融合视觉与几何推理的地下车位全场景检测算法

刘平; 马春蕾; 刘明杰; 朴昌浩; 黄龙航; 颜鲁鹏

doi:10.12263/DZXB.20260063

您当前的位置：

首页 >

文章列表页 >

融合视觉与几何推理的地下车位全场景检测算法

学术论文 | 更新时间：2026-07-30

- 融合视觉与几何推理的地下车位全场景检测算法
- A Full-Scene Detection Algorithm for Underground Parking Spaces Fusing Vision and Geometric Reasoning
- 电子学报 2026年54卷第4期页码：1692-1704
- 作者机构：
  
  1.重庆邮电大学自动化学院，重庆 400065
  2.重庆市建设信息中心，重庆 400010
- 作者简介：
  
  刘平男，1990年6月出生于重庆市。现为重庆邮电大学自动化学院副教授、硕士生导师。主要研究方向为无人系统轨迹规划、智慧泊车。E-mail: liuping_cqupt@cqupt.edu.cn
  马春蕾男，2001年3月出生于重庆市。2023年毕业于重庆邮电大学电气自动化专业。现为重庆邮电大学研究生。主要研究方向为机器视觉与图像识别。E-mail: S240333022@stu.cqupt.edu.cn
  刘明杰男，1987年出生于河南省辉县市。现为重庆邮电大学副教授、博士生导师。主要研究方向为智能网联汽车环境智能感知及控制、多源信息融合、机器学习。E-mail: liumj@cqupt.edu.cn
  朴昌浩男，1978年11月出生于吉林省延边市。现为重庆邮电大学教授、博士生导师。主要研究方向为智能网联汽车。E-mail: piaoch@cqupt.edu.cn
  黄龙航男，1997年9月出生于重庆市。2024年毕业于内蒙古工业大学电子信息系，现为重庆市建设信息中心科员。主要研究方向为智慧建筑。E-mail: 1511097975@qq.com
  颜鲁鹏男，1995年2月出生于甘肃省兰州市。2020年毕业于重庆交通大学建筑与土木工程专业。现为重庆市建设信息中心工程师。主要研究方向为住建领域数字化发展。E-mail: 954518155@qq.com
- 基金信息：
  
  国家重点研发计划(2022YFE0101000);重庆市建设信息中心项目(JSXXCQ-CY01)
- DOI：10.12263/DZXB.20260063
  中图分类号： TP391.41;U495
- 收稿：2026-03-11，
  
  录用：2026-03-25，
  
  网络首发：2026-05-06，
  
  纸质出版：2026-04-25
- 稿件说明：
移动端阅览
刘平, 马春蕾, 刘明杰, 等. 融合视觉与几何推理的地下车位全场景检测算法[J]. 电子学报, 2026, 54(04): 1692-1704.

LIU Ping, MA Chunlei, LIU Mingjie, et al. A Full-Scene Detection Algorithm for Underground Parking Spaces Fusing Vision and Geometric Reasoning[J]. Acta Electronica Sinica, 2026, 54(04): 1692-1704.
刘平, 马春蕾, 刘明杰, 等. 融合视觉与几何推理的地下车位全场景检测算法[J]. 电子学报, 2026, 54(04): 1692-1704. DOI：10.12263/DZXB.20260063

LIU Ping, MA Chunlei, LIU Mingjie, et al. A Full-Scene Detection Algorithm for Underground Parking Spaces Fusing Vision and Geometric Reasoning[J]. Acta Electronica Sinica, 2026, 54(04): 1692-1704. DOI：10.12263/DZXB.20260063

摘要

在智能停车管理系统中，复杂环境下高精度车位感知是实现自动化引导与资源优化的核心技术。传统检测方法成本高、维护困难且车位状态信息单一，而基于视觉的方法虽能提供丰富信息，但在地下停车场面临光照昏暗、强光源干扰、背景复杂、车辆遮挡严重等挑战。为此，本文提出一种融合视觉与几何推理的轻量化地下停车场车位实时检测网络（Real-Time Underground Parking Space Occupancy DEtection TRansformer， RT-UPSO-DETR）。首先，设计了膨胀空间注意力轻量级残差F-DBlock（Fusion DBlock）模块，在扩大感受野捕获全局模糊特征的同时显著减少冗余计算。其次，构建了多机制融合的Transformer编码层CSAL（Cognitive Spatial Attention Layer）模块，替换RT-DETR-R18中原有的AIFI模块，显著提升了模型在低光照、模糊图像中的特征判别能力。进一步，针对车辆相互遮挡导致的后排车位漏检问题，设计了几何推理透视拓扑补全模块（Perspective Topology Completion Module， PTCM），利用前排已检测车位作为几何锚点估计透视灭点，基于径向投影推断被遮挡车位的近似位置，有效恢复缺失车位框。为验证算法有效性，本文构建了地下停车场车位检测数据集以及遮挡补全专项测试集，并在公开低光照数据集ExDark上开展泛化实验。实验结果表明，RT-UPSO-DETR在自建数据集上相较于基线RT-DETR-R18模型，在精度、召回率、mAP@0.5和mAP@0.5-0.95指标上分别提升1.5%、3.3%、0.7%和0.9%；同时参数量降低22.1%，GFLOPs降低18.8%，推理速度为52.4 FPS，满足实时检测需求。在ExDark数据集泛化实验中，本文模型较RT-DETR-R18在精确率、召回率、mAP@0.5和mAP@0.5-0.95上分别提升2.1%、1.3%、2.1%和1.0%，验证了模型在低光照条件下的鲁棒性。车位补全方面，PTCM模块在遮挡补全实验中在漏检召回率、补全准确率和平均中心点误差指标上分别达到了82.48%、74.83%和23.27 pixels，证明其能有效降低漏检率。本文提出的RT-UPSO-DETR网络通过轻量化设计与多机制注意力融合，有效解决了地下停车场低光照、复杂背景和遮挡场景下的车位检测难题，透视拓扑补全模块进一步增强了系统对遮挡的鲁棒性，为实际智能停车系统提供了可行的视觉解决方案。

Abstract

In intelligent parking management systems

achieving high-precision parking space perception in complex environments is a core technology for automated guidance and resource optimization. Traditional detection methods suffer from high costs

difficult maintenance

and limited parking status information. Although vision-based approaches can provide rich information

they face challenges in underground parking lots

such as dim illumination

strong light interference

complex backgrounds

and severe vehicle occlusion. To address these issues

this paper proposes a lightweight real-time underground parking space occupancy detection network integrating vision and geometric reasoning

termed real-time underground parking space occupancy detection transformer (RT-UPSO-DETR). First

a lightweight residual F-DBlock (Fusion DBlock) module with dilated spatial attention is designed

which expands the receptive field to capture global blurred features while significantly reducing redundant computation. Second

a multi-mechanism Transformer encoder layer

cognitive spatial attention layer (CSAL)

is constructed to replace the original AIFI module in RT-DETR-R18

significantly enhancing the model’s feature discrimination ability in low-light and blurred images. Furthermore

to address the missed detection of rear parking spaces caused by mutual vehicle occlusion

a geometric reasoning perspective topology completion module (PTCM) is devised

which uses the detected front parking spaces as geometric anchors to estimate the vanishing point and then infers the approximate positions of occluded spaces via radial projection

effectively recovering missing bounding boxes. To validate the effectiveness of the proposed algorithm

we construct an underground parking space detection dataset and a dedicated occlusion completion test set

and conduct generalization experiments on the public low-light dataset ExDark. Experimental results on the self-built dataset show that compared with the baseline RT-DETR-R18

RT-UPSO-DETR improves precision

recall

mAP@0.5

and mAP@0.5-0.95 by 1.5%

3.3%

0.7%

and 0.9%

respectively. Meanwhile

the number of parameters is reduced by 22.1%

GFLOPs by 18.8%

and the inference speed reaches 52.4 FPS

satisfying real-time detection requirements. On the ExDark dataset

compared with RT-DETR-R18

the proposed model improves precision

recall

mAP@0.5

and mAP@0.5-0.95 by 2.1%

1.3%

2.1%

and 1.0%

respectively

verifying its robustness under low-light conditions. For parking space completion

the PTCM achieves a missed recall rate of 82.48%

a completion accuracy of 74.83%

and a mean center error of 23.27 pixels

demonstrating its effectiveness in reducing missed detections. The proposed RT-UPSO-DETR network

through lightweight design and multi-mechanism attention fusion

effectively addresses the challenges of parking space detection in underground parking lots with low illumination

complex backgrounds

and occlusions. The perspective topology completion module further enhances the system’s robustness to occlusions

providing a feasible vision-based solution for practical intelligent parking systems.

关键词

Keywords

references

Lang A H , Vora S , Caesar H , et al . PointPillars: Fast encoders for object detection from point clouds [C ] // 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2019 : 12689 - 12697 . DOI: 10.1109/cvpr.2019.01298 http://dx.doi.org/10.1109/cvpr.2019.01298

Chiang K W , Tsai S , Chen J A , et al . AI-driven mapping system for smart parking management applications using an INS-GNSS-solid-state LiDAR-monocular camera fusion engine empowered by HD maps [J ] . IEEE Open Journal of Intelligent Transportation Systems , 2025 , 6 : 995 - 1008 . DOI: 10.1109/ojits.2025.3587274 http://dx.doi.org/10.1109/ojits.2025.3587274

Yan S , O’Connor N E , Liu M M . U-park: A user-centric smart parking recommendation system for electric shared micromobility services [J ] . IEEE Transactions on Artificial Intelligence , 2024 , 5 ( 10 ): 5179 - 5193 . DOI: 10.1109/tai.2024.3428513 http://dx.doi.org/10.1109/tai.2024.3428513

Bazzaza T , Tohidypour H R , Wang Y X , et al . Accurate detection and localization of individual free street parking spaces using AI and innovative global motion estimation [J ] . IEEE Transactions on Intelligent Vehicles , 2025 , 10 ( 2 ): 1263 - 1272 . DOI: 10.1109/tiv.2024.3425811 http://dx.doi.org/10.1109/tiv.2024.3425811

Zhang Z S , He X M , Huang J W , et al . Parking detection using combined magnetic sensor and pulsed coherent radar [J ] . IEEE Internet of Things Journal , 2022 , 9 ( 18 ): 17210 - 17219 . DOI: 10.1109/jiot.2022.3151987 http://dx.doi.org/10.1109/jiot.2022.3151987

Zhang Y , Wu C , Liu S Y , et al . A parking detection algorithm based on multitransitory finite-state machine using magnetic wireless sensor network [J ] . IEEE Internet of Things Journal , 2024 , 11 ( 5 ): 8360 - 8372 . DOI: 10.1109/jiot.2023.3319340 http://dx.doi.org/10.1109/jiot.2023.3319340

Dheeven T A , Kumar P M , Venkatesh V , et al . IoT based sensor enabled vehicle parking system [J ] . Measurement: Sensors , 2024 , 31 : 100953 . DOI: 10.1016/j.measen.2023.100953 http://dx.doi.org/10.1016/j.measen.2023.100953

黄伟杰 , 张希 , 赵柏暄 , 等 . 基于视觉的停车场车位检测与分类算法 [J ] . 计算机系统应用 , 2022 , 31 ( 3 ): 234 - 240 .

Huang Weijie , Zhang Xi , Zhao Baixuan , et al . Vision-based parking space detection and classification algorithm [J ] . Computer Systems & Applications , 2022 , 31 ( 3 ): 234 - 240 . (in Chinese)

Sharma A , Kumar V , Longchamps L . Comparative performance of YOLOv8, YOLOv9, YOLOv10, YOLOv11 and Faster R-CNN models for detection of multiple weed species [J ] . Smart Agricultural Technology , 2024 , 9 : 100648 . DOI: 10.1016/j.atech.2024.100648 http://dx.doi.org/10.1016/j.atech.2024.100648

Wang S B , Chen R H , Wu H Y , et al . YOLOH: You only look one hourglass for real-time object detection [J ] . IEEE Transactions on Image Processing , 2024 , 33 : 2104 - 2115 . DOI: 10.1109/tip.2024.3374225 http://dx.doi.org/10.1109/tip.2024.3374225

Jrondi Z , Moussaid A , Hadi M Y . Exploring End-to-End object detection with transformers versus YOLOv8 for enhanced citrus fruit detection within trees [J ] . Systems and Soft Computing , 2024 , 6 : 200103 . DOI: 10.1016/j.sasc.2024.200103 http://dx.doi.org/10.1016/j.sasc.2024.200103

Zhao Y A , Lv W Y , Xu S L , et al . DETRs beat YOLOs on real-time object detection [C ] // 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2024 : 16965 - 16974 . DOI: 10.1109/cvpr52733.2024.01605 http://dx.doi.org/10.1109/cvpr52733.2024.01605

蒋智臣 , 胡俐蕊 . 基于改进RT-DETR的浅水海洋生物识别方法 [J ] . 电子测量技术 , 2024 , 47 ( 18 ): 155 - 163 .

Jiang Zhichen , Hu Lirui . Marine life identification method based on improved RT-DETR [J ] . Electronic Measurement Technology , 2024 , 47 ( 18 ): 155 - 163 . (in Chinese)

程鑫淼 , 张雪松 , 曹冰洁 , 等 . 改进RT-DETR的小目标检测方法研究 [J ] . 计算机工程与应用 , 2025 , 61 ( 15 ): 144 - 155 .

Cheng Xinmiao , Zhang Xuesong , Cao Bingjie , et al . Research on small object detection method of improved RT-DETR [J ] . Computer Engineering and Applications , 2025 , 61 ( 15 ): 144 - 155 . (in Chinese)

Feijoo D , Benito J C , Garcia A , et al . DarkIR: Robust low-light image restoration [C ] // 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2025 : 10879 - 10889 . DOI: 10.1109/cvpr52734.2025.01016 http://dx.doi.org/10.1109/cvpr52734.2025.01016

Yin D S , Hu L Y , Li B , et al . 5%>100 %: Breaking performance shackles of full fine-tuning on visual recognition tasks[C ] // 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2025 : 20071 - 20081 . DOI: 10.1109/cvpr52734.2025.01869 http://dx.doi.org/10.1109/cvpr52734.2025.01869

Lv H Z , Yang W Z , Yin Y B , et al . MDF-FND: A dynamic fusion model for multimodal fake news detection [J ] . Knowledge-Based Systems , 2025 , 317 : 113417 . DOI: 10.1016/j.knosys.2025.113417 http://dx.doi.org/10.1016/j.knosys.2025.113417

Chen S , Zhang H Z , Atapour-Abarghouei A , et al . SEM-net: Efficient pixel modelling for image inpainting with spatially enhanced SSM [C ] // 2025 IEEE/CVF Winter Conference on Applications of Computer Vision . Piscataway : IEEE , 2025 : 461 - 471 . DOI: 10.1109/wacv61041.2025.00055 http://dx.doi.org/10.1109/wacv61041.2025.00055

He K M , Zhang X Y , Ren S Q , et al . Deep residual learning for image recognition [C ] // 2016 IEEE Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2016 : 770 - 778 . DOI: 10.1109/cvpr.2016.90 http://dx.doi.org/10.1109/cvpr.2016.90

Wang C Y , Mark Liao H Y , Wu Y H , et al . CSPNet: A new backbone that can enhance learning capability of CNN [C ] // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops . Piscataway : IEEE , 2020 : 1571 - 1580 . DOI: 10.1109/cvprw50498.2020.00203 http://dx.doi.org/10.1109/cvprw50498.2020.00203

Nguyen H , Nawara D , Kashef R . Connecting the indispensable roles of IoT and artificial intelligence in smart cities: A survey [J ] . Journal of Information and Intelligence , 2024 , 2 ( 3 ): 261 - 285 .

闫旭东 , 钱莉 . 基于全景环视图像的停车位检测算法 [J ] . 华中师范大学学报(自然科学版) , 2024 , 58 ( 5 ): 526 - 532 . DOI: 10.19603/j.cnki.1000-1190.2024.05.004 http://dx.doi.org/10.19603/j.cnki.1000-1190.2024.05.004

Yan Xudong , Qian Li . Parking slot detection based on surround-view image [J ] . Journal of Central China Normal University (Natural Sciences) , 2024 , 58 ( 5 ): 526 - 532 . (in Chinese) . DOI: 10.19603/j.cnki.1000-1190.2024.05.004 http://dx.doi.org/10.19603/j.cnki.1000-1190.2024.05.004

王满利 , 张航 , 张长森 . 基于深度学习的低光照目标检测算法 [J ] . 北京邮电大学学报 , 2024 , 47 ( 5 ): 59 - 65 .

Wang Manli , Zhang Hang , Zhang Changsen . A low light target detection algorithm based on deep learning [J ] . Journal of Beijing University of Posts and Telecommunications , 2024 , 47 ( 5 ): 59 - 65 . (in Chinese)

邓冬冬 , 许建民 , 孟寒 , 等 . 基于蚁群算法与人工势场法融合的移动机器人路径规划 [J ] . 仪器仪表学报 , 2025 , 46 ( 2 ): 1 - 16 .

Deng Dongdong , Xu Jianmin , Meng Han , et al . Mobile robot path planning based on the fusion of ant colony algorithm and artificial potential field method [J ] . Chinese Journal of Scientific Instrument , 2025 , 46 ( 2 ): 1 - 16 . (in Chinese)

Wang D R , Tan J S , Wang H , et al . SDS-YOLO: An improved vibratory position detection algorithm based on YOLOv11 [J ] . Measurement , 2025 , 244 : 116518 . DOI: 10.1016/j.measurement.2024.116518 http://dx.doi.org/10.1016/j.measurement.2024.116518

Pokhrel A , Dao G . Optimizing YOLOv8 for parking space detection: Comparative analysis of custom YOLOv8 architecture [PP/OL ] . V1. arXiv ( 2025-05-23 ) [ 2026-03-04 ] . https://doi.org/10.48550/arXiv.2505.17364 https://doi.org/10.48550/arXiv.2505.17364 .

秦嘉奇 , 江泽涛 , 雷晓春 . 基于ICFIE-YOLO的低照度图像目标检测方法 [J ] . 电子学报 , 2025 , 53 ( 2 ): 514 - 526 .

Qin Jiaqi , Jiang Zetao , Lei Xiaochun . Low illumination image object detection method based on ICFIE-YOLO [J ] . Acta Electronica Sinica , 2025 , 53 ( 2 ): 514 - 526 . (in Chinese)

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于MAAM-LKN模型的稀疏阵列双子阵RCS缩减方法

面向K/Ka波段应用的宽带高线性CMOS低噪声放大器设计

低空无线网络中安全通信与感知一体化性能优化方法

基于脉冲响应压缩的卷积盲源分离方法

海空目标部分极化雷达回波动态建模与实验验证