I'm trying to compile zebra on Ubuntu 8.04. Looks like the configure script can't find 'QtCore' and 'QtGui'. I tried to satisfy this dependency but there are too many packages for QT. Is it possible to avoid QT and use GTK only?
Thanks.
Chet
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The program compiled fine. Just the right tool for my need.
Somehow the multiple code reading has some issue. I scanned the codes from a Seagate drive and a power supply and ran zebraimg on them. The EAN/ UPC codes were detected fine but the serial number code was not detected. Later I cropped the serial number (code 39) using gimp and re-ran it through zebraimg and it was detected okay.
-Chet
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
if the cropped image scans successfully without scaling (same pixels, different offset), it may be that a longer-narrower code did not intersect the scan grid? for Code 128 and Code 39 there must be a single horizontal or vertical scan line that traverses the entire symbol. the grid spacing is currently fixed to 8 pixels.
i'm actually having some trouble downloading these... the images i was able to grab seem like reduced size previews? 435x500 and 500x115? i don't think you would have picked up the Code 128 at that resolution...
if that is full size, it's too small: not all the bars/spaces can be distinguished... try scanning at higher resolution?
-spadix
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
the missing codes look like another case of JPEG artifacts affecting the scan: if you zoom in close you can see what look like extra bars in the spaces and extra spaces in the bars. this "ringing" of the sharp edges is introduced by the DCT used for JPEG compression.
the scanner used by the library is currently tuned a little on the sensitive side and has been known to pick up these shadows as extra "edges" that invalidate the decode.
i have a TODO to look into fixing this deficiency, but there are also a couple of workarounds you can try:
1. try saving your scans in a different format or reducing the compression level (higher quality). you may also try *reducing* the resolution - 300dpi may work better than 400 in this case(?)
2. smooth the artifacts in the JPEG before scanning. ImageMagick comes to the rescue here, eg:
so we found the missing codes... unfortunately i did not yet figure out exactly where the bonus EAN-13 decode came from. currently i'm thinking one of the other codes or even the text happens to look enough like a UPC-E that we pick it up... if i fix the missing check for trailing quiet zone it should help clean this out.
thanks again for the images, i'll add them to my test suite
-spadix
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I will try different resolutions with xsane. So far the program works with a few workarounds (with xsane). The images that I sent you were test case. They were scanned by Toshiba copier @ 400DPI and then cropped using gimp.
Let me know if I could be of any help in further testing.
Thanks.
-Chet
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
I'm trying to compile zebra on Ubuntu 8.04. Looks like the configure script can't find 'QtCore' and 'QtGui'. I tried to satisfy this dependency but there are too many packages for QT. Is it possible to avoid QT and use GTK only?
Thanks.
Chet
you should be able to build without support for the Qt4 widget using
# configure --without-qt
for reference, a list of available configure options may be obtained by
# configure --help
please let us know if the switch is not working for you,
-spadix
Thanks, let me try that.
-Chet
./zebraimg -v ~/Desktop/bar1.jpg
CODE-128:RP600PCAR0074941517
EAN-13:0082422000059
EAN-13:0763649002346
scanned 3 barcode symbols from 1 images in 0.29 seconds
./zebraimg -v ~/Desktop/bar2.jpg
CODE-39:9QM695K9
scanned 1 barcode symbols from 1 images in 0.05 seconds
where bar2.jpg is a subset of bar1.jpg
The program compiled fine. Just the right tool for my need.
Somehow the multiple code reading has some issue. I scanned the codes from a Seagate drive and a power supply and ran zebraimg on them. The EAN/ UPC codes were detected fine but the serial number code was not detected. Later I cropped the serial number (code 39) using gimp and re-ran it through zebraimg and it was detected okay.
-Chet
Chet-
if the cropped image scans successfully without scaling (same pixels, different offset), it may be that a longer-narrower code did not intersect the scan grid? for Code 128 and Code 39 there must be a single horizontal or vertical scan line that traverses the entire symbol. the grid spacing is currently fixed to 8 pixels.
the troubleshooting guide explains a few other common difficulties:
https://apps.sourceforge.net/mediawiki/zebra/index.php?title=Troubleshooting
if you attach your image to a support request we'll certainly take a look.
thanks!
-spadix
Thanks Spadix
I have been playing with the software and it is an excellent piece of work. :-)
I don't have any sophisticated code scanners other than a cue-cat. The cue cat is not astute enough to read some codes, zebra is.
Here are the two images that I scanned at 400 dpi
http://www.flickr.com/photos/barracuda-x/3324617780/
http://www.flickr.com/photos/barracuda-x/3324617774/
Thanks for a wonderful software.
-Chet
thanks for uploading your images!
i'm actually having some trouble downloading these... the images i was able to grab seem like reduced size previews? 435x500 and 500x115? i don't think you would have picked up the Code 128 at that resolution...
if that is full size, it's too small: not all the bars/spaces can be distinguished... try scanning at higher resolution?
-spadix
Hi Spadix,
bar1.jpg
Image Type: jpeg (The JPEG image format)
Width: 1344 pixels
Height: 1544 pixels
bar2.jpg
Image Type: jpeg (The JPEG image format)
Width: 1125 pixels
Height: 258 pixels
The second image bar2.jpg runs through zebraimg fine even it is smaller.
You are right, flickr doesn't allow the full size image if one is not signed in. Let me see if I can post them somewhere else.
Thanks.
Hi Spadix,
Try this link http://picasaweb.google.com/bsalinux/Zebra?feat=directlink
Download images using the download menu item right above the image.
Regards,
Chet
thanks! i got them
the missing codes look like another case of JPEG artifacts affecting the scan: if you zoom in close you can see what look like extra bars in the spaces and extra spaces in the bars. this "ringing" of the sharp edges is introduced by the DCT used for JPEG compression.
the scanner used by the library is currently tuned a little on the sensitive side and has been known to pick up these shadows as extra "edges" that invalidate the decode.
for reference, there is a related discussion here:
https://sourceforge.net/forum/forum.php?thread_id=2145706&forum_id=664595
i have a TODO to look into fixing this deficiency, but there are also a couple of workarounds you can try:
1. try saving your scans in a different format or reducing the compression level (higher quality). you may also try *reducing* the resolution - 300dpi may work better than 400 in this case(?)
2. smooth the artifacts in the JPEG before scanning. ImageMagick comes to the rescue here, eg:
# convert -blur 0 bar1.jpg MIFF:- | zebraimg MIFF:-
CODE-128:RP600PCAR0074941517
EAN-13:0763649002346
EAN-13:0076000005574
CODE-39:9BD648-550
CODE-39:9QM695K9
scanned 5 barcode symbols from 1 images in 0.05 seconds
so we found the missing codes... unfortunately i did not yet figure out exactly where the bonus EAN-13 decode came from. currently i'm thinking one of the other codes or even the text happens to look enough like a UPC-E that we pick it up... if i fix the missing check for trailing quiet zone it should help clean this out.
thanks again for the images, i'll add them to my test suite
-spadix
Thanks Spadix,
I will try different resolutions with xsane. So far the program works with a few workarounds (with xsane). The images that I sent you were test case. They were scanned by Toshiba copier @ 400DPI and then cropped using gimp.
Let me know if I could be of any help in further testing.
Thanks.
-Chet