Constructing an Urban Waterlogging Depth Grade Assessment Model Using Images Captured by the Public

XUE Fengchang; YANG Yunxia; CHEN Xiaojuan; LÜ Xin; CHENG Yannian

doi:10.13203/j.whugis20240290

您当前的位置：

首页 >

文章列表页 >

Constructing an Urban Waterlogging Depth Grade Assessment Model Using Images Captured by the Public

更新时间：2026-11-05

- Constructing an Urban Waterlogging Depth Grade Assessment Model Using Images Captured by the Public
- Geomatics and Information Science of Wuhan University Vol. 51, Issue 1, Pages: 46-54(2026)
- 作者机构：
  
  1.南京信息工程大学气象灾害地理信息工程实验室，江苏南京，210044
  2.中国气象局流域强降水重点开放实验室，湖北武汉，430205
  3.河北省气象灾害防御和环境气象中心，河北石家庄，050021
- 作者简介：
- 基金信息：
- DOI：10.13203/j.whugis20240290
  CLC： P208;P237
- Received：27 October 2024，
  
  Published：05 January 2026
- 稿件说明：
移动端阅览
薛丰昌,杨鋆霞,陈笑娟，等.利用社会公众拍摄图像构建城市内涝深度等级评估模型[J].武汉大学学报（信息科学版）,2026,51(1):46-54.

XUE Fengchang,YANG Yunxia,CHEN Xiaojuan,et al.Constructing an Urban Waterlogging Depth Grade Assessment Model Using Images Captured by the Public[J].Geomatics and Information Science of Wuhan University,2026,51(1):46-54.
薛丰昌,杨鋆霞,陈笑娟，等.利用社会公众拍摄图像构建城市内涝深度等级评估模型[J].武汉大学学报（信息科学版）,2026,51(1):46-54. DOI：10.13203/j.whugis20240290

XUE Fengchang,YANG Yunxia,CHEN Xiaojuan,et al.Constructing an Urban Waterlogging Depth Grade Assessment Model Using Images Captured by the Public[J].Geomatics and Information Science of Wuhan University,2026,51(1):46-54. DOI：10.13203/j.whugis20240290

摘要

衡量城市内涝严重程度的标准除了积水范围外，积水深度是更为重要的一个指标。面向城市内涝积水状态的传统监测方法存在成本高昂、监测范围有限以及仅能针对特定易涝区域进行监测的局限性，提出了一种利用社会公众拍摄的图像信息构建的城市内涝深度等级评估模型。利用爬虫技术，以“城市内涝”和“车辆受淹”等词语作为关键词，从互联网上收集社会公众拍摄的包含涉水小型车辆的图像和视频来构建样本集，采用车辆淹没深度作为参照标准，将内涝深度划分为安全（[0，20] cm）、不安全（（20

60] cm）和危险（>60 cm）3个等级，并为这些等级创建了相应的标签。应用YOLOv8模型构建了YOLOv8l、YOLOv8m、YOLOv8n和YOLOv8s 4组不同网络复杂度的模型，训练过程中，利用混淆矩阵的精确度和召回率、平均精度、平均精度均值4种精度指标对模型的准确性进行了综合评估。实验结果证实，YOLOv8算法能够有效识别不同场景下车辆的淹没状态，能够作为一种有效的工具监测和评估城市内涝的积水深度等级，不仅提高了内涝灾害评估的时效性和准确性，还拓展了社会信息在灾害管理中的应用，为未来的城市灾害管理提供了一种人工智能与社会公众信息相结合的解决方案。

Abstract

Objectives

In addition to the extent of water accumulation

the depth of water accumulation is a more important criterion for measuring the severity of urban waterlogging. In response to the limitations of traditional urban waterlogging monitoring methods

such as high costs

limited monitoring range

and the ability to only monitor specific flood-prone areas

a model for assessing the depth grade of urban waterlogging using images captured by the public has been proposed.

Methods

We use web crawling techniques with keywords like urban waterlogging and vehicles submerged to collect images and videos from the internet that include small vehicles in water

creating a sample set. Using the depth of vehicle submersion as a reference standard

we divide the waterlogging depth into three levels: Safe ([0

20] cm)

unsafe ((20

60] cm)

and dangerous (>60 cm)

and create corresponding labels for these levels. We apply the YOLOv8 model to construct four models of different network complexities that are YOLOv8l

YOLOv8m

YOLOv8n

and YOLOv8s. During training

we use a confusion matrix and four accuracy metrics that are precision

recall

average precision (AP)

and mean AP (mAP) to comprehensively evaluate the model's accuracy.

Results

(1) For the training set

all four YOLOv8 models converge after about 40 training epochs

with precision and mAP50 values stabilizing around 60%~70%

recall values around 50%~60%

and mAP50: 95 values around 35%. (2) In the test set

for the precision metric

YOLOv8m scores the highest for the safe class at 80.7%

YOLOv8l for the unsafe class at 55.4%

and YOLOv8s for the dangerous class at 84.8%. For the recall metric

YOLOv8m scores the highest for both the safe and unsafe classes at 66.3% and 68.8%

respectively

while YOLOv8n scores the highest for the dangerous class at 68.3%. For the mAP50 metric

YOLOv8s scores the highest for all classes

with 74.4% for safe

61.7% for unsafe

and 76.2% for dangerous. For the mAP50: 95 metric

YOLOv8m scores the highest for the safe class at 43.1%

while YOLOv8s scores the highest for both the unsafe and dangerous classes at 37.7% and 36.7%

respectively.

Conclusions

(1) The depth and width of the network

as well as feature extraction capability

have a significant impact on model performance. Although increasing network width and depth substantially increases the number of model parameters and computational load

not all evaluation metrics improve with the increase in network width and depth. Although the YOLOv8l model is the most complex

its precision and mAP50 scores are not superior

ranking fourth (69.3%) and third (68.0%)

respectively. Although the YOLOv8n model has the fewest parameters and computational load

its mAP50 score still achieves second place. Overall

YOLOv8m performs the best on the validation set

achieving the highest scores in precision

mAP50

and mAP50: 95

with precision and mAP50 being 2.2% and 2.6% higher than the second-ranked model

respectively

without significant sacrifice in recall

making it the optimal model in terms of comprehensive performance. (2) In the validation set

the detection accuracy of the four models for the unsafe class is lower than that for the other two classes. On one hand

this is due to the imbalance in sample sizes. On the other hand

it is also due to certain limitations in the model's detection performance. To address this issue

future research will focus on employing targeted data augmentation techniques for sample processing to improve the model's recognition capability for unsafe class samples

and introducing attention mechanism modules into the model's network structure to enhance its ability to learn key features.

关键词

Keywords

references

于朝阳 . 基于物联网的城市内涝在线监测与精细化治理系统建设思路探索［J］. 中国建设信息化， 2023 （ 24 ）： 63 - 67 .

YU Chaoyang . Discussion on the Construction of an Online Monitoring and Cause Analysis System for Urban Waterlogging Based on the Internet of Things ［J］. Informatization of China Construction ， 2023 （ 24 ）： 63 - 67 .

HUANG H B ， WANG X W ， LIU L . A Review on Urban Pluvial Floods： Characteristics， Mechanisms， Data， and Research Methods ［J］. Progress in Geography ， 2021 ， 40 （ 6 ）： 1048 - 1059 .

罗诗琦，罗斌，苏鑫，等 . 利用模拟数据进行遥感图像目标检测模型训练［J］. 武汉大学学报（信息科学版）， 2023 ， DOI： 10.13203/j.whugis20230149 http://dx.doi.org/10.13203/j.whugis20230149 .

LUO Shiqi ， LUO Bin ， SU Xin ， et al . Leveraging Synthetic Data for Object Detection in Remote Sensing Images ［J］. Geomatics and Information Science of Wuhan University ， 2023 ， DOI： 10.13203/j.whugis20230149 http://dx.doi.org/10.13203/j.whugis20230149 .

徐美，刘舒，孙杨，等 . 利用洪涝模型进行城市内涝风险快速识别与预警［J］. 武汉大学学报（信息科学版）， 2020 ， 45 （ 8 ）： 1185 - 1194 .

XU Mei ， LIU Shu ， SUN Yang ， et al . Rapid Identification and Early Warning of Urban Waterlogging Risk Using Flooding Model ［J］. Geomatics and Information Science of Wuhan University ， 2020 ， 45 （ 8 ）： 1185 - 1194 .

朱晓寒，李向阳，刘昭阁 . 大数据环境下城市内涝灾害的情景维度模型构建［J］. 武汉大学学报（信息科学版）， 2020 ， 45 （ 11 ）： 1818 - 1828 .

ZHU Xiaohan ， LI Xiangyang ， LIU Zhaoge . Constructing Scenario Dimension Model of City Waterlogging Under Big Data Environment ［J］. Geomatics and Information Science of Wuhan University ， 2020 ， 45 （ 11 ）： 1818 - 1828 .

薛丰昌，吕鑫，陈笑娟 . 基于改进DeeplabV 3+ 的城市内涝监测方法［J］. 测绘科学， 2023 ， 48 （ 10 ）： 216 - 224 .

XUE Fengchang ， LÜ Xin ， CHEN Xiaojuan . Urban Flood Monitoring Method Based on Improved DeeplabV3+ ［J］. Science of Surveying and Mapping ， 2023 ， 48 （ 10 ）： 216 - 224 .

薛丰昌，宋肖依，唐步兴，等 . 视频监控的城市内涝监测预警［J］. 测绘科学， 2018 ， 43 （ 8 ）： 50 - 55 .

XUE Fengchang ， SONG Xiaoyi ， TANG Buxing ， et al . Urban Waterlogging Monitoring Based on Urban Video Image ［J］. Science of Surveying and Mapping ， 2018 ， 43 （ 8 ）： 50 - 55 .

SWATHI Y ， CHALLA M . YOLOv8： Advancements and Innovations in Object Detection ［C］// International Conference on Smart Computing and Communication ， Denpasar， Indonesia ， 2024 .

SOHAN M ， SAI RAM T ， RAMI REDDY C V . A Review on YOLOv8 and Its Advancements ［M］//JACOB I J， PIRAMUTHU S， FALKOWSKI‐GILSKI P. Data Intelligence and Cognitive Informatics . Singapore ： Springer ， 2024 ： 529 - 545 .

TERVEN J ， CÓRDOVA‑ESPARZA D ， ROMERO‐GONZÁLEZ J . A Comprehensive Review of YOLO Architectures in Computer Vision： From YOLOv1 to YOLOv8 and YOLO‑NAS ［J］. Machine Learning and Knowledge Extraction ， 2023 ， 5 （ 4 ）： 1680 - 1716 .

央广网 . 北京发布城市积水内涝风险地图4个风险等级辅助市民汛期出行［EB/OL］. （ 2022-07-07 ）［ 2024-08-10 ］. https：//www.163.com/dy/article/HBMMGDHU0514R9NP.html https://www.163.com/dy/article/HBMMGDHU0514R9NP.html .

CNR Online . Beijing Releases Urban Waterlogging Risk Map with Four Risk Levels to Assist Citizens in Travel During the Flood Season ［EB/OL］. （ 2022-07-07 ）［ 2024-08-10 ］. https：//www.163.com/dy/article/HBMMGDHU0514R9NP.html https://www.163.com/dy/article/HBMMGDHU0514R9NP.html .

赵师震 . 基于改进YOLOv8的轻量型火灾检测算法研究［D］. 大庆：东北石油大学， 2025 .

ZHAO Shizhen . Research on Lightweight Fire Detection Algorithm Based on Improved YOLOv8 ［D］. Daqing ： Northeast Petroleum University ， 2025 .

刘瑞锦，何章鸣 . 基于YOLOv8的卫星遥感图像快速目标检测方法［J］. 空间控制技术与应用， 2023 ， 49 （ 5 ）： 89 - 97 .

LIU Ruijin ， HE Zhangming . A Fast Target Detection Method for Satellite Remote Sensing Images Based on YOLOv8 ［J］. Aerospace Control and Application ， 2023 ， 49 （ 5 ）： 89 - 97 .

张潇艺，杨胜龙 . 基于YOLOv8的民用船舶影像分类方法研究［J］. 工业控制计算机， 2024 ， 37 （ 4 ）： 72 - 73 .

ZHANG Xiaoyi ， YANG Shenglong . Research on Civil Ship Image Classification Method Based on YOLOv8 ［J］. Industrial Control Computer ， 2024 ， 37 （ 4 ）： 72 - 73 .

康睿，徐海祥，冯辉，等 . 基于改进YOLOv8的航海雷达图像目标检测算法［J］. 大连海事大学学报， 2024 ， 50 （ 3 ）： 23 - 30 .

KANG Rui ， XU Haixiang ， FENG Hui ， et al . Target Detection Algorithm of Marine Radar Image Based on Improved YOLOv8 ［J］. Journal of Dalian Maritime University ， 2024 ， 50 （ 3 ）： 23 - 30 .

SONG G L ， LIU Y ， WANG X G . Revisiting the Sibling Head in Object Detector ［C］// IEEE/CVF Conference on Computer Vision and Pattern Recognition ， Seattle， USA ， 2020 .

胡惠娟，秦一锋，徐鹤，等 . 面向无人机航拍图像的YOLOv8目标检测改进算法［J］. 计算机科学， 2025 ， 52 （ 4 ）： 202 - 211 .

HU Huijuan ， QIN Yifeng ， XU He ， et al . An Improved YOLOv8 Object Detection Algorithm for UAV Aerial Images ［J］. Computer Science ， 2025 ， 52 （ 4 ）： 202 - 211 .

程换新，矫立浩，骆晓玲，等 . 改进YOLOv8的遥感图像检测算法［J］. 无线电工程， 2024 ， 54 （ 5 ）： 1155 - 1161 .

CHENG Huanxin ， JIAO Lihao ， LUO Xiaoling ， et al . Improved YOLOv8 for Remote Sensing Image Detection ［J］. Radio Engineering ， 2024 ， 54 （ 5 ）： 1155 - 1161 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Advances in Artificial Intelligence-Driven Modeling， Prediction， and Applications of GNSS Atmospheric Effects

From Task-Specific Models to General-Purpose Model-Driven Paradigm Shift： A Review and Future Directions of GeoAI Research

Current Status and Trends of Artificial Intelligence Geodesy

A Brushstroke Style Transfer Method from Painting to Map

Safety Supervision Analysis and Technology Research of Intelligent Connected Vehicle High-Definition Map

Related Author

ZHOU Yaozong

ZHANG Zhenyi

REN Xiaodong

LUO Xiaomin

ZHANG Weixing

LIU Mengjie

BI Haohang

LOU Yidong

Related Institution

Institute of Geodesy and Photogrammetry, Eidgenössische Technische Hochschule Zürich, Zürich

School of Geography and Information Engineering, China University of Geosciences

School of Surveying and Mapping, Wuhan University

Hubei Luojia Laboratory

GNSS Research Center, Wuhan University

AI问答

⁰