 |
Claims  |
|
|
What is claimed is:
1. Apparatus for serially selectively collating each one of a plurality of
individual prime documents each having imprinted recipient address
information with respective individual non-addressed secondary documents
selected from a preset plurality of different individual said secondary
documents in which copies of individual said secondary documents are each
organized into respective sets, comprising in combination:
(a) means for serially advancing in a stop-and-go manner each one of said
plurality of prime documents along a processing route with each successive
stop position corresponding to a different work station on said route;
(b) means for executing a different work functions at each successive said
stop position, said work function executing means comprising:
(1) serial feeding means at a first said work station on said route for
serially advancing into said route individual ones of said prime documents
from said plurality of prime documents,
(2) intelligent optical character recognition means at a second successive
said work station on said route for reading said address information on
each successive said prime document and for converting the so read
information into representative output signals, said intelligent optical
character recognition means including
(a) first video camera means for imaging said address information in a
raster field and for converting said so imaged information into black and
white video data comprised of pixels, and
(b) an N-tuple classifier means that
(i) incorporates a plurality of discriminators each adapted to recognize a
respective class of a predetermined group of classes of said video data,
each said class being represented by a group of said pixels, and
(ii) presents each said pixel group to said discriminators in a
predetermined sequence, and
(3) a plurality of individually actuatable collating means, each one being
associated with a different respective successive said work station on
said route after said second work station, each one holding a different
said set, and each one being operable to combine one copy of its so held
said individual secondary document with each successive said prime
document when said prime document is in said associated work station; and
(c) computer means for regulating apparatus operation and for selectively
operating individual ones of said collating means responsive to said
output signals; thereby to produce from said address information on each
said prime document a document bundle comprised of that prime document and
selected individual said secondary documents of said secondary document
plurality.
2. The apparatus of claim 1 wherein said serially advancing means is
adapted to advance each said document bundle which results from passage
through said collating work stations to a subsequent envelope inserting
work station and wherein envelope inserting means is provided to insert
each succeeding said document bundle into a different respective envelope.
3. The apparatus of claim 2 wherein:
(a) said intelligent optical character recognition means further includes
second video scanning camera means associated with said N-tuple classifier
for reading said recipient address information on each prime document and
said N-tuple classifier means converts said so read information into
second output signals representative thereof; and
(b) an envelope imprinting work station is included at one stop position
and printer means is associated therewith, said printer means including
controller means, and said controller means is responsive to said second
output signals;
so that each said envelopes is imprinted by said printer means with said
recipient address information that occurs on individual respective prime
documents of each said document bundle.
4. The apparatus of claim 3 wherein said intelligent optical character
recognition means further includes third video scanning camera means for
reading said envelope imprinted recipient address information for
validating that each respective so imprinted envelope and prime document
of each said document bundle which is being inserted into such imprinted
envelope both have the same recipient address information.
5. An apparatus for inserting selected documents into an in situ addressed
envelope comprising in combination:
(a) means for serially advancing each one of a plurality of primary
documents along a processing route in a stop-and-go manner, each primary
document bearing a recipient name mailing address information and also a
secondary document selecting code;
(b) intelligent optical character recognition means including
(1) first video camera means at a first stop location along said route for
reading said selecting code on each primary document and for converting
said selecting code into video data comprised of pixels, and
(2) an N-tuple classifier that is in functional association with said first
video camera means and that
(i) incorporates a plurality of discriminators each adapted to recognize a
respective class of a predetermined group of classes of said video data,
each said class being represented by a group of said pixels, and
(ii) converts for each primary document said secondary document selecting
code into electrical output signals that are representative thereof;
(c) means at each of a series of subsequent second locations along said
route for holding and supplying selectively at each respective said
location one of a predetermined plurality of secondary documents with an
individual predetermined prime document when said predetermined prime
document is at such respective location, each said second location having
a different plurality of said secondary documents, and for combining
secondary documents so selected at individual ones of said respective
locations with said predetermined prime, document, each successive one of
said predetermined prime documents having a said selecting code which
corresponds to those said so selected secondary documents that are so
combined with each said predetermined prime document, thereby to form
after said predetermined prime document has progressed on said processing
route past said series of second locations a document set with each
individual said predetermined prime document;
(d) means at a subsequent stop location along said route for folding each
said document set into a bundle;
(e) said intelligent optical character recognition means further including
second video camera means at a subsequent stop location along said route
for reading said recipient name and mailing address information on each
primary document and for converting said recipient name and mailing
address information into respective electrical signal outputs
representative thereof;
(f) means at a subsequent stop along said route for advancing and
imprinting a mailing envelope with the so-read recipient mailing address
information responsive to each said signal output representative thereof;
(g) means at a subsequent stop location along said route for inserting each
said so folded document set into that respective so addressed envelope
whose imprinted address corresponds to the address appearing on the
primary document that is included in said bundle that is being inserted
therein; and
(h) computer control means for regulating and synchronizing the operations
of each of said serial advancing means, said intelligent optical character
recognition means, said holding and supplying and combining means, said
folding means, said envelope advancing and imprinting means, and said
inserting means, said computer control means being regulated by said
respective representative electrical signal outputs.
6. The apparatus of claim 5 wherein said optical character recognition
means further includes third video camera means for reading said recipient
name and address information on each said so imprinted envelope to
validate that each bundle is matched to each said so imprinted envelope.
7. A method for preparing a mass mailing comprising the steps of:
(a) serially advancing each one of a plurality of primary documents along a
processing route in a stop-and-go manner, each individual primary document
bearing recipient address information that includes a secondary document
selecting code that specifies which members of a predetermined secondary
documents plurality are to be collated with each said individual primary
documents, and carrying out the following successive steps:
(b) reading by intelligent optical character recognition each said
individual primary document, said reading being carried out by the steps
of:
(1) imaging said recipient address information with a video camera means in
a raster field and converting said so imaged information into black and
white video data comprised of pixels,
(2) organizing said pixels into groups,
(3) presenting said pixel groups to a plurality of discriminators in an
N-tuple classifier in a predetermined sequence thereby to recognize said
secondary document selecting code, and
(4) converting said so recognized selective code into representative
control electrical signals;
(c) selectively collating individual secondary documents with each said
individual primary document in accord with said control signals from among
said predetermined plurality of secondary documents thereby to produce
with each said individual primary document a document bundle comprised of
one said individual primary document and said individual selectively
collated secondary documents; and
(d) inserting each of said resulting document bundles into a different
individual respective envelope.
8. The method of claim 7 wherein each said primary document is folded after
being so read and before being selectively collated with said secondary
documents.
9. The method of claim 7 wherein in addition:
(a) each individual said envelope is imprinted with an address which
corresponds to the address shown on one said primary document; and
(b) each respective collated document bundle is inserted into the
so-imprinted envelope which bears the same imprinted address that is
associated with the primary document in said document bundle.
10. The process of claim 9 wherein said imprinting is carried out after
said inserting.
11. The process of claim 9 wherein said imprinting is carried out before
said inserting. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
FIELD OF THE INVENTION
This invention relates to improved automatic document selective collation
and envelope inserting apparatus utilizing intelligent optical character
recognition means.
BACKGROUND OF THE INVENTION
Various automatic document selective collation and envelope inserting
apparatus are known to the prior art. Such an apparatus includes an
optical sensor subassembly that detects a collation code which is
imprinted in a name and address field appearing on each one of a plurality
of serially advanced prime documents, such as, for example, a form letter
which has been previously imprinted with an individual name and address of
each one of a class of intended recipients.
The collation code, which is typically in the form of a mark sense code or
a bar code, incorporates machine readable instructions which, after sensor
detection, and conversion into electric signals, direct the selective
collation subassembly to associate secondary documents of a preset
plurality stored in the selective collator subassembly with each
individual prime document. Each resulting document bundle is inserted into
a mailing envelope that has a window through which the name and address of
the recipient as shown in the field of the prime document is readable.
Usually such apparatus is also equipped with a cooperating document
folding subassembly.
Recipients of mail which has been so processed have objected to the
presence of a collation code in association with their name and address.
In fact, recipients of such mail are believed to often regard the presence
of a collation code in association with their name and address when seen
through an envelope window as evidence that the contents of the envelope
constitute mass mailing advertising material which the recipients
sometimes discard without review. From the standpoint of, for example,
financial institutions which make periodic reports to customers,
stockholders and employees, this is an undesirable result.
In addition, the use of windowed envelopes is presently commonplace for
mailing of financial statements and the like. Confidential information,
such as the account number, balance, etc., may be unwittingly revealed by
minor displacement of the document set within the envelope. Such
information would be more secure if mailed in a non-windowed envelope.
There is a need for an automatic document selective collation and envelope
inserting apparatus which does not require the use of a (non-human)
machine readable collation code in association with the name and address
of the recipient. The present invention is directed to this need.
BRIEF SUMMARY OF THE INVENTION
This invention provides improved automatic document selective collation and
envelope inserting apparatus wherein intelligent optical character
recognition means is employed to read all information pre-imprinted in the
name and address field appearing on each one of the plurality of prime
documents being processed by the apparatus.
The intelligent optical character recognition (IOCR) means includes
associated interconnecting means and peripheral means which enable the
IOCR means to be functionally associated with existing selective collation
and envelope inserting apparatus without elaborate apparatus restructuring
or modifying. Indeed, if desired, the selective collation subassembly can
be operated with the same computer controller that was formerly employed
in using the prior art collation code sensor.
In a preferred embodiment, the IOCR means (when in such functional
combination with the other subassemblies of the selective collating and
inserting apparatus) make possible the processing of prime documents using
only the name and address of the recipient imprinted in a reading field
with no associated collation code.
In a presently preferred form, the apparatus of this invention is
additionally adapted to insert a duly collated document bundle comprised
of a prime document and selected collated secondary documents into a
windowless envelope which is imprinted by the apparatus with the name and
address of a recipient as shown on the prime document without any
associated imprinted collation code.
Although a plurality of IOCR systems are known to the prior art, these
systems generally are not adaptable for use with automatic document
selective collation and envelope inserting apparatus because of the
problems involved. These problems include reliability of recognition, the
time required for image processing and generation of output signals from
sensed images, compatibility with existing systems, and costs. The IOCR
means utilized in the present invention avoids and overcomes these
problems.
The IOCR means employed in this invention incorporates both a first
synchronous state machine for segmenting a number of images defined in a
bit map form into separate pixel groups and a second synchronous state
machine to which each pixel group is applied for classification. The first
and second synchronous state machines together comprise a recognition
engine. The IOCR means also includes new software for operating the IOCR
means in functional combination with other apparatus subassemblies.
Other and further objects, aims, features, advantages, purposes,
arrangements, embodiments, and the like will be apparent to those skilled
in the art from the following description together with the accompanying
drawings and appended claims.
BRIEF DESCRIPTION OF THE DRAWINGS
In the drawings, which comprise a portion of this disclosure:
FIG. 1 is a block diagrammatic view of one embodiment of an apparatus of
this invention wherein each one of a plurality of prime documents is
serially fed, address-field read by a camera equipped IOCR means, folded,
selectively collated with secondary documents, and envelope inserted;
FIG. 2 is a schematic diagram illustrating the intelligent optical
recognition system employed in the apparatus embodiment of FIG. 1;
FIG. 2A is a fragmentary schematic diagram illustrating the association of
more than one camera with the optical character recognition system of FIG.
2;
FIG. 3 is a diagram showing the interconnected interrelationship between
the IOCR system, the control computer and the peripheral components;
FIG. 4 is a flow diagram illustrating one embodiment of software for
operation of the IOCR system shown in FIGS. 1-4;
FIGS. 5A and 5B are a flow diagram illustrating one embodiment of software
for controlling operation of the apparatus shown in FIGS. 1-4;
FIG. 6 is a schematic diagram of another embodiment of IOCR-equipped
automatic document selective collation and envelope inserting apparatus of
FIGS. 1-5 wherein the imprinted recipient name and associated address of
each respective prime document is imprinted upon the sealed envelope
containing the respective inserted bundle comprised of that prime document
so imprinted and associated selectively collated secondary documents;
FIG. 7 is a simplified operational diagram of a high speed ink-jet printer
suitable for use in the practice of this invention;
FIG. 8 is a simplified vertical longitudinal sectional view through one
embodiment of the viewing station for image capture using camera 91 of
FIG. 6;
FIG. 9 is simplified diagrammatic view of one embodiment of the integrated
complex of processors and processing employed when collating, envelope
inserting and envelope imprinting with verification;
FIG. 10 is a flow diagram of an embodiment which is similar to the flow
diagram of FIG. 4, but which has been adjusted for "front end" handling
control of image capture and processing of recognized data from camera 43
of FIG. 6; and
FIG. 11 is a flow diagram of an embodiment which is similar to the flow
diagram of FIG. 4, but which has been adjusted for "back end" handling
control of image capture and processing of the recognized data from camera
91 of FIG. 6.
DETAILED DESCRIPTION
Referring to the drawings, FIG. 1 relates to an embodiment 21 of an
IOCR-equipped automatic document selective collation and envelope
inserting apparatus of this invention shown in block diagrammatic form.
The apparatus functions with a prime document and one or more secondary
documents. Each individual prime document, such as a one-page form letter,
is individually imprinted within a predetermined field that is located at
a predetermined position on the prime document with either (a) the name
and address of an individual recipient, or (b) both the name and address
of the recipient and an alphanumeric code. Each individual prime document
(not shown) is serially advanced as a workpiece by conveyor means through
apparatus embodiment 21 in a sequential stop-and-go manner through a
series of station stops with each station stop constituting a separate
station location at which a work function of apparatus 21 is executed by
one or more apparatus 21 subassemblies upon each prime document. The
residence time of each prime document at each work station is
substantially identical.
At prime document feeding station 22, a prime document at one end of a
prime document stack (not shown) is separated and advanced by a sheet
feeder or the like (not shown) to a reading station 23. In reading station
23, the name and address field, appropriately illuminated, is read by a
scanning video camera 43 in either one of two ways.
In one way, the address field includes, in addition to recipient name and
address information, an alphanumeric code that is conveniently printed in
adjacent relationship to the name and address information in the name and
address field (typically above or below). This alphanumeric code
represents specific data regarding particular secondary documents that are
to be combined with the prime document in the selective collator 28. This
alphanumeric code is read in reading station 23.
In the second way, the name and address information alone is used to
represent specific data regarding particular secondary documents that are
to be combined with a prime document in the selective collator. Here, the
name and address information in its entirety is read in reading station
23.
The so-read image is conveyed in the form of electrical input signals into
intelligent optical character recognizer system 42 which converts the
input electrical signals into output electrical signals that are
representative of selective collating information contained in the so-read
image.
These output signals can be used in various ways to control the operation
of the apparatus 21 and of its selective collator subassembly 28, as those
skilled in the art will appreciate. For one preferred example, these
signals are conveniently and preferably fed to a control computer or
collator/inserter controller (or control computer) 27 which produces
control signals commanding which individual ones of the predetermined
group of secondary documents that have been preliminarily loaded into the
selective collator 28 (as further described below) ar to be combined with
the individual prime document as it passes through selective collator 28
to produce a document bundle. These command electrical signals comprise a
serial signal set for each prime document. This set is conveniently
temporarily held by the control computer 27 until the prime document whose
so read information generated the particular signal set has advanced to
the selective collator 28.
After being so read, the prime document in reading station 23 is then
advanced to the next following station which in apparatus 21 is a folding
station 29. Upon arrival in folding station 29, the prime document is
folded in a predetermined manner. For example, a letter sheet can be
conventionally folded into three portions with two fold lines such that
the opposite end portions of the thus folded letter sheet overlap.
As those skilled in the art will appreciate, the folding station 29 can be
eliminated or can be located beyond the selective collator so that all
documents of a bundle are conveniently folded together.
From folding station 29, a prime document is advanced into the first
collating station 32 of the selective collator 28 which collator is
provided with a sequential series of collating stations that in this
illustrative embodiment consist of six in all. As a prime document is so
advanced, the control computer 27 can be programmed so that a first
collating command signal of the signal set for that prime document is
forwarded either directly or indirectly (as further described below) to
the selective collator 28 and to the first collating station 32 therein.
Each command signal of such a set is either a "go" (i.e., collate) or "no
go" (i.e., no collate) signal.
A "no go" signal received at first collating station 32 means that no copy
of a secondary document plurality that is stored for collation in the
first station 32 is selected for placement with the prime document
residing in the first station 32, while a "go" signal received at first
collating station 32 means that one copy of the secondary document
plurality that is stored for collation in the first station 32 is placed
with the prime document residing in the first station 32.
The prime document and, if commanded, a thus associated secondary document
are advanced from the station 32 to the second station 33 of the selective
collator 28, and another signal of the signal set for that prime document
(which signal is either "go" or "no go") is received at second collating
station 33. Depending upon the nature of the second signal of the signal
set, the second station 33 either does or does not combine with the
primary document in the second station 32 a copy of the secondary document
plurality that is stored in second collating station 33.
Next, the prime document in second station 33 together with any secondary
documents that have been thus combined therewith are then advanced to
third collating station 34 in the selective computer controller 28.
Another (third) control signal of the signal set for that prime document
(either a "go" or "no go" signal) is used to control whether or not a copy
of a third secondary document plurality that is stored in third collating
station 34 is combined with the prime document.
This process is successively repeated at each of the fourth collating
station 35, the fifth collating station 36 and the sixth collating station
37. Of course, a selective collator could have more or less than six
stations, if desired. An unused station in any actual collating operation
(that is, a collating station charged with no secondary document
plurality) can merely be given a "no go" signal for each document.
Conveniently and preferably, if the prime document is folded before
entering the selective collator 28, then each of the secondary documents
is conveniently and preferably pre-sized to form a document bundle or set
that will fit subsequently into the desired mailing envelope. Thus, the
secondary documents of each secondary document plurality can each be
preliminarily folded before being loaded into selective collator 28.
From the sixth collating station 37, the final resulting document packet or
bundle (not shown) is assembled by selective collation in collator 28. The
bundle which consists of primary and selected secondary documents passes
into envelope inserting station 39. As each document bundle is charged
into station 39, an envelope (not shown) is also charged thereinto from an
envelope feeder 40. In station 39, the document packet is inserted by
inserter fingers or the like (not shown) into the envelope. The so filled
envelope is preferably sealed by the inserter subassembly and is
discharged from the inserting station 39, thereby completing one complete
operational cycle of the apparatus 21.
The subassembly apparatus employed at each of the foregoing operating
stations of apparatus 21, except at reading station 23, is conventional
and known to the prior art. For example, suitable such apparatus is
utilized in commercially available automatic selective collation and
envelope inserting apparatus which is made and sold by the Mailcrafters
division of Inscerco Mfg. Inc. of Crestwood, Ill., the assignee of the
present patent application.
Preferably, the intelligent optical character recognition (IOCR) means that
is utilized in the practice of this invention and which is employed in the
reading at reading station 23 is adapted from that described in the
Etherington, Joslin and Newman patent application identified as
EP89910158-8, PCT GB 8901043 and WO90/03012 and also as U.S. Ser. No.
659,385 now abandoned. The teachings of those applications are
incorporated herein by reference. This technology and its application to
the present invention is now briefly reviewed and described:
A schematic diagram of a representative IOCR system 42 embodying this
intelligent optical character recognition technology is shown in FIG. 2.
IOCR system 42 incorporates a scanning or video camera 43 which in this
embodiment is preferably a charge coupled frame device that provides an
image picture that is preferably and exemplarily about 570 by about 450
pixels. Camera 43 is functionally associated through a video interface 46
with a so-called optical character recognition engine 44 which
incorporates a high speed trainable logic network that is controlled by a
32 bit microprocessor 56 or the like. In place of camera 43, other
embodiments could use a page scanner, a hand held scanner, or a line scan
camera.
The recognition engine 44 recognizes images represented by separate digital
pixel groups. It utilizes an N-tuple classifier to which each pixel group
is presented. This classifier includes a plurality of discriminators each
one of which is adapted to recognize a respective class of a predetermined
group of classes. The apparatus arrangement is such that each pixel group
is presented to the discriminators in a predetermined sequence.
Recognition means is provided by monitoring the output of the
discriminators. The presentation of each pixel group to the classifier may
be terminated as soon as the output from a discriminator satisfies a
recognition condition. If none of the discriminator outputs satisfies a
recognition condition, the presentation of a given pixel group to the
classifier may be terminated after a predetermined time interval.
The apparatus thus incorporates a first synchronous state machine for
segmenting a number of images defined in a bit map form into separate
pixel groups and a second synchronous state machine to which each pixel
group is applied for classification. Engine 44 is thus a synchronous state
machine. In a synchronous state machine, the stages of the operating
processes are stepped automatically under control of a system clock. A
synchronous state machine in applying the N-tuple method of pattern
recognition allows the use of a hardware implementation which achieves a
much higher speed of image recognition than can be achieved by a
predominantly software approach.
The use of the N-tuple method of pattern recognition allows for the
discriminators to be trained with different forms of the same shape thus
allowing multi-font recognition of the name and address fields.
Camera 43 captures for the name and address field and reference field (if
used) of each prime document indicia as images and converts same to
digitized video data representing black or white pixel image data captured
from the defined field area as a line by line sequence. A video interface
46 digitizes this data and orders such into a form suitable for subsequent
data processing. This data is then output from interface 46 to the logic
element 47-51 of recognition engine 44 which produces after data
processing an electrical output which represents character data that is
encoded to a suitable industry standard, such as ASCII (American Standard
Code for Information Interchange). The recognized character data is then
subjected to pre-determined decision criteria, utilizing the recognition
engine control microprocessor. The resultant decisions are then output to
a controller, such as control computer 27, through a host system interface
57. Control computer 27 uses this output to control operation of the
selective collator 28.
The camera 43 can be any convenient commercially available optical frame
camera. Preferably, the camera has an image resolution of at least about
300 dots per inch and a frame scan time of less than about 25 milliseconds
to efficiently capture as image so that a high speed of character
recognition of at least about 1,000 characters per second is achieved.
In actual fact, the character recognition speed is independent of camera
capture time. However, in terms of overall system throughput, it is
important that the camera frame capture time is short, as throughput is a
function of paper positioning and stop time + camera frame time +
recognition time (of the name/address and reference fields) + decision
process time. At a document input (feeder) speed of about 12000 sheets per
hour, this equates to approximately 300 milliseconds per sheet. Equally,
recognition speed has to be fast.
The recognition engine 44 provides for the segmentation and classificat | | |
