http://bit.ly/gEtc4W

Thank you for your response. I’ve tried out Vuzix AR products in the past. They are functional, but give the impression that you are viewing the world through a television set. For some niche applications, they may be adequate.

The system you described is quite feasible. The reason it can be done is that you are doing it in a “prepared” location where markers have been placed, and calibration data has been gathered. In fact, there have been a number of research demos of the years that accomplish what you are trying to do. The visual effects industry uses roughly similar techniques to allow directors and actors to see computer generated objects and effects while they are shooting on set.

Here a few AR projects you may find useful for reference.

http://singularityhub.com/2009/11/19/augmented-reality-does-time-travel-tourism/

As an alternative to a headmounted display, you could also consider using augmented reality binoculars or an LCD panel mounted on a tripod. With these displays, you would not need to track position after initial calibration. You would only need to track orientation. Using an LCD with a camera mounted on front would allow a group of clients to view the scene at once.

Here’s a system I prototyped a while back.

http://www.jarrellpair.com/augmented-reality-binoculars/

Good luck with your work, and if you build a system, I’d be interested in seeing some photos.

Best Regards,

Jarrell Pair

I noticed that Vuzix does not warrant a mention in your post. I am looking at the possible application of their AR 920 product in the field of architectural visualization. I did not get to see their product demo at CES so I have no idea what the world looks like through their glasses but it’s the only commercial product slated for release in the undetermined near future.

In my firm, we create BIM models using Autodesk Revit Architecture. Recently I started geolocating these models and displaying them in Google Earth. (Ok so I’m late to the game. Big deal, people still think that’s cool.)
Logically the next step is to be able to display a geolocated model in the field. Granted, once the user is in the field, the problem is tracking the user. GPS and compass are out, so what about photogrammetry ? I was wondering if there is the possibility of solving this through software rather than hardware.

Different existing commercial software packages allow the user to interact with imagery and 3D in ways that could (to me) be combined to allow the accurate positioning of a 3D building in an Augmented reality setting.

Autodesk Imagemodeler allows the user to “pin” known points of a 3D model to known points in a photo, for easy camera matching. The newly released VideoTrace Beta allows the user to “trace” a 3D model from one or more calibrated video segments. And of course existing camera matching software such as Autodesk Matchmover allows the user to extract a camera path from a video segment.

Well used Photogrammetry software like EOS Photomodeler uses coded targets for fine calibration of images for 3D extraction

Meanwhile LinceoVR and AR-Media use coded targets for AR display.

So riddle me this. Why can’t some brilliant software engineer combine these elements in the following way.

1. Head out to the project site with a geolocated survey.

2. Locate known coordinates in the field with coded targets.

3. Put on AR headset.

4. Walk a calibration route which allows the user to view a given number of targets at a time while recording the output of the AR glasses.

5. Export to a calibration software which uses elements of camera matching, photogrammetry, and a few parameters such as know distances between physical points to situate the user within the site.

6. Once this is done, slap the 3D model “onto” the site and “pin” the know points in the 3D site to the known points in the real world.

7. Keep updating the position of the user via the video feed from the headset using the coded targets, image recognition of the site (cruise missile style and existing x,y,z head tracking technology.

If the problem is weighing down the headset with active systems, take those systems out of the headset and put them into the world. The headset could be connected to a teeny tiny laptop carried by the user (MacBook Air or similar). The coded targets can be readily printed to whatever size is needed with any plotter and the AEC industry can have a reaaaaaaallllly cool new way of showing the client what they’re getting. Also a reaaaaaaallllly cool way of coordinating disciplines.

Granted it’s not a solution for people who want to walk around the street with it, but for my use where I’m displaying a building on a known, limited physical site it would be priceless.

I want to be able to see the building, walk around the building and walk into the building. I wholeheartedly believe that a software solution can get us there with existing technology.

Please, somebody make it so.

Best regards and thanks for a great post.

Jean-Frederic Monod

I am very familiar with Microvision. They have been around a long time having spun out of work at the University of Washington HITLAB in the 1990′s. Twelve years ago I worked for one of their board members, and I co-authored a white paper on applications for a fully realized, color, wide field of view Microvision retinal display. Since that time, I’ve used various red monochrome displays they’ve either prototyped or sold. About eight years ago, they released and sold the “Nomad” display. It was not widely adopted, and I do not think it is available today. When I used one, it delivered an excellent sense of imagery being overlaid on the real world. The field of view was relatively small though. I do recall feeling a slight tingling sensation in my eye. The military first looked at their display technology in the late 1990′s. Daylight readability was the key selling factor. Now, there are other technologies which are daylight readable. Microvision seems to have shifted their focus to micro video projectors. I get the sense that they really need a successful product on the market that takes advantage of their intellectual property. The biggest challenge to head mounted display (HMD) design is creating a wide field of display in a compact physical package. No one has solved this problem yet, though there are ideas that hold promise which could advance the technology forward given funding and time. There’s also the eyestrain issue which may never be fully resolved. On the other hand, stereoscopic 3D movies can also cause eyestrain discomfort. Regardless, they are now a major commercial success though they were initially introduced several decades ago. Head mounted display development is on a slow path, but I am always hopeful for research, technical, and market breakthroughs.