際際滷

際際滷Share a Scribd company logo
A Smart City-Smart Bay Project:
Establishing an integrated water monitoring system for decision
support in Dublin Bay
Fiona Regan, Timothy Sullivan, Ciprian Briciu, Helen Cooney, Dian Zhang*, Edel
OConnor*, Noel OConnor*, Alan Smeaton*
Marine and Environmental Sensing Technology Hub (MESTECH), National Centre for Sensor Research
Dublin City University
*CLARITY Centre for Sensor Web Technologies, Dublin City University
Dublin, Ireland
Project	
 Ra+onale	
 
Design,	
 deployment	
 and	
 integra2on	
 of	
 an	
 autonomous	
 real-足2me	
 
mul2modal	
 sensing	
 network	
 for	
 improved	
 decision	
 making	
 
Research	
 Objec+ves	
 
	
 
≒ Improve	
 Water	
 quality	
 monitoring	
 
≒ Improve	
 discrete	
 sampling	
 regimes	
 
	
 
≒ Iden+fy	
 and	
 Improve	
 detec+on	
 of	
 Security	
 threats	
 
≒ Iden2fy	
 threats	
 to	
 health	
 
(microbial	
 and	
 other	
 pollutants)	
 
≒ Enhanced	
 Signal	
 processing:	
 
Develop	
 surrogate	
 measurements	
 
≒ Produce	
 Baseline	
 datasets	
 on	
 water	
 quality	
 
	
 
	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Study	
 site	
 
	
 
Methods	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions
Sensorcomm3 t sullivan
Current	
 and	
 future	
 Network	
 Distribu+on	
 by	
 2014	
 
River	
 Li鍖ey	
 
Dublin	
 Bay	
 
Dublin	
 City	
 Centre	
 
2	
 km
Pilot	
 Sites:	
 Malahide	
 and	
 
Poolbeg	
 Estuaries

	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Study	
 site	
 
	
 
Methods	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions
In-足situ	
 sensors	
 
≒ Mul+-足parameter	
 sondes	
 equipped	
 with	
 real-足+me	
 telemetry	
 systems	
 
	
 
≒ IP66-足Rated	
 outdoor	
 network	
 camera	
 
	
 
≒ Ini+al	
 systems	
 deployed	
 in	
 October	
 2010	
 -足	
 August	
 2013:	
 
≒ Circa	
 2.5	
 million	
 images	
 have	
 been	
 collected	
 
≒ Circa	
 500,000	
 	
 individual	
 sensor	
 measurements	
 
	
 
	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Study	
 site	
 
	
 
Methods	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions

	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Study	
 Site	
 
	
 
Methods	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions	
 
Duc+ng	
 of	
 marina	
 structure	
 
220V	
 power	
 supply	
 
Commercial	
 telemetry	
 solu+on	
 box

Data	
 Analy+cs	
 
≒ Machine	
 learning	
 objec+ves:	
 automated	
 detec+on	
 and	
 trajectory	
 of	
 
vessels	
 
≒ Automated	
 Turbidity	
 event	
 detec+on	
 	
 pixel-足based	
 adap+ve	
 segmenter	
 
method	
 
≒ Salinity	
 predic+on	
 using	
 mul+ple	
 data	
 sources	
 (+de,	
 鍖ow,	
 weather	
 data)	
 
using	
 regression	
 tree	
 approach	
 
≒ Shipping	
 ac+vity	
 +	
 turbidity:	
 predic+on	
 of	
 sampling	
 +mes	
 and	
 microbial	
 
contamina+on	
 	
 separa+ng	
 natural	
 events	
 from	
 anthropogenic	
 events	
 
≒ Water	
 level	
 predic+on	
 
≒ Security	
 Threats:	
 Unauthorized	
 shipping	
 
	
 
	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Methods	
 
	
 
Study	
 Site	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions
1 Aug
2 Aug
3 Aug
4 Aug
5 Aug
6 Aug
7 Aug
0
5
10
15
20
25
30
Turbidity 2 m
Turbidity 4 m
Turbidity(NTU)
Date 2012
Detec+ng	
 and	
 automa+ng	
 turbidity	
 event	
 detec+on	
 
	
 
	
 
	
 
Introduc+on	
 
	
 
Ra+onale	
 
	
 
Methods	
 
	
 
Study	
 Site	
 
	
 
Instrumenta+on	
 
	
 
Data	
 analysis	
 
	
 
Results	
 
	
 
Conclusions
Sensorcomm3 t sullivan
Sensorcomm3 t sullivan
Sensorcomm3 t sullivan
Sensorcomm3 t sullivan
Sensorcomm3 t sullivan
Sensorcomm3 t sullivan
Conclusions	
 
≒ An	
 extensive	
 network	
 of	
 both	
 in-足situ	
 aqua+c	
 sensors	
 and	
 visual	
 sensing	
 
systems	
 have	
 been	
 and	
 are	
 in	
 process	
 of	
 deployment	
 in	
 Dublin	
 Bay	
 
	
 
≒ The	
 network	
 has	
 already	
 had	
 demonstrable	
 impact	
 on	
 monitoring	
 and	
 
understanding	
 dynamic	
 processes	
 in	
 Dublin	
 Bay	
 
≒ Incorpora+on	
 of	
 visual	
 sensing	
 nodes	
 into	
 the	
 network	
 has	
 proven	
 
advantageous	
 
≒ Machine	
 learning	
 and	
 increased	
 compu+ng	
 power	
 has	
 aided	
 in	
 data	
 
analysis	
 	
 future	
 work	
 will	
 emphasize	
 data	
 analy+cs	
 	
 
	
 
≒ Challenges	
 remain:	
 Increased	
 spa+al	
 coverage,	
 Biofouling!,	
 Cost,	
 
Transla2on	
 of	
 data	
 into	
 knowledge
Thank	
 You!	
 Ques+ons?	
 
Contacts:	
 +m.sullivan@dcu.ie;	
 鍖ona.regan@dcu.ie

More Related Content

Sensorcomm3 t sullivan

  • 1. A Smart City-Smart Bay Project: Establishing an integrated water monitoring system for decision support in Dublin Bay Fiona Regan, Timothy Sullivan, Ciprian Briciu, Helen Cooney, Dian Zhang*, Edel OConnor*, Noel OConnor*, Alan Smeaton* Marine and Environmental Sensing Technology Hub (MESTECH), National Centre for Sensor Research Dublin City University *CLARITY Centre for Sensor Web Technologies, Dublin City University Dublin, Ireland
  • 2. Project Ra+onale Design, deployment and integra2on of an autonomous real-足2me mul2modal sensing network for improved decision making Research Objec+ves ≒ Improve Water quality monitoring ≒ Improve discrete sampling regimes ≒ Iden+fy and Improve detec+on of Security threats ≒ Iden2fy threats to health (microbial and other pollutants) ≒ Enhanced Signal processing: Develop surrogate measurements ≒ Produce Baseline datasets on water quality Introduc+on Ra+onale Study site Methods Instrumenta+on Data analysis Results Conclusions
  • 4. Current and future Network Distribu+on by 2014 River Li鍖ey Dublin Bay Dublin City Centre 2 km
  • 5. Pilot Sites: Malahide and Poolbeg Estuaries
  • 6. Introduc+on Ra+onale Study site Methods Instrumenta+on Data analysis Results Conclusions
  • 7. In-足situ sensors ≒ Mul+-足parameter sondes equipped with real-足+me telemetry systems ≒ IP66-足Rated outdoor network camera ≒ Ini+al systems deployed in October 2010 -足 August 2013: ≒ Circa 2.5 million images have been collected ≒ Circa 500,000 individual sensor measurements Introduc+on Ra+onale Study site Methods Instrumenta+on Data analysis Results Conclusions
  • 8. Introduc+on Ra+onale Study Site Methods Instrumenta+on Data analysis Results Conclusions Duc+ng of marina structure 220V power supply Commercial telemetry solu+on box
  • 9. Data Analy+cs ≒ Machine learning objec+ves: automated detec+on and trajectory of vessels ≒ Automated Turbidity event detec+on pixel-足based adap+ve segmenter method ≒ Salinity predic+on using mul+ple data sources (+de, 鍖ow, weather data) using regression tree approach ≒ Shipping ac+vity + turbidity: predic+on of sampling +mes and microbial contamina+on separa+ng natural events from anthropogenic events ≒ Water level predic+on ≒ Security Threats: Unauthorized shipping Introduc+on Ra+onale Methods Study Site Instrumenta+on Data analysis Results Conclusions
  • 10. 1 Aug 2 Aug 3 Aug 4 Aug 5 Aug 6 Aug 7 Aug 0 5 10 15 20 25 30 Turbidity 2 m Turbidity 4 m Turbidity(NTU) Date 2012 Detec+ng and automa+ng turbidity event detec+on Introduc+on Ra+onale Methods Study Site Instrumenta+on Data analysis Results Conclusions
  • 17. Conclusions ≒ An extensive network of both in-足situ aqua+c sensors and visual sensing systems have been and are in process of deployment in Dublin Bay ≒ The network has already had demonstrable impact on monitoring and understanding dynamic processes in Dublin Bay ≒ Incorpora+on of visual sensing nodes into the network has proven advantageous ≒ Machine learning and increased compu+ng power has aided in data analysis future work will emphasize data analy+cs ≒ Challenges remain: Increased spa+al coverage, Biofouling!, Cost, Transla2on of data into knowledge
  • 18. Thank You! Ques+ons? Contacts: +m.sullivan@dcu.ie; 鍖ona.regan@dcu.ie