Provoking an Inspired Response 

Facebook Twitter Gplus LinkedIn YouTube RSS
magnify
Home Culture & Technology 3D Scanning, Capturing the 3rd Dimension

3D Scanning, Capturing the 3rd Dimension

3D Scan Process

3D scanning?

No, not the thing with special red and green glasses.

About 10 years ago I started developing 3D scanning technologies. It’s interesting to note that very few people could visualize what exactly 3D scanning was at that time.

Most people seemed to think that you needed special glasses and that holograms were involved. It did not matter how I explained it. It was only after I showed them a 3D scan of an object on a PC screen that their jaws dropped. Seeing truly is believing…and understanding.

3D scan

3D scan Allard Piersons Museum Amsterdam (Triangles 3D scanner)

So what is 3D scanning exactly?

3D scanning is a set of technologies, a tool box of solutions, to capture your 3D world.

3D Scanning is the process of capturing the geometry of a tangible object or scene and converting this into 3D data points.

Once captured, the data can be stored, shared, viewed and modified just as any other data on your computer. In addition, you can apply a photo (called textures) of the 3D scanned object or scene which can be overlaid onto the 3D geometry. This creates a higher sense of realism when viewing scanned objects on your PC screen. These days you can barely tell the difference. What is real and what is not is less distinct.

3D scanning could be seen as the reciprocate of 3D printing. 3D printing takes data and converts this into a tangible model in our real world. 3D scanning does the opposite by converting the geometry of a tangible object into data. This data can also be returned back to the real world using 3D Printing. You get the point.

3D scanners are not something new. And, while there have been many development advancements in recent years, the core technologies are, typically, old.

Actually “3D scanners” is not entirely the correct description for these technologies. “3D scanners” is a subset category. The formal name is “range-mapper”. But this post will refer to these technologies as “3D Scanners”.

The most recent milestone in the development of 3D scanning technologies was a period about 30-35 years ago. Laser scanning, “White Light” scanners as we now call them and several other technologies were being introduced. In fact, the White Light 3D scanner was first patented as far back as the late 70′s.

3D laser scanner

3D laser scanner

Laser type 3D scanners received a lot of development over the years and were in many cases the most popular. In part due the relatively simple, low-cost, good accuracy, wide range and flexibility properties that laser systems offered. In addition, the laser was an impressive device and implied something “advanced and sophisticated” to many people.

The workings of a laser scanner are simple. When a laser line, for instance, is projected onto an object you can see a profile of the object’s surface geometry when viewed at an angle. Using triangulation and knowing certain other parameters you can easily calculate the 3D position of each point along the illuminated laser line on the object. By sweeping the laser over the object you can do the same for each viewed step of the sweep, building a map of 3D points that represent the objects geometry.

White light 3D scanners are also based on triangulation to build a 3D pint map of an object or scene. They do something similar as laser scanners but in a different way. This type captures the geometry of an abject by projecting certain patterns onto the object. Typically these are black and white patterns and can be as many as 16 different patterns or more. The patterns are not random, they follow a strict code, usually binary. They sub divide the object as it were in numerically encoded sections. Each section is encoded or labeled with a numerical value and then further subdivided into smaller sections after each new pattern iteration. This is done until the sections can no longer be effectively divided into smaller sections due to limitations of the projection system and/or the camera used. White light scanners can be very fast in comparison to laser sweep type systems.

Laser and white light type scanners as well as many others are non-contact. They don’t touch the object during the scanning process.

Contact type scanners also exist and typically use a probe attached to some motion control apparatus to feel their way over a surface and map the 3D points of that surface.

Probe type “contact” scanners are less popular for common application. But they find important application in, for instance, measurement systems in manufacturing. They are much, slower than most “non-contact” types.

However, probe type scanners can be extremely accurate. Their data registration has a very high level of integrity. So much so that you can scan things as small as molecules and even atoms.

Here is a link to a short article on a self-build high resolution touch probe.

Probe type scanners are forming the basis to examining the micro and even nano world in micro- and nano-systems technologies. See, for instance, the home-built Scanning Tunneling Microscope. A really fantastic project. There are also pseudo type contact scanners that do not actually touch the sub micro structures during the scanning process.

“Today we can 3D scan everything from large surfaces of the earth to large molecules”

Today, 3D scanning is gaining traction on the market, in part, due to the sudden, but not surprising, popularity of 3D Printing technologies. This is stimulating the 3D scanning technology development. The perception and awareness of 3D technologies is becoming more common. The value of these technologies as well as the growing need is becoming more clear.

Viewing 3D Content

Another related progression is the viewing of 3D content. Viewing 3D content involves a lot more processing power than viewing pictures or even video. You typically need a good graphics card to handle the 3D content. Fortunately these days a good graphics card is relatively inexpensive and most PC’s ship with some pretty powerful cards. And, as long as it’s from companies like nVidia or ATI you will be able to number crunch millions of vertices per second.

But probably the biggest and most important progression here is allowing you to view that 3D content in your webbrowser. Sure there have been several ways to do this over the years but very few have really been able to tap into the power of your graphics card from your browser. In fact, only recently have solutions matured enough with the potential to become standards. Some are commercial plugins.

JIG 3D, simple O3D scene publisher

JIG 3D, simple O3D scene publisher

Two of the most important and free developments have been O3D and WebGL. O3D was a plugin based approach which worked in all major browsers. Its development was rapid and it offered very good performance and functionality. So much so that I took the risk and built a software application called JIG 3D to streamline the conversion and upload process of 3D data to your browser. Old video’s of that application can be found here and here. But O3D was shot down before it went into orbit due to its plugin based structure. Nobody wants to install plugins anymore. Unfortunately that meant that a progression to get 3D content to the webbrowser retarded. WebGL at that time offered much less in terms of performance and functionality. And, more importantly, stability. But it’s built-in to most all webkit based browsers these days. It’s now mainstream with an ongoing development cycle.

Below is an example of WebGL based 3D content viewer by SketchFab. Please note that this is still relatively new stuff and it might not work correctly in your browser and/or PC. In particular if your PC does not have the the necessary graphics processing power. Also, it won’t directly work in the IE webbrowser as this is not webkit based (see Google Frame). Best is to use FireFox or Google Chrome. If it does not show try refreshing the page.


 

Click to load and view the above 3D content.

That’s now, but what’s in store for the future?

Micro and nano applications aside, we can expect to see more and more “hybrid” type systems to appear. The merit of 2 or more different scanning technologies are consolidated into a single system.

“Disruptive” technology…

Recently we have witnessed a shift in the 3D scan arena with the introduction of some new technologies which allow real time 3D capture. These are turning out to be disruptive. They are game changers which will have serious impact on the industry and market.

These disruptive technologies will not necessarily over-throw all other 3D scanning technologies. But it will displace many. More importantly they will become the 3D scan technologies for the masses. For you and me. They will replace certain work processes as well as create new processes. Security, medical, manufacturing, CGI and many other industries will benefit and benefit greatly.

Traditionally, 3D scan technology development focused on scanning “accuracy” among other things. 3D scan accuracy was important and even paramount. Costs were usually very high for such systems. A basic 3D scanner could cost you thousands of Dollars.

ultra low cost 3D scanner introduced 2005, Triangles 3D scanner

ultra low cost 3D scanner introduced 2005, Triangles 3D scanner

Later on, low-cost system development became an equally important element.  A milestone was set in the mid 90′s by a French Cal-tech student who used nothing more than a pencil, a light source and a camera to 3D  scan in objects . This led to a number of patents which can be found here and here. Unfortunately the companies that purchased these important patents severely lacked the correct vision and direction. They were unsuccessful in the commercialization of these technologies.

High accuracy, low-cost systems seemed to be the focus in many cases. But in recent years, high-speed and even real-time systems have started to emerge and rapidly gain traction on the market.

One of the most important and recent real-time systems was introduced about 15 years ago. An advancement in this particular technology was patented as early as 2007.

It involves the use of a laser projector which produces a single static speckled pattern. The pattern is viewed by a camera which is placed along side the projector. Displacement of the pattern structure viewed by the camera provides an idea about the differences in-depth within the cameras field of view. The technology was developed and supplied through an Israeli company called PrimeSense.

If ever there was a goal to produce a “real” 3D camera then these real-time scanning technologies are becoming just that. You can imagine the profound effect that the microscope, telescope and, in particular, the 2D camera had on science, technology and society. Low cost, real-time 3D scanning systems will have the same effect and probably in an even bigger way.

Microsoft Kinect and Asus Xtion

Two of the most disruptive real-time 3D scanning systems on the market today are the Microsoft Kinect and the Asus Xtion. The interesting thing is that these devices were not introduced as 3D scanners but instead as Natural User Interfaces (NUI).

Other NUI technologies are on their way such as the Leap Motion (even the new Lytro camera could be used as 3D scanner and NUI device). However it is unlikely that the Leap Motion data output will allow use as a 3D scanner.

But the Xtion and, in particular, the lesser performing Kinect are the first movers in this arena. Both rely on the same technology. In fact, the core technology in both is from the same company, PrimeSense.

The scanning quality and accuracy of these devices is low.  What should be made clear is that their accuracy can be greatly improved on with minimal development effort. But what is really amazing, at this point, is that these devices are able to achieve results in a manner typically found in high-end systems costing thousands. Yet, both the Microsoft Kinect and the Asus Xtion cost less than 250$. Expect to find many hobbyists,  researchers and companies to supply some amazing 3D scanning tools for these devices very soon.

When it comes to 3D scanning the PrimeSense technology offers a low-cost, small size package with real-time scanning. The sum of these properties allows for a serious level of scanning flexibility. You can expect this technology to be repackaged by companies with different names and reside in $250-$500 range. For instance, the Sense Scanner from 3D Systems and the Structure Scanner from Occipital. Both employ the PrimeSense device as the core technology or scanning engine.

Whatever line of business you are in there is a good chance that you will, at one time or other, come across these devices and probably even want and need them.

That’s how important they are.

 
 
 Share on Facebook Share on Twitter Share on Reddit Share on LinkedIn
7 Comments  comments 

7 Responses

  1. Great article! There are too less articles on 3D scanning if you ask me, but thats only my opinion. :)

    If the Kinect 2 is being released with the higher accuracy you would expect i say a combination of Kinect + Reconstructme (or perhaps K-scan) and perhaps some post-processing software like Leios and Mudbox, will outperform a 14k+ Artec Eva System by price/speed easily, and in terms of shear scan quality will probably be nearing it. I already did a succesfull paid job with a Kinect 1,and Reconstructme (with Leios and Mudbox as post pro) so I anticipate some opportunities here… :)

  2. Hi Lucas,
    First, many thanks for your comment! You are the first here on this new blog; Hey…What’s the BIG idea. Hope that it provokes to inspire the response of others to do the same.

    Things are pretty clear about the Kinect and its potential. The only thing which is not clear is Microsoft’s vision. But I am sure that they are now getting the bigger picture. And, that’s probably part of their procrastination surrounding their marketing of the Kinect as a NUI as well as a 3D scanner. See also my post “Real Artists Ship”. I hope that they are on that boat.

    Another thing to take into account is that there are several other organizations doing parallel research to that of Kinect technology. Some have already produced some amazing results and even applied and received patents. Expect to see some serious competition in this arena. Not just at hardware level but also post processing tools.

  3. Andrew

    Hi again B. J – another great article !!
    your comment “What should be made clear is that their accuracy can be greatly improved on with minimal development effort…” – can you share any links ?

    • Hi Andrew,
      Thanks again for your comment. “The best software is hardware”. Some of the “brute force” methods to improve on the kinect type technologies is adding additional kinects, another camera, higher resolution camera, multi sampling etc. Here is an example:

      http://www.youtube.com/user/IAOag3DVermessung

      Note: what we are developing is different. While it also includes adding some low cost hardware, it combines technologies rather than just enhances existing kinect hardware.

  4. Andrew

    appreciate that B. J – you’re now a big bookmark :-)

    Just found this interesting paper:-

    http://research.microsoft.com/pubs/158080/paper-phil-Cha-ZZ-final.pdf

    Other –
    I really enjoyed your “background” write-up. I wish mankind was not such a “tribal” animal – we have a pretty poor setup here in Northern Ireland. Different groups of “christians” [protestant/catholic] fight each other at the drop of a hat – pathetic really. In discussion I always ask them to think what their allegiances may have been if born in (eg) Saudi Arabia…

    • Hi Andrew,
      Appreciated! :)
      Note: Following is off-topic:
      My heritage includes 3 different religions. I expect that these come together as one when a problem or a victory is common. Ireland is beautiful country, language and culture. Its history has always been turbulent and divided. I confess to know little about it. But I do know that it is one of the most globalized nations in the world! Something in common.