Label Application Sensing

[mark.duyvesteyn]

At my company we have a large automated packing system > putting product in bags > coding > robot stacking on pallets > pallets automatically stretch wrapped > pallet labeled with a barcode, automatic forklifts takes the pallets to the daystore. Its quite impressive to watch. The whole operation should involve no human intervention apart from replacing consumables, which is usually does. One issue we have some times is the Labeler. It applies the labels to the stretch wrap just prior to being wrapped around the pallet. There is no feed back to say if the label has been successfully applied. This can cause problems as sometimes 30 pallets can slip through with out labels. Ive been getting a little bored at work so have been trying to think of a way of sensing if there is a label on the stretch wrap.

Ultra sound.. would the sensor distinguish between a thin film of stretch wrap and a thickish paper/plastic label?

Light.. I was thinking about using the label as a mirror (its very white), shining bright LED's and looking for the reflection using a LDR or a phototransitor.

Barcode scanner.. Have a barcode reader set up that all ways looks for a barcode. The picaxe could be hardwired to listen to the "beep" The barcode move quite fast so might be hard to read it.

Its not possible to set up a light source on the other side of the stretch wrap and look for an interrupt caused by the label coming past.

Any ideas any one? from experience what method do you think would be best to pursue?

Cheers

 

[Andrew Cowan]

Reflected light sounds like a good idea.

Andrew

 

[BCJKiwi]

No feed back loop?
Presumably there is a computerised system determining the barcode on the label.
Shouldn't the label be read as the pallet leaves the labeling station to confirm the label is readable and the correct one? Standard barcode reader integrated into the work flow.
Also shouldn't the label be read again as it enters the day store (and leaves etc etc).

 

[mark.duyvesteyn]

there is a feed back for when the labler is out of labels

ribbon e.t.c., but nothing to say that the label is actually attached to the stretch-wrap correctly. The day store holds over a 100 pallets before are hand scanned in and manually taken away for long term storage.


Yeah its all computerized, the problems isn't the incorrect barcode on the label (when it puts the barcode on the label, then its going to be the right one) its the attaching of the label to the stretch wrap. There should be a integrated system for use with the printer we have.

do photo transistors care where they get their light source from? would bright white LEDs work ok?

 

[hippy]

If you can barcode scan that should be the best option, not just checking there's a label but that it's scannable.

Failing that, reflected light is a good idea. You might be able to find a colour / wavelength which is reflected well by the label but poorly by the stretch wrap.

 

[Jeremy Leach]

Maybe have two reflected light sensors spaced apart, one where the label never is and one where it should be - and compare the difference between the sensors, rather than just one. This might be more reliable because it would take account of variations in absolute light levels.

 

[mark.duyvesteyn]

thats clever

How would i go about making a cuircult that can be adjusted manually onsite?

A LDR changes resistance, could i put a variable resistor in series or parallel with it? Or is this best done Picaxe side?

 

[Jeremy Leach]

Actually having two sensors might not be necessary. But to use reflected light, it would be best to have a constant, ideally small, gap between the light source/light sensor and the label.

Are you saying that the labels fall off the shrink wrap stuff after they have been printed? I'm not too clear yet what's going on.

LDRs don't react quickly enough to light changes, so I think you'd need to use a photodiode. Probably go for Infrared too which would be less sensitive to visible light fluctuations. Might be able to find a complete source/sensor package ....I know they do them for slotted disc applications because I've used one before, but I'm sure other people will know more...

 

[papaof2]

Depending on how quickly the pallet needs to be processed, a photodiode or phototransistor will respond to changes (presence/absence of pallet or label) much faster than an LDR.

Some phototransistors bring out the base lead, so you can can adjust the bias to work with local conditions.

Jeremy's suggestion of two sensors is good, as it can adaprt to changes in lighting (presence/absence of daylight through a door/window, bulb failure in overhead lighting, etc).

You would probably be best served by doing some on-site testing of the light levels sensed by the detector (whichever type you choose) and seeing whether the differences in light levels can be handled in software. Measure the reflected light level seen by your sensor when "seeing" a label on a pallet at expected working distance, then when only seeing the plastic wrap. You may need to use lenses to limit the area of the pallet being observed.

A "higher tech" version would be to use your own light source with filters on both the source and detector. This may allow you to work with a light wavelength that is not reflected (much) by the plastic wrapping but does provide an obvious response from the label. Example - many types of paper "glow" under UV light. Use a wavelength of UV that the chemicals in the paper respond to but the plastic does not. Perhaps even pulse/modulate the light source to ensure that other light sources (flourescent tubes with UV in ther spectrum) don't affect the detector (the response speed of a photodiode/phototransistor is required in this instance).

John

 

[westaust55]

LDRs don't react quickly enough to light changes, so I think you'd need to use a photodiode. Probably go for Infrared too which would be less sensitive to visible light fluctuations. Might be able to find a complete source/sensor package ....I know they do them for slotted disc applications because I've used one before, but I'm sure other people will know more...

A LDR changes resistance, could i put a variable resistor in series or parallel with it? Or is this best done Picaxe side?[/QUOTE]

Funnily I quoted some data on LDR response time recently
http://www.picaxeforum.co.uk/showthread.php?t=10417&highlight=Decay+Time&page=15

But, here it is again as typical data:
Rise Time: (0~63%) 40mS- Decay Time: (100~37%) 10mS

 

[ciseco]

Hi,

The TCRT5000 by vishay isn't too fussy of acurate distances (at least an inch, maybe two absolute max with a white card), you can read the reflection easily with ADC 8 or 10 bit, less than a ?1 of bits and your sorted.

At 3.3v you need upwards of 82ohms for the IR LED, something like 150R at 5v, best to stick to 330R and treat it like a bog standard LED, it'll last longer. And 10K for the tranny part.

Got a circuit somewhere, but you'll find loads of examples for phototransistors on google.

Jeremy has a good point, add another to reference to, well worth the 60p extra

Miles
________
Ford Shelby GR-1 specifications

 

[rolsen]

Could the label detection be adapted to sensing wandering nursing home residents, say have a subtle reflective label on clothing. Most common "wandering Client Alert" systems at the moment is a resonator which activates a wrist transmitter the client is wearing, this in turn sends an alert to the facilities nurse call system.

These "wandering Client" systems can be quite expensive, over $A1,800 per doorway. Watch transmitters approx $A125 and need to be replaced approx each 3 months.

The system needs to sense specific "wandering" clients only at the same time letting visitors and staff pass. Say Picaxe processing interface.

 

[papaof2]

RFID chips might be better suited for the "wandering patient". Detectors are cheaper; chips require no power. Detection range could be a problem, depending on the hardware used. Some designs have a range of only a few inches; others can detect a pallet of goods being moved by a forklift.

A possible system design would have a unique RFID chip for each floor/ward and separate chips for staff and visitors. Detectors would be at each floor/ward entrance and exit (including emergency exits) and at the building's main and service entrances and exits.

I saw a relatively inexpensive (short distance) RFID sensor that could be interfaced to a PICAXE - sorry, don't remember where (narcotic pain killers do wonders for the pain of back surgery, but not always for clear thinking or short term memory)

There are some long range (450 feet in open space) readers but they start around $500US. They do have RS232 serial interface and can daisy chain up to 254 units (on a PC COM port, not a PICAXE). http://www.iautomate.com/r500sp.html

Smilar devices are available from other vendors.

John

 

[rolsen]

thanks John, well worth looking into.

Sadly there is an ever growing demand for such "client detection" systems.