TeamSurv presentation to UK Hydrographic Society, 2011
1 like213 views
This document discusses crowdsourced hydrographic surveying using data collected from boats equipped with standard instruments like GPS, depth sounders, and compasses. Many small boats can collectively build up good coverage over time. The data is processed to apply corrections and synthesize results, creating accurate depth models and charts. Quality is ensured through adaptive gridding, robust statistics, and outputting a quality measure. Initial testing found crowdsourced data to match well with multibeam data. The approach aims to cost-effectively map large areas by leveraging contributions from recreational boaters.
More Related Content
Similar to TeamSurv presentation to UK Hydrographic Society, 2011 (20)
![Course_List S Raisey-Skeats[1]](https://cdn.slidesharecdn.com/ss_thumbnails/060e891b-9c7c-4c60-b81f-ab207a2a2587-150522113649-lva1-app6892-thumbnail.jpg?width=560&fit=bounds)
TeamSurv presentation to UK Hydrographic Society, 2011
- 1. Tim Thornton – Smartcom Software
- 2. • Exceptional accuracy and coverage
- 3. • Exceptional accuracy and coverage • Quick to survey an area
- 4. • Exceptional accuracy and coverage • Quick to survey an area • Small number of highly skilled operators
- 5. • Exceptional accuracy and coverage • Quick to survey an area • Small number of highly skilled operators • Large investment in equipment and staff
- 6. • Exceptional accuracy and coverage • Quick to survey an area • Small number of highly skilled operators • Large investment in equipment and staff • Expensive for customers
- 7. • Exceptional accuracy and coverage • Quick to survey an area • Small number of highly skilled operators • Large investment in equipment and staff • Expensive for customers • Many surveys used older technology
- 8. • Exceptional accuracy and coverage • Quick to survey an area • Small number of highly skilled operators • Large investment in equipment and staff • Expensive for customers • Many surveys used older technology • New Zealand charts only stopped using data from Captain Cook (~ 1750) a few years ago!
- 9. • Many boats logging data as they go about their normal business
- 10. • Many boats logging data as they go about their normal business
- 11. • Many boats logging data as they go about their normal business • Build up good coverage with time
- 12. • Many boats logging data as they go about their normal business • Build up good coverage with time • Use standard instruments – GPS, depth sounder, compass
- 13. 0 10 20 30 40 50 60 -0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 %frequency Depth delta, m • Many boats logging data as they go about their normal business • Build up good coverage with time • Use standard instruments – GPS, depth sounder, compass • Accuracy and coverage good enough for most purposes up to ~ 100m depth, and better than old surveys
- 14. 0 10 20 30 40 50 60 -0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 %frequency Depth delta, m • Many boats logging data as they go about their normal business • Build up good coverage with time • Use standard instruments – GPS, depth sounder, compass • Accuracy and coverage good enough for most purposes up to ~ 100m depth, and better than old surveys • Low cost, low investment solution
- 15. • Can crowdsourcing be used as part of the hydrographer’s toolkit? • Where does it fit with other technologies? • How accurate is it?
- 16. Autonomous Remote Global Underwater Surveillance • US based, using small craft around Annapolis with WiFi link to send data ashore • Funded by US Navy to develop unmanned remote controlled survey vessels for military use Software package for commercial fishing • Users can upload logged data to the server for sharing with other users • Simple 5m sq. binning and averaging, no tidal or speed of sound corrections • Augmented with data from other sources
- 17. Open Street Map User Generated Content
- 18. Autonomous Remote Global Underwater Surveillance • US based, using small craft around Annapolis with WiFi link to send data ashore • Funded by US Navy to develop unmanned remote controlled survey vessels for military use Software package for commercial fishing • Users can upload logged data to the server for sharing with other users • Simple 5m sq. binning and averaging, no tidal or speed of sound corrections • Augmented with data from other sources
- 19. Software for leisure fishing and divers • Uses NMEA data or Lowrance/Eagle/Humminbird sonar log files • Can do bottom hardness from sonar data • Basic water level corrections • Simple data logging and map creation Software package for commercial fishing, with Bathy3D depth module • Users can switch on depth logging • Simple depth tracks, no tidal or speed of sound corrections • Depths used to augment charted depths
- 20. Different crowdsourcing projects use the crowd in different ways. Here are the key points for TeamSurv participants: Must have a boat and suitable instruments Motivated by altruism, fun and community Minimal skill or effort required to participate Data is aggregated and processed by us on server before publication Loggers engaged by free access to our data, other users will be charged when the FP7 project closes
- 21. Processing on server is key to data quality. The greater the data density, the more accurate our results. This differentiates us from projects like OpenStreetMap, where data from each user is basically treated discretely We have more in common with mass traffic speed models like TomTom
- 23. On Board •Calibration •Data Logging Track Ingest •Upload to web •Remove bad data •Ingest data Corrections •Geometry •Speed of sound •Sea level •(Vessel motions) Synthesis •Combine data •Remove outliers Chart Creation •Geostatistics •Gridding •Surface generation Error analysis Data display
- 24. Must be painless for user to install and use Hardware logger – 2 channels NMEA or Seatalk to USB Software logger works in parallel with PC navigation software Now getting support from equipment manufacturers:
- 25. GPS Tests carried out on 10 typical light marine and leisure receivers Even the cheapest GPS receivers accurate to 2m rms2 EGNOS found to have little or no benefit Depth Sounders No tests carried out as yet – on the to do list Manufacturers typically state depth accurate to 1% Transducer beam angle 6 – 15 degrees
- 26. Two levels of interpolation Primary: Tide height predictions Secondary: Residuals from tide gauges Speed of sound is similar, using salinity and sea temperature data
- 27. • much less critical than for multi-beam, as beam is ~ vertical • wider beam angle helps accommodate motions – we are measuring this • processing a large amount of data minimises the effect of motions • at present just modelled in error analysis, as no data from NMEA stream • analyses based on studies of vessel motions, plus environmental data Motion Vessel dynamics Waves Heave Sinkage in displacement boats and rising of planing boats, typically < 10cm Mean heave is centered on static WL so no nett effect Pitch Bow rises with speed, normally <5°, absorbed by beam angle More bow up than down causes slight over reading Roll In sailing boats causes over reading, mean error 1%. Power vessel zero sum In strong winds bias to one side causes over reading
- 29. Adaptive grid generated according to the data density and consistency Only generate depths where data meets a quality threshold Non-normal depth distribution, with fat tail to deeper depths, so use of robust statistics 0 10 20 30 40 50 4.0 4.5 5.0 5.5 Frequency Depth (m) Depth vs. Frequency Number of points =2 67
- 30. Calculate and output a quality of data measure as well as a depth value Quality of data is dependent upon number of points in the grid square and also the spread of the data
- 33. MultibeamTeamSurv Mean depth difference: 0.04m Standard deviation: 0.21m
- 34. Depth differences. 0.1m contours
- 35. 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00 4.00 5.00 6.00 7.00 8.00 TeamSurv Multibeam 0 10 20 30 40 50 60 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 -0.7 -0.8 %freq Depth difference (m) Depth difference histogram
- 36. Now: PDF Shape X, y, z CSV BSB Garmin Magellan Later: Web Server e.g. WMS/WFS Navico Fugawi S-57, S-100, S-102 …
- 37. It takes time to recruit the crowd Important to minimise barriers to participation A two way process – must be a benefit for the participant as well as for the project Be aware of cultural issues
- 38. Extend out of initial trials areas Change of depth with time, e.g. to monitor silting or scouring Additional data, e.g. tidal streams and currents, sea surface temperature, winds Commercialisation to support the project
- 39. Tim Thornton TeamSurv Project Coordinator [email protected] www.teamsurv.eu