What I see is a game changer. It will take a bit of time to filter through to us surveyos; but, when it does, wow.
To see what I am talking about to to the 10 minute mark of this podcast. The relevant part lasts around 4 minutes.
All About Android Video Podcast
The short version is Google is going to make a scanner that is part of a smart phone. No, not a "scan this document" scanner, a "scan this room and create a 3d model" scanner.
Larry P
I've been waiting for something like this ever since I first saw Photomodeler.
With a single photo you have millions of discrete "3D" pixels that can queried. A Surveyor here is using it for tough as-builts. He has tested it against tape measurements and it has very impressive accuracy.
With a few pics of a building or any other structure you have the WHOLE building. Third floor window widths? Got it. Sewer inverts on a box culvert with 3 other pipes entering 22' below the manhole? Got it.
There are so many opportunities for those with the proper training.
What is typical distance from camera to object? Is the distance dependant of camera's pixel specs?
ABSOLUTE BALONEY!!!
HORSE DOO DOO.
Takes multiple photos, and it's not as easy as it looks.
One photo only allows one plane solution. It's an eight-parameter projective transformation. Works for skid marks on a flat street.
Info for photomodeler: Tech specs
I've seen whole buildings from under 100' to art objects from inches.
What I meant to infer is that a single photo is made of millions of pixels. When a series of photos are "stitched", the results can be quite accurate. By no means point and shoot, hence the reference to training.
This is a good way to get experience in the scanning technology, that's for sure.
Who are the competitors (to Photomodeler)?
If you watch the video you'll see that this is not a different way to use photographs and cameras to do the modeling. Instead it is a 'LIDAR from your phone' type application.
The potential is huge.
Larry P
Photogrammetry has its place, but it's not a panacea. With a truly photogrammetric-quality camera, properly calibrated, XYZ coordinates of points can be determined to an accuracy of 1 part in 100,000. That is not possible with a digital camera you can buy off-the-shelf at Wal-Mart, but impressive accuracy is typical of 1 part in 10,000 of the distance of the camera stations to the object(s).
Auto correlation seems to work like magic, but it does have practical limitations in actual practice. Similar scenarios are starting to be marketed by Leica and Trimble for their high-end Total Stations and pole-mounted camera systems.
The stuff works, but the office work is still highly intensive, just like the laser scanners.
Be careful with your checkbook. It's a solution looking for a problem.
iWitness from Australia. I teach with it at LSU.
The checkbook might approve a $2,500 software and a Wal-Mart camera to see what this scanning technology is about.
Do a sensitivity analysis before you ask the boss to stamp a production job.
Image identification is a MAJOR CRITICAL criterion. Be careful!
Indeed.
I am curious about the QC routines available on these types of processing software.
What exactly is "image identification"?
Is this different much different form Autodesk's 123D?
http://www.123dapp.com/howto/catch
The whole terrestrial Photogrammetry thing seems very interesting. Remote sensing seems to be the next big place surveying is heading. Might as well get in on the ground floor and be on the cutting edge instead of getting left behind.
The cellphone scanner is similar to the handheld GPS IMHO.
It has its uses but not for accurate survey purposes. You get a 3D model but I would not
use output say for asbuilt survey. For reference purpose as additional data source after
the crew is done with the surveyusing a TS maybe.
...Or Gerry .... Google Earth....can occasional get a good fit with a cadastral overlay , but more than likely highly inaccurate...
RADU
Client: "So, will your 3D model include the steel on the underside of the bridge and the rail overhead catenary system at a quality level sufficient for design?
Game Changing Surveyor: "No problem dude...I got an app on my phone for that." :-O
> Is this different much different form Autodesk's 123D?
>
>> http://www.123dapp.com/howto/catch
>
> The whole terrestrial Photogrammetry thing seems very interesting. Remote sensing seems to be the next big place surveying is heading. Might as well get in on the ground floor and be on the cutting edge instead of getting left behind.
Watch the video. This has nothing to do with photogrammetry.
Larry P
> Client: "So, will your 3D model include the steel on the underside of the bridge and the rail overhead catenary system at a quality level sufficient for design?
>
> Game Changing Surveyor: "No problem dude...I got an app on my phone for that." :-O
I recall a time when these same things would have been said about GPS. I'm not saying in 6 months we will all have a laser scanner in our pockets.
I am saying in a few years these things will completely change how we do our business. Am also saying that those who scoff at the notion will be just like the guys who decided that "one of them fancy machines that claim to measure distance will never replace my guys and our steel tape".
It will certainly take time for the QA/QC side of things to prove themselves (if they ever do). The future is arriving every day and we need to be ready for it when it does.
Larry P
It's a special phone with two cameras. I just figured it had to be using photogrammetry for some of it's positioning.
