DAVID Laser scanner tests

Post by: Joris Weijdom Add comments
Bodycount

IMG_7853.jpgBased on the software called DAVID we got inspired for the Bodyscan concept. DAVID consists of two programms: DAVID Laserscanner and DAVID Shapefusion. The Laserscanner software is free and published under the GPL, DAVID Shapefusion can be bought for a very modest price. So what is this DAVID software?

“DAVID-Laserscanner is a freeware software for 3d laser range scanning. All you need to build a 3d scanner is a PC, a camera (e.g. a webcam), a background corner, and a laser that projects a line onto the object you want to scan. So everyone can use it to scan objects without high costs; this is the big advantage over commercial solutions which are rather expensive.
The concept of DAVID has been developed by the computer scientists Dr. Simon Winkelbach, Sven Molkenstruck and Prof. F. M. Wahl at the Institute for Robotics and Process Control, Technical University of Braunschweig, Germany and was published as a paper at the German Association for Pattern Recognition. ”

(Quote form DAVID website: http://www.david-laserscanner.com/)

So we downloaded everything and bought a license for Shapefusion. Then we followed the very simple instructions. Printing out the A4 patterns and putting them on cardboard was done in a few minutes. Also setting up the webcam was no problem (after downloading the latest drivers). The laser however was not such an easy thing to come by, so we ordered a cheap laser from a webshop that was advised on the DAVID-forum. But we couldn’t wait for this laser to arrive so we tried the option to work with the ‘Thin shadow line option’. For this you need a focused spotlight that can beam a sharp contrasting edge between light and shadow. But it is simply impossible to do that with lights that you may find in your house. Because ordinary lights have a very diffuse beam you end up having not a sharp contrasting shadow line at all.

IMG_7912.jpgAfter going to an electronics shop we where advised to go to a hardware store and look for a cheap laser-level. And so we did find a laser-level for € 3.79 (including batteries)! The quality of this laser is of course pretty bad, but for a first test actually surprisingly useful! Point is that you need a line and NOT a dot laser.

First you need to calibrate the camera, we used a Logitech Connect webcam at a resolution of 640×480. As the instruction mention, keep in mind to put all the automatic lighting settings off. Also be sure you can control the lighting situation in the room, any variation results quite quickly in bad scans.

IMG_7849.jpg This is a photo of the simple setup at my desk at home. To the left you see an external CRT monitor that is attached to my laptop on which we can see the DAVID scanning software. On the LCD screen of my laptop you see a preview of the 3D scan with the pixeldata directly mapped to the polygon object. In the middle is the webcam and to the right the two A4 cardboard grids put at a 90 degree angle. In front of this setup is a Buzz Lightyear toy (sorry, I couldn’t resist).

The scanning procedure is fairly simple. Once the camera is calibrated (the software needs to recognize the location of the black dots in the grid patterns), you simply move the laser over the object and boards at a 30 degree angle. In this manner the software can calculate the difference between the line on the boards and the line on the object. The scanning process happens in a dark environment so the laser can be easily seen by the webcam. Again, it is very important that you can control the light situation in your room.

IMG_7853.jpgIn this photo you can see what the problems are. Firstly shiny and transparent surfaces make the laser scatter all over the place which means that the software can’t register a clean line anymore. Another major problem is the color black. The laser doesn’t emit enough light on a black surface to be seen on the webcam.

Next step is to put on a light and make a snapshot of the objects pixeldata. What is enormously convenient is that the software automatically maps the pixeldata on top of the object!

IMG_7851.jpg On this photo you can see the result on the screen of my laptop. It looks like it is a 2D photo only, but believe me you can actually rotate the scan and you will see thet the ‘photo’ has actual depth. Point is that you need to make several scans from digfferent angles before you have enough data to create a 360 degree object. The scans that you make at different angles (by rotating your physical object) can be put together by the second piece of software from DAVID called Shapefusion.

In the next three photos you see the process of a series of scans of a Buddah sculpture. First foto shows the scanning moment. The second photo the results of three scans form different angles; front (with pixeldata mapped on it), left and right. The third photo shows the three scans aligned properly by Shapefusion.

IMG_7858.jpg IMG_7864.jpg IMG_7863.jpg

IMG_7865.jpgIn this last photo you see the final fused shape imported in 3D studio Max. This is not essential to make a 3D object by the way. The final fused object can be exported as an .OBJ-file that can be imported by many 3D software programs.

All together we are very happy with the first results. Having the setup ready within an hour and having 3D objects with three scans within 15 minutes is an amazing result. The quality of the 3D mesh is highly depended on the resolution of the camera and even more the quality of the laser. The line of our cheap laser has a width of at least 3 to 4 millimeter and is not very bright or solid. So we are confident that the quality of our scans, in terms of detai, will greatly improve with a better laser.

Next step, considering our concept is of course to scan much larger objects; human body’s. Interesting enough a post on this subjet was done during the time we where writing our concept. Have a look at the discussion on the forum of DAVID.

Down below you can have a look at some more photo’s I’ve made.

Comments are closed.

WP Theme & Icons by N.Design Studio
Entries RSS Comments RSS Log in