基于深度学习的车位检测方

doi:10.16180/j.cnki.issn1007-7820.2025.06.004

摘要/Abstract

摘要：

针对自动泊车系统中车位检测算法存在的任务单一、车位关键点遮挡导致检测效率下降等问题,文中结合目标检测和关键点检测提出了一种基于YOLO(You Only Look Once)的车位检测模型,可同时检测车位区域、车位使用情况以及车位关键点。在YOLOv7-tiny(You Only Look Once version7-tiny)模型输出端增加车位关键点回归分支并对其进行编码设计,通过改进损失函数提高模的识别和定位准确性。在主干网络融合注意力机制,提高网络的特征提取能力。实验结果表明,相较于YOLOv7-tiny,所提模型车位区域识别准确率均值提升了1.8百分点,车位使用情况判断准确率提升了1.1百分点,并且增加了车位关键点定位功能,定位成功率达到94.1%,具有较高的应用价值。

关键词: 自动泊车, 遮挡, 目标检测, YOLOv7-tiny, 关键点检测, 特征提取, 注意力机制, 损失函数

Abstract:

In view of the problems such as single task and occlusion of key points in parking detection algorithm in automatic parking system, a YOLO(You Only Look Once)-based parking detection model is proposed by combining target detection and key point detection, which can simultaneously detect parking area, parking use and key points. The key point regression branch of parking space is added to the output end of YOLOv7-tiny(You Only Look Once version7-tiny) model and its coding design is carried out to improve the accuracy of model identification and positioning by improving the loss function. The attention mechanism is integrated in the backbone network to improve the feature extraction capability of the network. The experimental results show that compared with YOLOv7-tiny, the average parking area recognition accuracy of the proposed model is increased by 1.8 percentage, and the parking space usage judgment accuracy is increased by 1.1 percentage. Moreover, the key point location function is added, and the success rate of location reaches 94.1%, indicating that the proposed method has high application value.

Key words: automatic parking, occlusion, target detection, YOLOv7-tiny, keypoints detection, feature extraction, attention mechanism, loss function

中图分类号:

TP391

唐玉良, 张轩雄. 基于深度学习的车位检测方[J]. 电子科技, 2025, 38(6): 23-29.

TANG Yuliang, ZHANG Xuanxiong. A Parking Space Detection Method Based on Deep Learning[J]. Electronic Science and Technology, 2025, 38(6): 23-29.

图/表 11

图1

图2

图3

图4

图5

图6

图7

表1

表2

图8

表3

参考文献 22

[1]	Jhang J H, Lian F L. An autonomous parking system of optimally integrating bidirectional rapidly-exploring random trees and parking-oriented model predictive control[J]. IEEE Access, 2020, 37(8):163502-163523.
[2]	Park W J, Kim B S, Seo D E, et al. Parking space detection using ultrasonic sensor in parking assistance system[C]. Eindhoven: IEEE Intelligent Vehicles Symposium,2008:1039-1044.
[3]	Jung H G, Cho Y H, Yoon P J, et al. Scanning laser radar-based target position designation for parking aid system[J]. IEEE Transactions on Intelligent Transportation Systems, 2008, 9(3):406-424.
[4]	Scheunert U, Fardi B, Mattern N, et al. Free space determination for parking slots using a 3D PMD sensor[C]. Istanbul: IEEE Intelligent VehiclesSymposium,2007:154-159.
[5]	Xu J, Chen G, Xie M. Vision-guided automatic parking for smart car[C]. Dearborn: Proceedings of the IEEE Intelligent Vehicles Symposium,2000:725-730.
[6]	Suhr J K, Jung H G. Fully-automatic recognition of various parking slot markings in around view monitor image sequences[C]. Anchorage: The Fifteenth International IEEE Conference on Intelligent Transportation Systems, 2012:1294-1299.
[7]	Zhao J, Gao H, Zhang X, et al. Ring fusion of fisheye images based on corner detection algorithm for around view monitoring system of intelligent driving[J]. Journal of Robotics, 2018, 39(2):1-9.
[8]	Zhang L, Huang J, Li X, et al. Vision-based parking-slot detection:A DCNN-based approach and a large-scale benchmark dataset[J]. IEEE Transactions on Image Processing, 2018, 27(11):5350-5364.
[9]	Huang J, Zhang L, Shen Y, et al. DMPR-PS:A novel approach for parking-slot detection using directional marking-point regression[C]. Shanghai: IEEE International Conference on Multimedia and Expo,2019:212-217.
[10]	Li W, Cao L, Yan L, et al. Vacant parking slot detection in the around view image based on deep learning[J]. Sensors, 2020, 20(7):2138-2150.
[11]	Suhr J K, Jung H G. End-to-end trainable one-stage parking slot detection integrating global and local information[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 23(5):4570-4582.
[12]	Min C, Xu J, Xiao L, et al. Attentional graph neural network for parking-slot detection[J]. IEEE Robotics and Automation Letters, 2021, 6(2):3445-3450.
[13]	Grbić R, Koch B. Automatic vision-based parking slot detection and occupancy classification[J]. Expert Systems with Applications. 2023, 22(3):120147.
[14]	Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]. Vancouver: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2023:7464-7475.
[15]	苏波, 柴自强, 王莉, 等. 基于姿态估计的八段锦序列动作识别与评估[J]. 电子科技, 2022, 35(12):84-90.
	Su Bo, Chai Ziqiang, Wang Li, et al. Eight-section brocade sequence action recognition and evaluation based on pose estimation[J]. Electronic Science and Technology, 2022, 35(12):84-90.
[16]	顾伟, 李菲菲, 陈虬. 基于多特征融合的行人检测方法[J]. 电子科技, 2021, 34(5):29-34.
	Gu Wei, Li Feifei, Chen Qiu. Pedestrian detection algorithm based on multiple feature fusion[J]. Electronic Science and Technology, 2021, 34(5):29-34.
[17]	赵轩, 周凡, 余汉成. 基于改进特征提取及融合模块的YOLOv3模型[J]. 电子科技, 2022, 35(7):40-45.
	Zhao Xuan, Zhou Fan, Yu Hancheng. Improved YOLOv3 model based on new feature extraction and fusion module[J]. Electronic Science and Technology, 2022, 35(7):40-45.
[18]	Woo S, Park J, Lee J Y, et al. Cbam:Convolutional block attention module[C]. Munich: Proceedings of the European Conference on Computer Vision,2018:3-19.
[19]	Rezatofighi H, Tsoi N, Gwak J Y, et al. Generalized intersection over union: A metric and a loss for bounding box regression[C]. Long Beach: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2019:658-666.
[20]	Zheng Z, Wang P, Liu W, et al. Distance-IoU loss:Faster and better learning for bounding box regression[C]. New York: Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7):12993-13000.
[21]	Zhang Y F, Ren W, Zhang Z, et al. Focal and efficient IOU loss for accurate bounding box regression[J]. Neurocomputing, 2022, 50(6):146-157.
[22]	Do H, Choi J Y. Context-based parking slot detection with a realistic dataset[J]. IEEE Access, 2020, 37(8): 171551-171559.

模型	mAP/%	FPS/frame·s^-1	定位成功率/%	判断准确率/%	FLOPs/10⁹
YOLOv7-tiny	95.6	43	-	97.5	19.7
改进 YOLOv7-tiny	97.4	42	94.1	98.6	20.2

序号	边框损失函数		车位关键点损失函数		CBAM	mAP/%	定位成功率/%	判断准确率/%	FLOPs/10⁹
序号	CIoU	EIoU	L₂	GS_LOSS	CBAM	mAP/%	定位成功率/%	判断准确率/%	FLOPs/10⁹
1	√	-	√	-	-	95.6	92.5	96.5	19.7
2	-	√	-	-	-	96.0	93.8	97.8	19.7
3	√	-	-	√	-	97.0	94.7	98.4	19.7
4	√	-	√	-	√	96.2	94.4	97.8	20.2

模型	mAP /%	FPS/frame· s^-1	定位准确率/%	判断准确率/%	FLOPs /10⁹
YOLOv5s	95.1	41	-	97.8	18.2
YOLOv7	97.9	14	-	98.2	100.9
YOLOv7-tiny	95.6	43	-	97.5	19.7
改进YOLOv7-tiny	97.4	42	94.1	98.6	20.2