| Author |
Message |
|
|
 Posted: Feb 21, 2012 - 01:30 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
I have a system similar to below.
my task is develop a system that measure mattress width and length. there is no need to take measurements while conveyor is running. it is ok that stop the conveyor and take measurements.
please suggest few methods for this task. |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 02:28 PM |
|
Joined: Feb 19, 2010
Posts: 507
Location: Montreal, QC, CA
|
|
|
|
|
|
|
|
Posted: Feb 21, 2012 - 02:33 PM |
|
Joined: Mar 27, 2002
Posts: 18520
Location: Lund, Sweden
|
|
What resolution is required? What accuracy is required?
I.e. a light above the conveyor, and photodiodes in between rollers will give you the length with perhaps 2 " resolution. Likewise, a row of photodiodes in one gap between two rollers could give you the width.
Accuracy of length and width would rely on many factors. One of them is how straight along the conveyor the matress is plaed. If it is placed diagonally then measurements will report longer langth and width. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 02:37 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
| 0.5mm at least |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 02:41 PM |
|
Joined: Mar 27, 2002
Posts: 18520
Location: Lund, Sweden
|
|
|
Quote:
0.5mm at least
Resolution or accuracy? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 02:50 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
there are other few things.
1.mattress is not parallel to conveyer
2.final output is calculate the density of the mattress. therefore currently conveyor is on 4 load cells and act as weighing scale.
a worker takes the W,H & L manually using tape and calculate the density.
calculated out put must 0.1 accuracy |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 03:10 PM |
|
Joined: Apr 16, 2001
Posts: 3522
Location: Phoenix, Arizona
|
|
| You're not going to get an accuracy of .5mm with a tape measure, let alone any kind of optical system. And if you measure the length on one edge and then measure the length on the opposite edge, that won't be within .5mm either. Why do you need such a tight tolerance? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 03:14 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
| project vendors ask the possibility of 0.5mm accuracy. but 1mm accuracy will ok. |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 03:22 PM |
|
Joined: Nov 11, 2003
Posts: 3849
Location: Chicago Illinois USA
|
|
| Is a mattress itself accurate to 1/2 mm? My initial thought it "no way!" It's going to vary more than that depending where you measure it. |
_________________
Discursive design,
Torby
Some days, it's just not worth chewing through the restraints.
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 03:33 PM |
|
Joined: Mar 27, 2002
Posts: 18520
Location: Lund, Sweden
|
|
|
Quote:
calculated out put must 0.1 accuracy
I still think you are having problems with confusing accuracy and resolution.
Apart from that 0.1 makes no sense at all unless you give a unit. (Is it a resolution of 0.1 kilometers? If so then just report 0 always - unless of-course you make huge mattresses  )
What is the resolution and accuracy today, when the measurements are taken manually? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 03:41 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
@ JohanEkdahl,
takes least 1mm with tape |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 04:24 PM |
|
Joined: Mar 27, 2002
Posts: 18520
Location: Lund, Sweden
|
|
| Last time: Accuracy or resolution? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 08:30 PM |
|
Joined: Aug 04, 2002
Posts: 1704
Location: Brisbane, Australia
|
|
Some basic arithmetic?
If output is to be calculated to an accuracy of 0.1 percent to what accuracy do You need to calculate volume?
Similarly to what accuracy do You need to measure the dimensions in order to achieve that calculation of volume?
And lastly to what accuracy does the load cell need to weigh in order to achieve final accuracy? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 11:44 PM |
|
Joined: Dec 11, 2007
Posts: 6841
Location: Cleveland, OH
|
|
For length and width one might use a camera munted on the ceiling. Then calculate the area by simply counting pixels that are white, against a dark background. In this case it doesn't matter if the mattress is tilted at an angle or not. Since you are after volume you don't really need the length and width, the area will suffice.
I'd probably consider using a CCD sensor such as in flat bed scanners, mounted vertically adjacient to the conveyor belt/rollers, to measure the heigth of the mattress.
With the center bludging up in some, and the ruffled, sewn edge sticking up on others, I'm not sure what the heigth really is.
You could define your own standard and have a robotic arm lower a 1 meter x 1 meter, X kg metal plat on the mattress, and then measure its heigth, (wire length, laser dots separation, ultrasound, etc.).
JC |
|
|
| |
|
|
|
|
|
|
Posted: Feb 21, 2012 - 11:50 PM |
|
Joined: Feb 19, 2010
Posts: 507
Location: Montreal, QC, CA
|
|
What kind of mattress is sized to 1mm accuracy???
Surely if the fabric fits the frame with no slack everything is good? |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 12:02 AM |
|
Joined: Dec 11, 2007
Posts: 6841
Location: Cleveland, OH
|
|
If it was a water mattress one wouldn't have to measure its desity, you could look it up in a book!
JC |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 12:17 AM |
|
Joined: May 26, 2004
Posts: 2538
Location: Las Vegas, Nevada
|
|
|
hugoboss wrote:
What kind of mattress is sized to 1mm accuracy???
Perhaps one ordered by my old drill sergeant. He knew exactly how a bed should be made. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 12:34 AM |
|
Joined: Feb 19, 2010
Posts: 507
Location: Montreal, QC, CA
|
|
| I might add that you will get much better accuracy measuring the hard edge of the bare frame than the soft and deformable edge of the finished mattress. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 01:21 AM |
|
Joined: Aug 04, 2002
Posts: 1704
Location: Brisbane, Australia
|
|
If the matras is on the conveyor use photo cell to sense begining and end of bed.
Use a rotary encoder on the conveyor drive and count the encoder pulses .
But that comes after You get real with Your specifications.
So GET REAL |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 02:25 AM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
@ JohanEkdahl,
it is 1mm resolution  |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 02:28 AM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
| I'm not familiar with CCD and image processing, i take image processing as my last option. |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 03:53 AM |
|
Joined: Feb 20, 2012
Posts: 3
Location: SLC, UT, USA
|
|
First off... I'm curious what the rest of the industry does to measure their products and what kind of tolerances they work to. I don't know the mattress industry, but along with others here I have a hard time believing that any of the major mattress manufacturers try to build to within 1mm of the final dimensions.
That said... what about using rails or rollers to square the mattress up on the conveyor before measuring it? That would be simple to rig up. In fact, if they were spring loaded or pneumatic so that they snugged up to the mattress they could even have provisions for taking the measurements.
Being the obsessive compulsive type, I was curious what the industry does about tolerances and measuring. I found the following quote on several sites relating to Simmons mattresses:
Quote:
Like all other bedding manufacturers, Simmons subscribes to the International Sleep Products Association standard for bedding products. Dimension sizes are allowed ± ½" tolerance. If a "bunkie" foundation is used, it shall have the same width, length and tolerances as the finished twin size box spring. Finished sizes are measured from seam to seam on the binding tape at the centerlines of the width and length.
I get the impression this only refers to the length and width and is for the purposes of fitting sheets and frames rather than calculating volume. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 08:21 AM |
|
Joined: Dec 01, 2003
Posts: 2500
|
|
A single camera mounted above might not have enough pixels for the desired resolution, but a pair of cameras mounted above and a known distance apart could sure do it.
Also, one could consider laser scanners such as made by Keyence Corp. |
_________________
Tom Pappano
Tulsa, Oklahoma
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 01:10 PM |
|
Joined: Dec 30, 2004
Posts: 8723
Location: Melbourne,Australia
|
|
| Bosch have some economical laser distance measuring devices with serial output. These have mm resolution. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 01:18 PM |
|
Joined: Mar 27, 2002
Posts: 18520
Location: Lund, Sweden
|
|
Let the matress come to stop at the end of the conveyor.
Use a moving photo detector running along the mattress, driven by a stepper motor and either a wire-wheel or a threaded-rod arrangement, to determine the length. A similar arrangement but running across determines the width.
Start outside the mattress. Run until light intensity falls. You are now at one endpoint. Keep running, ounting the number of steps the motor takes, until you see the light again. The measurement is the number of steps times the length of one step. |
|
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 02:33 PM |
|
Joined: May 03, 2011
Posts: 169
Location: Sri Lanka
|
|
|
Quote:
Use a moving photo detector running along the mattress
mattress it not parallel to bed. |
_________________
~ Chandana ~
a self learner
|
| |
|
|
|
|
|
|
Posted: Feb 22, 2012 - 02:50 PM |
|
Joined: Dec 11, 2007
Posts: 6841
Location: Cleveland, OH
|
|
You were given several options, some which require the mattress to be squared up, (parallel to the wall), for measurement, and at least one method which doesn't care what the orientation of the mattress is.
Both have trade offs, your choice.
If you don't like either of those options, then you have another one to consider:
Put the mattress in a large plastic bag, (which you probably do for storage and shipping anyway), and submerge it in a tank of water.
Measure how high the water mark moves up the side of the tank, which tells you how much water was displaced, which tells you the volume of the mattress.
As you have already weighed the mattress, you can now calculate its density.
This method, by the way, totally eliminates all of the "where, exactly, does the edge of a mattress begin?" issues.
JC |
|
|
| |
|
|
|
|
|
|
Posted: Feb 24, 2012 - 01:50 AM |
|
Joined: May 28, 2006
Posts: 716
Location: Toronto, Canada
|
|
I would install a row of light sensors each millimeter right under the path the mattress passes perpendicular to the direction. Put a laser (or two) on the ceiling that output a line (not a point) and which points to the sensors. Watch out when the mattress start to shadow the sensors and monitor the movement with a rotary encoder. Take many readings and do the math. Same you can do for the height - put the sensors on the wall.
No camera, and it is doable with AVR. You even can push the mattress by hand, no need for motor.
Personally I would use a small notebook to process the data or a tablet. All what the avr has to do is to scan many digital inputs (lets say 1000 times per second),read the encoder and to send serially to PC. One sensor could be one bit in a byte. If you can afford to read through an ADC, you can increase the resolution through interpolation.
If the mattress has to be fixed, then install the laser and sensors in a mobile device as in your scanner.
George. |
_________________
www.sofgel.ro bootloader for mega and xmega
www.elsofgel.com XME development board
|
| |
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
