Topic ID #13159 - posted 8/2/2011 1:47 PM
Jennifer Palmer
Webmaster
Kinect Hacked For 3-D Scanning Of Archaeology Site
Jennifer Palmer
Webmaster
Kinect Hacked For 3-D Scanning Of Archaeology Site
BY Kit EatonToday
University of California, San Diego students will be going to Jordan soon to take part in an archaeological dig that's decidely futuristic: As they uncover artifacts and structures in the soil, they'll be using high-quality 3-D scanning to record accurate positional details--rich data that could be incredibly useful in the future. Instead of using expensive and complex imaging systems like LIDAR, however, the team will use a hacked Microsoft Kinect to do the job for them.
It's actually using a system developed by the California Institute for Telecommunications and IT (Calit2), which taps directly into Kinect's streaming data feed that's a blend of 3-D positional data (achieved by projecting bright, invisible spots of infrared light onto objects and then observing them with an IR camera) and color video images. The Calit2 team has perfected this system so it's useful for making fast and accurate 3-D scans of objects that can then be inserted into a virtual world like Second Life--the trick is to correctly register all the images recorded of the object so they match up next to each other properly as you wave the Kinect around. Thus far the Calit Kinect hack uses an overhead video tracking device to do this, which limits it for indoor use--a tweak is already planned to let it work in an outdoors settings, however, and its inventor thinks it could even be used to scan whole buildings (at which point Google, with its penchant for doing this inaccurately for its Street View system, may get interested).
Read more here.
BY Kit EatonToday
University of California, San Diego students will be going to Jordan soon to take part in an archaeological dig that's decidely futuristic: As they uncover artifacts and structures in the soil, they'll be using high-quality 3-D scanning to record accurate positional details--rich data that could be incredibly useful in the future. Instead of using expensive and complex imaging systems like LIDAR, however, the team will use a hacked Microsoft Kinect to do the job for them.
It's actually using a system developed by the California Institute for Telecommunications and IT (Calit2), which taps directly into Kinect's streaming data feed that's a blend of 3-D positional data (achieved by projecting bright, invisible spots of infrared light onto objects and then observing them with an IR camera) and color video images. The Calit2 team has perfected this system so it's useful for making fast and accurate 3-D scans of objects that can then be inserted into a virtual world like Second Life--the trick is to correctly register all the images recorded of the object so they match up next to each other properly as you wave the Kinect around. Thus far the Calit Kinect hack uses an overhead video tracking device to do this, which limits it for indoor use--a tweak is already planned to let it work in an outdoors settings, however, and its inventor thinks it could even be used to scan whole buildings (at which point Google, with its penchant for doing this inaccurately for its Street View system, may get interested).
Read more here.
Post ID#18898 - replied 8/2/2011 6:17 PM
Dwarmour
That's pretty neat. I would like to see where this goes but after reading the first comment I don't think much will unless the use more advanced equipment. Hopefully the archaeologists will report their findings quickly.
Post ID#18900 - replied 8/3/2011 4:08 AM
Jennifer Palmer
Webmaster
The Kinect has been hacked for so many purposes. It's neat to read about someone doing it for archaeology.
Post ID#18901 - replied 8/3/2011 11:41 AM
elevenam
LiDAR has its place. It is accurate to the mm and has a very long range. However, the Kinect has advantages. Like a total station, a LiDAR scanner needs backsighting or fixed targets in order to sync up the scans. In an environment with a large amount of occlusion this can be difficult. The kinect is hand held in this example, allowing you to easily move around occlusion to get a complete scan. It has its limitations too. Range on the Kinect is limited to a few, maybe 10 meters, Less if you want the resolution of the scan to be good.
More importantly, the Kinect is $150. With a laptop and maybe some additional hardware (lights and a battery for the Kinect) you are looking at maybe $1000 for a purpose built system.The first commenter on the linked article gives away his bias when he states that he used to make and sell laser scanners... that business may be threatened in a small way if simple 3d cameras like those built into the kinect can do even 50% or 60% as good of a job for 1/100th or less the cost of a good laser scanner.
He reminds me of film photographers, extolling how superior the old ways are, without acknowledging the sheer convenience for the vast majority of people who can't or don't want to deal with the burdens inherent in the technique.
My question is how useful do experienced archaeologists consider this type of thing? I have my B.S. but I fully realize my experience is quite limited in the research arena (though I have worked on and presented papers at this point). I feel that this type of scanning and recording is extremely useful. The more context we can add to an object, and the more readily available we can make collections or sites to other Archaeologists, the better the field will be as a whole. Instead of relying on measured sectioned drawings of an artifact, you could download, manipulate and examine a fully textured 3d model. Nothing would beat examining the real thing of course, but conservation sometimes prohibits physical examination. Would a 3d model be better than the current methods or nothing at all?
I don't know the answers right now but I am trying to find out. I've been interested in the Kinect scanner as archaeological tool for a few months now, and I have some promising results, however if anyone has any input on this, I would greatly appreciate it. What other applications for the Kinect would you like to see?
More importantly, the Kinect is $150. With a laptop and maybe some additional hardware (lights and a battery for the Kinect) you are looking at maybe $1000 for a purpose built system.The first commenter on the linked article gives away his bias when he states that he used to make and sell laser scanners... that business may be threatened in a small way if simple 3d cameras like those built into the kinect can do even 50% or 60% as good of a job for 1/100th or less the cost of a good laser scanner.
He reminds me of film photographers, extolling how superior the old ways are, without acknowledging the sheer convenience for the vast majority of people who can't or don't want to deal with the burdens inherent in the technique.
My question is how useful do experienced archaeologists consider this type of thing? I have my B.S. but I fully realize my experience is quite limited in the research arena (though I have worked on and presented papers at this point). I feel that this type of scanning and recording is extremely useful. The more context we can add to an object, and the more readily available we can make collections or sites to other Archaeologists, the better the field will be as a whole. Instead of relying on measured sectioned drawings of an artifact, you could download, manipulate and examine a fully textured 3d model. Nothing would beat examining the real thing of course, but conservation sometimes prohibits physical examination. Would a 3d model be better than the current methods or nothing at all?
I don't know the answers right now but I am trying to find out. I've been interested in the Kinect scanner as archaeological tool for a few months now, and I have some promising results, however if anyone has any input on this, I would greatly appreciate it. What other applications for the Kinect would you like to see?
Post ID#18902 - replied 8/4/2011 6:26 AM
Jennifer Palmer
Webmaster
Interesting that this story is getting traction in some of the popular media outlets:
Archaeologists eschewing traditional whip / leather jacket combo for Kinect controllers
Archaeologists eschewing traditional whip / leather jacket combo for Kinect controllers
Post ID#18903 - replied 8/4/2011 8:29 AM
prisoner
I think the usefulness of it will be clear when people actually start testing and publishing the accuracy of these methods. I am currently doing 3D modeling of rock features using digital photogrammetry for my thesis, but several archaeologists always ask me why I don't just use LiDAR. My response is that LiDAR is several thousand dollars and my whole rig is about $700 (minus the software and computer). But how does the accuracy compare? I really don't know, since I am only using one method and have nothing to compare it too. On the other hand, archaeologists can get so caught up in the super duper accurate devices and then be fine with a centimeter or two of error on other things. For example, we spend thousands of dollars on total stations and some people don't level them correctly, don't calibrate the machines regularly, and are often times shooting the head of a nail which is a cm in diameter anyways. Then there are the submeter GPS units that are often used to record things that would be fine if they were recorded with a smaller, and cheaper, handheld. So, the real question is how much accuracy do you really need? If you just want a 3D model to oogle over you need very little. If you are trying to measure the width of incised lines on a ceramic or shell you might need a lot.
I am finding all of this 3D stuff pretty fascinating. LiDAR is obviously pretty nice, photogrammetry is nice and has the added advantage of also recording the surface image without having to stitch the image over it, and this Kinect thing is intriguing. Add to that the recent use of this RTI method that HP is doing and there is some real potential for serious 3D digital archiving of sites and artifacts for not too much money. I wish SAA would get a committee or something organized to start really comparing these methods and their potential.
Post ID#18904 - replied 8/4/2011 10:15 AM
elevenam
Prisoner: Would you be interested in comparing notes? I have been wrapped up in getting the RBG-D mapping to work, but my long term plan was to survey a single site with a Kinect, a LiDAR (if I can get someone to come operate their own it or loan it to the project) and photogrammetry, with the goal being a paper comparing the results of the methods on a single test site. At the very least I'd be interested in knowing which software you were using and how well it is working for you. If you want you can reach me at (elevenam AT gmail).
I share a lot of the some opinions as you on this subject... Would be interesting to discuss this further.
I share a lot of the some opinions as you on this subject... Would be interesting to discuss this further.
Post ID#18906 - replied 8/4/2011 11:59 AM
prisoner
Sure. I am pretty swamped currently with my thesis, course work, trying to find a job, my current job, and a new baby due next week. But maybe sometime this fall. My gmail: thnadner@gmail.com
|
Next topic: "UFO found on ocean floor? " |
|
Previous topic: "Aurora Native Title Internship Program - Australia" |
|
Looking for something else? Show recent posts in Discussion |
