Since my last update I've been learning how to use ReacTIVision with Flash. It's been hard because there's not as much information online to help as I'd like but I (think) I've finally figured it out - with a lot of help from the forums I've found.I downloaded the sample Flash client on ReacTIVision's site to get an idea of how to integrate the webcam and ReacTIVision with Flash. It's more complicated than I thought. I had to download (and figure out how to use) Flosc which acts as a gateway for Flash to talk with other software. This allowed me to run the Flash client which represents a fiducial symbol as a blue block on screen.
A screenshot of what this looks like: The two terminals run java and jar files which flosc and the flash demo need for this to work. Under the terminals is the flosc client. On the top right is the flash demo with blue blocks showing the location of the fiducials that ReacTIVision (bottom right) is picking up.
I'm still having a couple of problems with this. More often than not, the flash demo's terminal throws an exception which stops the flash from working. Other times, it simply stops working with no explanation! I'm going to look into this more to see if I can figure it out.
I'm also having some problems when I position the camera under a sheet of plastic and use that as a table to move the markers around on. Glare and reflections are sometimes blocking the camera's view of the markers. I'm going to try to get my hands on some frosted plastic or glass (or frost my own glass!) which will hopefully prevent reflections.
On Monday I'll be in Dublin and going shopping for some materials - the sheet of plastic/glass being the most important. I'll also need some card to make the outline of the map (and maybe models of the buildings if I think I'll have time).
I ended up using a different method than I had planned. I'm still using the same software, Swift 3D, but in a different way. Originally, I exported the model using Papervision3D, which resulted in a great interactive 3D flash file but because the model was generated after it was executed there was no way to add any information on top of the map.
Instead, I animated a walkthrough of the model in Swift 3D and exported a swft file to flash. This way, each frame of the animation has its own frame on the timeline in Flash and I can see the model on the stage.
I'm still looking into the original idea but I think this might end up being the best option.
Below is a video of the walkthrough (it's a lot less blurry in real life!). I want to go back to the map soon to add some more detail and make it look more realistic.
For the rest of the week I'll be working on tracking a marker's X and Y coordinates in Flash using reacTIVision and applying this to the model. I'll also be busy working on the next draft of the thesis for the end of the week.
For the past few days I've been thinking about how the user will interact with the map. I had planned on using RFID tags and a reader to translate the user's movements to the map on screen.
I decided to focus on the use of the webcam to track movement for the moment. I heard about fiducial markers and using them with a webcam to track their coordinates. While I was researching this I came across reacTIVision which Nora had talked to us about earlier this semester. It's open source software which is used to track fiducial markers and has been used for a lot of really interesting projects. The reactable seems to be the most successful use of the software.
I think this would be a great way to track the user's movement using the fiducial marker and map the X and Y coordinates to the Flash map.
So far, I've downloaded reacTIVision and printed some of the packaged fiducial markers to see how it works. I held up a random marker and it did recognise it! I'm really enjoying using it at the moment and think it could befinitely be something I would like to use. Unlike most of the other projects I've seen that have used it, I'm not planning on projecting anything onto the table. I'm not sure if this will make using it more or less complicated but I'll soon find out!
The two terminals run java and jar files which flosc and the flash demo need for this to work. Under the terminals is the flosc client. On the top right is the flash demo with blue blocks showing the location of the fiducials that ReacTIVision (bottom right) is picking up. 