Subj : Re: writing a software barcode reader
To   : comp.programming
From : Thad Smith
Date : Thu Oct 06 2005 12:13 pm

Eric.Medlin@gmail.com wrote:
> 
> I am writing software that will read a bar code from a 1bit tiff.  It
> will need to return the barcode numbers, magnification, and bar width
> reduction.  However I am having a hard time trying to determine the
> magnification and the bar width reduction.  The sample bar code I am
> reading starts with 30 1's then 60 0's then 29 1's.  This is the left
> hand guard bars.  I have no clue why the 1's are not the same length.
> Has anyone else noticed similar things when examing barcodes bit by
> bit?  Remeber this is being read from the actual 1 bit tiff file that
> will be printed this is not data from a hand held scanner.  I never
> expected to find any noise, but it looks like I might.  Or maybe it is
> just round off errors from the software that created the barcode.

I would assume round-off errors associated with scaling.

> Any
> way from this I thought I could compute the magnification and bar width
> reduction.  The file is 2400dpi so with no bar width reduction or
> magnification a module should be 31 dots.  So the first two modules
> should be 62.  The first two in the actually bar code are 90.  If we
> take 90/62 we get 1.4516.  I know the barcode was created with a 145%
> magnification, so this seems correct.  Now to find the bar code width
> reduction.  With 145% magnification the first module should be 45 dots.
>  So, (45-30)/2400 equals 0.00625.  I know the bar code was created with
> a bar width reduction of 0.007.  That is close, but not perfect.  Has
> anyone else done this and if so how?

Not here, but the closeness of the match suggests that you have the
correct approach.

Thad

.