基于深度学习的城市内涝区域车辆检测与分析

doi:10.16180/j.cnki.issn1007-7820.2024.05.003

摘要/Abstract

摘要：

在城市内涝场景当中,较多人与车辆被困于积水中,给大众生活带来不利影响。随着计算机技术的快速发展,深度学习在解决实际问题中的运用也越来越广泛。文中提出了一种利用TensorFlow深度学习框架搭建MaskR-CNN(Regions with Convolutional Neural Networks Features)模型的方法,对城市内涝场景的积水区进行检测,检测效果良好,mAP(mean Average Precision)值达到89%。同时,基于YOLOv5(You Only Look Once version 5 )模型,采用密集帧间差运算对处于积水区中的人车进行追踪,追踪精度达到90%左右,并使用YOLOv5外挂ResNet(Residual Network)实现了对内涝场景中的车辆进行淹没危险度分析。实验结果表明,文中所用模型的车辆危险度检测效果优于其他模型。

关键词: 城市内涝, MaskR-CNN模型, TensorFlow, 深度学习, 目标检测, YOLOv5, ResNet, 危险度分析

Abstract:

In the urban waterlogging scene, many people and vehicles are trapped in the water, which brings adverse effects to the public life. With the rapid development of computer technology, deep learning is more and more widely used in solving practical problems. This study proposes a method to build a MaskR-CNN(Regions with Convolutional Neural Networks Features) model using TensorFlow deep learning framework, which has achieved good detection results in the detection of waterlogging areas in urban waterlogging scenes, with the mAP(mean Average Precision) value reaching 89%. Based on the YOLOv5(You Only Look Once version 5) model, the dense interframe difference operation is used to track people and vehicles in waterlogged areas, and the tracking accuracy reached about 90%. Moreover, ResNet(Residual Network) attached to YOLOv5 is used to analyze the risk of submersion of vehicles in waterlogging scenarios. The experimental results show that the vehicle risk detection effect of the proposed model is better than other models.

Key words: urban waterlogging, MaskR-CNN model, TensorFlow, deep learning, target detection, YOLOv5, ResNet, risk analysis

中图分类号:

TU992.25

夏榕成, 刘德儿. 基于深度学习的城市内涝区域车辆检测与分析[J]. 电子科技, 2024, 37(5): 18-24.

XIA Rongcheng, LIU Deer. Vehicle Detection and Analysis in Urban Waterlogging Area Based on Deep Learning[J]. Electronic Science and Technology, 2024, 37(5): 18-24.

图/表 15

图1

图2

图3

图4

图5

表1

图6

图7

图8

图9

图10

图11

图12

图13

表2

参考文献 18

[1]	周宏, 刘俊, 高成, 等. 我国城市内涝防治现状及问题分析[J]. 灾害学, 2018, 33(3):147-151.
	Zhou Hong, Liu Jun, Gao Cheng, et al. Analysis of current situation and problems of urban waterlogging control in China[J]. Journal of Catastrophology, 2018, 33(3):147-151.
[2]	相广芳. 暴雨内涝后城市排水规划管理的策略研究-评《水利工程概论》[J]. 灌溉排水学报, 2022, 41(2):153-156.
	Xiang Guangfang. Research on the strategy of urban drainage planning and management after rainstorm and waterlogging-Review of Introduction to Hydraulic Engineering[J]. Journal of Irrigation and Drainage, 2022, 41(2):153-156.
[3]	张炜, 李思敏, 时真男. 我国城市暴雨内涝的成因及其应对策略[J]. 自然灾害学报, 2012, 21(5):180-184.
	Zhang Wei, Li Simin, Shi Zhennan. Formation causes and coping strategies of urban rainstorm waterlogging in China[J]. Journal of Natural Disasters, 2012, 21(5):180-184.
[4]	邬玲懿. 基于神经网络的排水管网预测模型的研究及应用[D]. 杭州: 杭州电子科技大学, 2011:23-50.
	Wu Lingyi. The study of forecasting model of urban pipe network based on BP neural network and its application[D]. Hangzhou: Hangzhou Dianzi University, 2011:23-50.
[5]	纪颖波, 钟炜, 姜腾腾, 等. 面向预案响应机制的城市内涝可控及推演机制研究[J]. 中国给水排水, 2015, 31(10):8-11.
	Ji Yingbo, Zhong Wei, Jiang Tengteng, et al. Research on urban waterlogging disaster control and inference mechanism based on emergency response mechanism[J]. China Water and Wastewater, 2015, 31(10):8-11.
[6]	赵冬泉, 李雪森, 唐兰贵, 等. SmartWater智能在线液位监测预警系统介绍及应用[J]. 给水排水, 2015, 51(9):95-100.
	Zhao Dongquan, Li Xuesen, Tang Langui, et al. Introduction and application of SmartWater intelligent online liquid level monitoring and early warning system[J]. Water and Wastewater Engineering, 2015, 51(9):95-100.
[7]	姜元军, 来丽芳, 李亚丹, 等. 基于GIS的城市内涝预警预报系统设计与实现[J]. 测绘与空间地理信息, 2016, 39(8):69-71.
	Jiang Yuanjun, Lai Lifang, Li Yadan, et al. Design and implementation of early warning and forecasting system for urban water logging based on GIS[J]. Geomatics and Spatial Information Technology, 2016, 39(8):69-71.
[8]	Yi F L, Bing H L, Jing Y P, et al. Research on the useof YOLOv5 object detection algorithm min mask wearing recognition[J]. World Scientific Research Journal, 2020, 6(11):276-284.
[9]	涂朴, 黄晨. 基于Tensorflow深度学习框架的手写体数字识别模型优化及应用[J]. 自动化技术与应用, 2020, 39(12):110-114.
	Tu Pu, Huang Chen. Optimization and application of handwritten number recognition model based on Tensorflow deep learning framework[J]. Techniques of Automation and Applications, 2020, 39(12):110-114.
[10]	何马均. 深度学习框架TensorFlow的高效分布式并行算法研究[D]. 成都: 电子科技大学, 2019:44-61.
	He Majun. Research on efficient distributed parallel algorithm of deep learning framework TensorFlow[D]. Chengdu: University of Electronic Science and Technology of China, 2019:44-61.
[11]	迪力夏提·多力昆, 张太红, 冯向萍. LabelMe标注核对系统的设计与实现[J]. 计算机技术与发展, 2022, 32(3):214-220.
	Duolikun Dilixiati, Zhang Taihong, Feng Xiangping. Design and implementation of LabelMe label checking system[J]. Computer Technology and Development, 2022, 32(3):214-220.
[12]	王鹏, 王玉林, 焦博文, 等. 基于YOLOv5的道路目标检测算法研究[J]. 计算机工程与应用, 2023, 59(1):117-125. doi: 10.3778/j.issn.1002-8331.2206-0316
	Wang Peng, Wang Yulin, Jiao Bowen, et al. Research on road target detection algorithm based on YOLOv5[J]. Computer Engineering and Applications, 2023, 59(1):117-125. doi: 10.3778/j.issn.1002-8331.2206-0316
[13]	邢姗姗, 赵文龙. 基于YOLO系列算法的复杂场景下无人机目标检测研究综述[J]. 计算机应用研究, 2020, 37(S2):28-30.
	Xing Shanshan, Zhao Wenlong. Review of UAV target detection in complex scenes based on YOLO series algorithms[J]. Application Research of Computers, 2020, 37(S2):28-30.
[14]	张伟, 刘娜, 江洋, 等. 基于YOLO神经网络的垃圾检测与分类[J]. 电子科技, 2022, 35(10):45-50.
	Zhang Wei, Liu Na, Jiang Yang. Garbage detection and classification based on YOLO neural network[J]. Electronic Science and Technology, 2022, 35(10):45-50.
[15]	杨陆野. 基于YOLO和帧间差分法的飞鸟检测算法[J]. 现代信息科技, 2021, 5(1):92-94.
	Yang Luye. Bird detection algorithm based on YOLO and inter frame difference method[J]. Modern Information Technology, 2021, 5(1):92-94.
[16]	刘金利, 傅志中, 周阳, 等. 基于帧间差分法和VIBE算法的运动目标检测[J]. 电子科技, 2015, 28(8):91-93,97.
	Liu Jinli, Fu Zhizhong, Zhou Yang, et al. Moving object detection based on interframe differencing and VIBE algorithm[J]. Electronic Science and Technology, 2015, 28(8):91-93,97.
[17]	成怡, 田文斌, 郑腾龙. 改进Inception-Resnet-V2网络的无人机航向识别[J]. 计算机工程与应用, 2022, 58(24):307-312. doi: 10.3778/j.issn.1002-8331.2106-0542
	Cheng Yi, Tian Wenbin, Zheng Tenglong. Unmanned aerial vehicles heading recognition based on improved Inception-Resnet-V2 network[J]. Computer Engineering and Applications, 2022, 58(24):307-312. doi: 10.3778/j.issn.1002-8331.2106-0542
[18]	黄静, 谢宣. 基于改进的SSD监理目标检测研究[J]. 电子科技, 2022, 35(5):7-13.
	Huang Jing, Xie Xuan. Research on supervision object detection based on improved SSD[J]. Electronic Science and Technology, 2022, 35(5):7-13.

类别标签	数量	图13(a)	图13(b)	图13(c)	图13(d)
safe	27	9	2	1	15
unsafe	1	1	0	0	0
danger	0	0	0	0	0
错检	2	1	0	0	1
漏检	0	0	0	0	0

模型名称	漏检数量	错检数量
RetinaNet	10	4
SSD	9	3
CenterNet-Hourglass	5	2
Faster-RCNN	4	4
Efficientdet-d1	3	2
YOLOv5	0	2