Project: Train detector

[David Hempy]

So I was sitting on my back porch watching a freight train rumble by, reflecting that there seems to be a lot more rail traffic these days. I postulated what that might reflect of the economy.  Then I wondered if there really is more traffic, or if I'm just noticing it more since I'm a new daddy?  My tykes LOVE the trains.  They stop whatever is going on to watch them pass before waving "bye-bye choo-choo!"

Now I'm thinking about a data logging train detector. How would that work?  Exactly what data would it record?  What would the data reveal?  Would the data be interesting to anyone else? 

Personally, I'm most interested in the detection part.  Here's my desired parameters:

Single point of observation, 100' from the tracks.  (No trespassing!) Power and net (or download) access available. Hobby budget.  All-season operation with three-year reliability, maintenance allowed.

I thought this might be fun for some Collexionaires.  I've got several ideas bumping around in my head, but wanted to use some of your cycles before prejudicing you with my thoughts.  So...think about how you might do this tonight, and what you would do with the data, and how you would share it.

Don't reply yet...discussion starts Tuesday!

-dave

[Jerome Hollon]

If you can setup something on each side of the tracks, it should be pretty easy to set a IR emitter/receiver and see when something breaks the line of sight. You'd want to set it diagonally across the tracks so openings in the carts don't set it off again.

If you can't do that, you could try tracking decibels, you'd need some trial to figure out just how loud a train is, and you might have some false positives, but we all know trains are loud. You could reduce false positives by requiring the loudness to be so loud for a minimum span of time, and prevent it from redetecting the train again by not listening for x number of minutes after a train was detected.

Just some ideas I'm throwing out, not actual implementations.

--
You received this message because you are subscribed to the Google Groups "Collexion" group.
To post to this group, send email to coll...@googlegroups.com.
To unsubscribe from this group, send email to collexion+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/collexion?hl=en.

[Chris Harn]

We looked into this just recently for one our quarries that loads trains.

All trains in the united states are equipped with 2 RFID cards (yes one for each side of the rail car), made by Transcore. (I think maybe all containers as well)

The cards have something like a 4 digit aphanumeric that represents the owner and then also are followed by 3-4 digit number.

I believe these RFID cards may have some read\write area on them as well.

In the case of one of my quarries we send off an email to CSX with each rail car codes and then a series of routing codes.

These people make a system: http://www.softrail.com/

To unsubscribe from this group, send email to mailto:collexion%2Bunsu...@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/collexion?hl=en.

[Cory Stargel]

Not much to add here other than I too am interested and seem to have observed more rail traffic lately.  However, I too may be biased as I just bought a house nearer the tracks.  (about two blocks away from Collexion actually!)

[Dave]

--- On Mon, 10/15/12, David Hempy <scoo...@gmail.com> wrote:

> From: David Hempy <scoo...@gmail.com>
> Subject: Project: Train detector
> To: coll...@googlegroups.com
> Date: Monday, October 15, 2012, 2:06 PM
>
> So I was sitting on my back porch watching a freight train rumble
> by, reflecting that there seems to be a lot more rail traffic these
> days. I postulated what that might reflect of the economy.  Then I
> wondered if there really is more traffic, or if I'm just noticing
> it more since I'm a new daddy?  My tykes LOVE the trains.  They
> stop whatever is going on to watch them pass before waving "bye-bye
> choo-choo!"

I've wondered that about the trains, too, and whether the number
reflects the economy. But, realistically, I also wonder whether
there are simply a higher number of shorter trains. The trains now
seem to be much shorter than they were decades ago, especially back
when they'd have helper locomotives in the middle of the trains.

> Now I'm thinking about a data logging train detector. How would
> that work?  Exactly what data would it record?  What would the
> data reveal?  Would the data be interesting to anyone else? 

The time of the train would obviously be the most important thing to
log, but I'm afraid that information would be mostly useless without
additional information on the length of the train, and, perhaps,
the number and type of cars, and, possibly their cargo capacity.

I think most cars are coded now with a type of bar code, such that
their number can be easily scanned as they go past. That might make
for an interesting database.

However, note that railroad companies are VERY territorial, and won't
permit you to mount anything on their right of way. They may not
even be happy with you doing the scanning remotely, but I doubt that
there's much that they can do about that.

The other concern would be whether any such information could
be accessed and used by an terrorists.

> Personally, I'm most interested in the detection part.  Here's my
> desired parameters:

> Single point of observation, 100' from the tracks.  (No trespassing!)

Good plan!

> Power and net (or download) access available. Hobby budget. 
> All-season operation with three-year reliability, maintenance allowed.

Hmm, ok.

> I thought this might be fun for some Collexionaires.  I've got
> several ideas bumping around in my head, but wanted to use some
> of your cycles before prejudicing you with my thoughts. 
> So...think about how you might do this tonight, and what you would
> do with the data, and how you would share it.

Are there any "train watchers" who would like to contribute, or whom
you could plunder ideas from?

Time, direction, speed, number of cars, types of cars, cargo capacity
of cars, serial number of cars, etc.

> Don't reply yet...discussion starts Tuesday!

This is a virtual Tuesday for me. :-)

> -dave

Dave

[Dave]

Wonder if a seismometer would give you a clue as to the cargo content
of the train? Might have to factor in speed, too. Does more ground
vibration correspond to more cargo? How would one determine
gross versus net weights?

Dave

[Chris Harn]

They make scales to read the weights.
The scales spit the data out RS232.

In my data the cars are approximately 25-30 tons empty and 110-130 tons
heavy. Rail cars full of rocks in my case.

http://en.wikipedia.org/wiki/Automatic_equipment_identification
Automatic equipment identification (AEI) is an electronic recognition system
in use with the North American railroad industry. Consisting of passive tags
mounted on each side of rolling stock and active trackside readers, AEI uses
RF technology to identify railroad equipment while en route.

-----Original Message-----
From: Dave
Sent: Monday, October 15, 2012 2:41 PM
To: coll...@googlegroups.com
Subject: Re: Project: Train detector

[Chris Harn]

lol maybe I should read the part about not commenting until Tuesday... doh

I just get so excited!

[Mike Andrews]

On 10/15/12 2:37 PM, Dave wrote:
> However, note that railroad companies are VERY territorial, and won't
> permit you to mount anything on their right of way. They may not even
> be happy with you doing the scanning remotely, but I doubt that
> there's much that they can do about that. The other concern would be
> whether any such information could be accessed and used by an terrorists.

Well, that rules out an inductive proximity sensor (like on roller
coasters)... so you're probably looking at ultrasound, or optical, or
vibration...

[Paul Eberhart]

Since I haven't seen it mentioned in the discussion, it might be
simplest to just point a webcam at the problem. I've read about much
deeper automated train inspection with CV systems, so it might* be
easy enough to count cars/duration/etc. You may have to work in the
IR range and/or provide your own light to mitigate light-level issues,
but most digital cameras are more sensitive there anyway.

Searching to see if anyone has done something similar with OpenCV is
rather difficult because "train" has a very different meaning in that
context.

*Based on my limited experience doing automated image processing, it
is _never_ as easy as it should be. The large grain of salt has been
supplied, you know what to do.

> --
> You received this message because you are subscribed to the Google Groups
> "Collexion" group.
> To post to this group, send email to coll...@googlegroups.com.
> To unsubscribe from this group, send email to
> collexion+...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/collexion?hl=en.
>



--

Paul Selegue Eberhart (PAPPP) - http://www.pappp.net/
Messages from this account may be in regard to PAPP...@gmail.com,
pse...@uky.edu or pse...@engr.uky.edu.

[timk]

I really like the sound level idea for its simplicity. A unidirectional shotgun microphone (even homemade) pointed at the tracks ought to show you big, long waveform spikes when a train goes trundling by. Two microphones aimed a fair bit apart would give you a good idea of the train's speed, which might even let you calculate its length and the number of cars.

My second thought was rangefinding with optics or ultrasound.

[Dustin Mays]

> My second thought was rangefinding with optics or ultrasound.

I first read 'rangefinding' as 'ragefinding', which I feel is appropriate as I HATE living near trains (lived on Waller Ave a couple years ago, and ways maybe 100ft or less from the tracks. It was awful).

Carry on with your discussion, gentlemen.

--
You received this message because you are subscribed to the Google Groups "Collexion" group.

To view this discussion on the web visit https://groups.google.com/d/msg/collexion/-/-x6fezPTcboJ.

[Dave]

--- On Tue, 10/16/12, Dustin Mays <dork.f...@gmail.com> wrote:

> From: Dustin Mays <dork.f...@gmail.com>
> Subject: Re: Project: Train detector

> To: coll...@googlegroups.com
> Date: Tuesday, October 16, 2012, 8:47 PM
>
>> My second thought was rangefinding with optics or ultrasound.
> I first read 'rangefinding' as 'ragefinding', which I feel is
> appropriate as I HATE living near trains (lived on Waller Ave a
> couple years ago, and ways maybe 100ft or less from the tracks.
> It was awful).

Could have been worse. At least the diesel-electric engines are fairly
clean. If you'd lived there in the age of coal-fired steam engines,
you'd get to watch a cloud of black soot falling over you every time
one passed.

Hmm, if coal-fired steam engines were to be used today (e.g., when the
world finally runs out of oil), will electrostatic precipitators be
required on the exhaust stacks? Or, will there be any engineers left
who know how to design coal-fired steam engines?

Or, will we finally have nuclear powered trains?

> Carry on with your discussion, gentlemen.

Well, it's sometimes more fun to make the discussion go off-track.
Err, sorry for that pun. :*)

Dave