RRS Discovery, Dundee: a 2014 laser scan baseline, re-measured in 2020

RRS Discovery, Dundee: a 2014 laser scan baseline, re-measured in 2020

The RRS Discovery was the last traditional wooden three-masted ship to be built in Britain. Designed for Antarctic research, she was launched as a Royal Research Ship in 1901. Her first mission was the British National Antarctic Expedition, carrying Robert Falcon Scott and Ernest Shackleton on...

Project at a glance

  • Client: Dundee Heritage Trust
  • Sector: Heritage and conservation, historic vessels
  • Location: RRS Discovery, Dundee
  • Date: 2014 baseline survey, report dated 08.04.14; 2020 re-survey, project 2399
  • Services: 3D laser scanning, control network survey, registered point cloud, 2D CAD sections, 3D hull mesh and external model, HDR virtual tour, 3D animation, dataset-to-dataset deviation analysis

The brief

The RRS Discovery was the last traditional wooden three-masted ship to be built in Britain. Designed for Antarctic research, she was launched as a Royal Research Ship in 1901. Her first mission was the British National Antarctic Expedition, carrying Robert Falcon Scott and Ernest Shackleton on their first, successful journey to the Antarctic, known as the Discovery Expedition. She is now the centrepiece of a visitor attraction in her home, Dundee.

Dundee Heritage Trust commissioned Luminous (then trading as Digital Surveys) to survey her. In the Trust’s words, the aim of the project was to create an accurate 3D representation of the vessel’s current state of repair, to provide a datum for monitoring potential crack and structural decay. The detailed 3D laser survey information could also be used to create useful visual models to aid in promoting and educating the public, as well as assisting in the ongoing preservation of the vessel.

The Trust has recorded that it is very grateful for support from Museums Galleries Scotland under the Recognised Collections Fund, which covered 100% of the costs of this project.

Two requirements sat inside one brief. The engineering requirement was a datum: a measured reference that later surveys could be compared against. The public requirement was material people could actually look at. Both come from the same capture.

The challenge

A wooden ship is a structure that never stops moving. Timber responds to moisture, temperature and load, and a hull held out of the water carries its weight in ways it was not designed for. Movement of that kind is slow, distributed and largely invisible to the eye. Nobody watching the vessel day to day would see it.

Traditional monitoring would fix permanent targets at key locations and read them periodically. That approach reports movement only where the targets are. On a hull whose deformation pattern is the unknown, choosing the measurement points in advance defeats the purpose.

The vessel also presented real capture difficulties. Internal areas could only be scanned where accessible, which on a 1901 research ship means low headroom, tight companionways and holds still in use as museum space. The masts and rigging are complex, elevated and impossible to reach with any contact method, so they had to be captured externally at long range.

Then there is the requirement nobody can retrofit later: whatever was captured in 2014 had to be good enough, and referenced well enough, for a survey six years later to be compared against it meaningfully. The accuracy of the baseline sets the smallest movement that can ever be detected.

Our approach

The 2014 baseline survey
A survey plan was agreed in advance to determine the optimum scanner setups before anyone arrived on site. On a vessel with restricted internal access, setup positions are the whole job. Get them wrong and you return for the gaps.

Two instruments were used, chosen for their environments. A Leica P20 captured the external hull, top decks, masts and rigging. The P20 is specified at ±1mm point accuracy, 1mm linear error and 0.8mm rms range noise at 10m, and its range and precision suit long shots up to rigging that cannot be approached. A FARO Focus 3D captured the internal areas, where the priority is a compact, portable instrument that can be carried through tight spaces and set up quickly.

Point density was approximately 3mm. All scans were full 360 degree and in full colour. A control network of temporary targets was established across the vessel to coordinate and register the scans. A control network is a set of reference points held common across the whole job, so that every scan is tied to one consistent frame rather than merely to its neighbours. It is the difference between a dataset that is internally tidy and one that can be trusted as a datum. Registration was carried out in Leica Cyclone. The external scan position plan runs to Station-037.

An iSTAR 360 degree HDR panoramic camera captured 20 external HDR images for the virtual tour, giving photographic imagery from the same positions as the scan data.

Our conclusion at the time was straightforward: a comprehensive 3D laser survey had been completed of the external areas of the vessel and internal areas where accessible. Using a range of technologies this would provide a digital archive of the vessel’s current state of repair. The information provides a snapshot in time and will form the basis of any future 3D conservation recording.

The 2020 re-survey

Six years later we returned as Luminous Group, to test that promise.

The re-survey used a Leica RTC360. Scans were taken around, below and inside the vessel to ensure a good comparison with the previous survey. Coverage was deliberately matched to the 2014 capture rather than optimised independently: a monitoring re-survey is only as good as its comparability, so the second dataset has to reach the same surfaces the first one did.

The two datasets were then merged in CloudCompare, which places the 2014 and 2020 clouds in the same coordinate frame so that they can be measured against each other directly. External sections were cut through the hull at nine positions along its length. Each section was colour-coded to show the shift between the two surveys and annotated with the deviations in millimetres.

Cutting sections rather than producing only a whole-surface colour map matters here. A section gives a conservator a drawing they can dimension, take to a shipwright and act on, with the deviation written on it in mm at the point it occurs.

What we found

The comparison returned measured movement.

Typical section deviations along the hull were 10.4mm, 12.7mm, 15.4mm, 21mm and 35.7mm. Sections were produced to show the deviation and confirm that the most movement has been to the port side of the vessel. Length sections produced from the two datasets and compared show an extension of 34mm.

Those figures are the point of the entire exercise. A hull that has moved by up to 35.7mm at a section, asymmetrically, favouring the port side, and lengthened by 34mm overall, is a structure behaving in a describable way. Conservators can now discuss a rate rather than an impression, target investigation at the port side rather than everywhere, and set the next re-survey knowing what magnitude of change the method resolves.

None of this could have come from fixed monitoring targets. The port-side bias is a whole-hull pattern, and it only appears because the whole hull was measured, twice, in all three dimensions.

Deliverables

From the 2014 baseline we issued the point cloud in raw Leica .imp and E57 formats, E57 being an open exchange format any third party can read; 8 2D CAD plans, comprising 3 long sections and 5 short sections; a 3D hull mesh model; an external 3D model; a virtual tour built from the 20 HDR panoramas; and a 3D animation.

From the 2020 re-survey we issued the merged dataset comparison, with colour-coded external sections at nine positions along the hull, annotated with deviations in mm, and length sections comparing the two surveys.

The visual outputs did the second half of the brief. The 3D animation, virtual tour and external model gave the Trust material for promotion and public education, drawn from the same capture that produced the engineering datum.

The outcome

Dundee Heritage Trust holds something few historic vessels have: a permanent record of RRS Discovery captured to approximately 3mm point density, and a second survey six years later that turns that record into quantified structural movement.

This is the project in our archive that closes the monitoring loop with real numbers. A baseline is a promise that a future comparison will be possible. In 2020 that promise was tested and paid out, in millimetres, on a 1901 wooden hull that had already spent a century being difficult to measure.

 

Back to Case Studies