Dear DAVID Community,

As you are probably aware, the entire DAVID company, team and product range has been integrated into HP Inc.
This forum is now read-only. Relevant content has been migrated to HP and merged into HP's Support Forums on November 1, 2016.

To start new discussions for 3D Scanning, please register and post your new topic at the HP Support Forums

Reflective scanning?!...

Questions, problems, comments and tips regarding the 3d scanning process.

Reflective scanning?!...

Postby RinksCustoms » Fri Jan 09, 2009 3:12 am

Hello all, first and foremost a thousand thank you's for everybody's efforts writing and experimenting with David software. In my application of David software does not involve reverse engineering but after market and custom design for local street & race cars looking for an edge or want to see how various body kits will look on their car before purchase or sign a contract for custom work. The cars I customize 80% of the time have paint jobs that are $10K+, and the customer and I would feel much better if there was a better way to obtain scan data than painting a car with even water based paints, (which involves a very tedious cleansing process to remove all the silicon based dressing and waxes used to protect the paint from UV damage).

I'd like to present a possible solution but i'm not a very experienced David user, so i would like to present my concept to all of you for your expert opinions before building an expensive motorized scanning platform. They say a picture is worth a thousand words.... NOTE: the light being reflected is not to scale due to youtube's video processing where the original one pixel width line was lost.
Image

And if a picture is worth a thousand words, a short animation should make the point that much better... a short animation and perspective view of the rig in action. For realism all the objects in this simulation are full size dimensions, ie. the front of the car is 3 meters wide.

click here for Animation

Here is the full 16 second clip i'm using as my simulation...

click here for Movie

The big question I have.. With Davidlaserscaner setup for background-less scanning and optimal camera settings, will it be able to detect the shape of the car by the reflection of the light in a similar way that a non glossy objects are detected when a laser line is passed over them? It is likely that i will need to use the mask feature to "hide" the windshield or similar surface which will create a double reflection which i think the David software will not cope with too good. I do not mind if i need to patch a few holes later in CAD.
Last edited by RinksCustoms on Sat Jan 10, 2009 9:27 pm, edited 1 time in total.
RinksCustoms
 
Posts: 8
Joined: Thu Jan 01, 2009 7:05 am

Re: Reflective scanning?!...

Postby MagWeb » Fri Jan 09, 2009 8:47 am

RinksCustoms wrote:
The big question I have.. With Davidlaserscaner setup for background-less scanning and optimal camera settings, will it be able to detect the shape of the car by the reflection of the light in a similar way that a non glossy objects are detected when a laser line is passed over them? It is likely that i will need to use the mask feature to "hide" the windshield or similar surface which will create a double reflection which i think the David software will not cope with too good. I do not mind if i need to patch a few holes later in CAD.


Hi RinksCustoms,

Interesting project, and a new dimension for DAVID!

Ok, some notes, (hope I didn´t miss some ideas):
-DAVID´s mask function works only in the calibration step and does not influence the scanresult (but such a thing would be a nice feature!).
If there is more than one laserline detected in one vertical pixel column, DAVID chooses the brightest.
-Think it doesn´t matter whether you scan planeless or with background, the problem of your application stays. If I got your idea right, you try to get rid of some mirrored beams by hitting the object perpendicular (so you choosed a linear movement). But for the car´s surface isn´t flat the direction of the mirrored beam isn´t predictable.
-To get light visible for the cam right on the surface itself, you´ve to break the mirroring or transparency effects of the shiny surfaces. So I cann´t imagine a methode, dealing with light, avoiding a treatment of the surface. Othwerwise the object will remain a "stealth fighter". I think you don´t have to paint the cars white. Maybe there´s a dusty road? Looking at my car, standing out there, scanning should be no problem :wink: oooh...and :idea: for the winter-salted roads: it is quite white and mate (maybe a sprayed on salt solution can work?)
-For I see no advantage of the linear movement in your application, I would use a pivoting laser movement. That should make the construction a lot easier and cheeper...

Gunter
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Reflective scanning?!...

Postby RinksCustoms » Sat Jan 10, 2009 9:24 pm

MagWeb wrote:...
-DAVID´s mask function works only in the calibration step and does not influence the scanresult (but such a thing would be a nice feature!).
If there is more than one laserline detected in one vertical pixel column, DAVID chooses the brightest...

That is very useful information on how david works, as i was not sure what David does with the "detected line"..

-Think it doesn´t matter whether you scan planeless or with background, the problem of your application stays. If I got your idea right, you try to get rid of some mirrored beams by hitting the object perpendicular (so you choosed a linear movement). But for the car´s surface isn´t flat the direction of the mirrored beam isn´t predictable...

I was thinking along that line too, that David would not know what to to with the multiple reflections of all the curves and isolated spots of those reflections as the scanner passed overhead.. Also if you were to take a fluorescent light (like a cheap 13" or so from Lowes/Home Depot/ect.) and hold it vertical to your car at night and position your eyes at a fixed point somewhat perpendicular to the light and parallel with the car while moving the light with your arm in a linear motion also parallel to the car you'll notice the phenomenon i poorly describe as the replacement laser line being detected by David. The reflected light from the neon tube (which is a line of light) should be detected in an identical way that the laser line is detected by david. But after your post and reviewing the animation in detail while imagining how David would position the plotted x,y,z points i realized that the reflective scanning may only be applicable for background-less scanning of optical lens faces (using th reflected light of a thin neon or CCFL lamp (2.0-5mm dia) to produce the lenses surface shape 100% without distortion. You made a good point about the secondary reflections (isolated pockets and other surfaces like the windshield) becoming a problem as David would confuse those as a false detections in x/y/z cordinates and make a very funny looking scan indeed.

I suspect that a reference scan would have to be done with a mirrored floor of the neon tube moving fully to it's endpoint and David capturing the reflected line as it moves along the track. This light line would be captured as a completely straight line unlike the standard angled line David currently uses. This would also require a whole new detection mode to be written in software for David to correctly plot the convex and concave surfaces in x/y/z space against the reference scan. This type of scanning would likely have the user to measure and input camera angles to the mirrored floor and distance from the center of the "mirrored" image that the camera see's. This would be a neato feature for David 3.0, and likely would only be able to be pulled off by the same geneouses that wrote this version of the software.

-To get light visible for the cam right on the surface itself, you´ve to break the mirroring or transparency effects of the shiny surfaces. So I cann´t imagine a methode, dealing with light, avoiding a treatment of the surface. Othwerwise the object will remain a "stealth fighter". I think you don´t have to paint the cars white. Maybe there´s a dusty road? Looking at my car, standing out there, scanning should be no problem :wink: oooh...and :idea: for the winter-salted roads: it is quite white and mate (maybe a sprayed on salt solution can work?)
-For I see no advantage of the linear movement in your application, I would use a pivoting laser movement. That should make the construction a lot easier and cheeper...

Gunter


The dust and salt solution is a very good idea. The salt would dry so that the car is coated evenly, matte in finish, and not affected by the waxes and silicones, but not to tell the owner exactly what the solution is as to not make them worry about the effects of the salt on the paint (possibly $10,000+ USD).

I have another possible way of scanning a full size car into David that occured to me only yesterday. In my profession that is autobody collision, when we goto refinish the cars after repair we use this plastic cover (much like that of a thin grocery store plastic bag) that is 3 or 4 meters wide when fully opened to cover areas of the car that isn't to be refinished. When in the spray booth and the exhaust fan is turned on, it pulls air that is around the car down through a grate covered pit in the floor and expels this to the atmosphere. This results in a slight vacuum under the car, sucking the plastic cover tight to the car and following the contours almost exact. Not sure if David would pick up the folds that occur when the plastic wraps around a contoured surface. The laser should be detected off this plastic easily as it can be ordered in white (translucent only, better), or blue (more transparent, maybe not so great) from an auto paint supplier. Using the spray booth i can generate near darkness, have mounting points for the laser(s) tracks and have room for the mounting of the camera positions. So the linear movement of lasers/neon becomes nearly as cheap as a rotational laser movement.

As always, I thank any insight and help with my problems,

Thank you Gunter for the insight.
RinksCustoms
 
Posts: 8
Joined: Thu Jan 01, 2009 7:05 am

Re: Reflective scanning?!...

Postby MagWeb » Sat Jan 10, 2009 11:19 pm

Hi again.

Aha, now I´ve got your idea,....
reflection of a line (neon) instead of projecting a line (laser).

I allready wondered about the specific charakter of the line in your animation.

The either function of the laser is not to project a line on the object, but to "cut" a plane "slice" of the object. All points enlighted by the laser are exactly on one plane (and they don´t have to, if you reflect a line as your neon). DAVID calculates the position of this plane by looking onto the lines on the calibplanes (or the reference scan). Once found this position, the 3D coordinates of the line on the object (in 2D) can be calculated.

Maybe one could build a scanning system for mirroring objects on the base of mirroring a line, but DAVID nowadays thinks another way - sorry.

You´re right: I can imagine that people might be not very happy covering their expensive cars with salt - was just a quick idea

Gunter
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Reflective scanning?!...

Postby florin.topala » Sun Jan 11, 2009 12:34 am

for covering the cars with salt :) ....

it is a ductape ... or maybe similar which is verry easy to remouve and it is used for painting ... I think it is usefull for XXXXXL projects

Florn

PS sorry for my bad english
florin.topala
 
Posts: 104
Joined: Mon Feb 19, 2007 12:47 am
Location: Timisoara, Romania

Re: Reflective scanning?!...

Postby MeatHEad » Sat Dec 11, 2010 12:12 pm

why were the youtube videos removed?

Does an ideal-like reflective tape exist that can be put over surfaces to get accurate point-cloud data that requires no background with David-Laserscanner?
MeatHEad
 
Posts: 14
Joined: Thu Dec 02, 2010 7:55 am


Return to 3D Laser Scanning

Who is online

Users browsing this forum: No registered users and 3 guests