CN103093218A

CN103093218A - Automatically recognizing form type method and device

Info

Publication number: CN103093218A
Application number: CN2013100130250A
Authority: CN
Inventors: 余建桥; 郭加旋; 况远春; 王迎
Original assignee: Southwest University
Current assignee: Southwest University
Priority date: 2013-01-14
Filing date: 2013-01-14
Publication date: 2013-05-08
Anticipated expiration: 2033-01-14
Also published as: CN103093218B

Abstract

The invention provides an automatically recognizing form type method and a device, wherein the automatically recognizing form type method comprises the following steps: extracting features of images of to-be-recognized forms; respectively matching the features of the images of the to-be-recognized forms with the features of images of a form base, and enabling the type of the forms which is matched from the form feature base to be the type of the to-be-recognized forms, and wherein the automatically recognizing form type device comprises an extracting module and a recognition module which are used for conducting the aforesaid procedures. The automatically recognizing form type method and the device have the advantages of being high in recognition efficiency and recognition accuracy and the like.

Description

Automatically identify method and the device of form types

Technical field

The present invention relates to form recognition technology field, relate in particular to a kind of method and device of automatic identification form types.

Background technology

Form document refers to the class image take word and form as main contents, is mainly the file and picture that the paper list archives is transformed by equipment such as scanners.After scanning paper list in system by equipment such as scanners, normally according to the type of form to the form document storage of classifying, be mainly to be undertaken by the staff to the identification of form types at present, therefore mainly have the ineffective problem.

Summary of the invention

In view of this, the invention provides a kind of method and device of automatic identification form types.Can solve the low problem of recognition efficiency in existing form types identification.

The invention provides a kind of method of automatic identification form types, comprising:

The characteristics of image of step a, extraction form to be identified;

Step b, the characteristics of image of form in the characteristics of image of described form to be identified and table features storehouse is mated respectively, the type of the form that will match from described table features storehouse is as the type of described form to be identified.

Further, described characteristics of image comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME, SUMX represents axis number in form, SUMA, SUMB, SUMC and SUMD represent respectively the axis number in A, B, C and D four zones, SUME represents the axis number in the E of rectangular area, wherein A, B, C and D four zones are the wide and high midpoint at form, the zone that four areas that 2 row 2 that form is divided into are listed as equate, rectangular area E has identical center with form, and widely and high is the wide and high by 1/3rd of form.

Further, described step b comprises:

In step b1, the SUMX that judges described form to be identified and described table features storehouse, whether the absolute value of the difference of the SUMX of form is less than first threshold, if in described table features storehouse, the absolute value of the difference of the SUMX of the SUMX of a plurality of forms and described form to be identified is all less than first threshold, execution in step b2, if the absolute value of difference of SUMX of the SUMX of a form and described form to be identified is only arranged less than first threshold in described table features storehouse, with the type of this only form in the described table features storehouse type as described form to be identified;

step b2, judge the SUMA of described form to be identified, SUMB, the SUMA of form in SUMC and SUMD and described table features storehouse, SUMB, whether the absolute value of the difference of SUMC and SUMD is respectively less than Second Threshold, the 3rd threshold value, the 4th threshold value and the 5th threshold value, if the SUMA of a plurality of forms in described table features storehouse, SUMB, the SUMA of SUMC and SUMD and described form to be identified, SUMB, the absolute value of the difference of SUMC and SUMD is all less than the threshold value of correspondence, execution in step b3, if the SUMA of a form is only arranged in described table features storehouse, SUMB, the SUMA of SUMC and SUMD and described form to be identified, SUMB, the absolute value of the difference of SUMC and SUMD is all less than the threshold value of correspondence, with the type of this only form in the described table features storehouse type as described form to be identified,

step b3, judge that whether the absolute value of the difference of the SUME of form in the SUME of described form to be identified and described table features storehouse is less than the 6th threshold value, if in described table features storehouse, the absolute value of the difference of the SUME of the SUME of a plurality of forms and described form to be identified is all less than the 6th threshold value, in will these a plurality of forms with the type of the form of the absolute value minimum of the difference of the SUME of the described form to be identified type as described form to be identified, if the absolute value of difference of SUME of the SUME of a form and described form to be identified is only arranged less than the 6th threshold value in described table features storehouse, type that will this only form is as the type of described form to be identified.

Further, first threshold be described form to be identified SUMX 1/11st, Second Threshold be described form to be identified SUMA 1/7th, the 3rd threshold value be described form to be identified SUMB 1/7th, the 4th threshold value be described form to be identified SUMC 1/7th, the 5th threshold value be the 1/7th, the 6th threshold value of the SUMD of described form to be identified be described form to be identified SUME 1/5th.

Further, described step a comprises:

Step a1, to form to be identified cut apart successively, binaryzation and filtering processes;

Horizontal line section and vertical line segment in step a2, the form to be identified of extraction after step a1 processes;

The horizontal line section that extracts in step a3, combining step a2 obtains the form framework with vertical line segment;

Step a4, the form framework that step a3 is obtained carry out negate and thinning processing successively;

Characteristics of image in step a5, the form framework of extraction after step a5 processes.

Further, described step a2 comprises:

To the form to be identified after processing through step a1, first corrode in the horizontal direction with horizontal direction straight line line segment structural element, then once expand in vertical direction take the expansion texture element as template, the length value of described horizontal direction straight line line segment structural element be described form to be identified width 3/5ths, described expansion texture element is:

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}];

To the form to be identified after processing through step a1, first corrode in vertical direction with vertical direction straight line line segment structural element, then once expand in the horizontal direction take described expansion texture element as template, the length value of wherein said vertical direction straight line line segment structural element be described form to be identified the cell height 5/7ths.

Correspondingly, the present invention also provides a kind of recognition device, is used for the type of identification form automatically, comprising:

Extraction module is for the characteristics of image that extracts form to be identified;

Identification module is used for the characteristics of image of the characteristics of image of described form to be identified and table features storehouse form is mated respectively, and the type of the form that will match from described table features storehouse is as the type of described form to be identified.

further, described characteristics of image comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME, described SUMX represents axis number in form, described SUMA, SUMB, SUMC and SUMD represent respectively A, B, axis number in four zones of C and D, described SUME represents the axis number in the E of rectangular area, A wherein, B, four zones of C and D are the wide and high midpoint at form, the zone that four areas that 2 row 2 that form is divided into are listed as equate, rectangular area E has identical center with form, and widely and high be the wide and high by 1/3rd of form.

Further, described identification module comprises:

The first judging unit is used for judging that whether the absolute value of difference of SUMX of the SUMX of described form to be identified and described table features storehouse form is less than first threshold;

The first recognition unit, be used for only having to described table features storehouse when the first judgment unit judges the absolute value of difference of SUMX of the SUMX of a form and described form to be identified less than first threshold, with the type of this only form in the described table features storehouse type as described form to be identified;

The second judging unit, when being used for absolute value when the difference of the SUMX of the SUMX of a plurality of forms in determining of the first judging unit described table features storehouse and described form to be identified all less than first threshold, judge that whether the absolute value of the difference of SUMA, SUMB, SUMC and the SUMD of form in SUMA, SUMB, SUMC and the SUMD of described form to be identified and described table features storehouse is respectively less than Second Threshold, the 3rd threshold value, the 4th threshold value and the 5th threshold value;

The second recognition unit, SUMA, SUMB, SUMC and the SUMD that is used for only having to described table features storehouse when the second judgment unit judges a form with the absolute value of the difference of SUMA, SUMB, SUMC and the SUMD of form to be identified all less than corresponding threshold value, with the type of this only form in the described table features storehouse type as described form to be identified;

The 3rd judging unit, be used for when the second judgment unit judges to the absolute value of SUMA, SUMB, SUMC and the SUMD of a plurality of forms in described table features storehouse and the difference of SUMA, SUMB, SUMC and the SUMD of described form to be identified during all less than corresponding threshold value, judge that whether the absolute value of the difference of the SUME of form in the SUME of described form to be identified and described table features storehouse is less than the 6th threshold value;

the 3rd recognition unit, be used for when the 3rd judgment unit judges to the absolute value of the difference of the SUME of the SUME of a plurality of forms in described table features storehouse and described form to be identified all less than the 6th threshold value, in will these a plurality of forms with the type of the form of the absolute value minimum of the difference of the SUME of the described form to be identified type as described form to be identified, and be used for the absolute value of difference of SUME of the SUME of a form and form to be identified only being arranged less than the 6th threshold value when described table features storehouse, type that will this only form is as the type of described form to be identified.

Beneficial effect of the present invention:

By the table features storehouse is set, store the characteristics of image of various types of forms in this table features storehouse, then form to be identified is extracted its characteristics of image, by the characteristics of image of form to be identified and the characteristics of image of the various types of forms in feature database are compared, just can identify the type of form to be identified.This process can realize by computing machine fully automatically, has therefore that recognition efficiency is high, identification accuracy high.

Description of drawings

The invention will be further described below in conjunction with drawings and Examples:

Fig. 1 is the schematic flow sheet of embodiment of the method for automatic identification form types provided by the invention.

Fig. 2 is the schematic flow sheet of the embodiment of the step S11 in Fig. 1.

Fig. 3 is the schematic flow sheet of the embodiment of the step S12 in Fig. 1.

Fig. 4 is the structural representation of form to be identified.

Fig. 5 is the pretreated structural representation of Fig. 4.

Fig. 6 is the horizontal line section that extracts from Fig. 4.

Fig. 7 is the vertical line segment that extracts from Fig. 4.

Fig. 8 is the structural representation of the form that obtains after Fig. 6 and Fig. 7 merge.

Fig. 9 is the structural representation after Fig. 8 negate.

Figure 10 is the structural representation of the axis that obtains after Fig. 9 refinement.

Figure 11 is the structural representation of embodiment of the device of automatic identification form types provided by the invention.

Figure 12 is the structural representation of the embodiment of the extraction module in Figure 11.

Figure 13 is the structural representation of the embodiment of the identification module in Figure 11.

Embodiment

Please refer to Fig. 1, is the schematic flow sheet of embodiment of the method for automatic identification form types provided by the invention.It comprises:

The characteristics of image of step S11, extraction form to be identified.

In the characteristics of image of step S12, form to be identified that step S11 is extracted and table features storehouse, the characteristics of image of various types of forms mates respectively, and the type of the form that will match from the table features storehouse is as the type of form to be identified.

In the present embodiment, by setting in advance the table features storehouse, store the characteristics of image of various types of forms in this table features storehouse, then form to be identified is extracted its characteristics of image, by the characteristics of image of form to be identified and the characteristics of image of the various types of forms in feature database are compared, just can identify the type of form to be identified.This process can realize by computing machine fully automatically, has therefore that recognition efficiency is high, identification accuracy advantages of higher.

Further, before step S12, need to set up the table features storehouse, store the characteristics of image of various types of forms in this table features storehouse.When setting up the table features storehouse, can adopt the method shown in step S11 or artificially to carry out image characteristics extraction to known various types of forms, then with the characteristics of image that extracts and form types corresponding stored in database, complete the structure in table features storehouse.And the table features storehouse of setting up is a dynamic base, can increase wherein or reduce at any time the characteristics of image of form types and correspondence.

Further, the characteristics of image of form comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME.Below in conjunction with Fig. 2, illustrate implication and the leaching process of SUMX, SUMA, SUMB, SUMC, SUMD and SUME.

Please refer to Fig. 2, is the schematic flow sheet of the embodiment of step S11 in Fig. 1.It comprises:

Step S21, form to be identified is carried out pre-service.

Wherein, pre-service includes but not limited to: cut apart, binaryzation and filtering processes.Particularly, at first form to be identified is carried out dividing processing extraction form wherein, namely remove form word segment on every side, obtain pure tabular drawing picture.Then to pure tabular drawing is looked like to carry out binary conversion treatment, obtain binary image; Preferably, adopt the local binarization method that pure tabular drawing is looked like to process, the step of local binarization method mainly comprises: the threshold value of the first, calculating every bit:

M = \underset{- d < l < d}{\max_{- d < k < d}} g (x + k, y + l),

N = \underset{- d < l < d}{\min_{- d < k < d}} g (x + k, y + l),

T (x, y) = \{\begin{matrix} (M + N) / 2, M - N &GreaterEqual; S \\ T^{'}, M - N < S \end{matrix},

The second, pointwise binaryzation:

b (x, y) = \{\begin{matrix} 0, g (x, y) \leq T (x, y) \\ 1, g (x, y) > T (x, y) \end{matrix} .

Wherein, g (x, y) denotation coordination (x, y) gray-scale value of locating, the result of b (x, y) expression g (x, y) binaryzation, T (x, y) expression binary-state threshold, (2d+1) * (2d+1) for asking for the template window of threshold value, S, T' are a certain critical value, span is [0,128].At last, the binary image of gained is carried out filtering process, remove the noise in the tabular drawing picture, obtain the denoising image; Preferably, adopt " spiced salt " noise in median filter method removal tabular drawing picture, certainly also do not get rid of the modes such as maximal value filtering, mini-value filtering and revised Alpha's mean filter of employing and remove noise.As shown in Figure 4, be the schematic diagram of form to be identified, Fig. 5 is the schematic diagram through the pretreated form to be identified of step S21.

Horizontal line section and vertical line segment in step S22, the form to be identified of extraction after step S21 processes.

Wherein, the extraction of horizontal line section and vertical line segment mainly comprises with straight line line segment structural element and corrodes and expanded for two steps with the expansion texture element.Through repeatedly testing and verifying, the below introduces a kind of the have horizontal line section of better effects and the extracting mode of vertical line segment: for the extraction of horizontal line section, corrode in the horizontal direction with horizontal direction straight line line segment structural element, then once expand in vertical direction take the expansion texture element as template; Wherein the length value of horizontal direction straight line line segment structural element is a relative value, is not absolute value, and the image of different machines, different batches scanning may be different, thus length value be one than the ratio value of form length; Consider that the character in form might connect together, in order to extract horizontal line section in the process of corrosion, therefore through experiment when to draw the length value of choosing horizontal direction straight line line segment structural element be 3/5ths left and right of form width effect better; The expansion of horizontal line section being done vertical direction is because the situation of line segment fracture may occur when corroding, and get up for the segment link that will rupture this moment, just need to carry out expansion process, namely the line segment overstriking that extracts; Line segment overstriking one circle is got final product, so the selecting structure element is:

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}] .

For the extraction of vertical line segment, corrode in vertical direction with vertical direction straight line line segment structural element, then once expand in the horizontal direction take the expansion texture element as template.On vertical direction, extraction and the horizontal direction of form straight line line segment are similar, and be better when the length value that draws vertical line line segment structural element through experiment is 5/7ths left and right of table cell height; Line segment and horizontal direction on vertical direction are similar, choose the expansion texture element to be:

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}] .

Fig. 6 is respectively the horizontal line section of form extraction to be identified and the schematic diagram of vertical line segment with Fig. 7.

Horizontal line section and vertical line segment that step S23, combining step S22 extract obtain the form framework, as shown in Figure 8.

Step S24, the form framework that step S23 is obtained carry out negate and thinning processing successively.The structural representation of the form framework of negate and refinement is respectively as shown in Fig. 8 and 9.

SUMX, SUMA, SUMB, SUMC, SUMD and SUME in step S25, the form framework of extraction after step S24 processes.

Wherein, at first calculate axis number (that is: table cell number) SUMX in form framework after refinement.Then, the wide W of computation sheet and high H, in wide and high midpoint, tabular drawing is looked like to be divided into the equal zone of four areas of 2 row 2 row: A, B, C and D, and the number of the axis in calculating A, B, C and four locals of D is respectively: SUMA, SUMB, SUMC and SUMD.At last, choose a rectangular area E in form inside, this rectangular area E and form have identical center, and height and width be form height and width 1/3rd, and calculate the number SUME of axis in the E of this rectangular area.Obtain thus the characteristics of image F=(SUMX of form to be identified, SUMA, SUMB, SUMC, SUMD, SUME).

Method through Fig. 2 has successfully extracted characteristics of image from form to be identified, above-mentioned characteristics of image can reflect the design feature of form well, and the below introduces a kind of method based on above-mentioned characteristics of image identification form types.

Please refer to Fig. 3, is the schematic flow sheet of the embodiment of the step S12 in Fig. 1.It comprises:

In step S31, the SUMX that judges form to be identified and table features storehouse, whether the absolute value of the difference of the SUMX of form is less than first threshold.

Wherein, if in the table features storehouse, the absolute value of the difference of the SUMX of the SUMX of a plurality of forms and form to be identified is less than first threshold, execution in step S32.If the absolute value of difference of SUMX of the SUMX of a form and form to be identified is only arranged less than first threshold in the table features storehouse, execution in step S36: with the type of this only form in the table features storehouse type as form to be identified.If in the table features storehouse, the absolute value of the difference of the SUMX of the SUMX of all forms and form to be identified all is not less than the first threshold values, recognition failures, this moment, most possible situation was the form identical with the type of form to be identified not in the table features storehouse, can cross a step prompting this moment is identified the type of form to be identified by the user, then in the type input system with form to be identified, system deposits the characteristics of image correspondence of the type of form to be identified and form to be identified in the table features storehouse in, to enrich the table features storehouse.

Preferably, first threshold can for 1/11st of the SUMX of form to be identified, describe this step below in conjunction with an instantiation.

1, referring to table one, suppose to have stored in the table features storehouse characteristics of image of the form of three types.

Form types	SUMX	The remaining image feature
			Type one	30	Slightly
Type two	25	Slightly
			Type three	31	Slightly

2, when the SUMX of form to be identified is 22, first threshold is 2.Hence one can see that, and in the table features storehouse, the absolute value of the SUMX of the form of three types and 22 difference is all greater than 2, so None-identified goes out the type of form to be identified.

3, when the SUMX of form to be identified is 31, first threshold is 2.8.Hence one can see that, in the table features storehouse absolute value of the difference of the SUMX of the form of type one and type three all less than 2.8, this moment execution in step S32.

4, when the SUMX of form to be identified is 25, first threshold is 2.2.Hence one can see that, the absolute value of the SUMX of form of type two and 25 difference only arranged less than 2.2 in the table features storehouse, and therefore the type of form to be identified is type two.

In step S32, the SUMA that judges form to be identified, SUMB, SUMC and SUMD and table features storehouse, whether the absolute value of the difference of SUMA, SUMB, SUMC and the SUMD of form is respectively less than Second Threshold, the 3rd threshold value, the 4th threshold value and the 5th threshold value.

Wherein, if when in the table features storehouse, the absolute value of the difference of SUMA, SUMB, SUMC and the SUMD of SUMA, SUMB, SUMC and the SUMD of a plurality of forms and form to be identified is all less than each self-corresponding threshold value, execution in step S33.If the absolute value of difference of SUMA, SUMB, SUMC and SUMD that SUMA, SUMB, SUMC and the SUMD of a form and form to be identified only arranged in the table features storehouse is during all less than each self-corresponding threshold value, execution in step S36: with the type of this only form in the table features storehouse type as form to be identified.If in the table features storehouse, the absolute value of the difference of the SUMA of the SUMA of all forms, SUMB, SUMC and SUMD and form to be identified, SUMB, SUMC and SUMD all is not less than each self-corresponding threshold value, recognition failures.

Preferably, Second Threshold can be 1/7th of the SUMA of form to be identified, the 3rd threshold value can be 1/7th of the SUMB of form to be identified, the 4th threshold value can be 1/7th of the SUMC of form to be identified, the 5th threshold value can be 1/7th of the SUMD of form to be identified, below in conjunction with an instantiation, this step is described:

1, referring to table two, suppose to have stored in the table features storehouse characteristics of image of the form of three types.

Form types	SUMX	（SUMA,SUMB，SUMC,SUMD）	SUME
				Type one	Slightly	(14,14,13,13）	Slightly
Type two	Slightly	(14,13,14,13)	Slightly
				Type three	Slightly	(7,7,7,7)	Slightly

2, as (SUMA, SUMB, the SUMC of form to be identified, SUMD)=(14,14,14,14) time, the second to the 5th threshold value is 2, the SUMA of the form of type one and type two in the table features storehouse, SUMB, SUMC, the SUMA of SUMD and form to be identified, SUMB, SUMC, SUMD divide other poor absolute value all less than the threshold value of correspondence, so execution in step S33.

3, as (SUMA, SUMB, the SUMC of form to be identified, SUMD)=(7,7,7,7) time, the second to the 5th threshold value is 1, and the SUMA of the form of type three is only arranged in the table features storehouse, SUMB, SUMC, the SUMA of SUMD and form to be identified, SUMB, SUMC, SUMD divide other poor absolute value all less than the threshold value of correspondence, and therefore the type of form to be identified is type three.

4, as (SUMA, SUMB, the SUMC of form to be identified, SUMD)=(21,21,21,21) time, the second to the 5th threshold value is 3, in the table features storehouse without any the SUMA of the form of a type, SUMB, SUMC, the SUMA of SUMD and form to be identified, SUMB, SUMC, SUMD divide other poor absolute value all less than the threshold value of correspondence, so the type None-identified of this form to be known.

In step S33, the SUME that judges form to be identified and table features storehouse, whether the absolute value of the difference of the SUME of form is less than the 6th threshold value.

Wherein, if the absolute value of difference of SUME that has the SUME of a plurality of forms and form to be identified in the table features storehouse is during less than the 6th threshold value, execution in step S35: with the type of the form of the absolute value minimum of the difference of that exist and SUME form to be identified in the table features storehouse type as form to be identified; This step S35 comprises two kinds of situations, a kind of is that the absolute value of difference of SUME of the SUME of a plurality of forms in the table features storehouse and form to be identified is all less than the 6th threshold value, in will these a plurality of forms with the type of the form of the absolute value minimum of the difference of the SUME of the form to be identified type as form to be identified, another kind is the absolute value of difference of SUME of the SUME of a form and form to be identified only to be arranged less than the 6th threshold value in the table features storehouse, and type that will this only form is as the type of form to be identified.When if in the table features storehouse, the absolute value of the difference of the SUME of the SUME of any form and form to be identified all is not less than the 6th threshold value, recognition failures.

Preferably, the 6th threshold value can for 1/5th of the SUME of form to be identified, describe this step below in conjunction with an instantiation.

1, referring to table three, suppose to have stored in the table features storehouse characteristics of image of the form of three types.

Form types	SUME	The remaining image feature
			Type one	16	Slightly
Type two	13	Slightly
			Type three	20	Slightly

2, when the SUME of form to be identified was 15, the 6th threshold value was 3.Hence one can see that, and in the table features storehouse, the SUME of the form of type one and type two and 15 difference are all less than 3, but the absolute value of the difference of the SUME of the form of type one and 15 is minimum, is only 1, and therefore the type of form to be identified is type one.

3, when the SUME of form to be identified was 22, the 6th threshold value was 4.4.Hence one can see that, the exhausted value value of the SUME of type three forms and 22 difference only arranged less than 4.4, and therefore the type of form to be identified is type three.

4, when the SUME in form to be identified is 10, the 6th threshold value is 2.Hence one can see that as can be known, in the table features storehouse absolute value of the SUME of the form of three types and 10 difference all greater than 2, so recognition failures.

In the present embodiment, by above-mentioned matching way, as long as the table features storehouse is enough comprehensive, just can identify rapidly, exactly the classification of form to be identified.

The below introduces the embodiment of device of the present invention.

Please refer to Figure 11, is the structural representation of the embodiment of recognition device provided by the invention.This recognition device can be identified the type of form automatically, and it comprises:

Extraction module 1 is for the characteristics of image that extracts form to be identified.

Identification module 2, the characteristics of image that is used for the various types of forms of characteristics of image and table features storehouse of form to be identified that extraction module 1 is extracted mates respectively, and the type of the form that will match from the table features storehouse is as the type of form to be identified.

In the present embodiment, by set in advance the table features storehouse in recognition device, store the characteristics of image of various types of forms in this table features storehouse, then extract its characteristics of image by 1 pair of form to be identified of extraction module, and by identification module 2, the characteristics of image of form to be identified and the characteristics of image of the various types of forms in feature database are compared, just can identify the type of form to be identified.This process can realize fully automatically, has therefore that recognition efficiency is high, identification accuracy advantages of higher.

Further, need to set up the table features storehouse in recognition device, store the characteristics of image of various types of forms in this table features storehouse.When setting up the table features storehouse, can adopt the mode shown in extraction module 1 or artificially to carry out image characteristics extraction to known various types of forms, then the characteristics of image that extracts and form types corresponding stored in the database in the recognition device, are completed the structure in table features storehouse.And the table features storehouse of setting up is a dynamic base, can increase wherein or reduce at any time the characteristics of image of form types and correspondence.

Further, the characteristics of image of form comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME.Below in conjunction with Figure 12, illustrate implication and the leaching process of SUMX, SUMA, SUMB, SUMC, SUMD and SUME.

Please refer to Figure 12, is the structural representation of the embodiment of extraction module 1 in Figure 11.It comprises:

Pretreatment unit 11 is used for form to be identified is carried out pre-service.

M = \underset{- d < l < d}{\max_{- d < k < d}} g (x + k, y + l),

N = \underset{- d < l < d}{\min_{- d < k < d}} g (x + k, y + l),

T (x, y) = \{\begin{matrix} (M + N) / 2, M - N &GreaterEqual; S \\ T^{'}, M - N < S \end{matrix},

The second, pointwise binaryzation:

b (x, y) = \{\begin{matrix} 0, g (x, y) \leq T (x, y) \\ 1, g (x, y) > T (x, y) \end{matrix};

Wherein, g (x, y) denotation coordination (x, y) gray-scale value of locating, the result of b (x, y) expression g (x, y) binaryzation, T (x, y) expression binary-state threshold, (2d+1) * (2d+1) for asking for the template window of threshold value, S, T' are a certain critical value, span is [0,128].At last, the binary image of gained is carried out filtering process, remove the noise in the tabular drawing picture, obtain the denoising image; Preferably, adopt " spiced salt " noise in median filter method removal tabular drawing picture, certainly also do not get rid of the modes such as maximal value filtering, mini-value filtering and revised Alpha's mean filter of employing and remove noise.As shown in Figure 4, be the schematic diagram of form to be identified, Fig. 5 is the schematic diagram through pretreatment unit 11 pretreated forms to be identified.

Line segments extraction unit 12 is for the horizontal line section and vertical line segment that extract the form to be identified after pretreatment unit 11 is processed.

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}] .

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}] .

Line segment merge cells 13, horizontal line section and vertical line segment for merging line segments extraction unit 12 extractions obtain the form framework, as shown in Figure 8.

Negate and thinning processing unit 14, the form framework that is used for line segment merge cells 13 is obtained carries out negate and thinning processing successively.The structural representation of the form framework of negate and refinement is respectively as shown in Fig. 8 and 9.

Feature extraction unit 15 is for SUMX, the SUMA, SUMB, SUMC, SUMD and the SUME that extract the form framework after negate and refinement unit 14 processing.

Wherein, at first feature extraction unit 15 calculates axis number (that is: the table cell number) SUMX in form framework after refinement.Then, the wide W of computation sheet and high H, in wide and high midpoint, image is divided into the equal zone of four areas of 2 row 2 row: A, B, C and D, and the number of the axis in calculating A, B, C and four locals of D is respectively: SUMA, SUMB, SUMC and SUMD.At last, choose a rectangular area E in form inside, this rectangular area E and form have identical center, and height and width be form height and width 1/3rd, and calculate the number SUME of axis in the E of this rectangular area.Obtain thus the characteristics of image F=(SUMX of form to be identified, SUMA, SUMB, SUMC, SUMD, SUME).

Successfully extracted characteristics of image from form to be identified through extraction module 1, above-mentioned characteristics of image can reflect the design feature of form well, and the below introduces a kind of mode based on above-mentioned characteristics of image identification form types.

Please refer to Figure 13, is the structural representation of the embodiment of the identification module 2 in Figure 11.It comprises:

The first judging unit 21 is used for judging that whether the absolute value of difference of SUMX of the SUMX of form to be identified and table features storehouse form is less than first threshold.

The first recognition unit 24, be used for determining the absolute value of difference of SUMX that only there is the SUMX of a form and form to be identified in the table features storehouse less than first threshold when the first judging unit 21, with the type of this only form in the table features storehouse type as form to be identified.The absolute value of difference that determines the SUMX of the SUMX of all forms in the table features storehouse and form to be identified when the first judging unit 31 all is not less than the first threshold values, the prompting recognition failures, this moment, most possible situation was the form identical with the type of form to be identified not in the table features storehouse, can cross a step prompting this moment is identified the type of form to be identified by the user, then in the type input system with form to be identified, system deposits the characteristics of image correspondence of the type of form to be identified and form to be identified in the table features storehouse in, to enrich the table features storehouse.

The second judging unit 22, the absolute value of difference of SUMX that is used for determining the SUMX of a plurality of forms in table features storehouse and form to be identified when the first judging unit 21 is during less than first threshold, and continuation judges that whether the absolute value of the difference of SUMA, SUMB, SUMC and the SUMD of form in SUMA, SUMB, SUMC and the SUMD of form to be identified and table features storehouse is respectively less than Second Threshold, the 3rd threshold value, the 4th threshold value and the 5th threshold value.

The second recognition unit 25, be used for when the second judging unit 22 determines the absolute value of difference of SUMA, SUMB, SUMC and SUMD that only there is SUMA, SUMB, SUMC and the SUMD of a form and a form to be identified in the table features storehouse all less than each self-corresponding threshold value, with the type of this only form in the table features storehouse type as form to be identified.The absolute value of difference that determines SUMA, SUMB, SUMC and the SUMD of SUMA, SUMB, SUMC and the SUMD of all forms in the table features storehouse and form to be identified when the second judging unit 22 all is not less than each self-corresponding threshold value, the prompting recognition failures.

The 3rd judging unit 23, the absolute value of difference of SUMA, SUMB, SUMC and SUMD that is used for determining SUMA, SUMB, SUMC and the SUMD of a plurality of forms in table features storehouse and form to be identified when the second judging unit 22 is during all less than each self-corresponding threshold value, and continuation judges that whether the absolute value of the difference of the SUME of form in the SUME of form to be identified and table features storehouse is less than the 6th threshold value.

The 3rd recognition unit 26, be used for when the 3rd judging unit 23 determines the absolute value of difference of SUME that the table features storehouse exists the SUME of a plurality of forms and form to be identified less than the 6th threshold value, with the type of the form of the absolute value minimum of the difference of that exist and SUME form to be identified in the table features storehouse type as form to be identified.Comprise two kinds of situations this moment, a kind of is that the absolute value of difference of SUME of the SUME of a plurality of forms in the table features storehouse and form to be identified is all less than the 6th threshold value, in will these a plurality of forms with the type of the form of the absolute value minimum of the difference of the SUME of the form to be identified type as form to be identified, another kind is the absolute value of difference of SUME of the SUME of a form and form to be identified only to be arranged less than the 6th threshold value in the table features storehouse, and type that will this only form is as the type of form to be identified.When the absolute value of difference that determines the SUME of the SUME of any form in the table features storehouse and form to be identified when the 3rd judging unit 23 all is not less than the 6th threshold value, point out recognition failures.

Preferably, first threshold be form to be identified SUMX 1/11st, Second Threshold be form to be identified SUMA 1/7th, the 3rd threshold value be form to be identified SUMB 1/7th, the 4th threshold value be form to be identified SUMC 1/7th, the 5th threshold value be the 1/7th, the 6th threshold value of the SUMD of form to be identified be form to be identified SUME 1/5th, to obtain optimum recognition effect.

Explanation is at last, above embodiment is only unrestricted in order to technical scheme of the present invention to be described, although with reference to preferred embodiment, the present invention is had been described in detail, those of ordinary skill in the art is to be understood that, can modify or be equal to replacement technical scheme of the present invention, and not breaking away from aim and the scope of technical solution of the present invention, it all should be encompassed in the middle of claim scope of the present invention.

Claims

1. automatic method of identification form types is characterized in that: comprising:

The characteristics of image of step a, extraction form to be identified;

2. the method for automatic identification form types as claimed in claim 1, it is characterized in that: described characteristics of image comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME, SUMX represents axis number in form, SUMA, SUMB, SUMC and SUMD represent respectively A, B, axis number in four zones of C and D, SUME represents the axis number in the E of rectangular area, A wherein, B, four zones of C and D are the wide and high midpoint at form, the zone that four areas that 2 row 2 that form is divided into are listed as equate, rectangular area E has identical center with form, and widely and high be the wide and high by 1/3rd of form.

3. the method for automatic identification form types as claimed in claim 2, it is characterized in that: described step b comprises:

4. the method for automatic identification form types as claimed in claim 3, it is characterized in that: first threshold be described form to be identified SUMX 1/11st, Second Threshold be described form to be identified SUMA 1/7th, the 3rd threshold value be described form to be identified SUMB 1/7th, the 4th threshold value be described form to be identified SUMC 1/7th, the 5th threshold value be the 1/7th, the 6th threshold value of the SUMD of described form to be identified be described form to be identified SUME 1/5th.

5. as the method for the described automatic identification form types of any one in claim 1-4, it is characterized in that: described step a comprises:

6. the method for automatic identification form types as claimed in claim 5, it is characterized in that: described step a2 comprises:

[\begin{matrix} 1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{matrix}];

7. a recognition device, be used for the type of identification form automatically, it is characterized in that: comprising:

8. the device of automatic identification form types as claimed in claim 7, it is characterized in that: described characteristics of image comprises: SUMX, SUMA, SUMB, SUMC, SUMD and SUME, described SUMX represents axis number in form, described SUMA, SUMB, SUMC and SUMD represent respectively A, B, axis number in four zones of C and D, described SUME represents the axis number in the E of rectangular area, A wherein, B, four zones of C and D are the wide and high midpoint at form, the zone that four areas that 2 row 2 that form is divided into are listed as equate, rectangular area E has identical center with form, and widely and high be the wide and high by 1/3rd of form.

9. the device of automatic identification form types as claimed in claim 8, it is characterized in that: described identification module comprises:

10. the device of automatic identification form types as claimed in claim 9, it is characterized in that: first threshold be described form to be identified SUMX 1/11st, Second Threshold be described form to be identified SUMA 1/7th, the 3rd threshold value be described form to be identified SUMB 1/7th, the 4th threshold value be described form to be identified SUMC 1/7th, the 5th threshold value be the 1/7th, the 6th threshold value of the SUMD of described form to be identified be described form to be identified SUME 1/5th.