Hello everyone, I'm very new here and to lasing in general. I am in need of a
massive amount of help / advice concerning a summer project I want to pursue.
I'm a computer science and chemistry student at a small undergraduate university
in Georgia. I'm very near to graduation, but I'm going to take the summer off
to cope with some sad personal "things". One of the projects I am very eager
to work on is building my own xy-scanner laser system. I don't know very much
at all about this field, but I'm determined to have something built and working
by summer's end.
My finals will wrap up in two weeks, and just as soon as they're done I'd like
to have some equipment already ordered and awaiting me.
I don't think anyone on my campus knows about laser scanners, but that said
there is an incredible amount of other expertise available at my disposal.
My Physical Chemistry professor specializes in laser systems and has promised
to help me on a number of issues. We also have a large engineering program and
plenty of equipment with which to build a custom housing for the laser. I can
get hands-on EE help, mechanical help, and so forth.
What I want to build
--------------------
I want to build a laser scanner system for use mostly outdoors. There
may be an occasional indoor use or two, but I mostly need to scale the
image to large dimensions as I'll be projecting onto the
buildings around campus, many of them quite large and quite tall.
(Nothing of the skyscraper variety, though having the capability to
project onto such a surface would be pretty nifty.)
I'm not sure what resolution requirements I'll have (bigger is better),
but scaling the scene up onto these buildings is what's most important. I'll
also need to be able to scale up and down depending on which building I'm
projecting onto. (I assume this has something to do with the galvo angles
being adjusted?)
I've been inspired by (and am in awe of) videos like these:
* [1] http://www.youtube.com/watch?v=EFWcAkxzkv4 (Laser graffiti 1)
* [2] http://www.youtube.com/watch?v=1M19vo0EYfI (Laser graffiti 2)
I'm not interested in simple "graffiti". I want to write applications
such as simple videogames controlled Kinect, Wiimote, or by another laser
(that I can chroma key detect with OpenCV and a webcam).
A Mario or Tetris clone would be easy to write.
I will definitely be doing 3D molecular visualizations because I simply
love cheminformatics. I could also use the system to run slideshow
presentations (indoors) since I give talks frequently.
It might also be fun in the long term to write an algorithm to
'vectorize' a movie like The Matrix, sync the video with the audio track,
and play it over the campus green. I'm not sure how feasible that is, but hey...
I just want to experiment with as much as I can.
Additional interesting demos:
* [3] http://www.youtube.com/watch?v=k6gismmL9yU (Kinect edge detect)
* [4] http://www.youtube.com/watch?v=h_Yu7bZiOig (Kinect control animation)
Video [3] for instance looks like it's just Canny or Sobel edge detect in
OpenCV. You don't even need a Kinect for that.
I expect that I'll be writing a lot of software this summer. Software to
communicate with and control the hardware, software to optimally control the
galvo positioning to achieve highest speed, a higher abstraction layer to
convert vector primitives into these hardware instructions, and then the
software applications themselves. For the most part, I don't mind low-level
programming and would rather not have to pay for any commercial software if
it's feasible, but if there is a chance that my programming could in any way
damage the equipment I may go with a pre-existing commercial solution if no
open source alternative exists.
(Any software I write will be hosted on github and available for all of you to
use as you see fit.)
Questions on Hardware, Theory, Practice, etc.
----------------------------------------------
1) Vector/Raster.
Galvos typically draw vector rather than raster images, correct? (I assume the
column-by-column, row-by-row demands of raster images are far too slow to reach
a framerate humans can percieve as video.) But are vector images still drawn
in row/column order, or is the image composed as the shortest distance "as the
crow flies" or by some kind of similar heuristic? I imagine such shortcuts
could vastly improve the framerate unless the scene itself is just too complex.
If there's any literature/theory out there, perhaps I can experiment with
writing different positioning algorithms depending on the application and
scene complexity--I think that will be really fun.
2) KPPS.
How expensive ($$) are they, and how high does it need to be to look good
on video without jitter? (Many of the videos above had a lot of jitter on
camera.) Does it scale inversely proportional to scene complexity? I want to
maximize KPPS (I think) and attain a very crisp yet complex image, but
minimizing cost is also really important to me.
3) Laser type, color, etc.
What colors project with the brightest intensity and can reach the farthest? If
I want to project onto the side of my campus dorms, or even a tall building,
what are the best choices? The standard green 532 nm laser seems to be well
known for distance and brilliance--is this the gold standard? How expensive are
others? If I order just the diode, how hard is it to build the housing?
4) Multicolor.
Can you mount different color lasers into the same galvo setup? That is, can
I use a single xy-galvo to display a multicolor scene by simply controlling
the software? If so, how much does that slow down the KPPS? (Linearly? Or not
at all?) Is it "better" to have only one color laser per xy-galvo? Or can I hack
multicolor?
5) Hardware setup.
Can you concisely describe the primary components of a typical scanner setup?
Lasers, xy-galvos, ILDA?, ... (there are lots of other components too?) What
about typical minimal costs for these components? Any open source software I
might be able to use? (Can you guys describe ILDA in general or point me to
something to read?)
6) Expected life of hardware.
How long do galvos last before burning out? How about lasers? Do different
types of lasers/galvos have longer life expectancies?
7) Clouds.
Are there any laser scanners that can project onto low stratus clouds? Is
this safe/legal to do given aircraft, etc? I know lasers are frequently used for
astronomical observations, but I certainly don't want to harm any pilots. If
these are legal, how much power would such a system require, and how expensive
would it be? It'd be pretty badass to play tetris on the clouds.
Conclusion
----------
This project _has_ to be done on the cheap. I'm considering dipping into my
scholarship and financial aid money to develop this system instead of eating
actual food. If I do a good job, I'm fairly certain that I'll make all that
money back from donations / additional scholarships. You see, my school
never sees students pull off stuff like this, and I'm pretty sure it will
resonate. (That's not why I'm doing this, but I do think my investment has a
high probability of being returned.)
That said, I need to minimize risk; I only feel safe using $2-3 k. That's the
budget, and I know it's probably unreasonably low. But I have to make it work
some how. Even if I have to cut some corners.
Everything I don't buy will have to be built on-site from materials
available around the campus. I expect that the galvos will eat the large
majority of the costs, and I think that's where I want to sink the money
anyway since I want a large, crisp, high KPPS image.
Also, if I impress the school with this, I might be able to get them to fund
future, larger-scale lasing endevors.
I'll keep checking back here, but I also intend to hang out in your IRC channel
as well. If you can walk me through the theory, a potential build that matches
my budget, and what I can expect, I would be extremely grateful.
Many, many thanks in advance!
(Sorry about the line wrapping at ~80 chars, I wrote this in vim before pasting here.)