Yet another newbie

[shadows44]

Hello !
I juste received my laser ( LC650-5-5-F(14x45)-NT ) and I made a few tries...
Looks like I have to improve the scanning !

Of course, since this was just a "I have the laser ! quick, I want to try it !" try, things are far from being perfect :

-not full black during scanning, I should wait night, I guess.
-I printed the patterns on some drawing paper ("canson" paper), but I'm not sure this is the best choice.
-I used walls and the “Calibration Corner” has an angle that would be something like 92°, not 90.
-elephant 1 is in wood but there are many holes, which is, I guess, far from beign the easiest thing to scan.
-elephant 2 is in plastic and there is, I think, some subsurface scattering effect
-laser was hand held.

[shadows44]

here is elephant 2 :

[shadows44]

My surface is heavily spiked. How can I reduce that?
Spikes are often caused by laser reflections at the surface. Try to reduce the camera exposure time / shutter / brightness… such that only one single laser line (without laser reflections) is visible. You can furthermore reduce noise by pressing the smoothing buttons multiple times and/or by sliding the laser more often over the surface. If this does not help, you might coat the object with a matt paint or a powder.

Since I have this issue on the paper, I guess I will have to choose another kind of paper...
I will try to get a good calibration corner this weekend : 90°, not weavy and so on.

[shadows44]

Another try ; I changed the webcam settings : saturation = 0, exposure & brightness = 0 (in order to have a darker picture and a thinner laser).
Everything else is just like yesterday.

It's already much better, especially on the head !

Next step : build a better corner !

[shadows44]

New corner : I user a drawer, pattern printed on standard A4 paper and adhesive tape.
Previous scans were made using a bigger pattern (2xA4).

By the way, to get an idea of the scale : small Ganesh height is 6.5cm.
There were issues with ears, which are translucent and tend to "glow" when the laser hits them.

Camera settings : same as before + contrast set to 100 (lower value, default is 105 and higher value is 200), sharpen set to 0 (I thought this postprocessing effect might not be a good thing for David, and maybe put it off is better)
Gamma = 66 (default value, I don't know if this can affect the scan)

[shadows44]

A "quick and dirty" fusion test !

This was just a try, I know that the scans are far from perfect (holes everywhere !).

I guess the next step now would be to build a system so that I can scan very slowly, without any shaking (this scan was handheld).

[MagWeb]

Hi, welcome

this is a good result for a hand-made-scan. The holes in the partial scans are no problem. Most will be filled by overlappings of others and if not: this is a job for poisson fusion.

Just a hint: Don`t use the smoothing on the scanning part at all, if you plan to join scans with Shapefusion. You loose details and , if used too much the fit of the partial scans might be worse. Use only interpolation. Filled depth discontinueties , these flashes are called "bad linkings" in this forum, do not interfere DAVID's allign process or fusion (with a little exception). These refere to the vertices (points) not to the faces. Poisson smoothes enough while fusing the meshes.

Keep on scanning
Gunter

[shadows44]

Just a hint: Don`t use the smoothing on the scanning part at all, if you plan to join scans with Shapefusion. You loose details and , if used too much the fit of the partial scans might be worse. Use only interpolation.

Ok, I will try this ; actually I diddn't use either smoothing or interpolation.

I also scanned a fossil ( a trilobite, length : 8 cm) ; I used 4 or 5 scans, and the fusion was made using meshlab (I used scans from David without any interpolation or smoothing).

[shadows44]

I also made a real-time version of this scan :
-using meshlab I made a low poly model (the high poly model is between 200 000 and 500 000 if I remember well) ;
-using Wings3D I created the UV-map and corrected the back of the rock.
-here are some screenshots made using Demoniak3D (http://www.ozone3d.net/demoniak3d/index.php?lang=2)

The next step will be to create a normal map using the high poly model, for the low poly one.

[shadows44]

I made this one to test the new David 2.0, and to try a new setting I've made...

By the way, I was so frustrated of being unable to save the objet that I just bought a license.

[shadows44]

Here is another scan, the model is a Colocynth.

I'm satisfied with the result, as it was a "quick and dirty" scan.

By the way , as you can see there's a TDT3D logo at the top... I created this scan for a TDT3D article which will be published very soon, basically it's a review (in French) of DAVID 2.0 !

[hal]

Great work Shadows44,

my compliments! your scan is perfect

Maybe you have already wrote it, but, sorry, how kind of camera are you using? Maybe can be usefull if you show us your hardware setup (camera, laser) just compiling your signature in your David Forum account (the text under every message). Sorry for disturb and many thanks for sharing your amazing results.

Mattia

[shadows44]

my compliments! your scan is perfect

Actually the result is nice, but I had to do a lot of tweaking : I had huge artefacts during the fusion step (with the same scans and resolution set as 300 instead of 600, poisson would basically give me a smashed vegetable... ).

My webcam is a "Hercule Deluxe Optical Glass" : only 20€, resolution can be set up to 800x600 (low framerate). Not very good for texture grabbing but perfect for any DAVID newbie.

I updated my signature, and I will post a picture of my setup tonight.

[shadows44]

Here's the setup !

I use a tripod, the laser line is on top. I guess I could improve this easily (laser & webcam on the same vertical line).
I sweep the laser line by hand : I just move the tripod head.

The new, lightweight corner : A4 sticker paper has been used. The two metallic parts are supposed to prevent any movement and to keep a good angle, however I just added them, There were not there yesterday when I scanned the colocynth.