Developing Robust Measurement Techniques for Vessel Sensor Offsets
Article

Developing Robust Measurement Techniques for Vessel Sensor Offsets

The world of hydrography is being opened up for new applications using a variety of boat-mounted sensors. Measurement of the relative positions of the sensors on board is vital for making the most of these new technologies. Stuart Leakey from the Port of London Authority and Geoff Wharton from Capture As-Built Ltd describe new techniques for the accurate measurement of relative sensor positions.

The Port of London Authority (PLA) is responsible for managing the tidal River Thames. Its Hydrographic Service has three dedicated survey vessels each permanently fitted with multibeam systems and with the ability to integrate a much wider variety of survey equipment as required. With the move to high-resolution multibeam systems it became clear that the organisation was in possession of a resource with value far beyond traditional bathymetry. Indeed, a large proportion of the PLA Hydrographic Service work now sees its vessels utilised to conduct surveys for monitoring and pre-works investigations as well as engineering-level surveys for asset management, and as-built surveys. With this demand comes the requirement to mobilise a variety of survey equipment, from vessel mounted lasers and photographic systems to towed arrays of sub-bottom equipment such as boomers, pingers and magnetometer arrays; all of which rely on accurate definition of sensor positions and layback.

The Demand

With improvements to GNSS positioning and sonar resolution it was becoming apparent that the accuracy of sensor offset data had become a significant factor in the system error budget, made more critical due to the increasing use of GNSS heighting to reduce bathymetric data to chart datum, and the temporary mobilisation of survey equipment to the vessels.

It was decided to re-survey the vessels to determine high accuracy sensor positions. Critically, the survey needed to define the positions of the primary sensors: the multibeam echo sounder (MBES), the inertial measurement unit (IMU) and the positions of the aiding GNSS antennae. The alignment of the IMU was also critical as it comprises arrays of gyroscopes and accelerometers, mechanically aligned orthogonally on the principal axes.

Surveying SV Yantlet

The PLA commissioned Capture As-Built Ltd to conduct a laser scan of the SV Yantlet, a 16m Catamaran with a permanently installed Reson 8125 MBES. In order to validate the data using a traditional total station, the PLA, exploiting its partnership with UCL to deliver a Category ‘A’ accredited MSc in Hydrographic Surveying, sponsored a dissertation entitled, “Investigating the use of Terrestrial Laser Scanner for Hydrography; Vessels Sensor Assembly”.

Ensuring the Yantlet was in her standard operational condition before being removed from the water, multiple measurements were made from fixed points around the hull to the waterline. This was done in calm water and the inclination, list and heel, was measured at the same time. This enabled the water plane to be ascertained within the point cloud at a later date.

Once clear of the water the boat was positioned in open space within the boatyard so as to allow access around the vessel and to limit the need for oblique vertical observations. The control survey network was simulated using StarNet to estimate the number of rounds of angles required to achieve the desired accuracy and then the observations were made using a Leica TS15 total station. In order to position the IMU within the vessel a double resection was required to locate the IMU within the starboard hull of the Yantlet. The total station survey used a combination of reflectorless shots to define the hull, and self-adhesive retro reflective targets when re-secting or when high accuracy repeatable observations were required, such as around the sensor assemblies. The StarNet least squares adjustment showed error ellipses of up to 17mm in the vertical.

The scanning was conducted in a conventional manner utilising a Leica C10 laser scanner with multiple circular planar targets within each scan. Scans were co-registered and processed using Leica Cyclone. When conducting the scanning, planar targets were fitted to the total station control network so that a comparison could be made between the two surveys. The method of establishing control within the boatyard when surveying the SV Yantlet meant that when the vessel was moved, the control no longer had any relationship with the boat.

The SV Verifier

Building on the success of the initial survey the PLA commissioned Capture As-Built Ltd to use refined techniques to scan the SV Verifier, a 21m mono-hull with a permanently installed Reson 7101 MBES.

The major change was to install recoverable control aboard the Verifier. Threaded inserts with M8 internal thread were permanently installed on board. Eight inserts were positioned such that at least three were visible from any required survey station location. The inserts accommodate spigot adaptors for round prisms and holders for spherical reflectors.

A Leica Nova MS50 Multi-station was used to observe the onboard control for all subsequent survey activities. It was also used to scan the external parts of the vessel at relatively low resolution and the MBES and mounting gondola at higher resolution to enable extraction of dimensional data. Total station observations were made with the MS50 to Leica round prisms installed in temporary spigot adaptors on the GNSS antenna mounting points.

The IMU itself was removed from the vessel and scanned off-site using a Romer Arm supplied by Hexagon Metrology, producing a very accurate high-resolution scan that could be independently positioned within the point cloud and defined the orientation of the IMU to sufficient accuracy to enable extrapolation of its alignment.

The IMU could not be accurately surveyed with the MS50 because it was located 700 mm below floor level and with restricted access. But both the IMU and its mounting plate are precision engineered to millimetric level. It was possible to position the IMU scan within the point cloud by surveying the mount points on the permanently installed mounting plate using a Leica Absolute Tracker AT960 and T-Probe. The tracker was set up on deck relative to four control points with a line-of-sight through the wheelhouse to a point below deck. Temporary targets were located using heavy duty double-sided tape such that they would be visible from a location below deck with clear sight of the floor access hatch for the IMU. The laser tracker was then set up below deck and positioned by observing the same four targets. The IMU was removed to reveal the mounting bolt holes and the bolt centres were surveyed using the Leica T-Probe with an 800mm extension.

It should be noted that as the control was already installed, it was possible to conduct these measurements whist the vessel was still afloat. With the vessel located in sheltered waters within Tilbury docks, reflector-less measurements of the water surface were taken with the MS50 relative to the onboard control to establish the waterline level under static trim whilst the survey with the tracker continued.

Deliverables

The acquired data was processed using Leica Cyclone software and compiled in AutoCAD. The vessel coordinate system was aligned relative to the IMU mounting bolt centres and the point data was exported using the data extraction function within AutoCAD to create the spreadsheet of offsets. A drawing was created showing the location of all relevant survey points and the waterline plane.

The primary deliverable was simply a set of orthographic drawings and a table of offsets which defined the positions of all the installed sensors relative to the IMU reference point. However, the point cloud itself has also proved a powerful deliverable. It has been used for fitting of equipment within the vessel, and installation of the equipment for positioning for sensor offsets and for management and maintenance of the vessel itself. For example, if a piece of fendering is lost it can be measured and cut to fit utilising the point cloud data or model and be waiting quayside for the vessel, thus reducing vessel downtime.

Additionally, we were able to define the centre of rotation of the vessel. This point needs to be defined relative to the IMU in order for the software behind the IMU to effectively uncouple the motions thus avoiding artefacts such as roll induced heave. As stated by Euler: “The oscillatory movement of a floating body (rolling or pitching) can be described as a rotation about the Centre of Flotation.”

Thus by defining the water plane within the point cloud, the water plane area is defined by the intersection of the two components and the centroid may be found by either using the facilities within the software if displaying within a CAD package, by numerical integration or, if the shape of the water plane is maybe sympathetically simplified, by taking moments.

Conclusions

When examining the error ellipses of the total station measurements, it is apparent that the total station method is weak when conducting oblique vertical observations, which are necessary when re-secting to establish a position within the confined spaces within the vessel, or when surveying within the confines of a busy boatyard. The registration technique of the laser-scan survey process removes this issue; and increasing the density of the registration targets adds redundancy and brings down residuals. The increased density of measurements on a particular feature which is facilitated by point cloud data aids accuracy by reducing the interpolated distances and inherent generalisations required with a less dense dataset such as that offered by conventional total station measurements.

The scanning method utilising the Leica ScanStation C10 has some advantages over the MS50 method, primarily the data capture rate is faster but at a lower quoted accuracy. However, with the MS50, the installation of the control and the scanning are carried out concurrently with one instrument and the separate scans are co-registered onboard the instrument, which streamlines the processing workflow. The MS50 data is of higher accuracy and has a lower noise level.

On vessels small enough that they may be considered rigid, the installation of permanent control aboard the vessel allows for repeat survey work on deck, and within the vessel without the requirement to remove it from the water. This remains true as long as the measuring instrument and observables move as one body. Consequently, although the initial survey may be costly, as the boat must be removed from the water, subsequent surveys aboard may be achieved whilst afloat, and the cost of the initial survey may be reduced if the work is conducted alongside scheduled maintenance.

Although the established control on the Verifier has already proved to be a valuable asset, the distribution of the control is poor in that it is very flat being primarily on the plane of the deck. The construction of the vessel dictated where the control could be positioned as one may not just drill a hole anywhere on a vessel. Using alternatives to drilled inserts may allow for a better control distribution to be established at a later date.

Both the surveys of the Yantlet and the Verifier proved to be a success, and allowed us to state the relative positions of the installed sensors to millimetric level, rather than the previous centimetric level. Along the way we have shown the weakness of total station measurements and the value of point cloud data and onboard control.

Further Work

The PLA has commissioned an 18.5m custom-built survey boat, which is being designed and built by CTruk in Essex. She has been designed to specification around the premise that she is a survey instrument that floats and is able to sustain an economical 20 knots.

Building on the lessons of the Verifier, she will have control installed strategically around her deck, coach roof, and internally within every compartment such that the distribution of the control varies through the principal dimensions. She will have two in-built IMU mounting points and is designed for permanent installation of the R2Sonic 2024 MBES (with the 700kHz upgrade package). The control will be established prior to her maiden voyage and the sensor position survey will form part of the commissioning process prior to her delivery at the end of this year.

The PLA is sponsoring another UCL MSc dissertation utilising the point cloud data of the Verifier. This will be jointly sponsored with Applanix and explore whether significant improvements can be made in removing vessel motion (heave) artefacts by accurately defining the centre of rotation (flotation) as enabled when using point cloud data with the observed water plane defined therein.

This article was published in Geomatics World September/October 2015

Geomatics Newsletter

Value staying current with geomatics?

Stay on the map with our expertly curated newsletters.

We provide educational insights, industry updates, and inspiring stories to help you learn, grow, and reach your full potential in your field. Don't miss out - subscribe today and ensure you're always informed, educated, and inspired.

Choose your newsletter(s)