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

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(Lightweight Axial Context Network, 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(Channel-and-Spatial Attention Block, 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模块(Lightweight Partial Context Transformer, LightPCT)，利用部分通道交互策略降低计算冗余，高效捕获长程依赖以修复道路拓扑连通性；并在解码器中集成双路径通道融合(Dual-Path Channel Fusion, DPCF)与轴向细结构增强(Thin Structure Enhancer, TSE)模块，旨在弥合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（India Driving Dataset, IDD）与越野空间检测数据集 (Off-Road Freespace Detection, ORFD)二次构建的非结构化道路数据集上的实验结果表明，AXON-Net在道路交并比指标上分别达到95.3%、88.1%，参数量仅为8.49 M，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。

Abstract: Unstructured road segmentation is a crucial component of environmental perception for autonomous driving, facing challenges such as the integrity of global topological modeling, the preservation of boundary details, and the trade-off between model efficiency and accuracy. To address these challenges, this paper proposes a Lightweight Axial Context Network (AXON-Net). Employing an encoder-decoder architecture, the network introduces a Channel-and-Spatial Attention Block (CASAB) in the encoder, which adaptively recalibrates feature weights by aggregating multi-dimensional statistical information to effectively suppress environmental noise, thereby enhancing feature discriminability in complex backgrounds. A Lightweight Partial Context Transformer (LightPCT) is designed at the bottleneck, utilizing a partial channel interaction strategy to reduce computational redundancy and efficiently capture long-range dependencies to restore road topological connectivity. Furthermore, the decoder integrates Dual-Path Channel Fusion (DPCF) and Thin Structure Enhancer (TSE) modules, aiming to bridge the feature semantic gap and explicitly enhance 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 the Off-Road Freespace Detection (ORFD) dataset show that AXON-Net achieves road Intersection over Union (IoU) scores of 95.3% and 88.1%, respectively, with only 8.49 M parameters, achieving a superior balance between segmentation accuracy and model efficiency. Ablation studies further validate the synergistic effectiveness of the proposed modules, demonstrating the network's potential application in unstructured road perception tasks.

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, doi: 10.19678/j.issn.1000-3428.0253402.

CAO Fu, XING Wenbin, ZUO Yong, ZHANG Ronghui, CHEN Junzhou. AXON-Net: Lightweight Axial Context for Unstructured Road Segmentation[J]. Computer Engineering, doi: 10.19678/j.issn.1000-3428.0253402.

参考文献

[1] YU F, CHEN H, WANG X, et al. BDD100K: A diverse driving dataset for heterogeneous multitask learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 2636-2645.
[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] 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. PAN K G, SHI C X. Vanishing point detection of unstructured road based on dominant orientation weighted voting[J]. Computer Engineering, 2017, 43(12): 237-241. (in Chinese)
[4] 杨雪, 范勇, 高琳, 邱运春. 基于纹理基元块识别与合并的图像语义分割[J]. 计算机工程, 2015, 41(3): 253-257. YANG X, FAN Y, GAO L, QIU Y C. Image semantic segmentation based on texture element block recognition and merging[J]. Computer Engineering, 2015, 41(3): 253-257. (in Chinese)
[5] 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, 222. YIN J X, MO L F. Measurement arithmetic for unstructured road based on boundary and region[J]. Computer Engineering, 2008, 34(15): 217-219, 222. (in Chinese)
[6] LIU Z, YU S, ZHENG N. A co-point mapping-based approach to drivable area detection for self-driving cars[J]. Engineering, 2018, 4(4): 479-490.
[7] HSU C M, LIAN F L, HUANG C M, et al. Detecting drivable space in traffic scene understanding[C]//2012 International Conference on System Science and Engineering (ICSSE). Washington D.C., USA: IEEE Press, 2012: 79-84.
[8] HERNANDEZ C, FERNANDEZ-LLORCA D, SOTELO M A. A hybrid vision-map method for urban road detection[J]. Journal of Advanced Transportation, 2017, 2017(1): 7090549.
[9] SLESS L, EL SHLOMO B, 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: 0-0.
[10] MA B, LAKSHMANAN S, HERO A O. Simultaneous detection of lane and pavement boundaries using model-based multisensor fusion[J]. 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[J]. 计算机工程, 2025, 51(7): 314-325. 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[J]. Computer Engineering, 2025, 51(7): 314-325. (in Chinese)
[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[J]. 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[J]. Journal of Sichuan University (Natural Science Edition), 2023, 60(1): 66-73. (in Chinese)
[13] LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA: IEEE Press, 2015: 3431-3440.
[14] RONNEBERGER O, FISCHER P, BROX T. U-Net: Convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer International Publishing, 2015: 234-241.
[15] 张新华, 黄梦醒, 张雨, 等. 基于深度学习的卫星图像道路分割算法[J]. 计算机工程, 2021, 47(10): 306-313. ZHANG X H, HUANG M X, ZHANG Y, et al. Satellite image road segmentation algorithm based on deep learning[J]. Computer Engineering, 2021, 47(10): 306-313. (in Chinese)
[16] SINHA D, EL-SHARKAWY M. Thin MobileNet: An enhanced MobileNet architecture[C]//2019 IEEE 10th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON). Washington D.C., USA: IEEE Press, 2019: 0280-0285.
[17] ZHAO H, SHI J, QI X, et al. Pyramid scene parsing network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE Press, 2017: 2881-2890.
[18] 张立恒, 王浩, 薛博维, 等. 基于改进 D-LinkNet 模型的高分遥感影像道路提取研究[J]. 计算机工程, 2021, 47(9): 288-296. ZHANG L H, WANG H, XUE B W, et al. Research on road extraction from high-resolution remote sensingimages based on improved D-LinkNet model[J].Computer Engineering, 2021, 47(9): 288-296. (in Chinese)
[19] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems. Long Beach, CA, USA: Curran Associates, Inc., 2017, 30: 5998-6008.
[20] VARMA G, SUBRAMANIAN A, NAMBOODIRI A, et al. IDD: A dataset for exploring problems of autonomous navigation in unconstrained environments[C]//2019 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]//2022 International Conference on Robotics and Automation (ICRA). Philadelphia, PA, USA: IEEE, 2022: 6542-6548.
[22] 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.
[23] WU T, TANG S, ZHANG R, et al. CGNet: A light-weight context guided network for semantic segmentation[J]. 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[J]. 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[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 7401-7416.
[26] HUANG P M, CHAO I T, HUANG P C, et al. BEVANet: Bilateral efficient visual attention network for real-time semantic segmentation[C]//2025 IEEE International Conference on Image Processing (ICIP). IEEE, 2025: 2778-2783.
[27] WANG Y, ZHOU Q, LIU J, et al. LEDNet: A lightweight encoder-decoder network for real-time semantic segmentation[C]//2019 IEEE International Conference on Image Processing (ICIP). Taipei, China: IEEE, 2019: 1860-1864.
[28] XU G, LI J, GAO G, et al. Lightweight real-time semantic segmentation network with efficient transformer and CNN[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(9): 10072-10084.
[29] CHAURASIA A, CULURCIELLO E. LinkNet: Exploiting encoder representations for efficient semantic segmentation[C]//2017 IEEE Visual Communications and Image Processing (VCIP). Washington D.C., USA: IEEE Press, 2017: 1-4.
[30] QIN X, ZHANG Z, HUANG C, et al. U2-Net: Going deeper with nested U-structure for salient object detection[J]. Pattern Recognition, 2020, 106: 107404.
[31] LO S Y, HANG H M, CHAN S W, et al. Efficient dense modules of asymmetric convolution for real-time semantic segmentation[C]//Proceedings of the 1st ACMInternational Conference on Multimedia in Asia. New York, USA: ACM Press, 2019: 1-6.

选择文件类型/文献管理软件名称

选择包含的内容