DK175729B1 - Hexagonal information encoding article, process - assigns optical properties to individual hexagons by ordering, hexagons in predetermined sequence - Google Patents

Abstract

The optically readable label (30) stores encoded information in the form of a data array of a multiplicity of information encoded hexagons (20) contiguously arranged in a honeycomb pattern, and having two different optical properties (e.g. black and white). The process assigns optical properties to individual hexagons in a predetermined pattern consists of ordering the hexagons (20) in a predetermined sequence and prints the hexagons (20) with two optical properties (e.g. black and white). A process for retrieving information creates an optical replica of the digital bit stream representative of the optical properties of the information encoded polygons, decoding that optical replica and retrieving the decoded bit stream.

Description

in DK 175729 B1

Merchandise, various components, letters, parcels, containers and an entire register of corresponding parts that are shipped or transported must often be identifiable with information on origin, means of transport, destination, name, price, part number and numerous others. information. In other applications, reading encoded information printed on labels affixed to such items may allow automatic registration of sales figures and inventory statements or the operation of electronic cash registers.

Other uses for such encoded labels include automatic routing and sorting of mail, packages, luggage and similar items, and placement of labels with production instructions on raw materials and components in a manufacturing process. Labels for these species of objects are usually marked with barcodes, one of which is the "Universal Product Code". Numerous other barcode systems are known in the art.

Commercially available barcoding systems typically lack sufficient data density to meet the 20 immediate and increasing need for encoding more and more information on the labels, which at the same time is getting smaller. Attempts to reduce the overall size and spacing between bars in different barcode systems to increase data density have not solved the problem. Optical 25 scanners of sufficient resolution to detect barcodes with contrast bars spaced at a distance of five mils (about 0.13 mm) or less are generally not economically suitable for manufacture due to the accompanying narrow tolerances at the label printing process 30 and the sophisticated optical apparatus required for reading bit-encoded lines with these dimensions. Instead, very large barcode labels must be produced to record the increased amount of data, with the result that such labels are not sufficiently compressed to be placed on a small object. Another significant factor is the cost of the medium from which the label Η I DK 175729 B1 I 2 I is manufactured, such as paper. A small label causes less

In paper costs, than a major label. This cost is, you

In a significant factor when operating with large quantities of objects. IN

I Instead of barcodes can be used: circular shapes

5 feeder using radially spaced wedge-shaped code parts

In terms, as disclosed in U.S. Patent No. 3,553,438, I

I or concentric black and white bit-encoded rings, such as I

As disclosed in U.S. Patent Nos. 3,971,917 and 3,916,160. IN

H I

In addition, networks with rows and columns of data can be used

In 10 encoded squares or rectangles, as disclosed in U.S. Pat

In document No. 4,286,146, or microscopic points I

i.cells, which form a network of uniform distances, as I

As disclosed in U.S. Patent No. 4,634,850, and tightly packed multi-I

In colored data fields or points or elements as mentioned in

In U.S. Patent No. 4,488,679. Certain of the above mentioned

In coding systems, and other known coding systems in the art

You are marked by errors in the data density, as is the case

I I

! that with the encoded circular patterns and the rectangular nets

In yellow or square fields. As is the case with I

20 the networks which include microscopic points or multicolor

For specific elements, a special orientation and special I are required

transport means, limiting their applicability to I

powerfully controlled reading environment. IN

Due to the magnitude of speed in modern over- I

H 25 guidance systems, where e.g. used conveyor belts with I

a width of 3 to 4 feet (about 90 to 120 cm) and where trans- I

the tire velocity approaches 100 inches (about 250 cm) per IN

H second or more where packages of different heights are kept, I

on which information-encoded labels are affixed and I

H 30 due to the need for the use of a small, inexpensive, I

compact label of approx. 1 square inch (approx. 6 cm2), I

there are great demands on the optical and decoding facilities which I

H is required to locate and read the data encoded I

H labels on such fast-moving packages and the like. IN

H 35 There are difficulties in simply collecting the label car

H led in the optical scanner. Further, when the label image I

Once recorded or identified, this must be accurately decoded before the next operation with the package in the transport system takes place, often within a fraction of a second.

These problems have led to a need to provide a simple, fast and inexpensive way to signal the presence of a data-encoded label within the field of view of an optical scanner, which scanner is mounted so that scanning of the entire conveyor belt can be performed. It is desirable that this property be associated with a data set of 10 high densities, as explained later in more detail.

Datasets with registration markings are known in the art, e.g. may be designed as concentric geometric shapes, with rings, squares, triangles, hexagons and numerous variants thereof, as explained in U.S. Patent No. 3,513,320 and No. 3,603,728. US Patent Nos. 3,693,154 and 3,801,775 also explain the use of symbols, including concentric circles, for identification and position, which symbols are attached to objects to be optically scanned. . However, in these systems two separate symbols are used to determine the identification of the data area and its location, which increases the complexity of the logic circuits required for the detection of the symbols, while also reducing the capacity of the associated data field for data content. . When two symbols are used, damage to one of these will also cause problems in locating the data field location and problems with the associated ability to provide information from the data field. In the last-mentioned system, separate 30 position and orientation markings are used at opposite ends of the data tracks with data-encoded linear markings with a limited data content capability.

The aforementioned systems are generally scanned with an optical sensor designed to produce a video output signal corresponding to the change in the intensity of the light reflected from the data cell as well as from the location.

I DK 175729 B1 I

I I

In the ring and orientation symbols. The video output signal from I

In such a facility, after being digitized to I

provided with a special bit pattern which can be assembled j I

In with a predetermined bit order. These plants, however, suffer

5 due to the disadvantage that two separate symbols are required

You first register the presence of the image and

Then, to determine its orientation. It is also pro- I

I ces, where the digitized output of the optical i

I 'sensor is juxtaposed with a predetermined bit order, I

10 representing both location and orientation symbols, I

You are of such a nature that it is more likely that you-

In -accurate readings are provided, as is the case

with the method and systems provided by I

In the present invention, due to the fact that prior art I

In 15 label registration systems, an inflexible I

sible characterization of the level of registration mark I

up signal. IN

U.S. Patent No. 3,553,438 discloses a circular I

data set with a centrally located registration mark, I

20 which includes a number of concentric circles. Registration I

the marking provides a means for recording I

the circular label at the optical sensor and for determination

H view of the label's geometric center, and thus the geometry I

be the center of the circular dataset. This is done by I

H 25 logic circuit operations to record the im- I

H pulse pattern representative of the center I design

of the registration mark. However, the data set, I

H as was also the case with barcodes, only a limited I

data capacity, and the system requires another circular scan

H 30 curing process. Use of both a linear and a circular I

H scanning in a system with such limited data capacity to I

in turn, creates an unwanted complexity in the plant

H for a slight increase in data capacity relative to I

H known barcodes. IN

H 35 To increase the data capacity of data sets there are I

H developed codes with a multiplied high density of colored I

In * DK 175729 B1 dots, as disclosed in U.S. Patent No. 4,488,679.

Systems of this kind disclosed in U.S. Pat.

4,488,679, however, requires the use of manual optical scanners, which scanners cannot record and decode 5 fast-moving data sets on packets carried on a fast-moving conveyor belt. Similarly, high density coding systems employing microscopic data-encoded dots, as disclosed in U.S. Pat.

4,634,850, special transport agencies to ensure that

10, the data set is moved in a designated direction and not with a random orientation, as would be the case with a packet carried on a conveyor belt or similar.

body. Thus, the encoded label must be read track I

for tracks using a linear scanner which is pre- I

15 bonded with the label conveyor to provide an I

suitable decoding of the information encoded on eti- I

decal. In the mentioned patent specification, the location of the card I

relative to the sensor is very carefully controlled to be I

readable. IN

20 Various colors have also been used within I

the area of manufacture of barcode systems, so that I

to overcome the optical problems of scanning a lot

small strokes. A barcode to which more than two I is used

optical properties for encoding data in a data set, by I

25 e.g. use of alternating black, gray and white lines

is disclosed in U.S. Patent No. 4,443,694. IN

GB-A-1 216 539 discloses objects with associated I

information, and also methods of preparing I

such articles, and methods of storage and / or I

30 capturing data using such objects, I

the items being especially marketing labels. On the label- I

the optical readings are printed at least two I

rows of black and white areas. A row is a time series of I

alternative black and white areas indicating successive I

35 information positions. The one or the other row is an I

information series and has a sequence of blacks and / or I

In DK 175729 B1 white areas located in line with the positions I of the time series and encoding information. Each black or white area in the I row indicates a bit in each row of information. A square set of information traces may be provided together with a time series along an upper edge, and a series of indicator columns with alternating black and white squares along the one side edge. GB-A-1 216 539 also suggests instead of I black squares, parallelograms, ovals, squares with I concave sides and X shapes. Reading is done using I10 transmitted or reflected incidence of light on any kind of photocell or photosensitive device, and H may include the use of buffer storage for retaining data from the label for subsequent recording, validation, or for loading in a computer. Reference is also made to EP-A-0 336 769 filed simultaneously by the same applicant.

According to one aspect of the present invention, there is provided a method for decoding a stream of digital signals representing an electro-optically scanned image corresponding to a plurality of polygons of the same shape, Η in which polygons are located with the geometric centers of adjacent polygons lying in the corners of a predetermined two-dimensional set, and each of these polygons has, in accordance with the encoding process, one of at least two mutually different optical properties, including the following steps: (a) performing two-dimensional time restoration of the scanned image by entering two-dimensional blocks from the scanned image at corresponding locations in a transformed image in accordance with a non-linear two-dimensional edge enhancement transformation and processing of the transformed image to obtain a re-established time signal, (b) using the re-established time signal from steps (a) to localising seizing the geometric centers of the polygons to identify the optical properties of the polygons, I (c) decoding the polygons by performing the inverse of the encoding method.

According to another aspect of the invention there is provided a device for decoding a stream of digital signals representing an electro-optically scanned image of a plurality of polygons of the same shape, the polygons being located with the geometric centers of adjacent polygons lying in the corners of the a predetermined two-dimensional set, and the polygons have, in accordance with an encoding method, one of at least two mutually different optical properties, including: scanned image in corresponding 15 locations in a transformed image according to a non-linear two-dimensional edge enhancement transformation and processing of the transformed image to obtain a re-established time signal, (b) means for using the re-established time signal to locate the geometric centers of polygons, (c) means for decoding the polygons by performing the inverse of the encoding process.

In view of the foregoing known drawbacks of optical coding systems, preferred embodiments of the present invention can be used for reading labels with compact, high density information readable which is optically readable, such as optically readable labels which may be encoded with approx. 100 highly error-protected alphanumeric characters 30 per square empty label area. Labels with compact high information density information that are optically readable can be read by an optical sensor when the label is attached to a package or to other similar items conveyed on a high-speed conveyor system, regardless of the orientation of the package thereon, or the varying heights for such packages on which the optically readable label is affixed. In a preferred decoding system combination for optically readable labels, the label can be reliably decoded even when it is tilted, curled, skewed or partially annihilated or partially torn.

IS It is also preferred that the optically readable labels with compact, high density information I be encoded by dividing information to be encoded in messages of high and low priority so as to provide a hierarchy of messages that are particularly error prone - 10 to protect the integrity of the encoded information. Optically readable labels with compact, high-density information that are enhanced can be encoded and decoded by methods that include error correction options,

so that erroneous or missing information is restored, I

15 and so that this is exercised with preference for the encoded messages of high priority.

Cheap optical readable labels can be produced by known printing devices and decoding them can be done with relatively inexpensive logic circuits.

Optically readable labels may include predetermined two-dimensional geometric polygon sets, the geometric center of such polygons being at the corners of the intersecting axes, as explained in more detail H later, in a predetermined two-dimensional set, with polygons 25 are provided with at least one of the at least two different optical properties. The polygons on such optically readable labels may be regular or irregular polygons, and the two-dimensional set of polygons on the optically readable H labels may be provided with two or more same or spaced apart axes in the plane of the label.

Optically readable labels may be printed on polygons that are completely opposite to each other, which are partially collisional or non-contacting. Thus, the last two structures will be provided with a number of intermediate regions on the optically readable label between adjacent polygons. Such intermediate regions may have the same or different optical properties relative to the two or more optical properties of the polygons. Two-dimensional sets of contiguous polygons having four or more sides are useful in conjunction with the construction of the optically readable labels provided by the invention.

Likewise, two-dimensional sets of either regular or irregular three or more sides polygons, which are either partially abutting or not touching each other when placed on predetermined axes in such sets, can be encoded and decoded according to the provided by the present invention.

In addition to the aforementioned various designs of geometric polygonal cells, the location of such polygonal cells, and the geometric design of the optically readable labels provided by such arrays of polygonal cells, the optically readable labels of the present invention can be provided by 20 designation includes a capture tag which consists of a series of concentric rings to aid in locating the optically readable labels on the objects to which they are attached, especially in dynamic label reading systems.

In a preferred embodiment of the invention, the data set includes a substantially square set of about 25 mm. 1 square inch (approx. 6 cm2) with adjacent hexagons forming rows and columns, and a centrally located capture mark with a geometric center indicating the geometric center of the data set. j The intercept mark may be provided with a design which is designated among a number of geometric shapes with in optical properties capable of producing a lightweight! recognizable video signal when scanning with an optical sensor along a linear scan line through the geometric center of the capture mark. In a preferred embodiment, the capture tag is designed with

I DK 175729 B1 I

I I

In a number of concentric rings with contrasting reflection I

In features which provide a periodic video signal, I

when scanning is done linearly. By using analog I

In filter means as part of the method of locating I

5 and decode the data set, the signal generated by I is compared

You are given directly by the optical sensor with a predetermined I

In frequency, thereby providing the opportunity for a fast and I

accurate comparison of frequencies with subsequent I

In determining the location of the data set attached I

10 for a substrate. The analog electrical output from I

the optical sensor representing the information input I

This is then digitized and decoded. When used I

the last of a step with an analog bandpass filter is created I

H allows the capture of the label to be exercised without I

In need of decoding the information-encoded label. IN

By locating the capture marking center, you can

a reference point is defined on the data set. If the center I

of the capture marking is located in the center of the label, I

a simultaneous determination of the center of op I can be exercised

In the 20 capture tag and for the data set. It must be preferred, I

In that the capture mark is located in the center of the label,

In practicing the present invention, it is not

In a claim. IN

The optically readable data set is designed so that I

25 can be encoded 100 or up to hundreds or even more

several fail-protected alphanumeric characters in an area I

of approx. 6.45 cm2 (1 square inch) when encoding the hexagons I

using three reflection properties such as the colors I

black, white and gray. For a sensor with a given optical resolution I

In accordance with the invention, the system provided by the invention enables an I

much closer packing of the information than is possible

by barcode cysts. For example, an optical sensor with I

H large solution is used in conjunction with that of the prior art

According to the present invention, several hundred I

35 alphanumeric characters are encoded on a 6.45 cm2 (1 l

square inch). Alternatively, one hundred characters per 6.45 I

In DK 175729 B1 cm2 (square inch) is easily detected with a sensor with a relatively low resolution.

Optically readable labels can be produced with varying data densities using as few as two or more 5 contrasting optical properties. Larger data densities and the inclusion of a capture tag in the system provided by the invention require a scanning apparatus of increasing complexity as well as the addition of more elaborate decoding algorithms for reading the encoded message, 10 compared to a barcode reading system.

Data encoding can be provided by encoding a plurality of bits from a binary bitstream in a set of opposite hexagons, each hexadec being provided with one of at least two different optical properties, although the encoding could alternatively be performed on a hexadec. by-hexagon basis. The digital bit stream may be generated by a computer based on manually entered data or data which is otherwise converted to a binary bit stream, or provision may have been made as a pre-recorded digital bit stream. Data to be encoded is arranged in bits in a predetermined order and in predetermined geographical areas of the data set to increase the number of transitions between hexagons with different optical properties.

In a preferred embodiment of the invention, the messages to be encoded are divided into high and low priority messages which are arranged separately in different geographical areas of the data set. High priority messages can be re-selected in low priority message areas to reduce the possibility of a high priority message being lost due to scanning errors that may be caused by fouling tears, creases and other types of damage of the data set. High priority messages are encoded in a centered area of the data set, near the intercept mark in the preferred embodiment to protect the message from damage, which damage is more likely SDK 175729 B1! ment. It must be "made possible by the cmation cap at the decode pixel network of: edges with jerking processes when the label axis edge of the re-encoder is well known: jerking machines with the present invention idled in bit hexagons, providing a prior aid of an electrically scannable sensor electrically is indexed at 5 by an analytical sensor ice size, cleaned from a le in the data label, and then inserted. At the same time, the analog signal is continuously digitized using an analog-to-digital converter and stored in an image buffer store. The stored digitized data representing the total label is available for further processing during the decoding process.

For stored logic program circuits, the digital data is transformed into an ordered series of transitions between hexagons with different optical properties. In a preferred embodiment of the invention, this is accomplished by calculating the standard deviation of the intensities of the reflectance properties recorded by the optical sensor in each pixel and a predetermined group of pixels surrounding the first pixel. Large standard deviations therefore correspond to transition ranges on the transitions between contrasting hexagons.

15 Additional data transformations involving filtering programs to determine orientation, direction and. distance between the hexagons is exerted on the digital data. The main steps in this process are: (1) Filtering the non-linear translated version 20 of the digitized image.

(2) Determining the orientation of the label, preferably by locating the three axes in the image (as shown in Fig. 2) and determining which axis is parallel to two sides of the label.

(3) Determining the center of each hexagon and defining the gray level in each center.

(4) Transforming the gray levels into a bit stream.

(5) After designation, implementation of error correction of the bitstream, 30 (6) After designation, change the bitstream to a predetermined character set.

It should be noted that although the method of the present invention is explained in association with hexagons having two or more optical properties, the method, and in particular the steps of adjusting the optical image for each slope of the labels, tear I DK 175729 B1 I 14 I and similar phenomena are used for other species of labels and other polygonal cells.

Other objects of the application of the present invention will be apparent from the detailed explanation of the invention.

It should be understood, however, that the detailed explanation of H's preferred embodiments of the invention is intended as an illustration only, and is not intended to constitute a limitation of the range of variations and modifications within the spirit of the invention. as will be appreciated by those skilled in the art.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described in more detail with reference to the accompanying drawings, in which: 1 is a plan view of a pickup mark formed with concentric rings; FIG. 2 shows part of an optically readable label with! hexagonally arranged for encoding data, fig. 3 is a plan view of an entire hexagonal optical readable label positioned against each other with three different optical characteristics for encoding binary data 20 and with a capture tag; FIG. Fig. 4 is a plan view of three cell-by-cell groups of hexagons positioned against each other which may act as a basic encoding unit; 5 is a group view showing a geographical H 25 representation of a data array including 33 H rows and 30 columns, thereby providing a network of 11 H rows and 10 columns of 3x3 encoding cell group units H with hexagons, H fig. 6 is a schematic view of a camera setting system in accordance with the present invention

H to adjust the position of the optical light sensor in I

H in accordance with the height of the package being recorded, I

H fig. 7 is an arrangement of the steps in it by the I

The present invention provided the decoding process, I

FIG. 8 is a flow chart illustrating progress

H the method of locating the capture mark, I

In DK 175729 B1 fig. 9 is a flowchart showing the structure of the encoding and decoding program and the data stream; FIG. 10 is a flow chart showing the sequence of auxiliary processing steps; FIG. 11 is a plan view of a group of adjacent hexagons arranged with the geometric center of adjacent hexagons located at the angular tips of a regular hexagon; FIG. 12 is a plan view of a group of up to 10 irregular hexagons placed with them

geometric centers of adjacent hexagons lying on I

the angular tips of an irregular hexagon, I

FIG. 13 is a plan view of a group of partial Is

adjacent polygons of mainly hexagonal I

15 in shape with the geometric centers of adjacent I

polygons located on the angular tips of a hexagon, I

FIG. 14 is a plan view of a group of polygo I

placed against each other and largely shaped like you

hexagons with the geometric center of adjacent poly- I

20 gones placed on the angular tips of a hexagon, I

FIG. 15 is a plan view of an optically readable I

label with opposite polygons mainly I

designed as hexagons with the geometric centers of the adjacent I

polygons located on the angular tips of a hexagon I

25 and including a capture mark, I

FIG. 16 is a plan view of a group of up to I

equilateral squares placed with the I

geometric centers in adjacent squares lying on wine- I

the tips of a hexagon, I

Fig. 17 is a plan view of a group of non-I

placed at right angles to one another at intervals between I

the rectangles, and with the geometric centers of adjacent I

rectangles placed on the angular tips of a hexagon, I

FIG. 18 is a plan view of a group of non-I

35 are pitted against each other at intervals between I

the pentagons, and with the geometric centers of adjacent I

In DK 175729 B1 I 16 In pentagres placed on the angular tips of a hexagon, In fig. 19 is a plan view of a group of adjacent rectangles positioned in staggered rows and columns with the geometric centers of adjacent rectangles H 5 located at the angular tips of a hexagon; 20 is a plan view of a group of partially octagonally spaced apart spaces between the octagonal and with the geometric centers of adjacent I octagonal located at the angular tips of a rectangular array.

H By encoding information using contrasting colors in juxtaposed hexagons or "cells" placed in a honeycomb pattern in a predetermined order and arrangement, the information can be stored on the label so that it can be recalled by an electro-optic sensor. Similarly, non-hexagonal polygonal cells located with the geometric centers of adjacent polygons disposed at the angular tips of a hexagonal or other pre-fixed array may similarly be used to encode information on an optically readable label. Such polygonal cells, when positioned with their centers at predetermined locations in a two-dimensional geometric array, and when encoded in a predetermined order using various optical properties H 25 in a number of such polygonal cells, can be "read" by an electro-optical sensor in H and then decoded in accordance with H with the method of the invention, which H is subsequently explained.

The polygonal cells are information encoded units 30 formed by a closed broken line, which cells are H placed in a predetermined two-dimensional pattern on an optically readable label. The label designs include a wide variety of different polygonal shapes and arrangements of various geometric shapes, such as hexagonal, rectangular and square arrangements, which are useful in practicing the invention. "Adjacent" poly-! B1 goneal cells may abut completely against each other, may to some extent lie against each other or do not lie against each other on the optically readable label provided by the invention.

5 "Completely Opposite Polygons" are polygons located with the geometric centers of adjacent polygons placed on the angular tips of a predetermined two-dimensional array, and with the edges of the polygons in contact with the edges of immediately adjacent 10 polygons, so that no are spaces. "Partially juxtaposed polygons" are polygons located with the geometric centers of adjacent polygons lying on the angular tip in a predetermined two-dimensional arrangement, the polygons being spaced apart along the respective edges from the surrounding polygons, thereby providing a number of spaces between the polygons on the optically readable label. "Non-opposing polygons" are polygons located singly with the geometric centers of adjacent polygons located at the angular points in a predetermined two-dimensional arrangement and without any contact between the edges of a single polygon and the polygons which surrounds this polygon. In addition, the polygonal cells and the predetermined two-dimensional grids or arrays on which the centers of adjacent polygons 25 are located may be irregular, be spaced apart between the axes, or be regular with equally spaced axes. Such two-dimensional alignment axes may be independent of the axes of symmetry, if any, in the polygonal cells.

30 Hexagons show significant advantages in encoding information on a label. These advantages are: (1) For a given optical resolution, the hexagons can be packed closer than other polygons. Eg. are, with a given resolution, the vertices of squares more difficult to solve, so that an otherwise unnecessarily large optical solution is required to "read" squares. Circles should be optimal in

I DK 175729 B1 I

I 18 I

Optical resolution but the space between adjacent

In circles are wasted and will complicate treatment and I

In printing the label image since there is a need I

I to give the spaces an optical property. Hexagonals I

5 enables maximum information packing compared to I

In the circles or with other polygons, including eight-

edges, squares, triangles, etc. Squares and triangles I

You cause problems because of the sharp corners. Circles and I

On octagonals create problems due to the unused flour

H 10 gap between adjacent circles or octagon. IN

(2) A mesh of hexagons located opposite each other

has three axes. Using a label of square I

In or in rectangular shape, the main axis of the hexagon can be localized

You look from its predetermined relationship to a page of I

In the label. This location of the main axis in a network of I

Hexagonals do the reading of the data encoded in I

hexagon easier, based on their relationship to the main axis. IN

H As used herein, "label" means a single I

unit, with a suitable adhesive backing to be attached to I

20 a package or product, the outer surface of a container I

or another object, upon which the optically readable information

mation is printed. IN

"Optically readable data set" or "data set" means you

in the present publication a pattern of juxtaposed I

25 hexagons or other polygonal cells with two or more I

optical properties such that in a retractable manner I

can be encoded a collection of data from the individual optical properties

creates of, as well as the spatial relations between the hexagons I

or whatever other polygonal cell shape now used. Sex I

H 30 the edges or polygons which are printed so that they I

H contains tangible information is hereinafter referred to as I

as "information-encoded" hexagons or polygons, on I

because of the way in which the information is encoded in eti- I

decal. IN

H 35 The pattern of hexagonally aligned hexagonal I

H the maximum number of transitions between the hexagons to the optimal I

In »DK 175729 B1> reading and for maximum information storage density is called a" honeycomb pattern ".

The contrasting reflective properties used in the printing of the individual hexagons or cells in the data set can be widely varied within the range.

the spirit of invention. In the present writing, you mean

"printing" placement of material with predetermined optical I

properties of a substrate, or alteration of the optical properties

creates as it does by "thermal" printing. "Pressure I

10 "also includes failure to secure a material I

having a predetermined optical property on a portion of the substrate,

wherein the substrate itself has a designated optical property. Eg. IN

when printing hexagonal cells in black and white, I

only the black cells are actually printed if sub- I

The 15 layer is white. As used herein, the white hexagonal I

cells thus also fall within the definition of the term I

"printing" or "printing". IN

"Optical Properties" as used herein means

properties such as light absorption, reflection and / or refraction

20 of the cells are printed on different media. If the cells I

is black (high density black ink), gray (halftones I

of black) and white (no printing on a white surface), I

as is the case in a preferred embodiment I

of the invention, the material is said to be provided according to I

25 with three optical properties. IN

"A number of concentric rings" or "concentric I

rings "10 herein means and with reference to I

FIG. 1 two or more concentric rings 12 of which one I

is constituted by the inner circular area 15, which is defined I

30 at the smallest radius "r" in the rings. IN

FIG. 2 shows a portion of an electro-optically scannable I

label of the present invention. As shown in FIG. IN

2, the label includes a number of adjacent printed six

edge-shaped cells arranged in a honeycomb pattern. IN

35 Each of the hexagons is indicated by the number 20, I

and includes six equal sides 22. The inner angles I

BDK 175729 B1: the two opposite edges provide a horizontal axis dt smaller than n regular six-label 30 with (about 2.5 x 2.5 squares or, in fact, drawn output 35, facing each other / ice horizontal sr 31, and vertical lines 33. In lxl inch, "R" and 30 vertical hexagon are ten.There are several irate boundary sequences of the up Ice pack.

6th hexagons were in staggered, alternately placed rtical side 22,

The axes at hexagonal edges 32 and: all reach the "R" hexagon 20. one produces hexagonal properties.

21 DK 175729 B1 e.g. in contrasting colors. These colors may be white 25, black 26, and also, optionally, however preferred, gray 27, as shown in FIG. 3, although other contrasting colors may be used. It is possible to use only two contrasting colors, such as white 25 and black 26, as shown in FIG.

2. In the preferred embodiment of the invention, three contrasting colors, white 25, black 26 and gray 27, are used, as shown in FIG. 3. The particular shades of white, black and gray are selected to obtain maximum contrast so as to facilitate identification by an electro-optic sensor. The gray level is so designated that its properties are approximately equal to the optical properties of white and black, as used in producing the label.

The label 30 of FIG. 3 may be designed as a single label having, in a preferred embodiment, an area of 1 square inch (about 6 cm 2), or, if a background of a suitable color (preferably white) is used, the label may be printed directly on the surface on a package, without providing a separate label. Due to the importance of providing a background with a controlled optical property as one of the contrasting colors, it is preferable to use a separate label, as it is easier to control the background color of the label.

It is important that the hexagons printed on the label are aligned with the sides of the label because of the subsequent determination of the main axis of the label, j as explained later. The label is printed so that the y-y axes of the hexagons constituting the honeycomb will be parallel to the vertical sides of the label, 32 and 34, as shown in FIG. Third

By "reading" the hexagonal data set for decoding the information contained in each hexagon, it is essential that a sharp color contrast is provided between adjacent hexagons. For reasons explained later, the scanning equipment and the necessary software for decoding the hexagons will be simpler.

I DK 175729 B1 I

I 22 I

The fewer optical properties used in encoding I

In the hexagons. However, fewer optical properties will also

In reducing the data density on the label. In a compromise between I

In the amount of decoded information that can be stored on the label, I

I 5 and the cost of scanning labels with many optical I

In properties, it has proved desirable to print the encoded I

In hexagons with three different optical properties, namely

the colors black, gray and white. If the substrate or label I

You have a good white background, the white hexagons can be produced

I 10 for lack of ink, and only black or gray hexagons I

You really have to press. IN

In a preferred embodiment of the invention, I

In, the gray hexagonal cells are weighed by printing cell I

In the clay with black ink, but only every 15 pixels in the pixel grid I

Thus, in a dot matrix printer, I is printed as described herein

provided the illustrative example. This is done using I

of a halftone algorithm in a manner known in the art

discipline. This causes a printer to print one beforehand

fixed portion of pixels so as to produce a designated I

20 gray hexagon, while a black hexagon requires printing of I

In each pixel within the hexagon. The special halftone algo- I

rhythm used to print labels in the preferred I

In the embodiment is listed in the list of source codes I

H marked "LABEL" in the program supplement page 29, lines 39-48. IN

The black hexagonal cells can be formed by printing I

In with a standard black ink. As explained below, you are exercised

which in the decoding process associated with the decoding process

In light software a rough distinction between the reflections I

from black, gray and white so that the exact color defi- I

30 nition is not necessary. On the other hand, if you

H colors other than black, gray and white are used or if I

Various shades of gray are used to create j I

of four or five color data sets, a more careful control occurs

Increase the contrast between the ink shades so that I

It is ensured that measurable optical properties are provided

cabinet differences between the individual colors. It will be able to

In DK 175729 B1 it is seen that the use of black ink is the simplest and easiest way to provide a honeycomb pattern with hexagonal cells with three different optical properties.

Since the preferred embodiment of the invention 5 is to provide a square label relative to the nature of the hexagonal cells, the edges of the honeycomb pattern will include incomplete hexagons 56. As shown in FIG. 3, these incomplete hexagons are not used to transmit useful information.

The label also includes an intercept mark.

The capture marking 35, as shown in FIG. 3, includes a number of concentric rings in contrasting colors (shown as black and white). The black rings are designated respectively. 42, 46 and j

48, and the white rings are designated respectively. 44, 50 and 52. Det. must you

15 it is preferred that the capture marking is located in the geometric center of the label so that it is less or less fully or partially damaged or damaged, such as if the circumference of the label is torn, soiled or damaged. In addition, the size of the image buffer store (as explained below) to be used for storing data from the label before identifying the tag's capture tag becomes as small as possible when the capture tag is located in the center of the label.

The number of concentric rings used in the capture mark may differ, but a number of six concentric rings 42, 44, 46, 48, 50 and 52 and the interfaces provided by the rings vary from white to black to white, etc. proved to be convenient and desirable.

30 A pattern correlation technique is used when comparing a calculated pattern for what could be read from the concentric rings with what is actually read.

When the comparison fits, the capture tag has been located, as explained in more detail later. The particular filter provided and used in conjunction with the preferred embodiment of the invention is set forth in DK 175729 B1 I 24 I of the Program Supplement page 41, lines 51-52, page 42 lines 1-8 and page 40 lines 19- 41 under the file name "FIND.C.".

In the Capture mark, may assume any diameter smaller than the data set to provide an area which may be 25% and preferably about 7% I of the data set area. It is preferable that the intercept marking is designed as small as possible, since the area recorded on the label cannot carry encoded H information. In the preferred embodiment, the diameter 10 of the printed rings is designated such that the outer boundary of the outer ring 52 is approx. 7.45 mm. That is, the area of the capture marking 35 in FIG. 3 records approx. 7% of the H surface area of a label 30 on a square inch (about 6 cm 2). In this way, a suitable capture mark 35 H 15 can be printed on a label 30 having an area of 1 square inch, without unduly affecting the amount of information which can be encoded in the sixty-square area surrounding the capture mark. As is the case with the incomplete hexagons along the outer edges of the label 55, the cut hexagons which lie against the outer limitation of the capture marking 56 are not used for encoding information. It is preferred that the width of each ring j be approximately the same size as the size of the distance between two sides of the hexagon (the x-x axis of Fig. 1), thereby facilitating resolution. Six rings are convenient. Six rings are a reasonable number to place on the least possible label area with the least likelihood of erroneous readings from "foreign" labels on the label and other "foreign" labels that are not on the label, such as labels on a trans -H 30 ports.

The intercept mark may be designed differently than with concentric rings. Eg. For example, squares, spirals or hexagons can be used to provide transitions between contrasting concentric shapes, as long as a linear section through the capture mark will provide uniform, predetermined and identifiable color transitions suitable for detection by an electro-optic sensor. , and measured by a suitable filter. It should be noted that although a spiral is not a collection of concentric circles, depending on the size and radius of the spiral, a close approximation to concentric circles can be provided. A concentric ring marking is preferred since the signal provided by a scan through their center has a frequency which is the same as if cut in other directions through the center of the concentric rings. This makes identification of the center simpler, as explained in more detail below, and allows identification of the location of the capture mark by a one-dimensional search with the analog or digital output of the scanning device, although the method provided by the invention alternatively or subsequently, a two-dimensional digital search can be used to provide increased accuracy when analyzing a digital signal.

Thus, the term "concentric rings" is used herein to include complete rings, partial rings in the form of semicircles, slices of concentric rings occupying between 180 ° and 360 °, and concentric spirals approximating concentric rings.

Since each hexagon can be encoded with three different optical properties, in the preferred embodiment, 1585 "bits" can be encoded with information in each hexagon (log 23). It is obvious that if fewer or more optical properties are used than three, the number of encoded bits in each hexagon will change accordingly.

The encoding algorithm is designed to obtain a density so close to the maximum data density and to increase the number of cell-to-cell transitions with different optical properties to support the two-dimensional clock repetition process which explained in 35 res later.

In FIG. 4 shows a group of 3 x 3 cells with nine in the SDK 175729 B1 unit, the unit. g, it is capable of detailing cells 60 s group in 13 bits,: hex position, gs mark bottom at fig. 4, thus: nth six-conductor can run 62 and [formation of 7 weighing bridges. 37 = 2187 ies certain white clay white asterious edge 66 Eor clockwise, stretcher set, slings 7 usable snows are provided - bits are encoded in each pair of hexagons, and 1 bit is encoded in each remaining single hexagon, as explained later in more detail.

It can therefore be seen that the label includes a highly efficient, easily readable (by means of suitable scanning equipment and analytical programs) label for encoding information with very high density on a relatively inexpensive label where the information is easily printed. As noted, in the preferred embodiment, an array of 10 hexagons in 33 rows and 30 columns is used on a 1-square inch label, with a capture marking representing approximately 7% of the total surface area of the label. In practice, 13 bits of information are provided from a group of 9 hexagons, so that 1.44 data bits per cell.

15 This is smaller than the theoretical number of 1585 pr. hexagon, which is due to other constraints in the encoding algorithm since all 37 patterns are not used and since the optically least desirable transitions between the cells are excluded.

For later explained reasons, it is desirable in the preferred embodiment of the invention to include a certain magnitude of error protection in the encoding of the label, so that the actual readable amount of information on the label is reduced in favor of a large degree of data error freedom in the decoding process.

It will be readily appreciated by one of ordinary skill in the art that the foregoing explanation of label embodiments used for hexagonal cells can be directly transferred to optically readable labels using 30 other polygonal cells. Similarly, the explained method of "printing 11" of optical properties in hexagons involves printing the optical properties of other polygonal cells, whether in black, white, gray (through halftoning) or in other colors. Corresponding to the 35 limitations and advantages of data density association applies to labels printed with other polygonal cells,

I DK 175729 B1 I

- I

than hexagons, when the optical properties are black and white and I

Optionally, gray is used when printing the polygonal cells. IN

I As is the case with labels containing hexagonal I

In addition, labels printed with other polygonal codes can be encoded

In 5 cells, "read" with scanning equipment of smaller size

In Plexity, if only two optical properties are used for I

In encoding information in the polygonal cells, especially optical I

properties such as black and white, since maximum I is obtained

In contrast to these colors. IN

I 10 They associated with labels with hexagons I

In prepared information encoding procedures and the associated I

In algorithm, can be used directly for labels printed

In with aberrant polygonal cells. In the same way as in the association I

In hexagon labels, incomplete poly-I is used

In 15 gonial cells that can appear at the edges of the optic I

Readable label, or provided by partial cut-off

I at the intercept mark, which includes a number of con-

In centric rings, not for encoding information. IN

In a "honeycomb pattern" includes a set of up against each other

H 20 second placed hexagons 310 whose geometric centers 311 I

Similarly, on the angular tips 311A lie in a "six-I

In angular mesh "or" a hexagonal arrangement "312, as shown in FIG

In FIG. 11. Regular hexagons, ie. hexagonals with six In equal sides and six equal internal angles, constitute

In 25 hexagonal arrays, which are also regular in out

In the molding, with three equally spaced axes (Al, I

In A2 and A3), the axes being positioned at an angle to each other

At a distance of 60 °. IN

If the hexagons 320 are irregular but symmetrical, I

For example, in 30 if the hexagons are extended along two parallel I

In pages 321, 322, the geometric centers 325 of adjacent I will

Hexagonals describe an irregular hexagonal array 327, I

As shown in FIG. 12. Such an irregular hexagonal arrangement I

You will still be provided with three axes (A1, A2 and A3), I

However, in 35 but the three axes will not be located with the same I

At a distance, that the three axes will not have a · I

In DK 175729 B1 spacing of 60 °.

Although in FIG. 12 is not a regular arrangement, it is nonetheless a two-dimensional geometric grid or arrangement of axes 5 with predetermined spacing. That is, the location and distance between the geometric centers of the hexagons, which are located at the angular points of the intersecting axes of the hexagonal array, are also predetermined. The geometry of the hexagonal array is thus used in the decoding process, as explained below. In particular, the filtration step performed on the transformed digi! speech data according to the image sensed by the optical sensor set to reproduce the predetermined label geometry so that the digital representation 15 of the sensed label can be used to accurately reconstruct the original grid. The reproduction process further provides the missing points from the hexagonal grid. The missing grid points are due to the fact that the transition between the optical properties does not occur between polygons with uniform optical properties.

With irregular hexagonal grids of the one shown in FIG. 12 it will be desirable to adjust the step of determining the major axis, step (3) (e) of FIG. 7 in the decoding process performed after the Fourier transform step 25 in determining the principal axis of the optically readable label.

The main axis of the label will be positioned so that the geometric centers of the polygons lie along this axis with distances that deviate from the distances of the other two axes.

In FIG. 13 shows a label design employing polygonal cells 330 which are substantially similar to hexagonal cells.

but which are twig-edged polygons instead of hexagonal I

ter. Correspondingly constructed polygons with more or less I

pages greater than 20 can also be printed. Polygons 330 lie only in I

partly opposite to each other as opposed to the imaginary hexagon I

35 cells 331 within whose boundaries the polygons I

330 is illustrated. ! IN

B) K 175729 B1's form as but toad: r one; jiven nene, jende: clay and in be> rme-: be- »t ered in the nothing» ring teres: r by: to thin 533 i if they G) i _ i 330> m is) isse 341, f ig .

! with Ite 175729 B1 an absolute requirement in the practice of the invention. The polygons 340 are shown with their geometric centers 342 lying on the angular tips of a hexagonal array 345. As with the polygons 330 in FIG. 13, the polygons 340 are substantially hexagonal and 5 they will appear as hexagons at moderate optical resolution.

FIG. 15 is a large-scale reproduction of a label as it will appear if printed with a dot matrix printer that prints 200 pixels per second. inch (approx.

2.54 cm). The polygons 360 in FIG. 15, the shape of 10 illustrates the geometric figure which will in fact be printed instead of a hexagon with such a dot matrix printer, due to the pixel density of the printer. Printers with higher pixel density would provide a greater approximation to a hexagon than the polygon 360 of FIG. 15. The polygons 340 in FIG. 14 and 360 in FIG. 15 are thus likely to result in configurations due to the accompanying limitations of certain printers, in connection with the hexagonal cell printing process, or as a result of deliberate attempts to print such polygons in the first instance essentially 20 as hexagons. The design of such polygons mainly in hexagonal form ensures that in practice they act as equivalents to juxtaposed hexagonal encoding cells.

As was the case in FIG. 3 contains the optically readable label of FIG. 15 also includes a capture tag 370 which includes a series of concentric rings 371 to 376.

As in connection with the hexagons of the FIG. 3, they are essentially hexagonally shaped polygons 3 60 in FIG. 15 arranged in columns "C" and rows "R", 30 as defined by the dotted lines 361 and 362, respectively. 363 and 364. As was the case with the hexagons of Figs. 3, the polygons of FIG. 15 with their geometric centers placed at the angular points in a hexagonal arrangement, as determined by the axes A1, A2 and A3, which are 35 at the same spacing. Labels according to the embodiment shown in FIG. 15, is thus encoded and decoded in BDK 175729 B1 below. such as a similar infor- mation: The filter steps correspond to the sensor range. In setting distances, preference is given to the fact that the 2 distances or mutually readable g. 16 are not equal to 20 at A3. It illustrates the points 420, squares in squares, placed placed at DK 175729 B1 33 the intersecting axes A1, A2 and A3. Again, the illustration of the hexagonal arrangement is supported by drawing with dotted lines of hexagons 431 in FIG.

17, which are positioned on top of the unrelated 5 rectangles 430, thereby providing spaces 435 between the rectangles 430. FIG. 18 similarly illustrates a label with pentagons 4 which are not in contact with each other and which are located with the geometric centers of adjacent pentagons 440 along the three axes A1, A2 and A3 spaced apart by the same spaced apart. The geometry of the non-touching pentagons is more easily illustrated by the drawing of the hexagons 441 with dotted sides, with spaces 445 being provided between the pentagons 440.

Alternative hexagonal arrays may be provided where the axes of the arrangement A1, A2 and A3 are spaced at uniform spacing without corresponding to the axes of symmetry of the polygonal figures themselves. Instead, the geometric centers of adjacent polygons are located at the angular points of the intersecting axes of the array. Such an arrangement is illustrated in FIG. 19, which includes a plurality of mutually interacting rectangles 450 with the geometric centers 451 in adjacent rectangles located along axes A1, A2 and A3.

Similarly, higher order polygons can be arranged in a predetermined two-dimensional grid. In FIG. 20, a plurality of partially touching octagon 460 is shown, defining a number of spaces 461 between the octagon 460. The centers 462 of adjacent octagon 460 are located at the angular points of the mutually intersecting axes A1 and A2, thereby providing an array of octagon 460 which can be used in practice according to the invention. The gaps 461 can be printed with an optical property which differs from the properties used for the octagonal edges 460. However, this is not an absolute requirement in the practice of the invention since it is the location, orientation and intensity of the optical property in the center of the octagonal region.

Η I

I DK 175729 B1: I

I 34 I

In the terminals 460, which are in a predetermined location in the I

In hexagonal arrangement formed at axes A1 and A2, I

You are of the utmost importance in the decoding process. IN

It is possible to provide many variations of I

In the 5 label. Eg. the label does not have to be a square I

I with a side of 1 inch (about 2.54 cm). This size you were

Designated as a reasonable size for a label to be obtained

I of an acceptable data density of 100 alphanumeric informa- I

I tion grades with a high degree of failure protection, without you

10 produce very large labels. It is desirable to add

You weigh a 1 inch page to reduce I

paper and other costs associated with printing, I

In shipping and handling such labels. Known lines I

code labels of similar size will exhibit a strong I

15 reducing the data density. Using 4, 5 or I

H more optical properties or colors to indicate six-! IN

the edges will be able to pack considerably more information

into a given area of hexagons of predetermined size I

but it will require an increase in the complexity of I

20 the software and the sensitivity of the scanning system, so that you

this information can be recalled. Therefore, it is in high I

degree desirable for practical purposes in using an encoder

three optical properties, black, gray and white. IN

Furthermore, the sizes of the hexagons and the interceptor I

The H 25 marking changes within wide limits. IN

Although a "grouping" of hexagons into groups with I

H 3x3 cells have been explained, other I can be used

patterns of grouping, or the grouping can be completely avoided, I

and the encoding algorithm can be directed to an individual six-I

30 border pattern. Furthermore, the relative size of the I

encoded information in the message in relation to error correction

the deed varies within wide limits. IN

The encoding process of I is explained below

H invention as used in connection with the invention

H 35 preferred embodiment of the label. IN

H The process can be started with a predetermined row of I

In DK 175729 B1 data that is wanted to be encoded on a label. In a preferred embodiment, the label is a shipping label and data is divided into two fields, identified as "high priority" messages and "low priority messages". It will be understood that the invention is not limited to two different messages or priority levels. Many messages and priority levels can be provided within the quantitative limits of a label of a given size and number of cells. ; 10 For example, where the label should be used as one

the shipping label, the "high priority message" include nine characters representing the postal code of the recipient of the package or letter. Nine numbers are referred to because nine-digit postcodes, though many single people and four

15 maes have postcodes with five numbers, because postcodes with nine numbers I

are used more frequently and more frequently. When processing packages for i

delivery is the postal code the most important information. Hereby I

the main destination of the package is determined, allowing you to

different scanning and packing control systems I are used

20 so that the packet is routed to the correct destination on I

trucks, aircraft, transport systems, etc. I

The low priority message can e.g. include name I

and shipping address with the zip code of the recipient of the I

package, as well as payment information. IN

25 The reason for providing a high priority j I

message and a low priority message is that high priority i

the message must be protected with a special error correction,! IN

so that the high priority message is placed {encoded) in i

a more central area of the label where it is smaller

30 probable that the message is damaged or destroyed, I

while at the same time providing opportunities for high-priority I

the drawing message is repeated and distributed in the low priority message

sharing, so that even if the high priority message alone I

destroyed, there will be a great possibility that high priority I

The low-priority message can be provided again from the low priority message

delelsen. By placing the high priority message in a center I

In set area it will only be necessary to decode the high priority message for certain purposes, so that only part I of the label needs to be processed, which reduces the processing time. This will e.g. This is the case when a package is placed on a transport system and only the postal code needs to be determined to identify it out of several transport routes that the package must use during the processing process.

The low priority message does not have to be provided twice on the label. However, as explained below 10, both the high priority message and the low priority message may include various error protection codes and! correction options such that the probability that H will accurately provide both messages is maximized.

The use of error protection characters as part of the encoded information may, in the preferred embodiment of the invention, together with a suitable stored program and an associated computer, cause the system to correct H an error during the decoding process in the manner explained below. . The use of error protection codes known in the H field is within the scope of the knowledge of one of ordinary skill in the art.

In practicing the invention, an operator generating a label can manually input data into a suitable data matter terminal, which, as explained below, is designed to activate a printer for printing a label with the high priority message and the low priority message encoded on suitable display in the hexagons of the label. It is not essential to the invention that a high priority message and a low priority message be provided, but it is desirable to maximize the likelihood that the most important data to be encoded can be recalled.

Η In the preferred embodiment, the label is also printed with H a centrally located pick-up mark consisting of H 35 a number of concentric rings with two contrasting colors H placed alternately, the colors being preferably two of the colors I DK 175729 B1 used for printing. the individual hexagons, and especially should be preferred that the colors are black and white to provide maximum contrast.

Upon the operator's manually entered data, a suitable programmed computer will be caused to encode each character in the loaded message and use suitable range statements to provide in the activated computer a binary bitstream representing the characters in the message, which bitstream is thus encoded by the region. that the high priority message and the low priority message are specified and that their relative location is stated. The operation is carried out by the program "TEXTIN.C" listed in the program supplement, page 1, lines 8-54, page 2, lines 1-54 and page 3 li-I

1-36, and as indicated in 110 in FIG. 9. A computer with I

The required characteristics may be Compaq Deskpro 386 (with a

16-MHz clock and an Intel 80387 math coprocessor chip). IN

Alternatively, the process can begin by informing it

mation to be encoded is already located in a binary I

bitstream, e.g. the information is received from a data file

20 germ medium or otherwise provided. Therefore you can

the message to be encoded is provided in an I

form which is manually converted (by electronic means) to I

a binary bit stream, or which begins as a binary bit stream. IN

Once the binary bit stream has been provided, I

25 or an error protected bit stream is provided at the I

step, which will be explained in detail later, the bit stream I

is arranged in accordance with a predetermined arrangement I I

pattern for encoding the present invention

provided hexagonal honeycomb. FIG. 5 shows a group op

30 position indicating the individual groups with 3x3 hexagon I

cells in each placed in a net or honeycomb with I

33 rows and 30 columns with hexagons. Each row is numbered and each column is numbered. The row numbers go from 1 to 33 and the column numbers go from 1 to 30. It can be seen that some

35 of the hexagons along the upper boundary and the right

delineation of the arrangement, and within the geometric network I

I DK 175729 B1 I 38 In the center area is indicated by X. This indicates that these hexagons do not contain bit-arranged information, this is because the outer hexagons indicated by X represent partial hexagons along the edge of the label, which causes, that each of these rows is provided with fewer hexagons. The inner I hexagons indicated by X represent areas occupied either by the capture mark or where there are incomplete hexagons along the outer boundary of the capture mark, such that these inner hexagons indicated by X , are not bit-lined. All the 1 hexagons that are not marked with X can record information. According to the preferred embodiment of the Invented Ice, each of the regions will be occupied by a black (B), white (W), or gray (G) hexagon. As noted above, while practicing the present invention, although | different grouping and arrangement techniques can be used, groups of 9 hexagons in 3 rows with 3H hexagons in each are used to delineate designated information H bits, and it is preferable that 13 information H 20 bits be encoded in each group with 9 hexagons.

Η In a data set of 33 rows and 30 columns of hexagonally placed hexagons, a grid H of 11 rows and 10 columns of hexagonal groups is provided, with H each group including a set of 3x3 cells of 25 hexagonal columns. , which can be seen in connection with FIG. 5. However, it will be appreciated that each row of groups of 3 x 3 cells within the network of 11 x 10 groups will include a group of either 7 or 8 hexagons due to the geometric arrangement of 30 hexagons and the number will change from row to row. in.

Thus, in this arrangement, six groups of 8 hexagons and five groups of 7 hexagons are provided. Further, the incomplete groups are provided at the centrally located capture mark. FIG. 5 is a graphical representation of useful hexagonal groups available for encoding information bits in a 33-row data set and 30 columns of hexagonal juxtapositions.

Groups of nine usable hexagons, as shown in FIG. 4, is encoded using the following algorithm:

Take eleven pieces of information and place them in a set I

with 7 hexagons identified as a, b, c, d, I

e, f and h

10 The hexagons g and i are used to represent each! IN

1 bit in such a way as to ensure that each of them I

differs from the hexagon h

Thus, the thirteen bits of information are encoded in an I

15 total group of 3x3 cells ie. nine against each other

placed hexagons. I I

For subgroups of 7 or 8 hexagons applicable:

20 Take eleven pieces of information and place them in the set I

with the first seven usable hexagons. IN

The eighth hexagon is used if available

named, to represent one bit. IN

For all subcells:

Place three pieces of information in so many hex pairs,; IN

as possible. j I

30 in I

Any single hexagon remained; IN

used to represent one bit. IN

Since placement of seven hexagons provides more I

35 combinations eleven bits (i.e. 37 = 2187 vs. 211 = 2048), I

certain compositions of hexagons must be rejected. They for- I

SDK 175729 B1 seals, e. For the place-Lvejebrinline fore-mmor.

• Opportunity positioning the eagles and the supply were communicated as 2t forde-digital ies for infor- • erity-as-for-and-for-attract, including 5. For example.

92 infor ->> mme in a south and with snow area, for further cause - DK 175729 B1 • 41 see code. In this example, it would also be desirable to use 120 check bits for the high priority message. This is determined by the desired magnitude of error correction capability.

Similarly, in this embodiment 5, 560 bits of low priority message are included. This includes 40 bits of high priority message which is inserted into the low priority message. In the example, 576 check bits will be added to the low priority message to maintain security and support the retrieval of the low priority message. This example illustrates the widespread use of check bits to preserve and enable the retrieval of the high priority message over the low priority message. It is to be understood that the foregoing is to be construed only, for example, and that the high priority message 15 may be both longer and shorter, and that the low priority message may be longer or shorter and that the number of check bits may be greater or less depending on the particular application.

In a "systematic code" a special message sequence is extracted and a delimited error check frequency is added. 'In a "non-systematic" code, a special message, sequence is extracted and the error check sequence is inserted so that the message is no longer delimited but can of course be recalled. It is within the scope of the present invention to use both systematic and non-systematic coding for error protection. In the following discussion a systematic code is used.

As used herein, the step of "inserting error detection symbols" includes systematic and / or non-systematic coding systems.

Various systematic linear cyclic error protection codes are known in the art, e.g. BCH-codes,

Reed-Solomon codes and Hamming codes. In a preferred embodiment, the Reed-Solomon codes are inserted separately to protect the error-free in high-priority messages and low-priority messages. The Reed-Solomon codes are very effective and are very useful when multibit characters are checked incorrectly. The Reed-Solomon codes are well known and it is to be understood that this is simply a preferred embodiment, although many other error correction codes could be used in conjunction with the present invention.

Reed-Solomon and other coding systems are discussed in e.g. Richard I E. Blahut, "The Theory and Practice of Error Control Codes," in 1983, Addison & Wesley, pages 174 and 175.

H Just as an example, some essential information 10 about the Reed-Solomon code is given below. Specific characteristics of the Reed-Solomon code can be specified by the following parameters: I m = number of bits in each symbol n = number of symbols in the block = 2m-lk = number of message symbols (number of 15 message bits = km) t® correction capability in number symbols = (n - k) / 2

A nine-digit zip code and a single alphanumeric character for further identification purposes require 36 H 20 bits without error protection in the example given below.

A Reed-Solomon code with the following parameters was designated H for the high priority message.

m = 6 (6 bit symbols) I

n = 26-l = 63; H = t = 10

Why: k = n - 2t = 43 I

H Since only six 6-bit symbols are required to I

H represent in the 36-bit message are the other 37 symbols I

H 3 0 (43-6) fill symbols inserted between encoding I

H means and decoding means which are not to be placed on I

H label. Ie the total number of bits needed

to the label for the high priority message is (63 - 37) x I

6, or 156 bits. IN

35 This error code scheme will be able to correct a maximum of 60 I

(10 x 6) bit errors, which makes up 38.5% of the bits used. IN

43 DK 175729 B1

As follows from the large number of completed fill symbols, I

the high error detection capability of this Reed-Solomon encoding makes it highly unlikely that the high priority message will be read incorrectly.

5 The low priority message was encoded with a Reed-

Solomon error protection code with deviating parameters, namely: m = 8 (8 bit symbols) n = 28 - 1 = 255 t - 36 10 k = n 2t = 183

Since there are 1292 bits available for encoding on the label of this example, a total amount of 1136 bits (1292 - 156 high priority message bits and check bits) 15 are available for encoding and check bits for the low priority message. Thus, the remaining 904 bits (255 x 8-1136) would have to include fill bits. This provides options for 560 bits (183 x 8 - 904) for the information content of the low priority message and 576 check bits.

20 To further ensure the re-provisioning of the high priority message, this is also included in the low priority message. By the Reed-Solomon error protection code used for the low priority message, this encoding enables an additional 86 6-bit alphanumeric characters with a maximum error correction capability of approx. 25.4%.

Using the aforementioned Reed-Solomon error protection encoding, the total number of 1292 information bits available is distributed on the visualized label as follows: 30 36 high priority information bits 120 high priority check bits 560 low priority information bits (including 40 bits of the high priority message, as is inserted into the low priority message) 35 576 low priority check bit

I I

I DK 175729 B1 in I 44 In the data bitstream, which includes the associated check bits for protecting the information, is transmitted to the single hexagons in the FIG. 5. It will be appreciated that many different distribution patterns can be used, emphasizing that the essential criteria to be determined are: (1) safe placement of the high priority message close to the capture mark (if one is provided H in the data set ),

In 10 (2) providing a pattern from which it is

reasonably easy to make readjustment. IN

The special error coding program used in it I

The example given here is included in the program supplement under j I

In the program "ERRCODE.C 'on page 15, lines 1-52 and page 16, I

I. 15 lines 1-50. IN

Encoding associated with Reed-Solomon codes requires you

multiplying the message code vector by a generator I

matrix. The matrix multiplication is performed using I

Galois field arithmetic. Adding two elements to field I

20 is provided by the performance of an exclusive operation I

H between the two elements. Multiplication is performed via a "log" - I

operation in the Galois field. log and antilog are provided

using lookup tables generated from I

primary polynomials, especially for the high priority message: 1 + I

25 x®, and for the low priority message: 1 + x2 + χ2 + χ4 + x8 I

Referring to FIG. 9, an auxiliary device is provided

program "GF.C" 126 for generating those for Galois field- i

the arithmetic required lookup tables. Utility I

"GF.C" is listed in the Appendix page 8, lines 1-53 and I

30 page 9, lines 1-32. The lookup tables are calculated and stored in I

file "GF.LUT" 127 for use during encoding and decoding- I

equipment. The generator polynomial g (x) for the Reed-Solomon code is fixed I

H is given by the following equation: I

H g (x) = (x + a) (x + a2) ....... (x + a2t) where a is I

H 35 the primitive element of the Galois field. IN

II

DK 175729 B1 j

Generator matrix for the Reed-Solomon code is provided by executing a long divison of each row in the generator matrix.

The known row in the generator matrix is obtained from the others by the exercise of a long division of xnmed with g (x).

5 The calculation of the generator polynomials g (x), as well as the generator matrices for both high priority messages and low priority messages, is implemented according to utility program "MKRSLUT.C" 125, which is listed in the program supplement page 10, line 1-52, page 11, line 1-53, page 12 , lines 1-54, page 10 13, lines 1-52 and page 14, lines 1-4. The lookup tables for the generator matrices are generated and stored in the file "RS.LUT"

Hexagon labels are printed with standard printing equipment that is widely available and inexpensive. A 15 printer with a 300 x 300 dot matrix per square inch (approx.

6 cm 2) will provide satisfactory results when printing labels in three colors (black, gray, white) with 888 hexagons plus a centrally located capture mark. A printer with such performance is the Hewlett Packard 20 Laser Jet Series II with 0.5 megabytes of data storage and a graphic resolution of 300 dots per minute. inch (about 2.54 cm). A 300 x 300 pixel grid with a density of 90,000 pixels per square inch produces approx. 90 pixels per hexagon of the preferred I

embodiment. Each pixel is assigned a value of 0 or 1, I

25 represented by a black or white pixel. This one

printer is used to print data sets in two colors with black I

and white hexagons. It can also be used to press I

datasets in three fax machines with black, white and gray hexagons I

if a halftoning algorithm is used for generation I

30 of gray hexagons, as previously explained. IN

Referring to FIG. 9 there is by means of an I

stored program "MKMA.PS.C", 140 provided an area lookup I

Table "REGIONS.LUT" 141 with 34 rows and 30 columns, which I

is analogous to that shown in FIG. 5, but adapted I

35 for the designation of black or white for the rings in the capture I

marking. The individual hexagons are encoded in black, I

SDK 175729 B1 jlagsta - a sub-listening aid for> axle per. Grinding and grinding is applied to the rings [t Covered Just by icing in 10 large referrals. Aging table: auxiliary: laid out: program page 17, car sharing and areas value on: slides in a storage right area close: smaller LC "180 n which is in and white: the icon is used when other color compositions (which is possible) will of course require the printing of other colors a considerable complexity over a solution using only three colors, black , gray and white or compared 5 to a solution with only two colors, black and white.

Thus, when each pixel of the printer has been assigned a black or white value, the label can be printed to produce an encoded format, as shown in FIG. 3, wherein certain hexagons are white, some are gray and some are black, and 10 wherein a capture tag, preferably provided with black and white concentric rings, is located in the geometric middle portion of the label.

After explaining how data is encoded and printed on the label, it is necessary to explain the subsequent translation or decoding process of the label. It will be appreciated that it is desirable to practice the label translation function at very high speeds, on the order of a fraction of a second, to increase the efficiency of handling the package (or other handling or label reading).

There are two different basic solutions that can be used in conjunction with recording the image when reading the label: The label can be replaced at a relatively low speed using a manual, fixed focus static scanning device. Alternatively, in order to achieve high operating speed, it is desirable to use an electro-optic sensor with a servo-controlled focusing mechanism to enable dynamic scanning of rapidly delivered packages of various sizes and heights. The decoding process and equipment explained later have been used in connection with a fixed focus scanning device. A process of the general possibilities explained herein with respect to a static focus device with fixed focus can be adapted to a dynamic scanning system with certain modifications in the optical system, as noted later. When processing packages with

I DK 175729 B1 I

43 I

At high speeds, it is desirable to have a high speed I

In scanning mechanism that can read labels that are being fed

with a linear velocity of approx. 100 inches (about 250 cm) per IN

For a second or more and carried under the place where the fixed I

In 5 scanning device is located. Imaging Function- I

Thus, in the following steps, the steps include, I

As the steps of the decoding process are shown in FIG. 7. I

H When a package or letter is conveyed at a high speed; IN

heat transport system, the illuminated area will be fairly j I

10 large, since the size of the packages to be fed 1 l

in the feed system, can be very large and is variable. IN

Eg. it is not unusual for an I to be provided

conveyor with a width of 42 inches (about 105 cm), by which I

In packages with a width of only a few inches and up to 3 I

In 15 feet {approx. 90 cm) (and the corresponding height) must be transported. IN

Therefore, a label of 1 square inch (about 6 cm2) may be I

located anywhere in the width of the carrier. The packages I

may also be located at oblique angles relative to the trans- I

the axis of movement of the door belt. Packages, letters and similar gen- I

H 20 stands will be provided with mutually different highs

In there, so that the labels to be scanned, e.g. IN

You can be placed 1 inch (about 2.5 cm2) or less above I

above the conveyor belt, and up to 36 inches (about 90 cm) or I

more above the conveyor belt, taking into account the I

25 maximum height of the packages that the plant can carry. IN

In order to properly illustrate the labels that you

is provided by the present invention, and in particular I

having regard to the wide variation in the width of the packages,

height and the angle at which the labels are displayed are I

It is desirable to use a high intensity light source, so that I

that the light may be well reflected from the I

identified two or more optical properties of the label. IN

The light may be infrared light, ultraviolet light or visible I

light, and the light spectrum of the applicable visible light can j I

35 vary. The technique of sensing the light includes pre-I

stepwise sensing of light reflected from the blacks, whites and I

49 DK 175729 B1 gray hexagons on the label.

The light source must produce light so that sufficient light is reflected to the light sensor (for example, a CCD device, as explained later), so that light sensor 5 can reliably distinguish between black, gray and white, or any other optical characteristics the hexagons are assigned. In a dynamic scanning system, a set of light-emitting diodes could be used to produce a light level of approx. 10 mW / cm2 in the area of the label to be illuminated, 10 at the level of the label. Such light-emitting diodes may be located over a certain area without the use of focusing lenses, or they may be arranged linearly with a cylindrical focusing lens. A laser light source passed through a suitable optical system to produce a linear illumination source could also be used in the practice of the present invention.

Light source selection and designation of properties.

one skilled in the art can be provided at the light source for the application herein. It should be recalled that since the label to be located has only a maximum size! of 1 square inch (approx. 6 cm2) and placed at heights · up to 36 inches (approx. 90 cm) on a 42 inch (approx. 100 cm) conveyor belt conveyed at speeds of f .g. up to 100 inches (about 250 cm) per second, it is very important to be able to illuminate the label, appropriately for ambient identification and location.

In connection with the static focus fixed sensor used in the illustrated example, it has been found that an illumination level of approx. 2 milliwatts / cm 2 is suitable for practicing the invention. Such illumination was provided by a fluorescent light source.

The second step in the recognition part of the decoding process is the optical sensing of the illuminated area with an electronically driven sensor. The camera / light sensor used in the illustrated example of a fixed focus static scanning system includes a color CCD television screen.

DK 175729 B1 I

50 I

! more of industrial quality, such as model No. WV-CD 130, which I

marketed by Panasonic Industrial Company, One Panasonic I

Way, Secaucus, New Jersey 07094 and provided with I

a 50 mm f: 1.3 C-mounted TV lens with a 5 mm pull-out I

5 tubes marketed by D.O. Industries, Inc. (Japan), 317 I

East Chestnut Street, East Rochester, New York 14445, and as I

is marketed under the trade name NAVITRON ™. The camera I

was connected to an image capture card called I

model DT-2803-60 marketed by Data Translation I

10 Inc., 100 Locke Drive, Marlboro, Massachusetts 01752.

Optical sensing may include an image transmission signal

applying the entire label using an area sensor;

such as the camera and image capture described above

or it may alternatively be provided with an I

15 linear scanning sensor with a chip with a charge coupled I

device ("CCD"), the scanning of the second label of the label

dimension is exercised by movement of the package (and label). And I

suitable CCD chip for this purpose is a linear CCD image sensor I

at high speed called Thomson-CSF THX 31510 00Z, 4096, I

20 marketed by Thomson-CSF, Division Tubes Electroni- I

ques, 38 rue Vautheir B.P. 305 92102 Boulogne-Billancourt I

Cedex, France. IN

In dynamic systems which include movement of I

the label carrying package in a transport system, it will be you

25 advantageously have a long optical path between the labels I

to be sensed and the light sensor. The main reason I

to produce a long optical path is thereby reduced I

the change in apparent size or magnification

of the label as sensed at a distance I

30 positioned light sensor. For example, the optical path is 4 feet I

(approx. 120 cm), the image size of the label will be 1 inch (approx

2.5 cm) across the conveyor belt has been very different, from I

the size of the labels 3 feet (about 90 cm) above the transport I

band. If a long optical path is used on e.g. 20 I

35 feet (about 6 feet), the image sizes of the same label- I

ut be approximately the same. This ensures that you

51 DK 175729 B1 sensing, regardless of its height, performs or nearly performs the entire range of the light sensor so as to provide a consistently high image resolution. If an area sensor is used instead of a linear sensor, the same principle 5 can also be used. This can be provided by a long optical path as illustrated in FIG. 6th

In order to focus on labels on packages of different heights, it is necessary to provide a height sensor. For this, an ultrasonic sensor can be used or the package can break a set of light rays. Any of these systems is applicable and the system can then activate a suitable adjustable focusing mechanism with an open or closed loop mechanism for sensing and sensing. positioning the optical sensing element (e.g., 15 lenses and sensor) in mutual relationship on a continuous basis, as shown in FIG. 6th

FIG. 6 is a schematic view of a camera focusing and adjustment system operated in accordance with the invention for adjusting the position of the camera light sensor 20 according to the height of the package being sensed.

In FIG. 6 shows a suitable lens 196, coil drive, height sensor and feedback loop according to the invention. In FIG.

6, the height sensor 206 may be an ultrasonic height sensor, or it may e.g. be a light beam which is refracted at each packet carried by the carrier. The output of the altitude sensor is transmitted to the microprocessor 204, which in turn activates the coil drive 202 for moving the coil 200 to which CCD 198 or another suitable light sensor is mounted.

An axis location sensor 208 senses the location of the coil 200, and the output thereof from it to the microprocessor 204 completes a feedback loop for sensing and adjusting the location of the coil 200.

The sensor must be capable of sensing the reflected light emitted by the illuminated label, and it must also produce an analog signal according to the intensity of the

I DK 175729 B1 I

52 I

In the reflection property of the label, such as these, I-

Brings at the individual pixels of the electro-optic sensor. IN

B A suitable light source, as explained above, may be I

Located on a mounting surface above a conveyor for

B 5 illumination of an area extending across the transport island

Clean the entire width with light of a predetermined quality and

B intensity. The reflected light from the label may become you

B deflected by a number of reflectors, then sensed by I

B electro-optical sensor. IN

B 10 The purpose of deflecting an optical path is to produce I

B a compact and therefore more rigid system. IN

B Filter the analog output of the sensor I

B hereafter. The analog electrical signal is used for i-I

B attachment to an analog bandpass filter to detect time

B 15 the absence of an intercept mark in the data set. The ten

B analog signal is then converted to digital signal I

B using a known analog-to-digital converter, I

B which is included in the image capture card, as below

B below explained, or by others known in the art

B 20 organs. Instead of an analog bandpass filter it is possible I

B to insert a digital filter circuit to determine I

B the presence of the intercept mark by comparison I

B of the digital data representing such a mark,

B with the digital output signal from analog-to-digital con- I

B 25 hosts, as explained later. IN

B An example of an area sensor with a CCD chip with I

B a number of detectors used in conjunction with the I

In the present invention, it is previously referred to as Panasonic I

B WV-CD 130 color CCD television camera. The analog output sig- I

B 30 nal from the sensor is transferred to the previously mentioned data area

B phrase DT 2803-60 image capture card, which includes-

Added a 6-bit monochrome video with A / D conversion to digital

B talization and later treatment. Using a suitable, I

B stored subroutine, the digits arranged in sequence are stored

B 35 speech output from the image capture card in a data I

Image orientation as an exact copy of the image which is I

In DK 175729 B1 been registered with the optical sensor. j

The most important part of the invention is the processing of the optically sensed image so that the original label structure can be recreated and oriented with accuracy, and so that the color (optical property) of each hexagon is reproduced. This is practiced using the steps mentioned below, according to which the known pattern by which the label was initially encoded and the bits were set up can be used to decode the information contained in the label.

Prior to using the above-mentioned CCD television camera and image capture card, as shown in FIG. 10, an introductory program "DTINIT.C" 250 is run such that the image capture table is placed in a known clear state, and so that the output color lookup tables followed by the program "DTLIVE.C" 255 are advanced, thereby activating the image capture card. The "DTGRAB.C" program then causes the image capture card to digitize the image for transfer to an image data store of 24 0 rows and 256 columns, with samples stored as 6-bit sizes divided by bytes. Two utilities, "DTSAVE.C" and "DTLOAD.C," establish the transfer of screens to and from a storage medium.

In the first recording of the label image, a known analog bandpass filter can be used to identify two or more optical properties of the capture mark concentric rings. It is preferable that these two optical properties be the colors black and white, since the greatest contrast will produce the greatest signal energy. In order to find a fixed pattern with transitions from black to 30 white to black, etc., it is preferable to perform a linear scan over the capture mark, which scan must be passed through the center of the mark, thereby providing a uniform frequency response independent of the label. orientation. That is, the marking rings give optimum effect when provided as contrast concentric rings. The output of the sensor is split SDK 175729 B1; one way along the lighthouse: the le le ener-center jkalise-: odelis-dpasfil-t underneath, it must lie diurnal rendition analogously. ring of FIG. in manual. Partial> suitable by corir ligvi s eybrin-: the condition.

analogously, the IBM w York, applied ring 1, is applied by a normalized digital bit stream, as explained later, through the filtration subroutines discussed below. The band that is filtered corresponds to the expected ringing frequency.

The one-dimensional digital bandpass filter is sized 5 to a sample size of 400 pixels per second. inch {approx. 2.5 cm) and to a length of 125 pixels (or 0.3125 inches - about 8 mm), designed based on the size of the printed capture marking rings, as shown in FIG. 3. The frequency was 300/16 line pairs per second. inch (about 2.5 cm), thereby providing a normalized frequency (where 400 line pairs per inch = 1) of 300/16 x 400 or 0.046875. A filter j with a band pass range that extends 5% below this frequency and 15% above was designated because distortions of the label typically result in image shrinkage and as a result! 15 of which an increased frequency. Stop bands from 15% below the frequency down to 0 and from 25% above the ring frequency to 0.5 (Nyquist's limit) were provided. The filter coefficients were stored in the file "IMPULSE.LUT" 275, cf. 10, for later use, omitting the first 62 coefficients since the filter is symmetrical. A flow chart is shown in FIG. 8th

A 25 pixel length filter was provided ΐ by sampling the bandpass filter at output intervals corresponding to the measured horizontal magnification. Eg. if the horizontal magnification of the image is 80 pixels 25 per inch (about 2.5 cm), every fifth sample of the filter used (400/80 = 5 pixels). For non integer steps, linear interpolation of adjacent filter samples was used.

in

Another two-dimensional filter of 25 x 25 pixels was also used. Sample values for this two-dimensional filter were based on the Euclidean distance of each point from the center of the filter, which was scaled for suitable horizontal and vertical magnification. Linear interpolation was then used for non-integer sample intervals.

The output of the above-mentioned one-dimensional filter was squared and smoothed by a first order i

I DK 175729 B1 I

I 56 I

In recursive low-pass filter, thereby providing an exponent

In tented window. When the smoothed output of the filter I

You exceeded a predetermined threshold, used a choice I

In designated two-dimensional filtration step for verification I

I 5 of the presence of the mark and to accurately determine- I

Given its location as explained later. The first one

Part of the two-dimensional filtration worked with a reduced I

10 pixel x 10 pixel filter size to reduce computation

In scope. This filter scans a rectangular area I

10 around the location detected by the one-dimensional filter

I tion. If the maximum two-dimensional correlation over- I

exceeding a predetermined threshold, the end I was used

In equal steps in the two-dimensional filtration with a full 25 x I

In 25 pixel filter for a narrow square window around the max

For 15 sims. If the best result from this filter exceeded I

a predetermined threshold, the center was detected. IN

If any of the threshold values were not reached

H exceeding, the program would partially postpone the smoothing file

and return to one-dimensional scanning. If you

The one-dimensional scan was completed without detection

by the presence of the intercept mark you would

the program sends out an error tag. IN

The reflected light intensities recorded i

at the optical sensor used may vary as I

H 25 due to variations in lighting, print density, paper I

H flashback, camera sensitivity and other reasons, herein

including a deterioration of the label, f.esk. folding, I

displacement, etc. As an optional (but desirable) step, you can

the reflected light which is sensed by the sensor and which over- I

30 is fed to the data store, normalized by known procedures. IN

H Using prior art, a stored nor I was used

malization program "NORM.C" 270, shown in FIG. 10 to I.

H analysis of the intensity level of it from the label back I

discarded light as detected by pixel blocks I

H 35 in the scanning device to find the minimum value

High and maximum values of reflected light intensity, I

57 DK 175729 B1 which was registered for the dataset. The set digital output signal from the scanning device mentioned above and from the image capture card was transmitted from the data store to the computer for further processing with said stored normalization program.

Using the equation y * mx + b, where, by inserting the minimum intensity into the density instead of x, a value of y = 0 will be provided and where, by inserting the maximum intensity of x, a value of y will be provided. = 63, the recorded intensities of the reflected light were adjusted for each pixel so that the blackest black color and the most white color present in the stored image were established by default and so that other shades of black , white and gray were adjusted according to these standards. The normalization step thus facilitates the processed image. The normalization was performed using the stored program "NORM.C". It will be appreciated that other more sophisticated normalization procedures known in the art will be applicable.

For the subsequent calculation, the stored reproduced label image is rescaled to provide an image of equal horizontal and vertical magnification.

This is again an optional step, but it facilitates the quick and accurate retrieval of the encoded information. The rescaling operation was performed to provide an image with uniform horizontal and vertical sample resolution of e.g. 150 pixels per inch (about 2.5 cm), as used in the illustrated static embodiment of. The invention has a fixed focus.

The re-scaling operation is provided by calculating the split row and column address e.g. specimens at 1/150 inch, based on the known horizontal and vertical magnification. Each point on the new uniformly rescaled image is then extracted from a suitable set of points on the rendered image in the storage medium. Bilinear interpolation is used for

DK 175729 B1 I

58 I

to approximate the value of the points to the sub-addresses. By re- I

the scaling is placed the center of the label in a known position I

in the data warehouse. The rescaled image is stored for later

use in the search step. In all subsequent process steps I

5, it is assumed that a re-scaled label image is centered

right in a known location in the web, but it should be noted, I

that no orientation of the label is hereby stated, * I

which label may still be inclined to ratio I

to the sensor. The rescaling operation is performed under control I

10 from a stored subroutine.

The next steps in the process are collectively referred to as "two-dimensional

The steps are performed by a suitable,

stored program and subroutines, referred to as "CLOCK.C" 290, I

as shown in FIG. 10 and listed in the program supplement of I

15 pages 44 to 51. This operation is performed in two dimensions I

on the rescaled image to accurately determine the plate I

the hex of each hexagon in the original dataset. Purpose I

with rate recall is to determine the sample location I

rings and to correct the effects of shear, curl- I

20 inclination or inclination of the label since the label does not I

necessarily is completely flat. This is an essential part I

of the process, and its use is not limited to by six-

edges encoded labels. It may also apply to others

processes of decoding the encoding tag, which includes I

25 is a regular two-dimensional grid such as squares, triangles I

etc. I

One-dimensional clock gene development is a known process, I

which is known in the field of signal processing. Two-dimensional I

beat recall is an extension of this process and you will

30 could be understood by the close reflection of those skilled in the art

the area. It should be understood that the term "bellwashing" I

seems somewhat confusing to the skilled person, since it is

has nothing to do with the temporal division. IN

The first step in performing the clock gene induction I

35 may be exercised by various non-linear placement operations

known in the art for providing I

DK 175729 B1 of signal components with a designated clock frequency, which is missing from the digitized image output signal from the optical H sensor and the image capture card. The purpose of the non-H linear placement operation is to provide the (pre-H 5 step) normalized and re-scaled image which H is provided at this point in the process and to shape it into a two-dimensional non-linear distribution that enhances

the transitions between adjacent contrasting hexagons. IN

In the preferred embodiment of the present invention 10 this is done by standard deviation distribution. This step I can also be performed by filtering with an image differentiation nucleus, which is known in the art, such as LaPlace or Sobel's nuclei, after which an absolute value is determined or a squaring of the results is performed. These processes are mentioned in Rafael G. Gonzalez and Paul Wintz, Digital I Image Processing, 1977, Addison Wesley.

I At the standard deviation distribution, image I is stored with undifferentiated cell-to-cell edges in the data store.

A standard deviation distribution is then provided for placement of the edges on adjacent, contrasting hex edges by determining the standard deviation of 3x3 pixel groups (which are not identical to the 3x3 I cell cell groups), to determine the standard deviation of pixel - In the tensities. The standard deviation calculations are performed to determine the pixel ranges, with a fixed color (with the lowest standard deviation), representing the interior of a hexagon, or representing the transition between two ► single-color hexagons, in contrast to the pixel groups provided with larger standard deviations , which represent 3 0-center transitions from. a hexagon with a color to an adjacent hexagon with a contrasting color. Since adjacent hexagons often have the same color, the standard deviation distribution will not completely delineate each hexagon. Missing edges between the hexagons will be provided by the fact that in the standard deviation process no distinctions can be made between hexagons with 175729 B1 ice propagation of those in the metric on a brought late metric or t for the tick: whether in an elec- trent-whole printed is stated: or oces- after <:> they are preceded.

six is lygo-> rt in benefit I I 175729 B1

late. Use of polygonal shapes with poorer packing properties, or sets of partially juxtaposed,

or not at all opposite polygons, instead I

for a pack of adjacent polygons,

5 will thus result in a poorer but nonetheless I

usable system performance in many applications. On certain I

however, points are provided as a result of the optical I

non-soluble high frequency components in polygonal encodings- I

lower order cells, ineffective cell packing and I

10 predetermined two-dimensional arrays, large spaces I

between the polygons, which is why the system performance will drop to an I

unacceptably low ability for information storage and retrieval- I

reduction. IN

Whether the system is acceptable depends on the quality of I

15 shows the signal provided by the electro-optical sensor. IN

When changing the sensor system, e.g. by increasing the number I

of samples per area unit above label surface, I

the signal detected by the sensor may be improved;

whereby the characteristics of information storage and repetition I

20 for such label arrangements, where polygons I

ne are partially in contact or not at all

in mutual touch improves. Such adjustments to

make less desirable label designs usable

lie within the knowledge of any professional within you

25 area is in possession of. IN

The process therefore created opportunities for large I

differences in the label article, bodies for providing I

»Optical signals and signal processing. Polygonal Cells I

of a regular or irregular form can be used as encoder I

30 optical units on readable labels according to the invention. IN

Furthermore, as long as the distance between each other and you

direction between the centers of the polygons is known in association I

to adjacent polygonal cells, the polygonal encoded I

cells could be in a predetermined pattern, such as I

35 differs from the hexagonal arrangement and the polygons can be

be positioned so that they are opposite each other so that they

BK 175729 B1 do not go to ssfor-: times> tical t will are and: monkeys.

inland

This giant urie-: snatching elbows-of 9 ides 8 half the time for weed-opera-> s too learned ^; now the eight technique is the product-: times j DK 175729 B1 63

The next subroutine, called creating a window, is optional. In practice, this subroutine is used in the invention to reduce the intensity of edges that are not attached to the hexagon's outline. These edges appear at two locations, at the marking rings, and the unmanaged image surrounding the label. A weighting function is used to reduce the intensity of these areas. The details of how the provision of a window acts as a precursor to a fast Fourier transformation are within the skill of the art.

A two-dimensional fast Fourier transform of the digital sizes corresponding to the (optional) standard deviation distribution processed by the window procedure is then performed under the control of a commercially available stored program. During the operation, a computer performs a rapid Fourier transformation of the image provided by the preceding step to produce a two-dimensional representation of distance, direction and intensity at the transitions between contrasting hexagons identified by the standard deviation distribution step. In short, the fast Fourier transformation is a measure of the distance, direction and intensity of the edges between the hexagons where these are known. That is, the regular distance and direction of the boundary of the hexagon will cause 25 certain points in the transformation area to have a high energy level. The brightest point will be 0.0 in the transformation plane corresponding to the DC component of the image. The six points surrounding the central point represent the distance, direction and intensity of the edges between the hexagons.

It will be appreciated by those skilled in the art that, as for hexagons, a two-dimensional representation of the distance, direction, and intensity of the transitions between contrasting polygons identified in the preceding standard deviation distribution step can also be calculated by performing a fast Fourier transformation of the digital data,

DK 175729 B1 I

64 I

which corresponds to the non-linearly distributed sensed label car I

missing. Thus, the distance and direction of the polygon boundary will

like to cause certain points in the transformation area I

has high energy. The number of points of higher energy which

5 surrounds the central point with the coordinates 0.0 in the transform I

the plane of the plane, will depend on the geometry of the particular I

polygonal encoding cells used for generation I

of the optically readable label. As for hexagons, you will

however, such points that surround the center point I-

10 time represent the distance, direction and intensity of the edges I

between the polygons or edges between polygons and intermediate I

spaces whose label structure is such that the polygons are I

in partial contact with each other or not in contact at all

together. IN

15 Since the image is assigned real values and not I

complex values, the transformation range is point symmetric I

around the point of beginning. As a result, it is only necessary

to calculate the half plane in the transformation area, I

thereby providing a saving I during the calculation time

20 on a factor of about two. Removal of these calculations

also reduces the required work in subsequent steps

with image filtering and reverse fast Fourier transform I

tion. The fast Fourier transformation program that uses I

in connection with the illustrated embodiment of I

25 a static system with fixed focus, it was commercially available

equal subroutine R2DFFT from package 87 FFT-2 from Microway, I

Inc. in Kingston, Massachusetts. IN

A filter process is now required to reconstruct _ I

the total outer constraint of all the hexagons in picture I

30, using the transformed digital I

data. This is done by eliminating all transformations

points of advice which do not correspond to the designated one

distance and direction of the boundaries of the hexagon, thus I

as these are identified in the standard deviation I step

35 distribution. Six prominent points are made in I

the transformation area due to the hexagonal I of the label

In DK 175729 B1 honeycomb structure. In fact, only three points are defined in the transformation area, since the image is point-symmetric about the starting point, and the other three points can be inserted from the first three. In the preferred embodiment,

5th form, the filtration is performed in three steps to remove the transitions from the standard deviation distribution step which are too spaced apart, too close together, and / or I

lies in the wrong direction. IN

First, high pass filtering is performed by assigning I

10 all points within a predetermined value around I

the starting point in the transformation range value zero, since they I

is at a distance of eight from the starting point, except I

six points designated in hexagonal shape I

in the graphical transformation area. These points correspond to I

15 distances which are greater than the hexagonal distances and carry sole

the information related to the missing transitions I

in the label image. To recreate the missing transitions in I

In the label image it is necessary to remove information about I

the missing transitions in the Fourier transform domain. IN

20 Next, all points are assigned outside a designated I

radius around the six prominent points of the transforma- I

tion range value zero. These correspond to foreign transitions, I

which are too close together. This operation is combined with I

the first such that a ring of residual I is formed

25 points. The production of this ring corresponds to the practice of I

mellemrumsbåndpasfiltrering. The inner and outer radii of I

the ring is determined from the expected distance between six

* perimeter of the edges. Since the "diameter" of the hexagon extends

to represent 5 pixels in the example explained here you should

30 hexagonal tips for a transformation length of 256 l

pixels lie 256/5 = 51.2 pixels away from the center. As a result I

of which, a ring having an inner radius of 45 l was used

pixels and an outer radius of 80 pixels, corresponding to six- I

edge diameters of 3.2 to 5.69 pixels. A filter which allows I

35 passage of higher frequencies was used because the deformerin I

of the label, such as shear and inclination causes

I DK 175729 B1 I 66

H

In image shrinkage. !

Following the performance of the interspace bandpass 1 filtering, as explained above, a

In ring with six prominent points, each point having

5 the same angular distance with respect to the center (point I 0.0) in the transformation area. To complete the task of removing unwanted information in the transformation area, a directional filtering step is used. Any point located at too great an angular distance from the prominent regions of the transformation region is assigned the value zero. This; has the effect in the auxiliary region that any edge which is not in one of the three directions determined by the six-! angled honeycomb pattern, removed. j

For the practice of directional filtering, it is necessary j

H 15 to find the most prominent point left after I

H the gap band pass filtering. This point is assumed to be one of the six prominent points in the transformation region, j H corresponding to the angular tips of a hexagon. Five other prominent points at the same radius from the center and at angular distances of 60 ° multiples are also prominent in the transformation area. Therefore, all other points with an angular distance greater than 10 ° are removed from a any of these points. Six wedges of the ring are left. At this directional filtering step, any information containing incorrect distance or direction in the image area is removed. The removal of incorrect distance information allows the recovery of a complete perimeter of each hexagon in the image area.

The aforementioned filtration steps are performed under the control of the stored subroutines.

The foregoing explanation of the filtration methods used in conjunction with the preferred label embodiment with closely adjacent hexagons requires modification when deviating 1 H 35 predetermined two-dimensional arrangements for the optically readable label are used. However, it will be appreciated by any skilled artisan in the art that only minor modifications are required in the filtration process to adapt to the divergent label arrangements previously explained in this specification, as illustrated in the accompanying drawings. attached figures.

Once the individual polygonal encoding cell is first determined, it is also predetermined that its delimitation will occur at certain angular distances, as they also have a designated number of sides of a specified length.

Next, it is necessary to determine the ratio of adjacent polygons, such as whether they should bump into each other completely, only partially bump into each other or not at all in contact with each other. The geometric arrangement according to which the geometric 15 centers of the polygons are to be set up must also be determined. Once the aforementioned label geometry is established, a person skilled in the art can build the appropriate filtration scheme for filtering the energy points of the transformation region so that only the brightest points corresponding to the appropriate distance and direction 20 of the boundaries of the polygons are treated with the inverse fast Fourier transformation.

As for the mentioned built-in filters, you will

realize that it is necessary to build a suitable dimension- I

rather space bandpass filter, which is based on the preceding I

25 between the polygonal encoding cells. IN

In addition, it is desirable to build a directional filter for I

filtering out energy points other than the most prominent ones

points corresponding to the axes of the predetermined two-dimen- I

zonal arrangement of the polygonal encoding cells. This I

30 removes any information associated with incorrect distance or I

direction of the polygonal encoding cells in the image area I

and in the spaces, if any. By knowing that

eliminating such incorrect information can a complete I

positioning the centers of the polygonal encoding cells I

35 be reconstructed in the image area by the inverse fast I

Fourier transformation according to the one below

In DK 175729 B1 I 68 I explained the process steps.

In order to be able to return to the image area at present and in conjunction with it to recover the outline image of the fully aligned hexagons in the data array, it is desirable to perform a two-dimensional inverse in fast Fourier transformation (2D-IFFT ) on the filtered I transformation domain data. The reverse transformation is accomplished by a standard two-dimensional reverse Fourier transformation subroutine (R2DIFT), available in package 10 87FFT-2 from Microway, Inc., Kingston, Massachusetts. Upon completion of the inverse transformation step, the perimeter hexagon is restored in the image domain. In the new image, the centers of the hexagons are provided with large size. The actual size of the points in hexagons.

15 centers depend on how many edges were provided in the neighboring area. Many edges produce greater H energy at the allowed frequencies and thus large size points. Fewer edges provide points with H smaller size. Size of the points is a good measure of the H 20 confidence level in the clock generation at any point.

The hexagonal image has now been recreated, but its orientation needs to be determined.

The hexagonal honeycomb pattern provided by the invention has three "axes" with an angular spacing of 60 °. The direction of these axes is determined by the brightest points of the transformation domain after the space-band pass filtering. It is now possible to assess which of these three axes is the main axis. This step is optional-30 free. If the step is not performed, the label must be decoded three times, using each of the three axes and only one of the axes providing a meaningful message. The main H axis is arbitrarily selected as the axis which runs parallel to two sides of the label, as previously explained and illustrated in FIG. 2nd

M If the boundaries of the square label are determined in DK 175729 B1 69 based on knowledge of the main axis, most of the energy placed in the restored hexagonal pattern will be within the boundaries of this square.

To determine the main axis, each of the 5 three axes is assumed to be the main axis. The resulting square label boundary is determined for each of the three axes, and the total clock generation pattern of energy before j in the square is recorded from the digital energy data output of the inverted transformation subroutine.

10 The correct outcome is the one with the greatest energy.

The angle of this main axis is stored for the launch step and other search operations. In this situation, it is not yet known whether the recorded angle lies in the right direction or is 180 ° away from the right direction.

15 The source code lists in the program supplement associated with this determination of the main axis are listed on page 49, lines 48-54, page 50 lines 1-53 and page 51, lines 1-5. It will be appreciated that all three label regions do not need to be fully defined, since the energy in the region common to all three squares does not need to be determined.

A stored program called "SEARCH.C" 300, which is illustrated in FIG. 10, combines the information about the transformed and retrieved hexagon center with the stored intensity levels in the original image so that the gray level value for each hexagon can be determined.

The search is carried out in such a way that the possibility of "getting lost" during the search is as small as possible. At the final result, a matrix for the 30 grayscale value for each hexagon must be provided in the data set. The source code lists for "SEARCH.C" are listed in the program extension on j pages 52 to 60. Four important information sets are built up in the first part of the SEARCH.C program. The set CVAL (clock value) stores a measure of the quality of the generated 35 clock signal for each hexagon, while the set GVAL stores the gray level values (0-63) for each center of each hexagon. The

I DK 175729 B1 I

I 70 I

In the other data sets IVAL and JVAL, the row and column locations are stored I

In for every center in every hexagon. IN

I From the main axis angle defined in step (e) and I

In the known distance of the hexagons (5 pixels) in the present

In the 5 example, the expected horizontal and ver- tical are calculated

In vertical displacement from the center of a hexagon to the centers I

In the surrounding six hexagons. IN

I After these calculations, the SEARCH.C program I works

I on the clock repetition signal generated from I

In the 10 data store, and the re-scaled label image, which also I

You have been pulled out of the data warehouse. The basic purpose

for this implementation routine set out in program I

In the insert on page 52, lines 13-54, page 53, lines 1-48, page I

I 56, lines 47-57 and page 57, lines 1-35, are mixing and I

In 15, the information from these two sources is contracted and I

You bring in a data matrix with the grayscale value for each hexagon. IN

In the Commencement step of the search is delimited by an I

square around the center of the label of approx. 1/3 inch (about 0.8 l

In cm). There is a good starting point in this area, I

20 as the point of greatest magnitude in the retrieved I

clock lineup. Next, the location of I

starting from the center of the label. This I-

point of departure is a point where the clock signal is strong and clear

and also a point which is relatively close to I

H 25 center of the label. A strong, clear signal is desirable

H to ensure that the search begins with an approved hexagon I

center, and it is desirable that this point be near I

center of the label so that its absolute location

can be determined without seriously affecting displacement or I

30 slope. A measure of the quality of the point in clock repetition

The slope pattern is the size of the point minus the size of the I

H eight surrounding points. The coordinates of the starting point I

H in the right-angled coordinate system is converted to polar I

H shape, and the polar coordinates are adjusted relative to the I

H 35 principal axle angle previously determined and this result I

H is converted back to coordinates in a right-angled coordinate

In DK 175729 B1 night system. These coordinates are scaled according to the expected row spacing (4.5 pixels) and column spacing (5 pixels) so as to find the insertion location in the hexagonal matrix. Rate quality, gray levels and positions corresponding to the initial hexagon are then inserted

in the sets of CVAL, GVAL, IVAL and JVAL. IN

The main search loop continues to locate I

the centers for the other hexagons. The loop ends when you

the expected number of hexagons has been located. The I

10 order in which the search for the centers of the hexagons I

is very important. The increased reliability of decode- I

the labeling process associated with label destruction stems I

from this particular search technique used, such as below

too explained. IN

15 Each iteration of the search loop begins with repetition

the provision of the location of the rate-elicitation tip I

of the largest size whose neighbors have not been sub- I

cast search for their largest sizes. From this I

known point, the search will be extended in the distance answer- I

20 to one hexagon in each of the hexagons. In this way I

there is a search pattern along a path from the best to the best

better provided tact quality. Thus, if there is an I

provided a poor quality clock range, for example I

in the middle of a label with a broken area, the search algorithm- I

25 but move around this area instead of searching

through it. By delineating such weak areas and reserves

if they are for the last treatment, the probability of you

to get lost in the net be substantially reduced. Since it, I

Getting lost is just as wrong as reading a gray level I

Inaccurate, this property of the search algorithm is very important

equal. IN

A subroutine is responsible for searching neighbors with I

best quality clock values found by head size I

voyage. The subroutine is looped six times, one for each neighboring sex

35 edge to the hexagon being examined. First, place- I is calculated

the ring of a neighbor. If this neighbor is outside the label I

SDK 175729 B1 subject, will be subjected to, as more lacquering of the star is accepted. On searching for il, as if weighing such, ICT, in order to [radiance search-necessary, s. Then it reverses the lucid toxes, polygonal in regions, running over two-dimensional axes with are case-3 conditions in .1 it be * _ orientation. s algorithm Long process-dimensional methods of edaxing for erection, legible readable I DK 175729 B1 label, as previously explained) and providing the necessary coordinates for the search subroutine.

Whether the label geometry is determined by such an optional step as explained above, or simply loaded into the decoding process by suitable modifications of the two-dimensional clock gene development process, the various label structures discussed and explained herein can be readily adapted by one skilled in the art. It will be appreciated that the number of axes on which the centers of the individual,

10 adjacent polygonal encoding cells are located and I

their respective angular orientation can be inserted into I

the step of determining the principal axis instead of the three I

axes of the six-I used in the preferred embodiment

edge lineup. Therefore, the major axis of the predetermined I

15 two-dimensional arrangement is determined without the exercise of the I

explained in step (e), experimental analysis. What you

relates to the hexagonal arrangement of the preferred embodiment

form can the information from the step of determining the main I

the axis and the known distance of the polygons are used to calculate I

20 the expected horizontal and vertical displacements from I

the center of one of the polygons to the centers of the surrounding I

polygons. After these calculations and after the necessary

adjustments in the search subroutine are performed, the search, herein I

including the preamble step and the main search loop step

25 fin, continue for this particular label design that I

is used. It will be appreciated that such minor adjustments

of the SEARCH.C search routine. 300 in the enclosed source code list I

may be exercised by a person skilled in the art. IN

After completion of the subroutine, the placement I is marked

30 of the present center so that it is not sought. IN

The intention is to exclude this point from being a point, I

whose neighbors should be sought. For loop iteration, add from I

0 to 6 new candidates and 1 candidate are removed. By an effect I

For the best use, a data structure could be used, I

35 in which the candidates are placed on the order of magnitude

as insertion and deletion operations are performed. Such an I

SDK 175729 B1 in id Ullman, 'L974, Ad-me requires: priori-lvor pos-can remem- ber the opera order n 3es if: eres to the 1 ication> g grenier has been ji> r to op-order> r eg.

:). This size-ί. Level- 't.

is still in the starting position none of the ^ center.

»The resemblance of a multi-way center.

: the ones for the; For failure, in space for passing DK 175729 B1 j in 75 closest to the label center. If a row above or below is closer than the row with the hypothetical center, appropriate shifts up or down are performed. If the left setting of short rows is exercised incorrectly, this is adjusted by switching the short rows one position to the higher re.

Horizontal location errors and reading the label the wrong way are investigated using the information laid down in the label, which is referred to as rough network information. The information is divided into hexagonal groups of 3x3 cells, as previously explained. For example, since the label may be provided with a grid of 33 rows and 30 columns, these groups form a grid of 11 x 10. The lower middle hexagon of each complete group of 15 x 3 cells has a particular property provided during encoding. A transition on each side of this hexagon is assured, as previously explained in connection with FIG. 4. For example, if the lower center hexagon is in black, the lower left hexagon and the lower right hexagon 20 must be either gray or white. A stored subroutine, listed on page 59, lines 27-52 and page 60, lines 1-33 of the program add-on utilizes this transition property to remove the completed two possible distortions. First, a set is provided, with each element of the set indicating whether a transition has occurred between two horizontally juxtaposed hexagons. Next, the set for each of the nine hypothetical cuts in the coarse grid is examined as a 3 x 3 pattern around the expected cut at 0. One of these cuts will show a better fit between the current 30 and the expected transition, and this cut position is maintained. Next, the same hypothesis is tested under the assumption that the label was read the other way around. This will occur if the main shaft angle is currently pointing from right to left relative to how the label was printed, 35 instead of from left to right.

If the label is simply the reverse, i.e. that they

I DK 175729 B1: I

IN

Η I

I - il

I 76 I

In higher rows, they are interchanged with the lower rows and that they

In higher columns are interchanged with lower columns, the result I

of the incision placement also be an inversion. However, I

You must undergo one significant transformation for correct I

In 5 ways to invert the label. While reading the short (length I

29) rows are defined on the left. Thus, when the label is I

In inverted, these labels need to be adjusted to the right. Just- I

The ring is being exercised, and it is this procedure that will cause,

In that the results from the incision hypotheses are anything but a simple one

In 10 inversion. In fact, the best results from I

In the cross-sectional test be better than any time- I

In the aforementioned test, if the label is in fact I

H is read inversely. IN

After determining whether the label is read

I or vice versa, and whether or not there is a cut in it

In absolute position, the label's matrix can now be decoded. IN

I With the correct determination of the image and cut, the image I

the processing functions completed, and the data decoding process- I

the commences begin. IN

In a stored program "RD.LABEL.C" 182 in FIG. 9, I

the file provided by the search appliance is read out and. IN

A bit stream file is generated with, in the preferred embodiment, I

In hex form, 1292 bits. A stored subroutine, CELL I, is used

DEC.C 183 in FIG. 9, which is listed in the program supplement to I

25 pages 63 to 66, for masking out useless hexagons, I

and for decoding, which is the reverse of I

In the coding program. IN

I The first step in the decoding process is to generate I

a bit stream from the hexagonal information, using I

30 is a hexagon-to-bit distribution process which is the inverse I

of the bit-to-hexagonal distribution process used I

at the encoding operation. IN

The bit (information) stream is then divided by pro- I

the gram in a bit stream with a high priority message and an I

35 bit stream with a low priority message, or in as many I

H bit streams used in encoding the label. IN

77; DK 175729 B1 j i

It is then necessary to apply error correction! to each bit stream using error coding techniques used in the label encoding process. If in e.g. When Reed-Solomon coding is used, error correction on the bitstream generated by the search program will produce an output signal set up in the same format as previously explained in connection with the encoding input file. Error correction can be performed in the following sequence (see the "Theory and Practice 10 of Error Control Codes" mentioned above).

1. Calculate syndromes 2. Calculate mislocation polynomial using Berlekamp-Massey algorithm 3. Calculate mislocation using Chien search 4. Calculate error sizes using Forney's algorithm.

The last step is performed only if a number of corrective errors have been detected in steps 2 and 3. The number of 20 errors detected is also calculated. If a non-correctable number of errors is detected, or if an error is found in the fill used (as explained above), a flag is inserted. The particular error coding procedure used in the present illustrative example 25 is referred to as ERRDEC.C 184 in FIG. 9. j

In controlling the package (by identifying its location in the carrier), the high priority message indicating the package's zip code can be used to activate suitable control arms or transport systems 30 for routing the package to the right truck, airplane or package carrier which carries the package to its destination.

Although the invention can be applied in a transport / diversion facility, it will be obvious that it can also be used in a wide range of information gathering, handling of parcels and the production of operations in which it is desirable reading a label on a package, a letter, a production part, a machine or similar devices and causing the system to perform packaging or production operations, e.g. with I 5 the object on which the label is placed. In accordance with the invention I, it is possible that these operations can take place with great speed, high accuracy, that a substantial amount of label information can be processed, and even that protection of a large part of information against I 10 is lost as due to damage to the label and similar phenomena.

Referring to FIG. 9 it can be seen that the decoded message can alternatively be displayed on a computer terminal, using the program TEXTOUT.C 185.

A program list for use in the practice of the present invention is found in EP-A-0 336 778.

Claims (17)

A method of decoding a stream of digital signals representing an electro-optic readable label image, corresponding to a plurality of polygons (20), of the same shape, the polygons (20) being located with the geometric centers of adjacent polygons on the angular tips of a predetermined two-dimensional arrangement, the polygons having, according to an encoding method, one of at least two mutually different optical properties, characterized by including the following steps: (a) performing a two-dimensional clock repetition on the sensed label image to obtain the a beat-evoked signal by placing two-dimensional blocks from the sensed image in corresponding locations in a trans-formed image in accordance with a non-linear two-dimensional edge enhancement transformation and processing of the transformed image to obtain a recalled rate signal, (b) use of the clock-triggering signal i step 20 (a) for identifying the geometric properties of the I polygons (20), I (c) decoding the polygons (20) by practicing the I opposite process of the encoding process. IN Method according to claim 1, characterized in that step (a) includes the following steps: I (I) performing a non-linear distribution operation I on the digital signals for identifying edge transitions I between the polygons (20), In (II) the practice of a Fourier transform on the non-linear distributed digital signals to obtain a two-dimensional representation corresponding to the direction, distance I and intensity of the optical properties of the transitions in the polygons (20), I (III) filtering with a space band pass filter of the trans-formed, non-linearly distributed digital signals to remove incorrect directions and distances in the optical properties of the transitions in the polygons (20), In (IV) performing an inverse Fourier transform I on the filtered, transformed non-linearly distributed digital signals to obtain the clock gene signal. The method of claim 1, characterized by further including a step before step (a) wherein the sensed label image is normalized to predetermined levels for each of the optical properties concerned. The method according to claim 1, characterized by further including a step before step (a), in which the image is rescaled to produce an image of equal horizontal and vertical magnification. The method of claim 2, characterized in that step (I) includes the generation of a two-dimensional distribution of the transitions between adjacent polygons having mutually different optical properties by calculating the standard deviation of the optical properties H of the image, which are detected for each pixel and for pixel 20 located close to each pixel by an electro-optic sensor (198), and larger standard deviation values correspond to transition areas at the polygon interfaces. The method of claim 5, characterized g-B mesh by further including a step of dividing into B-B threshold values of the sensed image in the center of each polygon (20) located in step (b). to determine the optical properties of the polygons (20) concerned. 7. A method according to claim 6, characterized in that the step of determining threshold values in the sensed label image is carried out by establishing histograms representing the optical H properties of the respective polygons (20). A method according to claim 1 or 2 or 7, H characterized in that step (b) includes: H 35 (I) performing an initiation step, wherein the two-dimensional H-rate clock gene signal provided in DK 175729 B1! step (a) in a predetermined region of the signal is sought to identify the location of the greatest intensity, (II) the pursuit of a step in a search continuation loop, in which the two-dimensional beat-recall signal 5 is searched over the entire beat-recall signal, beginning with the location of the greatest intensity. in step (I) and by looping to each adjacent location of the second highest intensity, each identified location corresponding to the center of the polygon. Method according to claim 8, characterized in that the image sensed by the electro-optic sensor includes a capture mark which is built up with a number of concentric rings (42, 44, 46, 48, 50. differing, alternately positioned optical properties, the first step of the process being to locate the intercept mark by filtering the digital signals and by correlating the digital signals with a predetermined frequency signal. Apparatus for decoding a stream of digital signals representing an electro-optically sensed label image in a plurality of polygons (20), with the geometric centers of adjacent polygons located at the angular tips of a predetermined, two-dimensional array, which polygons (20) in accordance with an encoding method, 25 is provided with one of at least two mutually different optical properties, characterized in that: (a) means for exerting a two-dimensional clock gene- In eliciting the sensed label image by placing j two two-dimensional blocks, from the emitted image in corresponding! in locations in a transformed image in accordance with a non-linear two-dimensional edge enhancement transformation and processing of the transformed image to obtain a recalled rate signal,! (B) means for using the recalled tactic! signal for locating the geometric centers of the polygons (20) and for identifying the optical properties of the polygons II, (c) means for decoding the polygons (20) by II exercising the reverse process of the encoding process. Apparatus according to claim 10, characterized in that the means for performing the clock gene induction include: II (a) means for performing a non-linear distribution II operation of the digital signals for the identification of I 10 transitions by the boundaries of the polygons (20), II (b) means for performing a Fourier transform I on the non-linearly distributed digital signals to provide a two-dimensional representation corresponding to the direction, distance, and intensity of the optical property transmitters; II 15 times in the polygons (20), I (c) means for filtering the transformed, I non-linearly distributed digital signals with a space pass filter II for removing incorrect directions and distances for the optical property transitions in the polygons (20), I 20 (d) means for effecting a reverse Fourier transform on the filtered, non-linearly transformed digital signals distributed to obtain the gene advance you called a clock signal, I Apparatus according to claim 11, characterized in that the non-linear distribution means include means I for producing a two-dimensional distribution of the transitions between adjacent polygons (20) having different characteristics when calculating the standard deviation of IH. the optical properties of the image recorded in I 30 each pixel and in pixels close to each pixel at the electro-optical sensor (198), with larger standard deviation values I corresponding to transition areas at the boundaries of the polygons (20). IN Apparatus according to claim 11, characterized by H 35 by including means for normalizing the filled-in label image to pre-determine the optimum for each of the optical properties of the image prior to performing the non-linear distribution operation. Apparatus according to claim 11, characterized by including means for re-scaling the sensed image to produce an image having the same horizontal and vertical magnification prior to the performance of the non-linear distribution function. j Apparatus according to claim 12, characterized by including means for generating threshold values 10 for the sensed label image in the center of each polygon (20) which has been located by the means in (b) for determining the optical properties of the polygons (20). ). Apparatus according to claim 15, characterized in that the means for generating threshold values further include means for constructing histograms I representing the optical properties of the polygons (20). IN Apparatus according to claim 10 or 11 or 16, ke η-I characterized in that the means used include: I 20 (I) triggering means for searching the two-dimensional re-evoked rate signal within a predetermined I region of the signal, for identifying the location of I the greatest intensity, I (II) a search continuation loop means for searching I 25 of the two-dimensional, re-evoked rate signal over the total re-evoked rate signal region, starting from the location provided with the greatest intensity in I (I) , and to extend the loop to each adjacent I position for the second largest intensity, each identified position corresponding to the center of a polygon (20). IN Apparatus according to claim 17, characterized in that the image sensed by the electro-optical sensor (198) includes a capture mark (35) consisting of a number of concentric rings (42, 44, 46, 48, 50) with In 35 different optical properties alternate with each other, as well as means for locating the capture mark (35) by I DK 175729 B1 I 84 In filtering the digital signals and by correlating the I digital signals with a signal of predetermined frequency. IN