[thelist] High-Speed ADF / Duplex Scanner

Jeff Wilhelm thelist at summit7solutions.com
Mon Feb 28 09:23:45 CST 2005


Good morning everyone! I know we've discussed this in the past, but I was
wondering if anyone had any feedback...

I was given a mission today by the president of the company. We need to be
able to scan in two-sided product registration cards at a high speed. We are
talking like 500,000+ a month. Assuming the scanner has a TWAIN interface I
will write software to automate that scanning, and will save the files with
a unique ID (datetime(14)_int(5)) as well as populate a database for
tracking. I need a recommendation on a highly-reliable, high-speed scanner
(SCSI or USB2 I suppose) that I can use for this.

Basically, once the cards have been scanned into the system we will hire
people to do data-entry on them; the table that stores the card information
(for processing) would look something like this:

UID
CARD_ID
FILENAME
SCANNED_TIMEDATE
PROCESSED_STATUS
PROCESSED_BY
PROCESSED_TIMEDATE

UID would be a UniqueID for EVERY CARD scanned into the system (Primary Key
basically) and could be a 14-char timestamp (YYYYMMDDHHMMSS) as well as a
random integer added to the end, in the event we scan more than one card per
second), as I noted above.

CARD_ID would be our Foreign Key relationship to the card information
metadata (for example, if we are scanning in CompanyX cards they may be ID
1, which tells the system what questions are included, and CompanyY may be
ID 2, etc... this way we could easily load up just the subset of questions.
So we will store MetaData on which CARD_ID has which questions. Also, before
the data entry begins, we will have the data entry tech enter both the
CARD_ID and COMPANY (store in a separate table) as a quick QA check to make
sure they match up.

FILENAME would be the name of the image (something like
\\STOR1\\20050227\%%UID%%.tif) so the system knows where to load the image
from, and we can also have a nice hierarchical file system on the server,
for easily backing up certain dates, doing stats on number of cards scanned,
etc...

PROCESSED_STATUS would be 0 by default, meaning "Scanned", 1 meaning "In
Processing", and 2 meaning "Complete". Or something.

PROCESSED_BY and PROCESSED_TIMEDATE would be the ID of whoever processed the
card and the time and date they processed it (good for metrics and
accountability). Plus, paired with the first 14 characters of the UID (the
date and time) we could give metrics on turnaround (how long between a card
is scanned until it is processed, etc...

Eventually maybe we'll want to barcode these cards so the scanner can
automatically enter the CARD_ID, but for now it's OK. So, if anyone has any
feedback on scanners, please send them to me!

Thanks,
Jeff



More information about the thelist mailing list