What is the theory behind old mechanical coin counter used in the banks? Can pattern recognition make it more cost effective and accurate? What is the algorithm for matching the picture of a coin shot when the coin is rotated some angle from the stored images in the memory?
Coin sorter and counter, using pattern recognition?
Started by ●April 7, 2005
Reply by ●April 7, 20052005-04-07
Acceed See wrote:> > What is the theory behind old mechanical coin counter used in the banks? > Can pattern recognition make it more cost effective and accurate? > > What is the algorithm for matching the picture of a coin shot when > the coin is rotated some angle from the stored images in the memory?Progressively differently sized holes. You cannot get better than that. If you have looked at coinage lately you know you cannot tell a quarter or a nickel from toy money. The Treasury Department hoped to take a huge amount of money out of circulation by making it collectible - 50 different kinds of quarters alone. Debasement of national currency is one of the BIG warning signs. -- Uncle Al http://www.mazepath.com/uncleal/ (Toxic URL! Unsafe for children and most mammals) http://www.mazepath.com/uncleal/qz.pdf
Reply by ●April 7, 20052005-04-07
In sci.physics Acceed See <invalicd@hotmail.com> wrote:> What is the theory behind old mechanical coin counter used in the banks? > Can pattern recognition make it more cost effective and accurate?It is a sieve, and no, mechanical coin counters are extremely fast, accurate, and the ones used by casinos run faster than people can pour coins into them.> What is the algorithm for matching the picture of a coin shot when > the coin is rotated some angle from the stored images in the memory?Google pattern recognition. -- Jim Pennino Remove .spam.sux to reply.
Reply by ●April 8, 20052005-04-08
In comp.dsp,sci.physics, On Thu, 07 Apr 2005 18:33:14 -0700, Uncle Al <UncleAl0@hate.spam.net> wrote:>Acceed See wrote: >> >> What is the theory behind old mechanical coin counter used in the banks?Certainly size, and I would think weight would also be fairly easy to measure, but they may not even measure that.>> Can pattern recognition make it more cost effective and accurate?It doesn't seem worth checking any more than size, since counterfeiting coins would be such a low-profit thing to do (especially considering the risk and possible consequences of getting caught) that no one would bother doing it. You can assume all coins are real, and reliably sort them by size. OTOH the technology to do this is getting cheaper by the minute, and if a bank believes extra security is needed (and it's cheaper than paying a teller to look through coins for fakes), a bank would buy such a thing. But don't base a business plan on this paragraph.>> What is the algorithm for matching the picture of a coin shot when >> the coin is rotated some angle from the stored images in the memory?One way is to rotate the the captured image a small amount, maybe one degree at a time, while comparing it to the stored image. There's probably faster and more reliable ways to do that, but that would be a discussion for another group, probably concerning AI. What does this have to do with DSP and physics, anyhow?>Progressively differently sized holes. You cannot get better than >that. If you have looked at coinage lately you know you cannot tell a >quarter or a nickel from toy money. The Treasury Department hoped to >take a huge amount of money out of circulation by making it >collectible - 50 different kinds of quarters alone. > >Debasement of national currency is one of the BIG warning signs.Warning sign of what? When I was growing up I heard about the Seven Warning Signs of Cancer... ----- http://mindspring.com/~benbradley
Reply by ●April 8, 20052005-04-08
Ben Bradley wrote:> One way is to rotate the the captured image a small amount, maybe > one degree at a time, while comparing it to the stored image. There's > probably faster and more reliable ways to do that, but that would be a > discussion for another group, probably concerning AI.I think that there are some properties of the Fourier transform of the image which are rotation invariant. This'll help you out. Cheers, Nicholas Sherlock
Reply by ●April 8, 20052005-04-08
isn't there a mechanism too that works with 'Eddy Currents' ( http://www.du.edu/~jcalvert/phys/eddy.htm ) Once, I was told that soda vending machines worked that way to distinguish between real coins and othezr metal disks of the same size regards, srtnxalt Acceed See wrote:> What is the theory behind old mechanical coin counter used in the banks? > Can pattern recognition make it more cost effective and accurate? > > What is the algorithm for matching the picture of a coin shot when > the coin is rotated some angle from the stored images in the memory? > > >
Reply by ●April 8, 20052005-04-08
In sci.physics SirThanxALot <sirthanxalot@spymac.com> wrote:> isn't there a mechanism too that works with 'Eddy Currents' ( http://www.du.edu/~jcalvert/phys/eddy.htm ) > Once, I was told that soda vending machines worked that way to distinguish between real coins and othezr metal disks of the same size> regards, > srtnxalt> Acceed See wrote: > > What is the theory behind old mechanical coin counter used in the banks? > > Can pattern recognition make it more cost effective and accurate? > > > > What is the algorithm for matching the picture of a coin shot when > > the coin is rotated some angle from the stored images in the memory?Most of those are simply a magnet that deflects the path of ferrous (the most common non-coin) objects into the reject slot. -- Jim Pennino Remove .spam.sux to reply.
Reply by ●April 8, 20052005-04-08
Acceed See wrote:> What is the theory behind old mechanical coin counter used in the > banks? Can pattern recognition make it more cost effective and > accurate?not sure if that would work on old or worn or damaged coins. might be an idea to have a sensor or similar which can read what metal the coin is made of. no idea if that's feasible. sammi
Reply by ●April 8, 20052005-04-08
jimp@specsol.spam.sux.com wrote:> In sci.physics SirThanxALot <sirthanxalot@spymac.com> wrote: > >>isn't there a mechanism too that works with 'Eddy Currents' ( http://www.du.edu/~jcalvert/phys/eddy.htm ) >>Once, I was told that soda vending machines worked that way to distinguish between real coins and othezr metal disks of the same size > > >>regards, >>srtnxalt > > >>Acceed See wrote: >> >>>What is the theory behind old mechanical coin counter used in the banks? >>>Can pattern recognition make it more cost effective and accurate? >>> >>>What is the algorithm for matching the picture of a coin shot when >>>the coin is rotated some angle from the stored images in the memory? > > > Most of those are simply a magnet that deflects the path of ferrous (the > most common non-coin) objects into the reject slot.I have a coin-operated time switch that I took out of a junked motel TV set in 1950. It accepts quarters and just about nothing else. An initial slot prevents too large a coin from getting in. A quarter is stopped in its fall and from rotating by two fingers just close enough together to keep it from falling through. The fingers are attached to a bar that pivots about an axis normal to the quarter. The quarter's weight tips the bar, the quarter tips out at a specific angle, and begins a free fall between walls that keep its axis horizontal. It has an initial velocity and rotation rate that depend on its weight. It passes through a magnetic field which slows it, altering the trajectory by an amount that depends on conductivity (or capturing it if it's magnetic). It then bounces off a pin and through another slot. Beyond the slot is a microswitch which starts a timer motor. (I wanted the motor to drive a telescope, but I hadn't the heart to tear the timer apart.) Coins that don't pass through the activation and collection slot come out of the coin-return slot. Pushing the coin-return lever spreads the constraining walls to free jams and wipes the magnet to eject iron slugs. It works with modern composite quarters too. They were designed to match the eddy-current braking characteristic of silver quarters. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 8, 20052005-04-08
jimp@specsol.spam.sux.com writes:>> Acceed See wrote: >> > What is the theory behind old mechanical coin counter used in the banks? >> > Can pattern recognition make it more cost effective and accurate? >> > >> > What is the algorithm for matching the picture of a coin shot when >> > the coin is rotated some angle from the stored images in the memory?>Most of those are simply a magnet that deflects the path of ferrous (the >most common non-coin) objects into the reject slot.Even rather simple coin processors use magnets to induce eddy currents in falling coins to deflect their path. Good coins will follow a known path while something like an aluminum slug will follow a different path. A cool demonstration is to drop a strong magnet through a piece of copper pipe and compare that time to dropping something nonmagnetic through it. (this has the "coin" and magnet reversed) -- -Mike
