Handling Interferometric Data: Streamlining the Processing Flow

Transcription

1 Handling Interferometric Data: Streamlining the Processing Flow Paper 5 at Hydro8, 4 th November 2008 Tom Hiller, Advanced Products Manager, GeoAcoustics Ltd. WORLD CLASS through people, technology and dedication

2 Objectives of this presentation Provide an introduction to interferometric technology and illustrate the survey results obtained Describe interferometric data processing with examples from native and third party software Discuss the processing challenges experienced and outline possible future developments / 2 / 10-May-10

3 1. An Introduction to Interferometric Sonar Technology / 3 / 10-May-10

4 The Basics: what the interferometer does Side scan transmit geometry Simultaneous bathymetry and amplitude data products / 4 / 10-May-10

5 Also called: Phase Measuring Bathymetric Sonar (PMBS) Interferometric Multibeam Wide Swath Sonar Bathymetric Side Scan or just Interferometer Collects a swath of data under the vessel, and the survey area is painted in by running a survey pattern of parallel run-lines. / 5 / 10-May-10

6 How does an interferometer work? Multi-element receive array measuring phase differences. In the GeoSwath case the primary array consists of two transducers mounted to a V plate, each with multiple staves: a e d c b Bottom stave is transmitter, multiple receive elements. Uses phase differences to measure angle. Result: time series of angles (and amplitudes) / 6 / 10-May-10

7 Sonar Geometry a b c d e / 7 / 10-May-10

8 Comparisons with Beamformer: Raw data is a time series of angles vs angle series of times Very wide-swath bathymetry, even in shallow depths view angle over 240 degrees: sees shoal obstacles co-registered true digital side scan Swath width insensitive to roll Compact and robust transducers and electronics / 8 / 10-May-10

9 Principles of operation Transmit is short in time (a few cycles), wide across track and narrow along track (like a side scan) Multiple receiver staves within each transducer Phase measurement based on differential time Phase difference (Ф) / 9 / 10-May-10

10 Acoustic Theory 1 Phasor diagrams (right) illustrating the operation of a phase measuring sonar (left). Phase of signal on stave A (θa) results from path length Transmit (Tx) Scatterer (Sc) Receive A (RxA). / 10 / 10-May-10

11 Acoustic Theory 2: Noise Sources Phasor diagrams illustrating sliding footprint and sea noise effects / 11 / 10-May-10

12 / 12 / 10-May-10 Looking at the Raw Data in Detail: key feature: 1000s of data points/ping

13 / 13 / 10-May-10 The GeoAcoustics GeoSwath Plus

14 GeoSwath Deployment Examples Large and small vessels. Fixed, semi-permanent and temporary. Pole mount, hull mount. AUV and ROV mount. / 14 / 10-May-10

15 Suitable for very small survey vessels for lakes and rivers / 15 / 10-May-10

16 Hull mount examples Above image courtesy of Semantic / 16 / 10-May-10

17 / 17 / 10-May-10 Installations on larger vessels

18 / 18 / 10-May-10 GeoSwath on Minerva ROV

19 AUV module electronics for Hydroid Remus cm / 19 / 10-May-10

20 Producing Survey Data for Nearshore engineering works... nautical charting... River channel monitoring and dredge works... Environmental mapping / 20 / 10-May-10

21 Mapping in Shallow Water: 0-3m / 21 / 10-May-10

22 / 22 / 10-May-10 Port surveys

23 Harbour post-dredge survey using an interferometer mounted on a man-portable AUV 100m / 23 / 10-May-10

24 / 24 / 10-May-10

25 Deeper water application: mapping Swiss lakes Images courtesy of NGU, Norway / 25 / 10-May-10

26 Complex Seamless Geological Mosaics: combining backscatter and bathymetry information from the GeoSwath system using GeoTexture software / 26 / 10-May-10

27 Key capability: survey to the shoreline. Example from GeoSwath Plus Survey of Portsmouth Harbour, NH / 27 / 10-May-10

28 / 28 / 10-May-10 Run Lines in 3 areas; near shore, southern and northern

29 Looking at the Near shore Area 2km / 29 / 10-May-10

30 / 30 / 10-May-10 1m contours

31 / 31 / 10-May-10 A profile from the 1m depth contour

32 / 32 / 10-May-10 The GeoSwath Mosaic

33 2. Processing of Interferometric Data / 33 / 10-May-10

34 GeoSwath Native Data Processing summary / 34 / 10-May-10

35 GeoSwath Data Processing Key steps: Amplitude filtering Statistical filtering Binning / 35 / 10-May-10

36 / 36 / 10-May-10 Unprocessed, Filtered and Within-Ping Binned Data

37 Every point shown ( in each 1m bin, 80m swath, 7m deep) 1m Gridded data no interpolation or smoothing / 37 / 10-May-10

38 Reasons for the line-line and bin-bin centimetric repeatability: Key Characteristic: lots of data with accurate range and uncertainty in angle In each depth bin may be 100s of measurements with approximately independent noise and random distribution Mean of this many depths gives a very repeatable and reliable number Standard Error of the Mean = Standard Deviation of data divided by the square root of the number of data points This gives an estimate of the bottom using all the data / 38 / 10-May-10

39 Results: High Data Density Ensures Survey Quality Raw Data 1m Grid Bin size sonar footprint min. feature size. Data density > (or >>) 10 per bin. / 39 / 10-May-10 39

40 3. Interferometric Data Handling in Third Party Software / 40 / 10-May-10

41 Current GeoSwath Third Party Software Dedicated Interfaces Hypack Fledermaus QINSY CARIS SABRE any GSF-compatible software / 41 / 10-May-10

42 Routes From Ping to Processing Software Real-time control GeoSwath Plus On-line processing Transfer of flagged raw data via Ethernet (real-time) Hypack QINSY GeoSwath Plus Acquisition Hardware Data storage and transfer GeoSwath Plus Off-line processing Data Flagging and conversion to GSF Data filtering and conversion to reduced raw file Fledermaus SABRE Other GSF reader CARIS Further GS+ processing and export as xyz Other vendors / 42 / 10-May-10

43 Examples of Interferometric Surveys with Third Party Data Processing Lake Bourgne, Louisiana: 173 sq. nm, 1m-5m deep GeoSwath and SABRE (SAIC) Lyme Bay, UK: 80km wide bay surveyed from the 1m depth contour to 1km out to sea GeoSwath and Fledermaus (Halcrow & NetSurvey) Cape Fear, USA: Repeat surveys of shipping channel and shoreline to beach GeoSwath and CARIS (US Army Corps of Engineers) River Meuse, the Netherlands: River monitoring surveys GeoSwath and QINSY (RWS) / 43 / 10-May-10

44 Examples of Interferometric Surveys with Third Party Data Processing Lake Bourgne, Louisiana: 173 sq. nm, 1m-5m deep GeoSwath and SABRE (SAIC) Lyme Bay, UK: 80km wide bay surveyed from the 1m depth contour to 1km out to sea GeoSwath and Fledermaus Cape Fear, USA: Repeat surveys of shipping channel and shoreline to beach GeoSwath and CARIS River Meuse, the Netherlands: River monitoring surveys GeoSwath and QINSY / 44 / 10-May-10

45 Lake Borgne Debris Mapping: Overview SAIC, Marine Science and Technology Division (MSTD) under contract to NOAA Office of Coast Survey (OCS) / 45 / 10-May-10

46 Lake Borgne Debris Mapping: Overview Item detection and debris mapping Approx. depth range: 4 ft. to 20 ft. Delineate up to the 4 ft. depth curve Approx. survey area: sq. nm / 46 / 10-May-10

47 42 feet, Wooden hull Laffite skiff, inboard diesel, 2 feet draft (Vessel provided by Campo Marine) Retractable bow mount for GS+ V4 POS/MV, DGPS positioning SAIC s ISS-2000 Integrated Survey System / 47 / 10-May-10

48 / 48 / 10-May-10 SABRE processing (SAIC)

49 Examples of Interferometric Surveys with Third Party Data Processing Lake Bourgne, Louisiana: 173 sq. nm, 1m-5m deep GeoSwath and SABRE (SAIC) Lyme Bay, UK: 80km wide bay surveyed from the 1m depth contour to 1km out to sea GeoSwath and Fledermaus Cape Fear, USA: Repeat surveys of shipping channel and shoreline to beach GeoSwath and CARIS River Meuse, the Netherlands: River monitoring surveys GeoSwath and QINSY / 49 / 10-May-10

50 GeoSwath and Fledermaus bathymetry uncertainty / 50 / 10-May-10 Acknowledgements to Duncan Mallace, Netsurvey

51 / 51 / 10-May-10

52 Examples of Interferometric Surveys with Third Party Data Processing Lake Bourgne, Louisiana: 173 sq. nm, 1m-5m deep GeoSwath and SABRE (SAIC) Lyme Bay, UK: 80km wide bay surveyed from the 1m depth contour to 1km out to sea GeoSwath and Fledermaus Cape Fear, USA: Repeat surveys of shipping channel and shoreline to beach GeoSwath and CARIS River Meuse, the Netherlands: River monitoring surveys GeoSwath and QINSY / 52 / 10-May-10

53 GeoSwath in CARIS HIPS Implement processing and QC tools that reduce acquisition to processing ratios (Rate of Effort) Error Modeling and Propagation Apply Corrections Tide, Geodetic, Sound speed, Motion Surface Creation to Locate Errors Data Cleaning CUBE, Statistical, IHO, Area based *Designate Soundings Quality Control 3-D Fly Thru, Profiles, IHO QC / 53 / 10-May-10

54 Study Location: North Carolina Cape Fear / 54 / 10-May-10 Acknowledgements to Mike Forte, Field Research Centre, Duck NC

55 S/V Sea Quester 8 meter Aluminum Almar Draft 0.75 meter Geo Acoustics Swath Sonar (side -mount) Applanix POS MV wetmount Odom digibar velocimeter Hypack line navigation RTK logging / 55 / 10-May-10

56 Swath coverage 4 7.5kts / 56 / 10-May-10

57 / 57 / 10-May m

58 Examples of Interferometric Surveys with Third Party Data Processing Lake Bourgne, Louisiana: 173 sq. nm, 1m-5m deep GeoSwath and SABRE (SAIC) Lyme Bay, UK: 80km wide bay surveyed from the 1m depth contour to 1km out to sea GeoSwath and Fledermaus Cape Fear, USA: Repeat surveys of shipping channel and shoreline to beach GeoSwath and CARIS River Meuse, the Netherlands: River monitoring surveys GeoSwath and QINSY / 58 / 10-May-10

59 GeoSwath and QINSY -real time interface -into native QPS structure / 59 / 10-May-10

60 Data Comparison Same boat, same ancillaries, 2 fixed sonar mounts: GeoSwath 250kHz and Reson Seabat 8101 RTK GPS positioning and height control, Octans motion sensor Surveys run alternately, same 4-line pattern; GeoSwath-8101-GeoSwath-8101 Processing separately: GeoSwath via GS+ software 8101 via QPS QINSY GeoSwath via QINSY Data compared in final grid / 60 / 10-May-10

61 Interferometer vs Beamformer Comparison on the River Meuse (1m bins) / 61 / 10-May-10

62 Detailed comparisons of 300m profiles GeoSwath1 vs GeoSwath2 Beamformer1 vs Beamformer2 GeoSwath1 vs Beamformer1 GeoSwath2 vs Beamformer2 Conclusion: GeoSwath and Beamformer results are as repeatable as each other via either processing route / 62 / 10-May-10

63 GeoSwath and Hypack GPS PPS/ZDA, Position Hypack PPS box GeoSwath Data Ethernet Hypack Helmsman s display Control Port & Stbd Sonar, Attitude Sensor, MiniSVS Data stored in HS2 & HSX files / 63 / 10-May-10

64 A generic processing path: GeoSwath and CUBE With acknowledgements to Brian Calder, UNH / 64 / 10-May-10

65 The General Context Of course these also apply to any data processing: MBES, SBES, LIDAR, etc / 65 / 10-May-10

66 Challenges with third party approaches Implementing appropriate processing for interferometric data (and new data attributes) Data volume & data transfer: conversion overhead Data density & file sizes: storage issues Re-processing issues Traceability and quality control Handling uncertainty / 66 / 10-May-10

67 Approaches being taken to ease the pingto-chart data flow Sonar-specific attributes should be added by the acquisition system There should be appropriate attribute handling by the third party software There needs to be agreed data transfer & storage conventions (i.e. GSF) Vendors must implement smooth conversion to appropriate formats Availability of appropriate re-processing tools Proving and implementation of the uncertainty models / 67 / 10-May-10

68 Conclusions: Interferometers have been shown to be capable of providing very high resolution survey data Interferometric data is being merged into the standard processing paths Users are now driving the further streamlining of the processes / 68 / 10-May-10

69 Kongsberg Maritime / 69 / 10-May-10