Surveillance Camera-based Rainfall Estimation
Xing WANG (Ph.D. Candidate) 1,2,3
Prof. Xue-jun LIU1,2
(Head of Video-GIS team: http://schools.njnu.edu.cn/geog/)
1. School of Geography, Nanjing Normal University, China
2. Key Laboratory of Virtual Geographic Environment (Nanjing Normal University),
Ministry of Education, China
3. Department of Geography and Regional Research, University of Vienna, Austria
Masaryk University
10/03/2022
1. Background
2. Video-based rainfall estimation
3. Audio-based rainfall estimation
4. Conclusion
2021.07.10 Ahrweiler，German
Links:
https://www.dw.com/zh/%E5%BE%B7%E5%9B%BD%E8%A5%BF%E9%83%A8%E6%B4%AA%E7%81%BE%E6%83%A8%E7%83%88%E9%98%B2%E7%81%BE%E7%B3%BB%E7%BB%9F%E5%93%AA%E5%8E%BB%E4%BA%86/a-58316881
https://www.climate.gov/news-features/event-tracker/global-warming-increased-risk-intensity-louisianas-extreme-rain-event
2016.08.15 Louisiana，USA2021.07. 20 Zhengzhou，China
2020.08. 18 Chongqing，China
(a) number of disasters, (b) number of deaths and (c) economic losses during 1970–2019 from all hazards.
published by World Meteorological Organization (WMO).
Reasons for more frequent extreme rainfall events:
1) Global warming
2) Urbanization
https://www.worldweatherattribution.org/heavy-rainfall-which-led-to-severe-flooding-in-western-europe-made-more-likely-by-climate-change/
Ground-based rainfall observation
Sheffield J, Wood E F, Pan M, et al. Satellite remote sensing for water resources management: Potential for supporting sustainable development in data‐poor regions[J]. Water Resources
Research, 2018, 54(12): 9724-9758.
Ben H.P. Maathuis. Constraints and opportunities for Water Resources Monitoring and Forecasting using the Triple Sensor approach. 2018
https://public.wmo.int/en/resources/bulletin/hydrological-data-exchange
About 9000 at 2005, 10359 at 2021.
Data collected by：WMO
Problems:
1). Reduction in the number of observation resources.
Ground-based rainfall observation
WMO Secretary-General Prof. Petteri Taalas:
“In an era of cutting-edge satellite technology and artificial intelligence, there are countries which still
lack basic rain gauges”
2). Unevenly distribution of observation resources;
Shortcomings: insufficient spatial representation
Radar & Satellite
Remote Sensing-based rainfall observation
Shortcomings:
1) need for ground-based measurements for correction
2) not satisfied for the fast hydrological response applications
Lack Spatial and Temporal resolution
Rainfall audio
Microwave links
Rain gauge from car
Ground-based, low cost, high resolution
WMO, Intergovernmental Hydrological Programme (IHP), NOAA, NASA
> 400 million surveillance cameras in ChinaHardware
Spatial resolution
https://www.comparitech.com/vpn-privacy/the-worlds-most-surveilled-cities/
Video/audio transmission by 4G and 5G
Video/Audio analysis and processing
Software
Temporal resolution
Video-GIS team
j
i
y
x
z O(i0,j0)
(x0,y0,f)
H
X
Y
y
x
AQI & PM 2.5 estimation
Video-GIS team
Air pollution modeling in urban street canyons
Vehicles are the main source
of air pollution
1. Background
2. Video-based rainfall estimation
3. Audio-based rainfall estimation
4. Conclusion
2.1 Rainfall calculation based camera collaboration
Rain streaks: the blurred pixels by raindrops
Rainfall intensity: is defined as the ratio of the total amount of rain (rainfall depth) falling during a
given period. It is expressed in depth units per unit time, usually as mm per hour (mm/h).
2.1 Rainfall calculation based camera collaboration
Step 1: Rain streaks detection from videos;
Step 2: Raindrop size and speed calculation;
Step 3: Rainfall intensity estimation.Camera imaging model
(Pinhole imaging)
rain streaks: the blurred pixels by raindrops
Shortcomings of surveillance Video-based Rain Gauge (VRG):
Non-cooperative surveillance scenarios Moving objects
Precision control model of rainfall inversion based on VRGs collaboration:
Experiments (in Nanjing Normal University)
Wang X, Wang M, Liu X, et al. A novel quality control model of
rainfall estimation with videos–A survey based on multi-surveillance
cameras[J]. Journal of Hydrology, 2022, 605: 127312.
Simulation Experiments (in Nanjing)
Wang X, Wang M, Liu X, et al. A novel quality control model of rainfall
estimation with videos–A survey based on multi-surveillance cameras[J].
Journal of Hydrology, 2022, 605: 127312.
1. Background
2. Video-based rainfall estimation
3. Audio-based rainfall estimation
4. Conclusion
3.1 Parallelizing rainfall level classification network
(i) The acoustic features of different rainfall-level
audio are similar;
(ii) Background noise in surveillance sound space
significant.
3.1 Parallelizing rainfall level classification network
(i) The proposed algorithm achieves optimal
performance compared to some existing relevant
models, indicating that the proposed algorithm can
effectively determine the rainfall level from ordinary
surveillance audio.
(ii) However, there is still much room for
enhancing. In particular, the classification of “no_rain”
and “small rain” scenarios and the distinction of
“violent rain” and “heavy rain” need to be improved.
Wang X, Wang M, Liu X, et al. Rainfall observation using surveillance audio[J]. Applied Acoustics, 2022, 186: 108478.
1. Background
2. Video-based rainfall estimation
3. Audio-based rainfall estimation
4. Conclusion
4 Conclusion
GIS
Video
Video-GIS
2) Video + GIS Keeping GIS active
1) Single camera Surveillance camera-sensor network
Xing WANG
Email: jwangxing0719@163.com
Masaryk University
10/03/2022