i'm not going to give you a manufacturer/model #.

some sort of webcam would probably be the first thing to try and would be the easiest to set up since all the streaming is set up for you. You wouldn't have to get the focus perfect, but you would probably have to tinker with it a little bit. And you'd have to make sure that you could see the entire board - shouldn't be a problem if the through-holes are cut through thin enough material (tin? etc etc).

You'd have to also tinker with your lighting. but, like I said, these are trivial problems that can be worked through in a day at the longest.

I'm not going to answer any more of these general "how is the camera set up" questions. This post was a freebie given in good faith.

Give me specific questions related to specific problems you are having with your comprehension of this solution and I will be happy to give you answers.


[Edited on November 15, 2005 at 5:33 PM. Reason : s]

11/15/2005 5:29:43 PM

Quote :

"some sort of webcam would probably be the first thing to try and would be the easiest to set up since all the streaming is set up for you. You wouldn't have to get the focus perfect, but you would probably have to tinker with it a little bit. And you'd have to make sure that you could see the entire board - shouldn't be a problem if the through-holes are cut through thin enough material (tin? etc etc). You'd have to also tinker with your lighting. but, like I said, these are trivial problems that can be worked through in a day at the longest. I'm not going to answer any more of these general "how is the camera set up" questions. This post was a freebie given in good faith. "

Like I said, you just assume something will work, use your (about to be painfully obvious) poor engineering intuition, and call something simple.

Lets start here:
http://jpbrown.i8.com/cubesolver.html

A logitech camera is used here. We initially see that an SDK is provided to aid in this recognition task that we have before us. Things are looking up. Then we read a little more and we find out that we are in for a world of hurt. Specifically:

Quote :

"The longest part of the this project involved writing the color-recognition software"

Keep in mind it was harder to do this than to construct a robotic apparatus to turn the cube (and you can read all about the engineering details of this).

Furthermore, in this instance, the camera is at close range to this cube, and still errors once per cube (when well calibrated)and required a confirmation step be included into the control.

How do you think the camera will fare at a large enough distance to capture the entire board? My guess, pretty shitty. And this is before we start considering that you will have to have some type of shitty looking chess board with holes big enough for the camera if it is even at all possible.

So what are our other options? One could be to make the camera move over the bottom of the board surface. I saw this was done by some college kids with a checkers board (I think I mentioned it in the thread earlier) but then you have to have a whole mechanism to do this, which is by no means simple.

Another option I think could work would be a high resolution photo camera. But then you have to have some sort of interface to the camera to take pictures AND get them over to the machine for processing. All in a timely enough fashion so that the music that plays in reaction to the game is realistic, and not seconds later.

The point being, there is nothing simple about a camera recognition scheme at all. I am relatively certain I could take one of these $20 boards that have the "sockets" for the peices, and hack it apart, tap into the wiring for it (maybe even the logic) and produce a 56k serial upstream interface to a PC quicker and cheaper than any camera scheme.


I wasn't even trying to make this an owned thing. You gotta understand I'm gonna call out bullshit when I see it.

[Edited on November 15, 2005 at 8:18 PM. Reason : x]

11/15/2005 8:13:48 PM

i want to see this $20 socket board you are talking about

11/15/2005 8:22:51 PM

Oops, meant $40

http://www.chesssets.com/cart/shopper.cfm/action=view/key=CL008?source=affiliates&bid=4042&aid=CD1723&opt=Kasparov_Travel_Chess_Computer

11/15/2005 8:31:43 PM

how does it distinguish between the parts, seeing which peice has moved?

11/15/2005 8:42:30 PM

I don't know for certain but as the peices have to go down in pegs, I assume it detects that if its whites turn to move then 2 peices have to be picked up to make the move happen and that whichever hole is filled at the end of the move must have a white peice now. It knows the current state and valid next states.

There could be something "in" the peices that helps in legal move detection but I'm just speculating.

[Edited on November 15, 2005 at 8:47 PM. Reason : x]

11/15/2005 8:46:00 PM

please write up a formal proposal and the team will take it into consideration

eh that sounded too sarcastic
seriously, I'm interested

tell me more

[Edited on November 15, 2005 at 9:02 PM. Reason : -]

11/15/2005 9:00:57 PM

Quote :

"The longest part of the this project involved writing the color-recognition software" Keep in mind it was harder to do this than to construct a robotic apparatus to turn the cube (and you can read all about the engineering details of this). "

its completely different situation. the cube was not in an enclosed, color controlled area. with the chess board there will be very little color variation beyond the pieces moving.

if you actually read your own link, you'd see that the problems they described with the color recognition software are due to lighting and color variances - both of which will be GREATLY minimized in the chessboard setup.

Conclusion: you haven't demonstrated any problem with using a camera for this problem. it is still the easiest and quickest solution. not to mention the most elegant proposed so far.

[Edited on November 15, 2005 at 9:34 PM. Reason : s]

11/15/2005 9:25:32 PM

Just stop posting.

You are a two bit troll and nothing more.

11/15/2005 10:13:00 PM

are you fucking shitting me. whatever.

11/15/2005 10:26:41 PM

Quote :

"Conclusion: you haven't demonstrated any problem with using a camera for this problem. "

Yes I did. Now stop posting, your idea is dead.

11/15/2005 10:29:28 PM

no you didn't. just because they used a webcam doesn't mean the situations are at all similar and its pretty obvious that you're avoiding the fact that the chessboard would be a fully enclosed and color/light controlled environment, UNLIKE the example you posted. which is significant

but look

i've already gone through a lot of these problems in classes and other projects. i know what i'm talking about and i know the significant difference that a closed environment can have on this sort of solution.

11/15/2005 10:54:46 PM

No you don't.

11/15/2005 10:56:15 PM

alright well its never going to be built so you will never be able to prove that its impossible and i will never be able to prove that its possible.

but, i do find it telling that you never came up with a workable solution in this thread.

11/15/2005 11:03:16 PM

No you don't.

11/15/2005 11:04:56 PM

that's more like it

11/15/2005 11:08:49 PM

so i guess this idea is slowly dying now?

11/17/2005 1:26:59 AM

Updates?

11/18/2005 3:11:57 PM

there's not much to say. snewf asked how to do something he'll never do. three solutions were proposed.

now snewf is off thinking up a new problem and soon enough we'll begin posting in that thread.

11/18/2005 3:33:44 PM

Quote :

"i will eat my own shit in the middle of the brickyard if you follow through with this 100%"

11/18/2005 3:49:22 PM

I think we'll have to wait until the weekend is over for the new problem.

You know, after the rounds of bong hits and circle jerking, etc.

11/18/2005 3:58:28 PM

if i were doing it i'd use the contact/gap idea

for each move you know what piece it was because it was first lifted (contact off) and then placed (contact on).

[Edited on November 18, 2005 at 7:24 PM. Reason : i]

11/18/2005 7:24:21 PM

^ Uh, do you know how to play chess? I fail to see how that would work.

11/18/2005 7:31:10 PM

then you're retarded

11/18/2005 7:59:13 PM

yeah i think i would do the contact gap now after thinking about it some
i thought distinguishing between the peices would be tough but you dont have to for the reason just mentioned

11/18/2005 8:40:23 PM

Real Name : George
AIM Name : G021578
Age : 27
Sex : M
Hometown : Bremerton, WA
Class : Alumnus
Major : CSC


I just died a little

11/19/2005 12:43:24 AM

if I build this thing I feel like I've got license to slap most of you in the face

some of you with my dick

11/19/2005 2:43:11 AM

did you ever find out and/or reveal how they built it the first time?

11/19/2005 4:44:33 PM

Updates

11/23/2005 1:51:30 PM

With your simple explantion of how you would solve this problem it leaves many details to whether this would work or not. If you put as much effort into your idea as how you explained it then I'm inclined to say it probably wouldn't work.

Instead of going into further detail you just wanted to attack me? That is fine, but your suggestion that my age or major would imply I should know how to play chess is fucking retarded. In none of my classes did they once teach me how to play the game. You can claim to have died a little, but as it stands you have only shown that you are quick to attack someone's character.

11/23/2005 2:18:48 PM

He was just doing some pretty weak trolling man, let it slide. He'll eventually check the thread and craft some weak shit about how he wasn't trolling, which is just bait for another response. Ignore it, the kid is lame.

11/23/2005 2:38:52 PM

Last update was 8 days ago?

Project must be too heated to update about.

11/27/2005 4:07:50 PM

shut up fag

i MEANT

that is EXACTLY the kind of thing a CSC graduate should IMMEDIATELY think of. It boggles my mind that you don't know how it would work. Since you know where every single piece starts on the board, and assuming only one piece may be moved at a time, if a contact is broken and then made in a different place, you know that the piece that was on the broken contact is now on the new one. You could really use anything. Pressure pads, elec gap, rfid, whatever the hell you want. It's a solution to the problem of knowing which piece is where and when. The rest would be handled programatically depending on how you want to make the music. I don't care about that part of the problem.

11/27/2005 4:45:02 PM

you mean this isn't done yet?

11/27/2005 5:05:19 PM

^^ That as you have described it isn't anywhere near a complete specification to define the problem which is why it looks like a baseless troll.

So far I got:

gephelps 1
you 0

You need to flesh it out because right now there are some open ends and its really lame that you are attacking this guy based on something as vaguely as you have stated it and him being confused.

And be my guest in trying to add me to either the troll or attack list, whichever your agenda is. Because I'm pretty sure I have you dominated in this thread at this point.



[Edited on November 27, 2005 at 5:32 PM. Reason : x]

11/27/2005 5:31:39 PM

Quote :

"that is EXACTLY the kind of thing a CSC graduate should IMMEDIATELY think of. It boggles my mind that you don't know how it would work. Since you know where every single piece starts on the board, and assuming only one piece may be moved at a time, if a contact is broken and then made in a different place, you know that the piece that was on the broken contact is now on the new one. You could really use anything. Pressure pads, elec gap, rfid, whatever the hell you want. It's a solution to the problem of knowing which piece is where and when. The rest would be handled programatically depending on how you want to make the music. I don't care about that part of the problem."

and that sort of short-sightedness is EXACTLY the kind of thing an engineering student should immediately be suspicious of. with the sort of solution you've described, there's no way to recover from an error and there's no tolerance for ANYTHING whatsoever - maybe the user would like to relax the rules a little bit and play a more casual game.

but the real killer of your idea is the possibility that a user would knock over another piece while moving one of his pieces. Or, that a few pieces would somehow get knocked over independently. In short, this described solution is THE WORST one to be proposed in this thread.

11/27/2005 8:58:45 PM

updates?

11/30/2005 8:11:57 PM

i'll up my original 100 by at least double

12/7/2005 4:51:48 PM

Any updates?

2/2/2006 2:17:06 PM

anxiously awaiting to hear the progress of this idea

2/2/2006 2:28:36 PM

dakota_man's method combined with some sort of output displaying the current status of the board would be best.

The board status is known.

its dude1's turn

dude 1 picks up a piece, moves it to another slot.

The board knows which piece ws moved.

dude1 then hits a button to verify that he approves the move and record time.

the board then validates the legality of the move and records the new state of the board.

Even if you dump all the pieces, the original state of the board is still recorded in software.

The only con here is that different "game modes" would have to be programmed into the board.

2/2/2006 2:49:24 PM

forcing a requirement to add an LCD output and forcing the user to confirm a move is too much overhead. Most importantly it adds unnecessary cost. Less important, but still noteworthy is the unnecessary wrinkle it creates in the standard chessboard user-interface. it is unnecessary bloat considering the far superior solutions that have been suggested in the thread

[Edited on February 2, 2006 at 3:22 PM. Reason : s]

2/2/2006 3:20:51 PM

it can be an lcd, an audible readout, a text print out, or as the original topic suggested: musically, whatever. if you're talking webcam then we already need some sort of PC. Output is not out of the question.

As for overhead in confirming the move, thats part of chess. So its not an addition.

you would need to create a custom chess board, but it can be done without changing the user experience.


The initial setup cost would be the only issue here.

With some sort of visual idenification system there is much much more room for error.



and as far as which one is superior

Which way do you think those electronic chess boards work?

I dont see a camera.

[Edited on February 2, 2006 at 3:37 PM. Reason : .]

2/2/2006 3:35:49 PM

this thread rulez #1

2/2/2006 4:17:25 PM

Quote :

"if you're talking webcam then we already need some sort of PC."

errrr.... once again, its sad that you guys are engineers

2/2/2006 4:24:16 PM

im comp sci.

if you wanted to do this very simply, you'd buy the $30 electronic chess board.

2/2/2006 5:32:25 PM

There is a $30 board out there that plays music on chess moves?

2/2/2006 8:41:44 PM

no, but who says you can't use the output generated by one of those boards to play music.

http://www.thechessstore.com/product/CS975EE

If you could figure out how to get the piece locations off of the board the way the handset does you'd be set.

2/2/2006 9:41:45 PM

I already made that statement in this thread.

2/2/2006 9:55:24 PM

i read like the first post and then part of this page so there you go.

2/2/2006 10:38:30 PM