CN101093553A - 2D code system, and identification method - Google Patents

Abstract

A 2-D code system is prepared for forming code character by arranging square unit blocks, dividing code character region to positioning region and positioning space region as well as data region, using four the same dark-color square positioning patterns to from said positioning region, applying coding method with unique identification-direction to generate data in said data region and using light-color unit block as positioning space region being used to isolate four square positioning patterns from said data region.

Description

A kind of 2 D code system and recognition methods thereof

Technical field

The present invention relates to information stores, location and recognition technology field, the extremely recognition methods of particularly a kind of comprehensive 2 D code system of distinguishing.

Background technology

Bar code is called for short bar code, and it is a kind of optical mask with special shape and specific arrangement mode.Common bar code can be divided into bar code and two-dimensional bar code at present.

Bar code (hereinafter to be referred as the one dimension sign indicating number) is all widespread uses and known by people in the whole world.The one dimension sign indicating number is only represented information in the horizontal direction, does not then represent any information at vertical direction.Since coming out the seventies in 20th century, the development of one dimension sign indicating number is very fast, has been widely used in every field such as logistic storage, the manufacturing, bibliogony thing, postal service, communications and transportation, production automation management.Code is wherein a kind of more famous one dimension sign indicating number for 39 yards.

Though bar code is widely used, its storage data volume is less relatively, mainly relies on the support of background data base, and bar code has only provided a kind of index to article.In addition, the one dimension sign indicating number is very poor to the support of encode Chinese characters for computer, and size is relatively large, and confidentiality is low.At these weak points of one dimension sign indicating number, press for a kind of more large storage capacity, higher confidentiality, stronger error correction, the bar code system that can be independent of the database support, have multiple coding even scalable coding mode.At these problems a kind of new bar code system---two-dimensional bar code that arisen at the historic moment.

Two-dimensional bar code is called for short two-dimension code, and it adopts the graphic element of different gray scales or different colours to represent specific data message, and graphic element is arranged according to matrix form.Because canned data simultaneously in the horizontal and vertical directions, so quantity of information is bigger.The lifting of information capacity provides the space for the application of error coding algorithm and cryptographic algorithm, and a lot of classical algorithms are widely applied in the two-dimension code, have improved the fault-tolerance and the robustness of system.

The principal feature of two-dimension code is:

(1) information capacity is big: two-dimension code has been increased to two dimension with the space dimensionality of information representation from one dimension on the basis of one dimension sign indicating number, thereby capacity increases greatly.For example, in the GB/T1828-2000 national standard, the QR sign indicating number of edition 40 can hold 1817 Chinese characters, and this information capacity is that bar code is inaccessible far away.

(2) strong error correcting capability: the one dimension sign indicating number is not considered error correction.Two-dimension code has been introduced error correcting code, makes two-dimension code when being partially damaged, and still can some or all of distinguish out information.The error correction algorithm of two-dimension code adopts general error coding algorithm, and for example, the QR sign indicating number in the GB/T1828-2000 national standard adopts the RS algorithm, and when error-correction level was H, error correction capability was 30%.

(3) coding range is wide: two-dimension code not only can be represented conventional arabic numeral, English alphabet, ASCll symbol, pass through extended coding, can also represent multiple character set such as Chinese character, Japanese character, and the scalable coding mode set makes it can adapt to following more character set encoding mode.

(4) introduce encryption mechanism: two-dimension code super large information capacity provides the performance space for numerous outstanding cryptographic algorithm, can prevent the forgery of various certificates, card etc. like this.

Two-dimension code has many different code systems, is broadly divided into following two types:

(1) stack two-dimension code: simply a plurality of one dimension sign indicating numbers are got up at vertical stack, and placed the border, the left and right sides that flagpole pattern is represented whole codeword respectively at the two ends of sign indicating number.Wherein the PDF417 with the U.S. is the most famous.

(2) matrix type two-dimension sign indicating number: the different permutation and combination of the square-shaped patterns unit by different gray scales or different colours in the rectangular area on planar graph are represented information.The location of code word often relies on the positioning pattern that has special shape in the code word figure.Typical matrix type two-dimension sign indicating number comprises: QR Code, Data Matrix, Visual Code, purple light sign indicating number or the like.

The QR sign indicating number is in exploitation in 1994 by the Nippondenso company of Japan.In upper left side, upper right side and the lower-left side of QR sign indicating number figure, there are three special complicated positioning pattern symbols, as shown in Figure 2.By special positioning pattern, system can locate the whole codeword graphics field rapidly, and differentiates the recognition direction of code word figure.The QR sign indicating number uses famous Reed-Solomon algorithm (being called for short the RS algorithm) to be used for mistake decoding, and it can carry out the Capacity Selection of three kinds of grades.The QR sign indicating number is a kind of two barcode technologies of technical grade, has obtained sufficient development and application in Japanese this country.

Fast development along with the personal radio communication business, the smart mobile phone of band camera is popularized day by day and is popular, use mobile phone cam to become the new bright spot that two-dimension code is used as the scanning device of two-dimension code, the just common mobile phone two-dimension code of saying, this is an application that has development space of two-dimension code.By the special recognition software of installing on the smart mobile phone, the user can carry out web page browsing, download easily and appreciate audio-video frequency content or the like.All pre-installed two dimension code reading software on the mobile phone of a lot of countries and regions, all being printed with on many publications and the plane media advertisement can be for the two-dimension code of mobile phone recognition software use.

But traditional 2 D code system is applied to mobile phone two-dimension code some shortcomings are arranged, QR sign indicating number for example, it is a kind of bar code system of technical grade.Large-sized QR sign indicating number needs the recognizing apparatus of specialty, for example laser scanner or the like.In the mobile phone application,, therefore can only use undersized bar code because the visual precision that mobile phone cam is taken is lower.Practical application shows, and is very poor greater than 30 * 30 the effect of QR sign indicating number on mobile phone.But the finite capacity of undersized QR sign indicating number, and do not provide size less than 20 * 20 version, its error correcting capability can not get demonstrating fully.In addition, the positioning pattern of QR sign indicating number is also complicated, and area is bigger, and this has just taken more space, has reduced information capacity.In other words, in the mobile phone application, advantages such as the capacity of QR sign indicating number and error correction can not get demonstrating fully.

With mobile phone two-dimension code as one independently branch research and develop, emerge large quantities of outstanding mobile phone two-dimension code systems.The Visual Code of the Michael Rohs invention of Deutsche Telekom AG has adopted the method for strip positioning pattern, and it is by a vertical long navigation bar and a short-range missile navigate bar as positioning pattern, as shown in Figure 4 mutually.

In number of patent application 200610086966.7, the purple light sign indicating number has been announced in " a kind of 2-D barcode system capable of rapid positioning based on positioning pattern " lining that the Beijing UNISITS Technology Co., Ltd announces.The quick location core of this yard is to have used and is tending towards 0 navigation bar target, eccentricity by eccentricity and is tending towards 1 navigation bar anchor, the positioning pattern that the navigation bar space is formed jointly, and can form more station-keeping mode and figure by navigation bar anchor, navigation bar target, navigation bar space, as shown in Figure 5.In fact this bar code has used for reference the notion of the navigation bar in the Visual Code 2 D code system, all belongs to the bar shaped locator meams.

The common feature of two-dimension codes such as Visual Code and purple light sign indicating number all is: adopt the flagpole pattern locator meams.This positioning pattern feature is more obvious, can improve location and recognition speed to a certain extent, but the strip positioning pattern has taken more barcode position, makes the bar code information capacity reduce; And more light interval region is arranged around the bar-shaped zone, these light interval regions have also taken the more information position, and large-area light areas very easily polluted and influenced the recognition result, even causes the erroneous judgement in whole bar code zone, even the location failure.

China is since 1991 become a full member of international EAN tissue, and by the end of 2004, existing two kinds of bar code code systems were by national standard, that is:

(1) the PDF417 sign indicating number of the U.S. (standard numbering: GB/T 17172-1997);

(2) Ri Ben QR sign indicating number (standard numbering: GB/T 18284-2000);

These two standards are not " mandatory " standard, belong to " recommendatory " standard.

Summary of the invention

The objective of the invention is a kind of 2 D code system, positioning pattern is combined with coded system, and realize location, raising capacity, comprehensive recognition fast.

The object of the present invention is achieved like this: a kind of 2 D code system and recognition methods thereof, and code word is rearranged by the square shaped cells piece, and the code word zone is divided into positioning area, location spacer region and data field; Positioning area is made up of four identical dark square positioning patterns, and the central point of four square positioning patterns is positioned at four vertex positions of square just; Data in the described data field generate by the coding method with unique recognition direction; Septal area, circle, described location is light cell block, and it separates four square positioning patterns and data separation.

The present invention sends out the recognition methods that another purpose provides aforementioned two-dimension code.

Another object of the present invention is achieved in that a kind of 2 D code system recognition methods, distinguishes according to the following steps:

1) use image acquisition equipment to gather picture intelligence, with the analog signal conversion of gathering and be output as digital image signal;

2) digital image signal that step 1 is obtained carries out the pre-service of gradation conversion, histogram equalization, denoising, sharpening, edge maintenance;

3) pretreated signal is carried out the binaryzation operation: the gray scale image that will obtain after will transforming according to the single threshold partitioning algorithm is converted into binary picture;

4) employing is carried out zone marker based on the iteration zone growth algorithm of four connected domains, obtains a plurality of connected domains, and extracts the edge of connected domain;

5) adopt Hough (Hough) transfer pair to obtain the connected region edge and carry out SHAPE DETECTION, the row filter of going forward side by side according to form factor, filters out a plurality of candidates' square positioning pattern;

6) finish screening in groups between the square positioning pattern: select 4 successively in a plurality of candidate points that from step 5, filter out, if the relative position of its central point satisfies the relative position relation of square vertices, these 4 points just can be in groups so, and it can determine an accurate code word zone; Otherwise selecting other 4 points more successively screens in groups; Like this, can from a plurality of candidate points, pick out the candidate point of a few composition groups;

7) carry out spatial sampling: according to wherein one group of 4 anchor point that obtain in the step 6, the specified data district according to the position of 4 anchor points, obtains M * N spatial sampling position, take out the value of the bianry image of M * N position correspondence in the bianry image, and be put in order among the data matrix D;

8) finish determining and obtaining information bit sequence of recognition direction: the data matrix D that obtains for step 7, distinguish from the upper and lower, left and right four direction, according to the code Design rule, there is and only have the recognition result of a direction can pass through the decoding checking algorithm, carry out data check successively, during by verification, this direction is exactly correct recognition direction, and can obtain bit stream information from this direction as a direction; If the verification of four direction, is then returned step 7 all by checking algorithm and is chosen next group and handle; If the data after the anchor point that all can be the in groups sampling are all by checking algorithm, so just decoding failure;

9) finish character set mappings: carry out further data-switching according to the bit stream that obtains in the step 8, be mapped as specific character set, thereby obtain original character data information.

The invention provides a kind of 2 D code system and recognition methods thereof.Code word is rearranged by the square shaped cells piece, and the code word graphics package on the painting canvas contains frontier district, positioning area, data field.The frontier district is that bar code is all around around the white space or the light areas of bar code graphics.Positioning area comprises four identical rectangle positioning patterns and corresponding location spacer region.The rectangle positioning pattern is dark square-shaped patterns, and square interior does not comprise any light figure, and the foursquare length of side is the odd-multiple of the cell block length of side.The location spacer region of light color is separating between rectangle positioning pattern and the data field.The figure of positioning area can not comprise the recognition directional information.Comprise several square shaped cells pieces in the data field, several cell blocks are distributed in the data field according to certain plane geometry rule.

Compare with existing 2 D code system, beneficial effect of the present invention is: positioning pattern is simple more, the code word capacity is bigger, is particularly suitable for the mobile phone two-dimension code application.

Existing 2 D code system based on the bar shaped location feature, for example Visual Code, purple light sign indicating number or the like, its positioning pattern is the strip figure, its structure is bigger, and interval region is bigger, and this just makes contaminated possibility very big, and has reduced the code word capacity.In addition, be similar in the QR code system of localization method of the present invention, the positioning pattern more complicated, three identical positioning patterns have been comprised in the code word, each positioning pattern can be regarded as by three overlapping concentric squares and form, and they are respectively 7 * 7 dark module, 5 * 5 light module and 3 * 3 dark module (referring to GB/T18284-2000).This positioning pattern is bigger than complexity and size, and the pattern space that takies is more, and is special in the small size version of mobile phone application need, information capacity very little (version less than 20 * 20 is not provided).In the QR of little version sign indicating number, advantages such as high capacity, strong error correction can not get demonstrating fully.

" ditetragon criterion " that the present invention proposes is the features location figure, and be simple in structure.When each positioning pattern adopts 1 cell block, compare the size minimum of positioning pattern with existing all 2 D code systems.When the small size version, the location efficiency of positioning pattern is very high, needing to be particularly suitable for the mobile phone two-dimension code of small size version to use.When the large scale version, can use 3 * 3 or 5 * 5 positioning pattern, to improve bearing accuracy.In addition, the positioning pattern that the present invention proposes is simple in structure, and the location spacer region is also very little, and the two contaminated possibility reduces greatly, thereby has strengthened the location anti-interference.In addition, the pattern space that positioning pattern and spacer region take still less can be more for the space of information representation.Under equal codeword size condition, can represent more information, thereby improve the codeword information capacity.With the purple light sign indicating number is example, and when not considering check information, the information capacity of code word with 10 * 10 versions of two orthogonal navigation bars is about 2 ⁶⁰, and the information capacity of 10 * 10 versions of the two-dimension code in this invention is 2 ⁸⁴, be about the former 1,600 ten thousand times.

It mainly is to rely on the asymmetry of the positioning pattern in the code word to determine that 2 D code system in the past solves the recognition direction.For example in the QR sign indicating number, three positioning patterns are positioned at wherein three of four summits of code word matrix, definite recognition direction that the arrangement mode on these three summits can be unique.Orthogonal navigation bar is positioned at the lower right corner of code word among the Visual Code, so the lower right corner can be considered as the code word origin, and long navigation bar direction can be considered as the y direction of principal axis, and short navigation bar can be considered as the x direction of principal axis.

2 D code system among the present invention has proposed a kind of method of definite recognition direction of novelty, its recognition direction dependence is separated code check and is determined, that is: codeword information is only when unique direction is distinguished, could pass through the checking of decoding checking algorithm, thus the correct recognition result who determines.Sign indicating number type design concept in according to the present invention, possible recognition direction has only 4, i.e. upper and lower, left and right.The develop rapidly of current electronic technology, the ability that hardware calculates improves greatly, in the short time (hundreds of millisecond) 4 directions is carried out verification calculating and has been easy to realize.

Description of drawings

Fig. 1 is according to two-dimension code code word synoptic diagram of the present invention;

Fig. 2 is the synoptic diagram of a preferred embodiment designed according to this invention;

Fig. 3 is existing QR sign indicating number synoptic diagram;

Fig. 4 is existing Visual Code sign indicating number synoptic diagram;

Fig. 5 is existing purple light sign indicating number synoptic diagram;

Fig. 6 is the program flow diagram of two-dimension code recognition methods of the present invention;

Fig. 7-Fig. 9 is the synoptic diagram that two-dimensional bar code of the present invention adopts different positioning patterns;

Shown in Figure 10 is the synoptic diagram of another embodiment of two-dimension code of the present invention.

Embodiment

Now reaching embodiment in conjunction with the accompanying drawings is described in further detail 2 D code system among the present invention and decoding process thereof.

Fig. 1 has described the numeral structure of two-dimension code.Among Fig. 1,1, the rectangle positioning pattern; 2, location spacer region; 3, data field; 4, frontier district.

Three parts have been comprised in the figure: frontier district, positioning area and data field.Wherein frontier district 4 is the light areas around the code word.Positioning area is made up of four square positioning patterns, location spacer region 2 is separated positioning area and data field 3, four square positioning patterns are positioned at the place, four summits of code word rectangle, and the line between the central point of four square positioning patterns constitutes a square just.

The codeword size of this two-dimension code can be any size.The area of square positioning pattern can be any size.

Cell block in the data field of this two-dimension code is by different color (white two kinds of colors generally darkly) expression information specific.

Square positioning pattern in the positioning area of this two-dimension code satisfies " ditetragon criterion ", and promptly each square positioning pattern is dark square, and the line of the central point of four square positioning patterns constitutes a bigger square just.

The data field of this two-dimension code can be arbitrary shape.The number of data field can be one or more.Can contain pattern area in the data field.

It is inner or outside that the positioning area of this two-dimension code can be positioned at the data field.

Consider in the digital picture after CCD sampling and the conversion, the marginal existence ambiguity of black unit piece, must cause the corrosion of black unit block edge after binaryzation, its extent of corrosion is according to factors such as ambient lighting, CCD susceptibility, binaryzation algorithms and difference.Suppose the extent of corrosion of black unit piece on the four direction of upper and lower, left and right identical (this hypothesis is near truth), the geometric center of the positioning pattern after the corrosion still is in the geometric center of former figure.

When the length of side of location segment is the odd-multiple of the cell block length of side just, the geometric center of positioning pattern is in the geometric center of unit segment just, even after the corrosion this geometric center also can be directly as the position of image space sampled point, also be the reference point of determining other sampled points, the error minimum of introducing.

On the contrary, if when being even-multiple, the geometric center of positioning pattern is positioned at the edge of unit segment, from spatial sampling point farthest, the error maximum of introducing when determining spatial sampling point is easily misread, particularly under the situation of image serious corrosion, be difficult to determine sampled point, very easily cause and misread.

This two-dimension code can be by being provided with the mask mode or the encoding filter mode reduces the quantity that occurs certain special pattern in the data field, and this special pattern is the inner dark square-shaped patterns that does not comprise light figure, and the adjacent position all is light segment all around.

This two-dimension code can be got rid of some code word figures by the mode of encoding filter, the figure of " ditetragon criterion " has appearred in their data field satisfying, be that four isolated zones are arranged in the data field is the dark square-shaped patterns that do not contain any light figure, and these four foursquare mid points are positioned at foursquare four vertex positions of another one just.

The encoding filter mode refers to: if the code word that obtains by codeword generator does not satisfy the filtration requirement, then this code word is passed into disuse as the forbidding code word.

In the YUV color space, Y is a gray scale, the brightness of expression color, and described light color is meant Y less than 50%, described dark color is meant that Y is more than or equal to 50%.

The positioning area of this two-dimension code can not comprise the recognition directional information, and determining of recognition direction can be by means of the coding of data message in the data field (as adopting CRC8 algorithm coding, the generator polynomial G of CRC8 (X)=X ⁸+ X ⁵+ X ⁴+ 1) decoding checking algorithm.That is: all separate code check in whole possible recognition directions, have and only have the recognition result of a direction can pass through checking algorithm, thereby find correct recognition direction, realize that omnirange is readable.

A plurality of single two-dimension codes can cascade form bigger two-dimension code, add cascaded message in the data field of these two-dimension codes, so that multi-group data is recombinated in order during decoding.

With reference to the process flow diagram of Fig. 6, two-dimension code shown in Figure 2 to be distinguished, its detailed process is as follows:

In step 1, at first carry out image acquisition 601.Gathering image need be with the image acquisition equipment that has charge-coupled device (CCD).Usually the CCD device is output as the voltage or the current signal of simulation, according to the actual conditions needs to these analog signal samplings and quantification, thereby be converted into digital image so that follow-up digital image processing algorithm is handled and is distinguished.At present, a lot of digital cameras or mobile phone cam have also appearred on the market, its output signal for adopted and quantize after colour or gray scale image.In the present embodiment, adopted mobile phone cam as collecting device.

In step 2, mainly be that the digital image that step 1 obtains is carried out pre-service 602, comprising operations such as gradation conversion, histogram equalization, denoising, sharpening, edge maintenances.These pre-service belong to the general-purpose algorithm of digital image processing, do not belong to content of the present invention.

In step 3, mainly carry out binaryzation operation 603, gray-scale map is converted to the black and white binary picture.At first, according to the single threshold partitioning algorithm gray scale image is converted into binary picture, that is: with the pre-service gray scale image that obtains in the step 2

$f(x,y)=\left\{\begin{array}{c}1,f(x,y)>T\\ 0,\mathrm{else}\end{array}\right.$

Wherein (x is that (x y) locates the gray-scale value of pixel to coordinate, and T is a segmentation threshold y) to f.This segmentation threshold can be tried to achieve by the iteration threshold algorithm based on grey level histogram, and this iterative algorithm formula is as follows:

$T(i+1)=\frac{1}{2}(\frac{\underset{k=0}{\overset{T\left(i\right)}{\mathrm{\Σ}}}{h}_k\·k}{\underset{k=0}{\overset{T\left(i\right)}{\mathrm{\Σ}}}{h}_k}+\frac{\underset{T\left(i\right)+1}{\overset{L-1}{\mathrm{\Σ}}}{h}_k\·k}{\underset{k=T\left(i\right)+1}{\overset{L-1}{\mathrm{\Σ}}}{h}_k})$

In the present embodiment, total L=256 gray shade scale of gray scale image.In the formula, T (i) is the threshold value that the i time iteration obtains, h _kFor gray-scale value is the pixel number of k.

The intermediate value of at first getting effective gray scale in the grey level histogram is used the iterative algorithm of following formula, till T (i+1) ≈ T (i) repeatedly as initial threshold T (0).The value of the T (i) of this moment just can be used as optimal segmenting threshold and carries out binary conversion treatment.

In step 4, the main employing based on the iteration zone growth algorithm of four connected domains carried out zone marker 604.If the present picture element position is that (x, y), then (x-1, y), (x+1, y), (x, y-1), (x y+1) is called its four connected domain.Its iterative algorithm is described below:

void?RegionTag(int?x，int?y)

int?visit_flag＝3；

If (f (x, y)==0) // 0 corresponding black color dots

{fx，y)＝visit_flag；

RegionTag(x，y+1)；

RegionTag(x+1，y)；

RegionTag(x-1，y)；

RegionTag(x，y-1)；}

In step 5, mainly adopt Hough (Hough) transfer pair connected region border to carry out the SHAPE DETECTION row filter 605 of going forward side by side.

After in step 4, obtaining connected region, obtain the border of connected region by the binary picture rim detection.

In the parameter space that Hough transformation obtains, can extract the length and the width value in zone, the ratio of the two is called form factor, and it is the parameter of a good inspected object shape.For example, foursquare form factor equals 1.

According to form factor, can filter out a plurality of candidates' square positioning pattern.

Step 6 is mainly finished the screening in groups 606 of positioning pattern.What is called refers in groups: the line of the central point of 4 candidate's positioning patterns constitutes a bigger square.

Select 4 successively in a plurality of candidate points that filter out from step 5, if the line of its central point can constitute a bigger square, these 4 points just can be in groups so, and it can determine an accurate code word zone.Otherwise selecting other 4 points more successively screens in groups.Through algorithm 6, can from a plurality of candidate points, pick out the candidate point of a few composition groups.

So-called accurate code word zone, referring to this zone, to belong to the possibility in real code word zone bigger, but its validity finally also needs to determine by the decoding checking.

Step 7 is mainly finished spatial sampling 607.According to 4 anchor points of a wherein composition group that obtain in the step 6, can the specified data district.According to the position of 4 anchor points, can obtain M * N spatial sampling position, take out the value of the bianry image of M * N position correspondence in the bianry image, and be put in order among the data matrix D.

In this example, according to prior imformation, code word is 10 * 10 size versions, and therefore 10 * 10 spatial sampling positions are arranged on bianry image.Suppose in 4 positioning patterns that the coordinate of upper left, upper right, lower-left, 4 points in bottom right is respectively (x _P1, y _P1), (x _P2, y _P2), (x _P3, y _P3), (x _P4, y _P4), then according to following formula can the specified data matrix in the capable n column element of m value corresponding to coordinate in the bianry image be [x (and m, n), y (m, n)] the value of pixel, 0≤m≤M-1 wherein, 0≤n≤N-1.

N＝10，M＝10

x ₁(n)＝(N-n-1)x _p1+nx _p2

y ₁(n)＝(N-n-1)y _p1+ny _p2

x ₂(n)＝(N-n-1)x _p3+nx _p4

y ₂(n)＝(N-n-1)y _p3+ny _p4

x ₃(m)＝(M-m-1)x _p1+mx _p3

y ₃(m)＝(M-m-1)y _p1+my _p3

x ₄(m)＝(M-m-1)x _p2+mx _p4

y ₄(m)＝(M-m-1)y _p2+my _p4

$\left[\begin{array}{c}A\\ B\\ C\\ D\end{array}\right]={\left[\begin{array}{cccc}{x}_1\left(n\right)& y_1\left(n\right)& 0& 0\\ x_2\left(n\right)& y_2\left(n\right)& 0& 0\\ 0& 0& x_3\left(m\right)& y_3\left(m\right)\\ 0& 0& x_4\left(m\right)& y_4\left(m\right)\end{array}\right]}^{-1}\left[\begin{array}{c}N-1\\ N-1\\ M-1\\ M-1\end{array}\right]$

$\left[\begin{array}{c}x(m,n)\\ y(m,n)\end{array}\right]={\left[\begin{array}{cc}A& B\\ C& D\end{array}\right]}^{-1}\left[\begin{array}{c}1\\ 1\end{array}\right]$

Step 8 is finished determining of recognition direction and is obtained information bit sequence 608.Data matrix D for step 7 obtains may distinguish from the upper and lower, left and right four direction, according to the code Design rule, has and only have the recognition result of a direction can pass through the decoding checking algorithm.Therefore, the data matrix is distinguished from the upper and lower, left and right four direction successively, and carried out data check successively, during by verification, this direction is exactly correct recognition direction, and can obtain bit stream information from this direction as a direction.If the verification of four direction is not all by checking algorithm, so just decoding is failed.

The element of tentation data matrix D is arrangement as shown below

$D=\left[\begin{array}{cccc}{d}_{00}& d_{01}& L& d_{0,N-1}\\ d_{10}& d_{11}& L& d_{1,N-1}\\ M& M& O& M\\ d_{M-\mathrm{1,0}}& d_{M-\mathrm{1,1}}& L& d_{M-1,N-1}\end{array}\right],$ For this example, N=10 is arranged, M=10, promptly

$D=\left[\begin{array}{cccc}{d}_{00}& d_{01}& L& d_{09}\\ d_{10}& d_{11}& L& d_{19}\\ M& M& O& M\\ d_{90}& d_{91}& L& d_{99}\end{array}\right],$

When reading from top to bottom, the bit sequence that obtains is

[d ₀₀，d ₀₁，d ₀₂，L，d ₀₈，d ₀₉，d ₁₀，d ₁₁，d ₁₂，L?d ₁₈，d ₁₉，L?d ₉₀，d ₉₁，d ₉₂，L?d ₉₈，d ₉₉]

When reading from left to right, the bit sequence that obtains is

[d ₉₀，d ₈₀，d ₇₀，L，d ₁₀，d ₀₀，d ₉₁，d ₈₁，d ₇₁，L，d ₁₁，d ₀₁，L?d ₉₉，d ₈₉，d ₇₉，L，d ₁₉，d ₀₉]

When reading from the bottom up, the bit sequence that obtains is

[d ₉₉，d ₉₈，d ₉₇，L，d ₉₁，d ₉₀，d ₈₉，d ₈₈，d ₈₇，L，d ₈₁，d ₈₀，Ld ₀₉，d ₀₈，d ₀₇，L，d ₀₁，d ₀₀]

Turn left when reading from the right side, the bit sequence that obtains is

[d ₀₉，d ₁₉，d ₂₉，L，d ₈₉，d ₉₉，d ₀₈，d ₁₈，d ₂₈，L，d ₈₈，d ₉₈，L?d ₀₀，d ₁₀，d ₂₀，L，d ₈₀，d ₉₀]

Then, successively these 4 sequences are separated code check, can be ordered sequence, can send into algorithm 9 and shine upon by the sequence of checking algorithm.The sequence that obtains when reading with " from top to bottom " is an example now, describes proof procedure in detail, and the processing of other direction sequences in like manner can get.

In 100 bit sequences, because the position of following 16 lucky corresponding positioning patterns of pixel does not comprise data message.Therefore, in bit sequence, these 16 pixels are removed.

[d ₀₀，d ₀₁，d ₁₀，d ₁₁，d ₀₉，d ₀₈，d ₁₉，d ₁₈，d ₉₀，d ₈₀，d ₉₁，d ₈₁，d ₈₈，d ₈₉，d ₉₈，d ₉₉]

The sequence of 84 remaining bits is data sequence, can be expressed as

[b ₀，b ₁，b ₂，...b ₈₃]

[b wherein ₀, b ₁, b ₂... b ₇₅] be data bit, [b ₇₆, b ₇₈... b ₈₃] be check bit.According to prior imformation, the CRC8 algorithm has been adopted in the digital coding in the present embodiment, when therefore decoding, can judge [b by the CRC8 checking algorithm ₀, b ₁, b ₂... b ₈₃] in whether mistake appears.

Concrete steps are: [b ₀, b ₁, b ₂... b ₇₅] back supplies 4 zero, and to form new length with check bit be 88 data sequence, that is:

[b ₀，b ₁，b ₂，...b ₇₅，0，0，0，0，b ₇₆，b ₇₈，...b ₈₃]

Corresponding 1 byte of per 8 bits is converted to 11 bytes with sequence.With 11 decoding verification algorithms of directly sending into CRC8; If the checking result is 0, expression is proved to be successful, and this sequence is an ordered sequence, otherwise authentication failed.

The generator polynomial of CRC8 is G (x)=x ⁸+ x ⁵+ x ⁴+ 1.

Step 9 is mainly finished character set mappings 609.Carry out further data-switching according to the bit stream that obtains in the step 8, be mapped as specific character set, thereby obtain original character data information, as arabic numeral, ASCII character or the like.

In the present embodiment, the known set of digits that is mapped as.With [b ₀, b ₁, b ₂... b ₇₅] sequence is 3 decimal numeral rules according to per 10 bits switch, can obtain final numerical information.

t _i＝b _i9×2 ⁹+b _i8×2 ⁸+b _i7×2 ⁷+L+b _i0×2 ⁰，t＝0，1，L?6

p ₇＝b ₇₃×2 ³+b ₇₂×2 ²+b ₇₁×2 ¹+b ₇₀×2 ⁰

P wherein ₇It is 1 decimal number.t _iBe 3 decimal numbers, as data t _i, need supply 0 in front at＜100 o'clock.

Final decoded data is [t ₀, t ₁, t ₂... t ₆, p ₇].

Above-described embodiment and implementation method are embodiments of the invention only, and the invention is not restricted to the foregoing description.For one of ordinary skill in the art,, all belong to the protection domain of design of the present invention and claims in any conspicuous change of under the prerequisite of the principle of the invention, it being made.

Claims (10)

1, a kind of 2 D code system, code word is rearranged by the square shaped cells piece, and the code word zone is divided into positioning area, location spacer region and data field; It is characterized in that: described positioning area is made up of four identical dark square positioning patterns, and the line of the central point of four square positioning patterns constitutes a bigger square just; Data in the described data field generate by the coding method with unique recognition direction; Septal area, circle, described location is light cell block, and it separates four square positioning patterns and data separation.

2, a kind of 2 D code system according to claim 1 is characterized in that: described coding checkout method adopts the CRC8 algorithm; The generator polynomial of CRC8 is G (X)=X ⁸+ X ⁵+ X ⁴+ 1.

3, a kind of 2 D code system according to claim 1 is characterized in that: the length of side of described each square setting figure is the odd-multiple of the cell block length of side.

4, according to claim 1 or 2 or 3 described a kind of 2 D code systems, it is characterized in that: described positioning area is positioned at the outside or inner of data field.

5, a kind of 2 D code system according to claim 4, it is characterized in that: when occurring certain special pattern in the described data field, can be by the mask mode being set or the encoding filter mode changes, this special pattern is four isolated dark squares, and the central point of these four figures is positioned at foursquare four vertex positions, and the adjacent position all around of these four figures all is light segment.

6, a kind of 2 D code system according to claim 4, it is characterized in that: the data field of described two-dimensional bar code comprises cascaded message, can cascade form bigger two-dimension code, during decoding multi-group data is recombinated in order.

7, a kind of recognition methods as power 1 described a kind of 2 D code system is characterized in that: distinguish according to the following steps:

1) use image acquisition equipment to gather picture intelligence, with the analog signal conversion of gathering and be output as digital image signal;

2) digital image signal that step 1 is obtained carries out the pre-service of gradation conversion, histogram equalization, denoising, sharpening, edge maintenance;

3) pretreated signal is carried out the binaryzation operation: the gray scale image that will obtain after will transforming according to the single threshold partitioning algorithm is converted into binary picture;

4) employing is carried out zone marker based on the iteration zone growth algorithm of four connected domains, obtains a plurality of connected domains, and extracts the edge of connected domain;

5) the connected region edge that adopts Hough (Hough) transfer pair to obtain carries out SHAPE DETECTION, and the row filter of going forward side by side according to form factor, filters out a plurality of candidates' square positioning pattern;

6) finish screening in groups between the square positioning pattern: select 4 successively in a plurality of candidate points that from step 5, filter out, if the relative position of its central point satisfies the relative position relation of square vertices, these 4 points just can be in groups so, and it can determine an accurate code word zone; Otherwise selecting other 4 points more successively screens in groups; Like this, can from a plurality of candidate points, pick out the candidate point of a few composition groups;

7) carry out spatial sampling: according to wherein one group of 4 anchor point that obtain in the step 6, the specified data district according to the position of 4 anchor points, obtains M * N spatial sampling position, take out the value of the bianry image of M * N position correspondence in the bianry image, and be put in order among the data matrix D;

8) finish determining and obtaining information bit sequence of recognition direction: the data matrix D that obtains for step 7, distinguish from the upper and lower, left and right four direction successively, according to the code Design rule, there is and only have the recognition result of a direction can pass through the decoding checking algorithm, carry out data check successively, during by verification, this direction is exactly correct recognition direction, and can obtain bit stream information from this direction as a direction; If the verification of four direction, is then returned step 7 all by checking algorithm and is chosen next group and handle; If the data after the anchor point that all can be the in groups sampling are all by checking algorithm, so just decoding failure;

9) finish character set mappings: carry out further data-switching according to the bit stream that obtains in the step 8, be mapped as specific character set, thereby obtain original character data information.

8, a kind of 2 D code system according to claim 7 and recognition methods thereof is characterized in that: described image acquisition equipment is mobile phone cam.

9, a kind of 2 D code system according to claim 7 and recognition methods thereof, it is characterized in that: to the data matrix D successively from upper and lower, left and right four recognition directions obtain bit sequence, and adopt the CRC-8 checking algorithm to separate code check one by one, when the checking algorithm success, promptly find correct recognition direction, realize that omnirange is readable.

10, a kind of 2 D code system according to claim 7 and recognition methods thereof is characterized in that: determine that according to 4 anchor points of upper left, upper right, the lower-left of 4 positioning patterns, position, bottom right M * N spatial sampling position adopt following formula to carry out:

x ₁(n)＝(N-n-1)x _p1+nx _p2

y ₁(n)＝(N-n-1)y _p1+ny _p2

x ₂(n)＝(N-n-1)x _p3+nx _p4

y ₂(n)＝(N-n-1)y _p3+ny _p4

x ₃(m)＝(M-m-1)x _p1+mx _p3

y ₃(m)＝(M-m-1)y _p1+my _p3

x ₄(m)＝(M-m-1)x _p2+mx _p4

y ₄(m)＝(M-m-1)y _p2+my _p4

$\left[\begin{array}{c}A\\ B\\ C\\ D\end{array}\right]={\left[\begin{array}{cccc}{x}_1\left(n\right)& y_1\left(n\right)& 0& 0\\ x_2\left(n\right)& y_2\left(n\right)& 0& 0\\ 0& 0& x_3\left(m\right)& y_3\left(m\right)\\ 0& 0& x_4\left(m\right)& y_4\left(m\right)\end{array}\right]}^{-1}\left[\begin{array}{c}N-1\\ N-1\\ M-1\\ M-1\end{array}\right]$

$\left[\begin{array}{c}X(m,n)\\ y(m,n)\end{array}\right]={\left[\begin{array}{cc}A& B\\ C& D\end{array}\right]}^{-1}\left[\begin{array}{c}1\\ 1\end{array}\right];$ 0≤m＜M，0≤n＜N

Wherein, (x _P1, y _P1), (x _P2, y _P2), (x _P3, y _P3), (x _P4, y _P4) be respectively the coordinate of 4 anchor points of upper left, upper right, lower-left, position, bottom right;

M, N are respectively the line number and the columns of data matrix.

Images