文瓶颈模块LightPCT结构

Fig.3 Structure of the context bottleneck module LightPCT

图4 结构细化解码器示意图

Fig.4 Schematic diagram of the structure refinement decoder

图5 保留样本与剔除样本对比

Fig.5 Comparison between retained samples and excluded samples

图6 消融实验模块激活热力图

Fig.6 Activation heatmaps for the ablation experimental module

图7 不同模型分割结果可视化对比

Fig.7 Visualization comparison of segmentation results from different models

参考文献 31

1	YU F, CHEN H, WANG X, et al. BDD100K: a diverse driving dataset for heterogeneous multitask learning[EB/OL]. [2025-11-01]. https://arxiv.org/pdf/1805.04687.
2	CORDTS M, OMRAN M, RAMOS S, et al. The Cityscapes dataset for semantic urban scene understanding[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2016: 3213-3223.
3	林畅, 郭伟, 任哲聪, 等. 基于Transformer的目标跟踪与分割统一算法. 计算机工程, 2024, 50(9): 130- 141. doi: 10.19678/j.issn.1000-3428.0068414
	LIN C , GUO W , REN Z C , et al. Unification algorithm for object tracking and segmentation based on Transformer. Computer Engineering, 2024, 50(9): 130- 141. doi: 10.19678/j.issn.1000-3428.0068414
4	叶宝林, 孙瑞涛, 李灵犀, 等. 基于融合状态预测的深度强化学习A2C的交通信号控制. 计算机工程, 2025, 51(5): 33- 42. doi: 10.19678/j.issn.1000-3428.0069478
	YE B L , SUN R T , LI L X , et al. Traffic signal control based on deep reinforcement learning A2C integrated with state prediction. Computer Engineering, 2025, 51(5): 33- 42. doi: 10.19678/j.issn.1000-3428.0069478
5	朱思远, 李佳圣, 邹丹平, 等. 基于半监督学习的非结构化道路缺陷检测算法. 计算机工程, 2025, 51(9): 14- 24. doi: 10.19678/j.issn.1000-3428.0069534
	ZHU S Y , LI J S , ZOU D P , et al. Unstructured road defect detection algorithm based on semi-supervised learning. Computer Engineering, 2025, 51(9): 14- 24. doi: 10.19678/j.issn.1000-3428.0069534
6	LIU Z Y , YU S Y , ZHENG N N . A co-point mapping-based approach to drivable area detection for self-driving cars. Engineering, 2018, 4(4): 479- 490. doi: 10.1016/j.eng.2018.07.010
7	HSU C M, LIAN F L, HUANG C M, et al. Detecting drivable space in traffic scene understanding[C]//Proceedings of International Conference on System Science and Engineering (ICSSE). Washington D.C., USA: IEEE Press, 2012: 79-84.
8	FERNÁNDEZ C , FERNÁNDEZ-LLORCA D , SOTELO M A . A hybrid vision-map method for urban road detection. Journal of Advanced Transportation, 2017(1): 7090549.
9	SLESS L, SHLOMO B E, COHEN G, et al. Road scene understanding by occupancy grid learning from sparse radar clusters using semantic segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops. Washington D.C., USA: IEEE Press, 2019: 101-106.
10	MA B , LAKSHMANAN S , HERO A O . Simultaneous detection of lane and pavement boundaries using model-based multisensor fusion. IEEE Transactions on Intelligent Transportation Systems, 2002, 1(3): 135- 147.
11	沙宇洋, 陆京涛, 杜浩凡, 等. 适用于导盲场景的多尺度特征融合轻量化道路图像分割算法. 计算机工程, 2025, 51(7): 314- 325. doi: 10.19678/j.issn.1000-3428.0068674
	SHA Y Y , LU J T , DU H F , et al. Lightweight road image segmentation algorithm with multi-scale feature fusion for blind guidance scenarios. Computer Engineering, 2025, 51(7): 314- 325. doi: 10.19678/j.issn.1000-3428.0068674
12	金汝宁, 赵波, 李洪平. 一种轻量化非结构化道路语义分割神经网络. 四川大学学报(自然科学版), 2023, 60(1): 66- 73.
	JIN R N , ZHAO B , LI H P . A lightweight neural network for semantic segmentation of unstructured roads. Journal of Sichuan University (Natural Science Edition), 2023, 60(1): 66- 73.
13	SHELHAMER E , LONG J , DARRELL T . Fully convolutional networks for semantic segmentation. IEEE Transactions on Intelligent Transportation Systems, 2017, 39(4): 640- 651.
14	RONNEBERGER 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. Berlin, Germany: Springer, 2015: 234-241.
15	栾庆磊, 毛宜东, 陈梓华, 等. 基于YOLOv8-AFDP的红外道路目标检测. 光电工程, 2025, 52(8): 250079.
	LUAN Q L , MAO Y D , CHEN Z H , et al. Infrared road target detection based on YOLOv8-AFDP. Opto-Electronic Engineering, 2025, 52(8): 250079.
16	SINHA D, EL-SHARKAWY M. Thin MobileNet: an enhanced MobileNet architecture[C]// Proceedings of the 10th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON). Washington D.C., USA: IEEE Press, 2019: 280-285.
17	ZHAO H S, SHI J P, QI X J, et al. Pyramid scene parsing network[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2017: 2881-2890.
18	安玉, 葛海波, 何文昊, 等. 基于补偿注意力机制的Siamese网络跟踪算法. 计算机工程, 2024, 50(4): 187- 196. doi: 10.19678/j.issn.1000-3428.0067601
	AN Y , GE H B , HE W H , et al. Siamese network tracking algorithm based on compensated attention mechanism. Computer Engineering, 2024, 50(4): 187- 196. doi: 10.19678/j.issn.1000-3428.0067601
19	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2017: 5998-6008.
20	VARMA G, SUBRAMANIAN A, NAMBOODIRI A, et al. IDD: a dataset for exploring problems of autonomous navigation in unconstrained environments[C]//Proceedings of IEEE Winter Conference on Applications of Computer Vision (WACV). Washington D.C., USA: IEEE Press, 2019: 1743-1751.
21	MIN C, JIANG W Z, ZHAO D W, et al. ORFD: a dataset and benchmark for off-road freespace detection[C]//Proceedings of International Conference on Robotics and Automation (ICRA). Washington D.C., USA: IEEE Press, 2022: 6542-6548.
22	BADRINARAYANAN V , KENDALL A , CIPOLLA R . SegNet: a deep convolutional encoder-decoder architecture for image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481- 2495. doi: 10.1109/TPAMI.2016.2644615
23	WU T Y , TANG S , ZHANG R , et al. CGNet: a light-weight context guided network for semantic segmentation. IEEE Transactions on Image Processing, 2020, 30, 1169- 1179.
24	ROMERA E , ALVAREZ J M , BERGASA L M , et al. ERFNet: efficient residual factorized ConvNet for real-time semantic segmentation. IEEE Transactions on Intelligent Transportation Systems, 2017, 19(1): 263- 272.
25	SHI M , LIN S , YI Q , et al. Lightweight context-aware network using partial-channel transformation for real-time semantic segmentation. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 7401- 7416. doi: 10.1109/TITS.2023.3348631
26	HUANG P M, CHAO I T, HUANG P C, et al. BEVANet: bilateral efficient visual attention network for real-time semantic segmentation[C]//Proceedings of IEEE International Conference on Image Processing (ICIP). Washington D.C., USA: IEEE Press, 2025: 2778-2783.
27	CHAURASIA A, CULURCIELLO E. LinkNet: exploiting encoder representations for efficient semantic segmentation[EB/OL]. [2025-11-01]. https://arxiv.org/pdf/1707.03718.
28	LO S Y, HANG H M, CHAN S W, et al. Efficient dense modules of asymmetric convolution for real-time semantic segmentation[EB/OL]. [2025-11-01]. https://dl.acm.org/doi/epdf/10.1145/3338533.3366558.
29	WANG Y, ZHOU Q, LIU J, et al. LEDNet: a lightweight encoder-decoder network for real-time semantic segmentation[C]//Proceedings of IEEE International Conference on Image Processing (ICIP). Washington D.C., USA: IEEE Press, 2019: 1860-1864.
30	QIN X B , ZHANG Z C , HUANG C Y , et al. U²-Net: going deeper with nested U-structure for salient object detection. Pattern Recognition, 2020, 106, 107404. doi: 10.1016/j.patcog.2020.107404
31	XU G A , LI J C , GAO G W , et al. Lightweight real-time semantic segmentation network with efficient transformer and CNN. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(12): 15897- 15906. doi: 10.1109/TITS.2023.3248089

[1]	朱思远, 李佳圣, 邹丹平, 何迪, 郁文贤. 基于半监督学习的非结构化道路缺陷检测算法[J]. 计算机工程, 2025, 51(9): 14-24.
[2]	朱永清, 赵鹏, 赵菲菲, 慕晓冬, 白坤, 尤轩昂. 基于深度学习的生成式文本摘要技术综述[J]. 计算机工程, 2021, 47(11): 11-21,28.
[3]	潘奎刚,石朝侠. 基于主方向加权投票的非结构化道路消失点检测[J]. 计算机工程, 2017, 43(12): 237-241.
[4]	尹建新;莫路锋. 边界与区域相融合的非结构化道路检测算法[J]. 计算机工程, 2008, 34(15): 217-219,.

选择文件类型/文献管理软件名称

选择包含的内容

面向非结构化道路的轻量化轴向上下文分割网络

Lightweight Axial Context Segmentation Network for Unstructured Road

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 31

相关文章 4

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

面向非结构化道路的轻量化轴向上下文分割网络

Lightweight Axial Context Segmentation Network for Unstructured Road

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 31

相关文章 4

编辑推荐

Metrics

本文评价