KAPSID – Neko's Blog

Form Follows Function

Fiducal Tracking Works w/o IR Bandpass Filter

Hmmm…my fiducials are now getting tracked, when i remove the IR bandpass filter. Replaced my my webcams lens with the 4mm wide angle.

Need to fiddle ’round a bit more, i guess.

picture-101

w/o bandpass filter

picture-111

with IR bandpass filter

Filed under: Fiducials

Fiducial Recognition almost there

picture-41

7.30am and no coffee yet. No fiducial recognition either…but almost there. I’ve played around with the contrast, brightness and sharpness adjustments in the camera settings dialog and at least achieved reacTABle to noticing blobs inside the fiducal markers. Means it can seperate darker parts from white spots. I’m quite sure now, that it is only about my lens/sensor issue. Tried my 4mm wide angle lens as well (not on the table though) and get super sharp images, when hovering sightly above the plastic case surrounding the sensor. So same problem, I’ll have to make up for the missing distance between both parts. But first i’ll get a coffee… 😛 Post my results later this day.

Cheers

Filed under: Fiducials

DSI SetUp Improvement

0.26am: 4.5 hours and 2 Alley McBeal DVDs later, I find myself in a pretty messy garage with a few fiddly improvements on my DSI setup that have been worth the effort. Firstly did i raise my Endlighten about a few millimeters and also brought my leds closer to the acrylic. Both done with 2 extra layers of white foam tape and added an uppermost layer of aluminium tape to reflect as much IR light as possible back into my setup. Secondly I’ve exchanged my drafting vellum projection plane with a regular tracing paper. Last and most time consuming procedure of all has been the re-soldering of pcb conductors that happened to break whilst step 1. Next time I will buy those quite costy little connection modules to avoid breakage of conductors and allow a reassembly by just screwing those pcbs apart. Now I know that they would have been worth it. In any way…I’m relatively satisfied with my blobs, that glow like WOW! Relatively, because my fiducial recognition still didn’t work out, even though the images became quite more readable. I assume its my dirty hack camera I’ve posted about recently, that still causes some minor blur and therefore probably prevents tracking. I guess I call it a day and tackle this problem tomorrow. G’night

picture-11

Filed under: Hardware

DSI first results

It is 01.40 in the morning and I just got my first results, which are presentable but still far from good.  There’s a massive delay of about half second between finger/blob input and projector output. I’m also loosing my blobs, whenever I move my fingers to fast. Not quite sure what exactly causes it. Could be my very old HP presentations projector I’m using and the configs in addition. Fiducials are recognizable for the human eye, but reacTIVision still doesn’t track them, which is a bugger. The fiducials are a bit blurry and their contrast whilst capturing to low. Will have to clarify the delay and fiducial issue until this week ends, otherwise I may have to consider dumping DSI for the presentation. I’m still interested how the results in FTIR would look like, even though I start to dislike the idea of a compliant surface. Either way…enough for today. G’night.

Filed under: Hardware

Fiducial Tracking with reacTIVision

picture-22

My first look into reacTIVision’s fiducial tracking application. Tested it on my builtIn cam only, but seems to be straight forward. Can’t wait to get my table setup sorted to run proper tests. I really have to boost the tabe building process. It takes way to long for I need to start coding end of next week.

Can’t wait to have a reliable setup.

Gabor

Filed under: Fiducials

Endlighten Acrylic

Finally I’ve got my huge piece of Endlighten cut into smaller pieces. Can fit it now into my IR-frame. Did test the setup already and unfortunately got rather less than average results. This has probably the following reasons:

  1. My IR frame still needs to be adjusted to the new acrylic measurements. 4mm less thickness cause the IR leds to sit slightly above the upper edge and have now, due to the 45 angle cut edges of my Endlighten, almost 1cmspace between acrylic and leds.
  2. My IR frame is still not “sealed”, which lets heaps of IR light escape into the surrounding. Does not do me much of a favor 😉
  3. My camera is slightly out of focus. I’ve replaced the now missing IR filter with a tiny, clear piece of plastic to compensate the additional space between lens and sensor. works well on shorter distances but is useless when trying to capture detail that is more than 30cm away from the camera.
  4. I probably have to get another mirror. The one I’m utilizing right now doesn’t seem to do it’s job to well.

Nevertheless did I actually use my MT device today for the first time. This is because I could run the fire demo .swf, which did made my fingers emit fire particles on the touch screen. pretty cool but still to inaccurate to show pictures or flicks at this point. But this will change very soon.

Gabor

Filed under: Hardware

Compliant surface on hold

The compliant surface is on hold for at least another day.

This has two reasons:

  1. In middle of the process I realized, that I haven’t had enough silicone. In fact it lasted exactly for two layers, which is not enough.
  2. I’ve got my Endlighten Acrylic cut into smaller pieces. Each at 80x60cm to fit my IR-frame setup.

Will continue on either wednesday or thursday. Inhaling this stuff is nasty and i can not claim looking forward to the second session. Needed a whole day to get these fumes out of my garage.

Gabor

image0391

Filed under: Hardware

Hello world!

Today I’ve decided to set up a blog to documenting my progress in writing a multiple touch application for a FTIR & DSI set up in AS3.0.

This is because i managed to place my 80x60x1cm acrylic sheet inside my IR-Frame, which against all fears turned out to be a smooth fit.

Today I will start creating the compliant surface and make my mind up about where to get my piece of Endlighten acrylic cut. The tight time frame until end of this semester and my low budget force me to stick with my 80x60cm setup. But everything’s gonna be fine.

Here are the first snaps…it works 🙂

picture-8IR FrameFTIR SetUp

Filed under: Hardware