Administrator, 1016304ar
Administrator, 1016304ar
Administrator, 1016304ar
By Doug Cronin, Mel Broadus and Barbara Reed, Naval Oceanographic Office, Stennis Space
Center, USA, John Shannon Byrne and Walter Simmons, Science Applications International
Corporation, USA and Lindsay Gee, Interactive Visualization Systems Inc., Canada
The U.S. Naval Oceanographic Office’s must be validated for use in bathyme
(NAVOCEANO’s) six T-AGS 60 class and try, hydrography and imagery products.
one T-AGS 51 class survey vessels pro The amount of data to be processed
vide multi-mission survey capabilities will continue to increase to more than
using the latest generation high-resolu- 2000 times today’s data quantities.
tion multibeam and digital side-scan To meet these challenges, the NAV-
sonar systems, along with state-of-the- OCEANO developed area-based-editing
art positioning, orientation, oceano (ABE) tools have been integrated with
graphic profiling and ancillary sensors. Science Applications International
Each of these platforms accommo Corporation’s (SAIC’s) Survey Analysis
dates two 34-foot Hydrographic Survey and area Based EditoR (SABER) soft
Launches (HSLs) outfitted with shallow ware package and the Interactive
water multibeam and digital side-scan Visualization System s Inc. (IVS)
systems. These capabilities provide us Fledermaus software package. These
with an unprecedented ocean observ integrated packages have fielded the
ing and mapping capability that pro tools for data cleaning, validation, cor
duces massive amounts of data that rection and quality control. Three-
Gbvtes/Year
2500
2000-1
1500
1000
500
ISS-60 Overview
Introduction
The Integrated Survey System (ISS-60) is a distrib
High-resolution multibeam and digital side scan uted network-based software system installed in all
sonar systems, along with state-of-the-art position NAVOCEANO’s oceanographic and hydrographic
ing and attitude sensors and other ancillary sen vessels. It supports a variety of needs including
sors on our ships and HSLs, have provided the high-resolution seafloor mapping, physical and
NAVOCEANO with one of the best equipped survey chemical oceanography and precision station keep
fleets in the world. With this tremendous increase ing accommodating over-the-side sensor deploy
in bottom-mapping capability comes a correspon ments. The ISS-60 is easily configured to meet the
ding increase in the data that must be validated for needs of various user groups, provides a flexible
inclusion into the shallow and deep water hydrog foundation for present operations and allows easy
raphy, bathymetry and imagery products that we expansion to support potential requirements for
produce. With our ships operating 24 hours a day, future missions.
7 days a week and 10 months a year or more, NAV ISS-60 is an established, reliable, easily modified
OCEANO will soon continually collect more data shipboard computer system for navigation, hydro
than anyone else in the world will. We face a poten graphy, bathymetry and oceanography data acqui
tial of a 22-fold increase in the amount of bathy sition and processing. It is an integrated system
metric data to be processed - a maximum of over hosted on PC-compatible computers providing a
2.75 terabytes per year versus the recent level of combination of mission planning functions, timing
125 gigabytes per year. This figure rises to roughly coordination, real-time data acquisition, integrated
300 terabytes per year when multibeam imagery error checking and diagnostics, sensor and ship
and digital side scan sonar are included. Figure 1 controls, real-time data quality assurance and
depicts the increase in bathymetric/hydrographic sharable data control and display.
data volumes. The 30 years of multibeam data The modular ISS-60 software design allows for
prior to 1997 barely registers on the graph! [1] scalable configurations with the T-AGS 60 and
NAVOCEANO’s survey fleet conducts world-wide T-AGS 51 class ship configurations distributed
survey operations. Our approach for most effi across four computer systems and with the HSL
ciently generating products and information from configuration fully functional on a single computer
this survey data includes; (1) the integration of system. Each of these configurations is operated
onboard systems to centralise operation, data by a single operator who controls survey opera
management and QC, (2) the integration of pro tions and monitors the quality and coverage of the
cessing and Quality Assurance (QA) into the real navigation and environmental survey data.
time data acquisition environment, (3) the automa The operator interface is consistent across the
tion of historically labour intensive processing with T-AGS 60 class, T-AGS 51 class and HSL system
integrated quality assurance practices and (4) the configurations. This commonality across NAV
smart data-basing of the information produced, OCEANO platforms allows the use of one set of
while maintaining an effective approach for storing operational procedures across the fleet and sim
the full-resolution supporting datasets. plifies hydrographic personnel training.
The T-AGS Mission Electronic Suite (MES) is the col
Under current NAVOCEANO procedures, the major lection of shipboard oceanographic and hydrograph
blocks of the hydrographic workflow are: (1) defini ic related computer systems, network systems, sen
tion of survey requirements, (2) creation of project sor systems and peripherals. The ISS-60 system
plan, (3) definition of survey collectables, (4) col software and computer hardware provide the opera-
WS 2 WS 6 WS 8 W S 12 B rid g e X term
(L inu x) (Linux) (Linux) (L inux) RD I A DCP
uss
G eneral-P urpose Stations
O ff Shin C O M M S
Ocean graphic ÇXBT, C TD . .. )
— Q IC G eoD A S S S S
K lein 5000 S SS
Prin ters
Plo tters
H SL «1 plug-in
I l S L n l plug-in
Figure 2: ISS-60 and Mission Electronics Suite (MES) block diagram for T-AGS 60 class ship configuration
Plu g
c o n n e c tio n
to ship
M E S LA N
Figure 3: High-level
ISS-60 and Mission
Electronics Suite
(MES) block
diagram for HSL
configuration
tional focal point for the personnel conducting sur operational procedures are the same between the
vey operations. Figure 2 provides a high-level ship and the HSL configurations, even though the
overview of the T-AGS 60 class ship MES. mix of sensors is different. The processes and pro
cedures for tasks are basically identical between
In this configuration, the ISS-60 resides on four the ship and the HSL.
computers refered to as the central suite.
Redundancy in real-time data recording is support In addition to supporting a wide range of real-time
ed by simultaneously recording two copies of all related data acquisition and survey control
acquired data, one to a file system on c c l and a requirements, ISS-60 includes a survey mission
second to a file system on cc2. planning module that can be used in both the ship
ISS-60 interfaces with the majority of the perma board and the office environment.
nently installed scientific sensors and provides
generic data distribution and remote display serv
ices for roll-on/roll-off systems. Two Network- Mission Planning
Attached Storage (NAS) systems, each with 1.3
terabytes of disk capacity, provide the storage for To plan a survey operation effectively, the hydro-
all acquired data. The ISS-60 system can be con grapher must be able to view available information
figured to automatically archive logged data files about the survey area in a Geographic Information
from the primary logging disk to the NAS on a con Systems (GlS)-like environment. The ISS-60 Survey
figurable periodic basis, or on operator request, Planning module provides for viewing information
such as at the end of each survey line. from low- resolution sources such as the World
Each NAVOCEANO ship accommodates two HSLs. Vector Shoreline and Digital Bathymetric Data Base
These boats typically support daylight operations contours of the ocean and from high-resolution
and are brought back onboard at the end of the sources such as raster digital charts, digitised
day. Figure 3 provides a high-level overview of the shoreline and soundings, prior surveys and gridded
HSL system configuration. Plug-in network connec depth layers. Survey transects can be planned
tions facilitate data transfer on and off the HSLs. directly over these sources.
The ISS-60 Graphical User Interfaces (GUIs) and In some cases, large datasets are viewed best in
Figure 4: Areas
(blue) and tidal
zones (red) over
chart
Figure 5: Survey lines planned
over chart (Green indicates the
line has been run)
GIS or drafting environments (Arc Info, ArcView, software package or created interactively in ISS-60
AutoCAD etc.). These packages allow rapid review Survey Planning by typing in coordinates or by
of the existing data, such as side scan sonar selecting points on the screen. The World Vector
mosaics, for planning additional survey lines to Shoreline is shown in green.
complete required coverage. Survey transects
planned in these systems can be imported directly The hydrographer evaluates the requirements of
into the ISS-60 for completion of planning, sched the survey for coverage, object detection and the
uling and surveying. accuracy of position and depth against the antici
Initial planning for a survey operation is normally pated water depth and configuration of the survey
carried out in the office with updates and revisions area when planning the spacing and orientation of
being done in the field as necessary to fulfil the survey lines to be run. The ability to plan while
requirements as the survey progresses. An area’s overlaying an existing chart or prior survey
tidal characteristics are analysed and tidal correc enhances efficient and safe vessel operations
tion zones and parameters are defined for applica through selection of orientation, equipment and
tion to observed water levels for various portions speed. At the same time, the hydrographer can
of the survey. Oceanographic analysis may also plan and visualise cross lines necessary to prop
result in planning of sound speed profile zones. erly evaluate survey adequacy. Figure 5 illustrates
The pre-survey analysis includes expected tempo planned survey lines.
ral/spatial temperature and salinity variability and
their potential impact on the sound velocity struc It is seldom possible to have a water level gauge in
ture. This will include freshwater runoff, wind pat every portion of the survey, so a scheme must be
terns, tidal influences and diurnal warming and devised for correcting the observed values to rep
cooling. resentative values for each zone. Historic tide data
Figure 4 illustrates the layout of the survey opera and oceanographic analysis are used to determine
tion area (blue), special interest areas (blue) and co-tidal and co-range lines from the planned water
tidal zones (red) overlain on a chart. These areas level stations. Zones are defined with parameters
and zones can be imported from an external GIS to keep the water level corrections for each zone
within the allocated accuracy
budget.
The planner reviews historic
survey and charting data,
selects items for special
investigation, estimates the
abundance of obstructions
and features (manmade de
bris or natural), selects tide
station locations and makes
trade-offs to optimise the
equipment assigned and the
survey effort required.
During the planning stage,
Hydrographic Project Speci
fications are developed. This
document provides guidance
for planning the survey to
meet the data standards
required. The specifications
include but are not limited to
charts of the area, prior sur
veys, families of survey lines,
line sequence plans, line Figure 6: Survey schedule and distance estimate
spacing, survey classification,
water level gauge installation plan, tidal zone bound The ISS-60 system control program, shown in Figure
aries and associated correctors, reference bench 7, provides access to the system user configurable
mark and horizontal control mark descriptions and parameters. A separate configuration file is main
values, navigation and positioning plan and the qual tained for each survey platform to manage details
ity control quality assurance plan. The Specifications such as which sensors are active, the sensor allo
are assigned to the hydrographic party for execution. cation to network and/or serial ports, frequency of
data logging, frequency of file name change interval
Figure 6 illustrates survey lines and a line-sequenc and various tolerance monitoring parameters includ
ing plan with adequate spacing for turns and pro ing GPS Dilution Of Precision, data timeouts, across-
vides the on-line and off-line distance required to track error, etc. The lower portion of the screen
complete the schedule. As the survey progresses shows active sensors and active real-time process
the schedule also shows the on-line distance com ing programs. Each icon on the lower portion of the
pleted. screen is colour-coded to indicate the current sta
tus. Green icons mean the task is operational with
expected inputs and output. Red icons mean the
Real-time Survey Operations task is running but has encountered some type of
failure. The type of failure is reported to the ISS-60
The ISS-60 supports underway, station keeping central message display and recording facility. The
and line-following operations. Each type of opera error messages can be referenced to assess and
tion may have a number of different activities; trouble shoot any faults that may be occurring.
these include deep-water bathymetric, hydrograph Yellow icons mean the task is operational with
ic, geophysical, acoustics, biological and physi expected inputs and outputs, however the task has
cal/chemical oceanographic activities. To support been configured to not record any data.
these operations, the ISS-60 can be configured in
various ways for specific data collection rates, real The bathymetry data acquired from the multibeam
time processing and data visualisation required by sonar systems are saved in a Generic Sensor
the mission scenario. Format (GSF). [2] All other logged data files are
&
F ile Survey C o n tro l Conf i g u r a t io n S e n so rs L ogging
V e s se l F i l e : t 6 0 .v s l
C o n fiq F i l e : d e fa u lt_ 6 0 _ A .c fg Data Gap: 2 0 0 .0 0 m
saved in space delimited ASCII format. All logged igation manager module. An example layout of the
data file names are produced by the ISS-60 soft real-time navigation manager module is shown in
ware and follow a standard naming convention that Figure 8. Currently enabled display layers include
includes a two-character platform designator, a survey line waypoints, survey lines, survey line
two-digit year, a three-digit day, a three-character labels, latitude/longitude graticule, ship icon, cur
sensor identifier and a two-digit sequence number. rent line, next line, tide zones (red) and existing
File names can be automatically changed based on multibeam bathymetric coverage. A configurable
elapsed time, current file size and operator status area at the bottom of the display depicts
request. Two identical copies of the acquired data pertinent navigation information.
are logged in real-time. The primary dataset is The line-sequencing plan (Figure 6) defines the
recorded to the data disk on c c l. The secondary intended ship routing as a series of waypoints and
dataset is simultaneously recorded to the data survey transects that specify the order in which the
disk on cc2. The separation across two computer survey will be conducted. Once the plan has been
systems provides for a high level of fault tolerance. loaded in the real-time environment, the line
The outputs of the ISS-60 survey planning effort schedule can be adjusted as required by the
described above load into the ISS-60 real-time nav events of the survey on a line-by-line basis. All the
P»i«æ.|j|§i5s
Figure 8: ISS-60
Navigation
Manager Display
tApst s X doiktop (Ml t lL s Æ M Æ M k z LOOP.
«13 03.831(1
■
1 1 2
O' ,
43 03.82
F i
43 03.81
Wf
ii
WGS-84 U ÎM , Z one: 19 070 42.28W j 0/0 42.2/W H 070 42.26VV 0/0 42.25W |<1/0 42.24W j 0/0 42.23W
information developed in the planning phase can Sound velocity zone definitions from the planning
be layered on the real-time navigation screen. phase can be layered on the navigation manager
and helm displays. Sound Velocity Profiles (SVP)
The ISS-60 helm display (Figure 9) is hosted by a PC are reduced from profiling CTDs, XBT and XSSVs
or an Xterm located near the helmsman on the bridge. with the results written into the ISS-60 dataset.
A synchronised copy of this display is also visible and The resulting profile is viewed within ISS-60 and
fully controllable by the ISS-60 operator. This feature can be applied to the appropriate sonar systems
is particularly useful in assisting with communication and long-baseline navigation solutions on operator
between the bridge and the survey operation. request. Once an SVP has been applied, a record
of the profile is maintained with the applicable data
ISS-60 interface to the shipboard dynamic position for traceability. The current value from the sea sur
ing system and to the autopilot on the HSL provides face sound velocity probe is compared continually
steering guidance to maintain platform adherence to with the current sound velocity profile at the appro
the intended survey track. This is a key feature for priate depth. If the difference between the current
hydrographic survey operations, where line spacing surface velocity and the value from the last cast is
is stretched to most efficiently cover the planned greater than the set threshold, an operator alert is
area while ensuring that the required seafloor cover generated to suggest that a new profile should be
age criteria are satisfied. With reliable wide-area taken.
DGPS and a suitably configured interface with the The colour-filled bathymetry data (Figure 8) are gen
ship control systems, transect line following will typ erated during data acquisition to assess seafloor
ically be maintained to better than 5 metres of coverage and overall quality of the bathymeric
across-track error and in general is maintained to data. As data are received from the multibeam, the
better than 3 metres of across track error. soundings are binned into a regularly-spaced X, Y
The tide zone boundaries can be layered on the grid file supporting several update modes. The
navigation manager and helm displays. During data cells can be updated based on first value, last
acquisition predicted water levels can be applied to value, average value, minimum value, or maximum
the bathymetric data with applicable parameters value. In general, for real-time operations the last
for each tide zone. These predictions can be value option is most appropriate. Any gaps in cov
replaced with observed water levels during pro erage can be identified in the real-time environ
cessing in SABER. ment and in the processing environment down-
stream of data cleaning. Fill-in lines can be added graphic operations required from the DSSS sys
in either environment to complete coverage. tems is object locations, bottom composition
In addition to the geospatial coverage display, sev changes and coverage assurance.
eral swath oriented displays of the bathymetry data Targets are selected from the imagery data both in
are supported during data acquisition. These real-time and during post-time review. The target
include a scrolling colour fill, scrolling waterfall of location and other information about the target
across-track depth profile and scrolling 3D display such as dimensions, orientation, classification and
of the Total Propagated Error (TPE) estimates. operator comments are stored in a separate ‘tar
Horizontal and vertical components of the TPE get file’. A small portion of full resolution imagery
estimate [3] are computed for each beam during around the area of the identified target (‘snippet’)
data acquisition and saved with the bathymetry is also stored with the target information.
data. Onboard, all side scan sonar data are converted to
The contents of the ISS-60 dataset are incremen the Unified Sonar Image Processing System
tally archived from the central suite to the Network (UNISIPS) format for analysis. The imagery is
Attached Storage unit. This can be done manually, archived, processed and QC’d on the shipboard
for example at the end of each survey line, or it can systems on a daily basis.
be done automatically based on file name
changes. Data from the HSL are automatically
archived after the HSL is brought onboard and con Processing Throughput - Objectives and
nected to the ship network. Approach
mated and less manually intensive, shifting the pri the automated data cleaning filter, cross compare
mary manual interaction to the validation and QC different sources of bathymetry and perform inter
phases of the workflow.[l] Software improvements active editing. This approach provides the analyst
for reaching these objectives include three major with a seamless transition between the full-resolu-
components: (1) an Area-Based-Editor (ABE), (2) an tion swath-oriented data, geospatial-oriented pre
automated data cleaning filter and (3) the incorpo sentations of the data and interactive 3D visualisa
ration of these two capabilities with COTS software tion of the data. As shown in Figure 10, the link
packages to provide a complete integrated data pro between the software packages is the Pure File
cessing solution. These components are designed Magic (PFM) file structure, which allows users to
to work together to achieve the 4:1 objective. move easily and quickly between the visualisation
Two Co-operative Research and Development surfaces, the complete set of contributing data
Agreements (CRADAs), with IVS of Fredericton, points and if necessary, back to the supporting full-
N.B. and with SAIC of Newport, Rl resulted in the resolution swath oriented files. The PFM file struc
integration of NAVOCEANO's tools with each com ture includes a multi-surface binned file with fixed
pany's commercial product(s). Under the CRADAs, spatial extents, units and cell size that supports
SAIC's SABER processing and analysis product and direct access by position to the minimum depth,
IVS’s Fledermaus 3D visualisation product have maximum depth, average depth, standard deviation
both been integrated with NAVOCEANO's Area and number of observations for each cell in the bin
Based Editor. These tools allow the analyst to view file. In addition, the file structure supports direct
the bathymetry and imagery data geospatially, access by position of all (filtered and unfiltered)
review and assess the data points invalidated by measured data points that contribute to each cell.
Raw U N IS IP S S o ftw a r e
Sonar So nar Form at C onversion
GEO D AS GEODAS Imagery QC
C ollection ► QC/Targel C ontact Id cn tification/ Taget Processing
System System M osaic Production
Target C hange D etection
F ile v// B ottom f to v in c in g (m osaic)
snippet Coverage Assessm ent (m osaic)
Pointers are maintained to allow direct indexing acquired on the T-AGS 51 and T-AGS 60 class plat
back to the source data files in order to access sen forms and their HSLs. SABER’S hydrographic man
sor specific or data type specific information. agement system reads the major survey milestones
from the ISS-60 message file to assist with populat
ing report templates and to summarise the progress
Shipboard Processing Analysis and of the survey and the processing.
Quality Control Interfaces are also supported for data that origi
nate from other acquisition and processing sys
Onboard data processing begins as soon as data tems, allowing data from all assets involved in NAV
files have been copied from the data acquisition OCEANO surveys to be processed with the same
systems to the Network Attached Storage unit. data structure and processing tools. Various multi
Four workstations are available for onboard post beam sonar formats - including Reson, Simrad,
processing and validation (Figure 2). SeaBeam, L3 and XTF - can be converted directly
to GSF allowing for full reprocessing if required. In
These processing seats each access data stored on addition, SHOALS/CHARTS LIDAR, CARIS HDCS,
the Network Attached Storage unit via a one-gigabit HTF and ASCII x,y,z formats may be loaded directly
switched Local Area Network. Through the PFM soft into PFM to support data cleaning, sounding reduc
ware libraries, the SABER and Fledermaus software tion and product generation.
products support the GSF bathymetry format output Initial processing tasks include line/file/event log
from ISS-60 and the UNISIPS side scan sonar target generation and maintenance, graphic review of the
and mosaic formats. SABER directly integrates with navigation data for quality; review and/or update of
the ISS-60 dataset directory structure, eliminating correctors to the data; and record keeping.
any need for copy, import, or translations steps. Navigation problems identified at this stage can be
Figure 11 illustrates the dataflow pipeline for data addressed by editing the offending fixes from the
Fa* E(« vww Se&ngt Process Analyst*
Display Formai
inset)
X(m) 3618S0 M
iC o d f t k ; Y (IB) 4769633 29
■OspJ) (m) Min Depth (m) 0 00
Max Depth (m> 00
Avg Depth (m) 00
MlnFilt Depth (in). o0
» PORTSMOUTH H/VRE Ma^ilt Depth (m) 00
;vv •p,FG52fij HORN ■AvgFilt Oeplh (m) 00
» Sndgs 0
ANCHORA W I N IS W SWDov 0.0
IIO.IO -W 2 :
' n o te A ) jJ l
\\23 ^
Stielman Rks
Navigation QC
raw position data and reconstructing the track line data format. This consists of binning the depth val
using a two-pass forward/backward Kalman filter, ues into a pre-defined grid that computes and
or by selecting an alternative positioning source. saves the minimum, maximum, average, standard
The reconstructed trackline can then be replaced deviation and the number of data points for each
in the GSF file over the time frame of interest. cell of the grid file. The bin cell size is determined
Figure 12 shows the navigation from a family of by the resolution of the sonar so the full resolution
survey lines displayed using SABER. This screen data can be properly Q C’d. For a hydrographic sur
provides the interface for initial navigation QC. vey, the extents of the PFM file are typically defined
Once the navigation has been verified and any by the sheet boundaries. All depth values that fall
issues resolved, known corrector updates can be into each cell are maintained and can be directly
applied. Since all corrections are applied during accessed by position. During the load process, a
data acquisition, this step is typically limited to statistical ‘area filter' can be used to identify erro
improvement of tide corrections with the applica neous soundings that need to be invalidated from
tion of preliminary observed water levels if avail further processing. Note that no data points are
able. If observed water levels are not available deleted; the outliers are simply flagged as invalid.
until some time later, then processing continues The area filter provides a level of automation
using the predicted value applied during acquisi intended to replace the labour intensive ‘first-pass’
tion. When necessary, the full range of systematic of manual line-based interactive editing. For each
corrections including antenna and transducer off data point, the area filter compares the difference
sets, alignment biases, SVP, draft and tides can be between the observed depth and the mean value
reapplied. Traceability is supported by maintaining for the cell with a selectable tolerance (typically 3
all the offsets and correctors, including the applied sigma). If the tolerance is exceeded, the data point
SVP, with the raw data, allowing for reapplication of is identified as invalidated by the area filter. When
the corrections as needed. available, targets identified from review of the side
Following the flow presented in Figure 11, the next scan imagery can be used to establish a ‘no-filter’
step is to ‘load’ the bathymetry data from the radius around the target location, so depths within
swath-oriented full- resolution data files to the PFM the radius are not invalidated by the automatic fil
ter. The no-filter radius is operator selectable and province delineation. UNISIPS also provides more
is based on the accuracy of the target location capabilities critical to mine warfare activities.
from the DSSS. Loading the targets into the PFM Figure 11 outlines the functions and relationships
data structure can occur simultaneously with load of DSSS imagery collection and processing.
ing the bathymetry, or the bathymetry can be Figure 13 shows a 2D shaded relief visualisation of
loaded first and targets loaded later. In the ship the minimum filtered depth surface from a PFM file
board environment, this flexibility is necessary to containing Reson 8125 data from Portsmouth, NH.
allow bathymetry cleaning and review to occur par This PFM file was created in UTM units with a cell
allel with the post-time review of the side-scan size of 0.5 meters and a depth precision of 1 cen
imagery for targets. timetre. The survey vessel track lines are superim
Side-scan targets are identified during acquisition posed on the bathymetry. Similar presentations
and the imagery files are reviewed post-acquisition are available for the maximum depth surface, aver
to ensure that all significant contacts have been age depth surface, standard deviation surface,
identified. Target files, containing the target posi number of observations surface and for the unfil
tion, a snapshot of the imagery for the target and tered versions of the minimum, maximum and aver
comments entered during the target identification, age. The standard deviation surface visualised
are loaded into the PFM file following completion of with the superimposed track lines can be particu
imagery review. The bathymetry and imagery data larly useful for locating potential problem areas as
around each target can then be analysed in SABER identified by patterns that are aligned parallel or
and/or Fledermaus for hydrographic decision-mak- perpendicular to the data collection azimuth. This
ing. surface can also help identify small-scale seafloor
The full-resolution side scan sonar data are con features that may not be readily apparent when
verted to UNISIPS format for further processing viewing the minimum, maximum, or average sur
and archiving. The UNISIPS software suite creates faces.
mosaics, allows bottom province delineation and
provides access to the full-resolution scan-line The human visual system has an enormous capaci
data from the mosaic. The mosaic is used in con ty for receiving and interpreting data quickly and effi
junction with bottom samples to assist in bottom ciently and therefore must be an integral part of any
Bounding Coords
Figure 14: 3D
visualisation of
same Reson
Map Display
8125 data
r Checked
presented in r Show Dataset (on.*off)
\ Pala Set Control / 1
Figure 13
effort to understand complex data. The key is to be concurrent SABER and Fledermaus displays
able to present the data in as intuitive a fashion as enhance the analyst’s perception of the area and
possible. The more intuitive the presentation, the rapidly focuses their attention to the significant fea
more rapidly data are interpreted and the more new tures or anomalies. This results in overall improve
information can be extracted. Integrating interactive ment in data interpretation and processing decision
3D visualisation into the overall processing system making.
has allowed the analyst to take full advantage of
data density and, simplifies analysis and interpreta After the data have been loaded and the first pass
tion by allowing interaction and exploration of com cleaning completed by the area filter, each of the
plex multidimensional data. Bringing 3D visualisa PFM bin surfaces are visualised over a ‘bin sub-
tion up to the front end of the processing pipeline area' of the extents of the PFM. Any outliers not
provides a powerful capability to assess the quality invalidated by the area filter will be visible in either
of the data throughout the processing timeline. Even the minimum filtered or the maximum filtered sur
small-scale anomalies can be identified and deci faces. The unfiltered minimum and the unfiltered
sions regarding anomaly resolution can be made maximum surfaces can also be visualised to
earlier in the process. This provides an essential review the data points invalidated by the area filter.
capability for data interpretation and decision mak When issues are noted, an ‘edit sub-area’ of the
ing during the processing workflow. The integration displayed surface is selected for viewing all con
between SABER and Fledermaus allows the analyst tributing data points. The small highlighted area
to start one application from the other, visualising west of the sand waves in Figure 13 was selected
the same extents of the data. The 3D scene pre to launch SABER’S Multi-View Editor (MVE). Figure
sented in Figure 14 was generated simply by select 15 shows all data points that exist within the
ing the Fledermaus option from SABER to load the selected edit sub-area. The MVE allows the analyst
same section of the PFM. Fledermaus presents to review all data points by individual and cumula
complex, multidimensional datasets in a natural and tive validity criteria, assess and modify invalida
intuitive manner, allowing integration of multiple tion’s made by the statistical area filter, manually
components from various sensors without compro validate or invalidate data points as appropriate
mise to the quantitative aspects of the data.[4] The and review sounding selections. From the MVE, any
structure. When review of the
edit sub-area completes, the
sub-area is identified as
‘checked’. Checked sub-
areas can be highlighted in
SABER and Fledermaus to
indicate which sub-areas
have undergone detailed
review. When all fliers have
been resolved for the current
view of the ‘bin sub-area’,
this region can be marked as
checked and the analyst
advances to the next bin sub-
area. This approach allows
the analysts to review rela
tively large sections of the
PFM bin surfaces at a time
and focus their efforts on the
Figure 15: Multi-view interactive editor showing all contributing soundings areas that need additional
selected from a subset o f the PFM file cleaning, while requiring mini
mal (if any) efforts for areas
sounding can be traced back to its location in the were no additional cleaning is required.
source data file. Targets derived from the side scan
imagery are displayed in MVE and can be correlat The Fledermaus 3D editor (Figure 16) has a similar
ed with the appropriate least depths in the multi interactive exploration interface as the main visu
beam bathymetry. When appropriate, these least alisation window and allows the soundings to be
depth values are identified as ‘feature’ selections coloured by depth or attributes - file, line, ping and
for hydrographic data products. The edits are beam. Deleted soundings can be shown or hidden
saved as changes to Boolean flags in the PFM data in the display, as can the various binned surfaces.
The editor also allows selec
tion of soundings by various
criteria and selection or dele
tion as significant features.
Further details of the select
ed soundings can also be
obtained by drilling down far
ther to the native files for
mats of the original data.
Fledermaus displays the side-
scan targets as point objects
in the 3D scene and allows
concurrent display of any
side scan sonar image snip
pets, to aid in the verification
and feature classification
(Figure 17).
In the shipboard-processing
environment, new data are
Figure 16: Fledermaus 3D editor showing same three features viewable in Figure continually loaded into the
15 current PFM as files become
Figure 17: A
scene of
Portsmouth
Harbour, New
Hampshire. The
main Digital
Elevation Model
is coloured by
depth from a
multibeam sonar
survey. Draped on
the surface are a
number o f side
scan sonar
images
available from ISS-60 and the side-scan acquisition points from the near vertical beams to produce a
systems. As new data are loaded, the checked flag statistical assessment of system repeatability as a
is cleared for each cell of the bin file that has been function of angle (or beam). In addition to providing
updated to clearly indicate the updated cells of the an evaluation of each beam, these results also
bin file. For large survey areas, multiple analysts can allow the hydrographer to evaluate the adequacy of
work on a common PFM simultaneously. The edits all corrections applied to the data. This approach
and review work of one analyst become visible to has proven useful for evaluating newly installed or
other analysts as the work progresses. At any time, upgraded systems prior to starting survey opera
the progress of the data cleaning effort, with tions and for assessing the suitability of all correc
respect to the area covered, is visualised easily by tions that have been applied to the data during sur
displaying one of the PFM surfaces showing the vey operations. [1]
checked status. The sounding reduction algorithm uses an area-
While Figure 11 implies that the QC processes based approach to reduce the full dataset to a sub
occur at the end of the workflow pipeline, QC set that preserves the significant seafloor features
processes are performed throughout the acquisi using shoal-biased down-sampling. Soundings that
tion and processing timeline to ensure that any have been selected are highlighted in the displays
problems are identified in a timely manner. As an allowing for detailed review and QC against all
example, junction analysis at the intersection of neighbouring data points. Selected soundings may
main-scheme and cross-lines and along adjoining be displayed on the screen in text form superim
sheet boundaries, is typically performed on a regu posed on the current chart or highlighted in the 3D
lar basis during the performance of a survey and view for further review and QC.
then repeated at the completion of a survey. It On completion of sounding reduction and QC analy
should also be noted that performing the data sis, the edits that have so far been saved to the
cleaning review using an area-based process PFM data structure are ‘unloaded’ back to the GSF
makes any horizontal or vertical offsets between data files. Primary product generation occurs from
adjacent lines discernable, whereas, with the pre these updated GSF data files. This includes the
vious line-based approach these problems would exporting of selected soundings and soundings
not be easily seen. marked as features to Electronic Navigational
For detailed sonar system performance metrics, a Chart (ENC) and Digital Nautical Chart (DNC®) pro
reference surface is established using only the duction packages. For certain data products, the
near vertical beams from the multibeam sonar cleaned and validated full resolution native data
from data acquired with small line spacing. All files are input into various databases.
soundings from all beams are compared with all It should be noted that the while Figure 11 shows
a stepwise processing approach, the combined the field. Then, all that is required when the data
SAIC and IVS tools do not require that the steps be arrives in-house is to exercise Quality Assurance
completed in a specific sequence. This is an impor procedures.
tant point, as it may be useful to assess cross
check comparisons prior to completing data clean
ing. Likewise, final water level corrections may not Product Generation
be available until 30-60 days following the survey;
so final tides may be applied after data cleaning is The output of the sounding selection module of the
completed. The ability to generate 3D visualisa Area Based Editor goes directly into the CARIS
tions of the bathymetry early in the process great chart production software. In addition to the final
ly assists with data interpretation and with identifi selected soundings, additional digital data sources
cation of areas that need to be investigated. can be input to the CARIS production software.
These include historical data from the NAVOCEANO
Data Warehouse, satellite imagery for shoreline
In-house Processing and Quality updates and Notice to Mariners for navigational
Control aides. Outputs from the CARIS production software
suite include Digital Nautical Charts (DNCs),
At the end of the survey, all survey planning files, Electronic Navigation Charts (ENCs) and Tactical
raw hydrographic/bathymetric GSF files and ancil Electronic Chart Overlay Products (TECOs) such as
lary ASCII data files, processed GSF hydrograph Additional M ilitary Layers (AMLs). TECOs are
ic/bathymetric data files, full- resolution DSSS data designed to enhance the user’s understanding of
files, target files from the DSSS, mosaics and PFM his surroundings by depicting militarily significant
data files are archived to 4mm DAT tapes. The DAT information directly on the electronic navigational
tapes, Report of Survey and all field products are system. These products include area definition,
shipped to NAVOCEANO. Upon arrival at NAV- dense sounding layer, high-resolution contour layer,
OCEANO, the data package is sent to Data Ingest mine warfare objects and bottom sediment classi
for cataloguing and copying to the Data Warehouse. fication layers. Figure 18 illustrates a generalised
Once the data have been loaded into the Data production flow for NAVOCEANO.
Warehouse, the following QC processes are NAVOCEANO utilises CARIS chart production soft
checked and/or verified (1) all data and support ware to generate both DNC® and ENCs. The CARIS
files are present and readable; (2) survey stan software modules include the CARIS Object
dards established in the planning phase were cor Manager for DNC® (CARIS DOM) and ENC (CARIS
rectly followed, including verifying that horizontal HOM), Digital Terrain Model (CARIS DTM),
and vertical control were properly established; (3) Geographic Information System (CARIS GIS) and
tides and sound velocity files are correct; and (4) CARIS Suppress Soundings. Other COTS software
bathymetry data have been completely processed tools used for source evaluation, data extraction
and observed tides have been applied. If predicted and product displays are IV S ’s Fledermaus,
tides were applied during data collection, these ERDAS’ Imagine and ESRI’s ArcView.
files need to be replaced with observed tides when In support of the fleet NAVOCEANO is now a co-pro-
they become available. NAVOCEANO’s Analysis and ducer of electronic DNC charts under a memorandum
Validation Branch then performs a rigorous in-depth of agreement with the National Imagery and Mapping
data appraisal. It is important to note that this step Agency (NIMA). Under this agreement, NAVOCEANO
is not a re-processing step. Each data type, such as is responsible for compiling DNC libraries directly
bathymetry, SSS imagery, water levels, calibration affected by NAVOCEANO surveys. Upon completion
data, etc. has its own QA/QC process. Check-off of the DNC library compilation using the CARIS soft
lists have been established to monitor the data flow ware and after quality assurance using the NIMA VPF
through the validation process. The final step is to Validator software suite, the libraries are sent to
assign the final quality assessment of the data and NIMA for general dissemination as part of the NIMA
then archive the data set. chart maintenance and update program. The NAV
With the improvements to processing techniques OCEANO compilation may be used to produce an ENC
(both software and hardware), NAVOCEANO’s goal in partnership with the International Surveys Program
is to have all the survey completely processed in and foreign host nation hydrographic offices.
Production
Flow
Request
from
Customer
Feedback
from
Customer
Figure 18:
Data product
production flow