(Underwater Intervention 1994, San Diego, California, February 1994, pp34-38.)

Experiences of Using Stereoscopic Video with an Underwater Remotely Operated Vehicle

Andrew Woods 1
Tom Docherty 2
Rolf Koch 3

1 Centre for Marine Science and Technology,
Curtin University of Technology, GPO Box U1987, Perth 6845, AUSTRALIA

2 School of Electrical and Computer Engineering,
Curtin University of Technology, GPO Box U1987, Perth 6845, AUSTRALIA

3 School of Mathematical and Physical Sciences,
Murdoch University, South Street, Murdoch 6150, AUSTRALIA

ABSTRACT

Stereoscopic video offers a number of advantages to the operation of Underwater Remotely Operated Vehicles (ROVs).

This paper discusses the experiences of using the Curtin University Stereoscopic Video System (SVS) on the 'Triton' Underwater ROV operated by Woodside Offshore Petroleum. The trials have been performed at the NorthRankin `A' gas production platform, 130km off the coast of Western Australia.

The ROV pilots' perceptions of the SVS and how it affected their use of the ROV are also discussed. Operators reported that it was perceptually easier to perform manipulator tasks, it provided them with a better feel of the work-site and gave them confidence in their actions. Greater dexterity could also be achieved which will conceivably reduce damage and down-time. Time trials have also revealed the potential to significantly improve task performance.

Overall, the experience we have gained with this system on the Triton ROVindicates a very positive future for stereoscopic video on underwater ROVs.

INTRODUCTION

Stereoscopic or binocular vision is the main means by which the brain obtains depth information about the world around us. The fact that the two human eyes are positioned side by side, separated by a distance of approximately65mm, means that each eye sees a slightly different view of the world. The brain processes the slight differences between the two eye images and thereby mental impression of depth is perceived.

Conventional television does not reproduce this effect to the observer because both eyes see the same image, therefore only a two-dimensional or flat image is seen by the viewer. Stereoscopic video is the process by which video and television technology are used to produce three-dimensional images. A stereoscopic video display (or 3DTV) presents a realistic three-dimensional image to the observer by displaying a different image to each eye of the observer. The two images are obtained from two video cameras mounted side-by-side. The brain is able to process these two images to perceive a three dimensional image, the same way it would normally process real world images.

STEREOSCOPIC VIDEO SYSTEM

We have developed a stereoscopic video system for use with underwater ROVs. This system consists of an underwater stereoscopic camera which is mounted on the ROV and a stereoscopic display which is located in the ROV control room. The stereoscopic video camera (shown in Figure 1) consists of two video cameras mounted side-by-side in an underwater housing. When it is fitted to the underwater ROV, one of the ROV's existing video channels passes the stereoscopic video signal to the control room. In the control room, the stereoscopic display (shown in Figure 2) displays the left and right images from the stereoscopic camera to the viewer who wears a pair of 'sun-glasses' style 3D glasses to see the underwater image in full-colour flicker-free 3D.

Stereoscopic Camera   stereoscopic display
Figure 1: The Underwater Stereoscopic Video Camera. Figure 2: The Stereoscopic Display

The stereoscopic video system works by storing left and right images from the stereoscopic video cameras onto alternate images of a single video signal. The PAL video standard (used in Australia and Europe) displays 50images per second. In this system 25images per second are from the left video camera and the other 25 images per second are from the right video camera in a left, right alternating sequence. The NTSC video standard (used in North America) displays 60images per second, therefore 30 images per second are from the left camera and the other 30 images per second are from the right camera. At this point the stereoscopic video signal is compatible with the PAL or NTSC standards.

When the stereoscopic video signal is received by the stereoscopic display, internal electronics store the incoming video signal and display it on the monitor at twice the original rate. This means that the 50Hz images (60Hz for NTSC) from the video camera are displayed flicker-free at a rate of 100Hz (120Hz for NTSC) on the screen (ref 1).

The advantages of this technique are that the system is compatible with existing conventional video equipment. Only a single video channel is used in the ROV's umbilical cable and the stereoscopic video signal can be recorded using a single unmodified video recorder.

FIELD USE

The Stereoscopic Video System has been used in the field by Woodside Offshore Petroleum on a Perry Tritech Inc. `Triton' ROV. The ROV is operated at the North Rankin gas production platform 130km off the coast of North-Western Australia (see Figure 3). The water depth at the platform is 125 metres. The ROV is fitted with a Schilling manipulator arm which is used to perform a wide range of tasks. A range of field experiments have been carried out to assess the performance of the stereoscopic video system on the ROV.

Triton ROV
Figure 3: The Triton ROV fitted with the
Stereoscopic Video System at the North Rankin Platform.

BENEFITS OF STEREOSCOPIC VIDEO

The benefits of stereoscopic video are quite wide ranging (ref 2) - the most obvious being the increase in depth perception. Depth is perceived by a range of different visual cues which include perspective and relative size (things are smaller the further they are away from the camera), shadows and shading (these can show contours and the relative positioning of objects), etc. Unfortunately, many of these cues are missing or reduced in the underwater environment and the underwater environment is not structured like our usual environment where walls are vertical and floors are horizontal - a larger fish is not necessary closer than a smaller fish, shading and shadows can be muted by murky water and lighting is not necessarily from above like the sun or lights in a room. All of these factors make the depth cue of stereoscopic vision more important than in normal everyday viewing.

Our study has observed the following benefits of stereoscopic video on underwater ROVs:

Our experience to date indicates that the use of stereoscopic video (3Dviewing) will improve task performance and reduce task time (when compared to 2D viewing). The amount of improvement will depend upon a range of factors, including how depth-dependent the particular task is. It is, however, very difficult to quantify this improvement in the field situation because of the large number of external (uncontrollable) factors involved in field trials. It is therefore difficult to compare the two situations (2D vs. 3D) accurately. It is possible to set up a controlled and repeatable task, but this is not necessarily representative of the field environment. Subjective assessments are considered to be more valuable for field evaluation (refs 4,5). Several laboratory based studies (refs 6,7), and indeed our own (ref 8), have measured improved task performance with manipulative tasks.

DRAWBACKS

There are some drawbacks associated with the use of stereoscopic video. The most obvious one is the increased cost of the equipment. The stereoscopic camera and the stereoscopic display are both specialised equipment and need to be specially procured. The stereoscopic video system we have developed is, however, compatible with all of the existing video equipment of the ROV. It is compatible with the existing umbilical cable, video MUX, video recorders and video routing. The amount of specialised equipment is therefore kept to a minimum.

There are also some image distortions associated with stereoscopic video. These distortions will depend upon the camera and display configurations being used (ref 9). It was noticed by the operators that when approaching a stationary object at a constant velocity, the speed of approach would appear to increase as the ROV came close to the object. This is a known effect of stereoscopic video and in this instance was also partially due to the wide angled lenses being used on the stereoscopic video camera. This effect can be reduced by the appropriate choice of camera parameters, however, we also expect that if the operator is made aware of the nature of this effect, any problems will be reduced.

Another strange effect which we have noticed is that if the observer moves while looking at the stereo image, the image will appear to `follow' the observer. In one particular situation, the operator had `parked' the ROVagainst part of the structure. When the operator moved his head, the image appeared to move and as a result the operator thought the ROV had moved when in fact it had not. This effect may or may not have any detrimental effects, however we expect that any problems will be reduced if the operator is made aware of the presence of the distortion.

When things come close to the stereoscopic cameras, the image may appear blurred or as double images. This also happens if one tries to look at something which is too close to the eyes. If something is placed too close to the cameras, the stereoscopic display will attempt to reproduce the image very close to the observer's eyes and similarly this will be difficult to view. This effect can be overcome simply by closing one eye or switching the display to2D when things come too close to the cameras.

Eye strain is something we have been very careful to monitor and document in the field trials of the stereoscopic video system. Many people may be aware of the eye fatigue and headache problems associated with the 3D movies of the 1950s era. Much of the problems of these movies were due to the red/blue technique used to achieve 3D and also bad alignment of the stereoscopic images. The stereoscopy and 3D of today has evolved to a much more sophisticated level. Needless to say, eye fatigue and headaches can still be a problem if good quality control and stereoscopic alignment are not maintained. In our field trials of the stereoscopic video system, operators have reported that they have not experienced any headaches they would associate with using the system. Some people have reported some eye fatigue especially when the system is badly aligned, however, we believe that eye fatigue can be reduced if good alignment of the cameras is maintained and appropriate camera parameters are chosen.

CONCLUSION

We believe that the benefits of using stereoscopic video for the operation of underwater remotely operated vehicles far outweigh the disadvantages.

The operators have reported that the stereoscopic video system would be primarily useful for manipulator tasks - a task which requires high accuracy and is highly (3D) depth-dependant. The operators also indicated that the SVS was useful when `flying' and navigating the ROV. As mentioned previously, the SVS allowed the operators to better identify the structure and layout of the platform - one of the operators commented that it would be valuable for inexperienced operators for this reason.

Overall, the experience we have gained with this system on the Triton ROV indicates a very positive future for stereoscopic video on underwater ROVs.

ACKNOWLEDGMENTS

The authors wish to thank Woodside Offshore Petroleum for their support of this project.

REFERENCES

1. Woods,A.J., Docherty,T., Koch,R., "The Use of Flicker-Free Television Products for Stereoscopic Display Applications", Stereoscopic Displays and Applications II, Proceedings of the SPIE volume 1457, 1991.

2. Merritt,J.O., "Evaluation of Stereoscopic Display Benefits", in Introduction to Stereoscopic Displays and Applications, Short Course Notes, L.Hodges, D.McAllister, J.Merritt, editors, SPIE - The International Society for Optical Engineering, Washington, 1991.

3. Pastoor,S., Schenke,K., "Subjective Assessments of Dynamic Visual Noise Interference in 3D TV Pictures", Proceedings of the Society for Information Display, Vol. 30, No. 3, 1989.

4. Drascic,D., Grodski,J.J., "Using Stereoscopic Video for Defence Teleoperation", Stereoscopic Displays and Applications IV, Proceedings of the SPIE volume 1915, 1993.

5. Booher,H.R. (editor), "MANPRINT, An Approach to Systems Integration", U.S.A., Van Nostrand Reinhold, 1990.

6. Cole,R.E., Parker,D.L., "Stereo TV Improves Manipulator Performance", Three- Dimensional Visualisation and Display Technologies, Proceedings of the SPIE volume 1083, 1989.

7. Touris,T.C., Eichenlaub,J.B., Merritt,J.O., "Autostereoscopic Display Technology in Teleoperation Applications", Telemanipulator Technology, Proceedings of the SPIE volume 1833, 1992.

8. Woods,A.J., Docherty,T., Koch,R., "Field Trials of Stereoscopic Video with an Underwater Remotely Operated Vehicle", Stereoscopic Displays and Applications V, Proceedings of the SPIE, 1994 (in press).

9. Woods,A.J., Docherty,T., Koch,R., "Image Distortions in Stereoscopic Video Systems", Stereoscopic Displays and Applications IV, Proceedings of the SPIE volume 1915, 1993.


Copyright on this document is retained by Curtin University. This document is not public domain. Permission is hereby given to reprint this paper on a non-profit basis for scholarly purposes provided the document is unaltered and this notice is intact. This paper may not be reprinted for profit or in an anthology without prior written permission. If you wish to reprint this paper on this basis, please contact the primary author at the address shown on the first page of this document.


GO BACK to Andrew's Home Page

Last modified: 30th August, 1995.
Maintained by: Andrew Woods