CN101388079B - Title block identification method - Google Patents
Title block identification method Download PDFInfo
- Publication number
- CN101388079B CN101388079B CN2008102167954A CN200810216795A CN101388079B CN 101388079 B CN101388079 B CN 101388079B CN 2008102167954 A CN2008102167954 A CN 2008102167954A CN 200810216795 A CN200810216795 A CN 200810216795A CN 101388079 B CN101388079 B CN 101388079B
- Authority
- CN
- China
- Prior art keywords
- cell
- lattice
- straight line
- template
- key element
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000000605 extraction Methods 0.000 claims description 9
- 230000009021 linear effect Effects 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 4
- 239000012634 fragment Substances 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 6
- 238000012545 processing Methods 0.000 abstract description 5
- 238000012163 sequencing technique Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 10
- 238000011960 computer-aided design Methods 0.000 description 7
- 238000013461 design Methods 0.000 description 7
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a recognition method for icons, the steps of the method comprise reading characters and straight lines in an electronic figure file, merging the characters and the straight lines and then sequencing, taking a next template, finding key words and key word cells in the electronic figure file according to key words set by the template, searching total cells in the icons according to the key word cells and an icon researching range set by the template, matching the total cells according to the key word cells defined by the template and the topological relation among total value cells, mixed cells and group cells, reading the information of the value cells, the mixed cells and the group cells, matching all templates like this, merging and processing the recognition result of each template, and removing repeated parts. The invention realizes accurate location for cells with different types in the icons, and avoids influence brought by the change of the icon formats on recognition.
Description
Technical field the present invention relates to a kind of title block identification method, relates in particular to a kind of based on the title block identification method on the CAD design drawing of cell topological relation.
Background technology is along with computer-aided design (CAD) (CAD) popularizes, and the effect of electronic drawings and archives performance in the production of design companies and management is also increasing, and the degree that comes into one's own improves day by day.Many owners require design companies that corresponding with it electronic drawings and archives is provided toward contact when receiving design drawing.Electronic drawings and archives is having the incomparable advantage of paper drawing aspect utilizing again.If design companies can effectively be managed the electronic drawings and archives of accumulation for many years, can in a large amount of historigram shelves, find figure shelves relevant and effectively utilization with current engineering design, just can save the time of redesign greatly, bring considerable benefit.But the electronic drawings and archives of magnanimity has also brought a managerial difficult problem.For these figure shelves effectively being classified effective informations such as the engineering in the necessary extraction figure shelves, specialty, deviser.
Key message in the figure shelves often is present among " title block " (Title Box).As shown in Figure 1, be the synoptic diagram of the title block in the drawing.Title block 10 is similar to a form, has write down the information of design companies, deviser's information, and job number, and the engineering information that the map title etc. are concrete has illustration to figure grade purposes.The figure shelves that nearly all formal printing is paid all have title block.Therefore, title block information has crucial effects for retrieval figure shelves.Extracting title block information at present is a repetition and loaded down with trivial details manual typing process, and in the face of the figure shelves of magnanimity, workload is very big, and occurs artificial mistake easily.
The research of extracting title block information has automatically had some achievements.These researchs all concentrate on by custom built forms and realize reading of title block information.Because the title block of same designing unit all has similar layout, so the mode of custom built forms can solve the problem that reads of considerable part title block.Yet these recognition methodss so far all are at relatively-stationary title block form, mainly discern by the feature literal in the title block, can't solve the title block variation issue that occurs in actual applications.These problems comprise:
● the variation of title block layout.Though title block has layout roughly, often there is not strict unified standard; Owing to the diversity of CAD plotting mode, the designer is also arbitrary easily to carry out some modifications to title block simultaneously.Such as, because literal is too many, the cell of depositing literal is elongated, widens or is shifted.The variation of this class layout can cause recognition failures.
● the repetition identification problem that similar title block causes.Many title blocks seem identical, in fact exist to change, and particularly prompt text-the key word of information changes, and makes that a template can't the multiple title block of simultaneous adaptation.If change template of formulation at each, duplicate the problem of identification again easily, promptly some figure can be applicable to a plurality of templates simultaneously.When particularly the title block form is many, be difficult between the different templates distinguish fully, this situation is very easy to take place.
● there are not the identification and the location of the cell of key word guide.Some cell does not have tangible key word nearby, but accurately identification again.
Summary of the invention the objective of the invention is to overcome the deficiencies in the prior art, proposes a kind of title block identification method, by the extraction to the title block general character, formulate the rule of a cover coupling, cell in the title block is accurately located, and the literal in the reading unit lattice extracts title block information.
A kind of title block identification method disclosed by the invention reads title block information in the electronic drawings and archives with the computing machine Automatic Program, comprises the steps:
The first step reads literal and straight line in the described electronic drawings and archives, more described literal of merger and straight line, ordering then;
In second step, take out a template;
In the 3rd step,, seek key word and key element lattice thereof in the described electronic drawings and archives according to the setting of described template; Repeat following steps A again, till can't in described electronic drawings and archives, finding new cell or hunting zone to exceed the scope of described template definition;
Steps A: with the key element lattice that found is the center, seeks the described key element lattice that found all cells on every side according to the title block hunting zone that described template is set in described electronic drawings and archives;
Write down described all cells topological relation each other that has found, and with described template in topological relation between the different key element lattice mate, if the match is successful then enter next step, otherwise got back to for second step, in second step, take out next template;
In the 4th step,, seek described value cell lattice, mixed cell lattice and group cell and extraction information wherein according to the key element lattice of described template definition and the topological relation between other value cell lattice, mixed cell lattice and the group cell;
The 5th step, if also have other templates in the described computer program, then got back to for second step, from described other templates, take out next template, otherwise merge the recognition result of handling each template, eliminate the part that wherein repeats.
The method of described literal of merger and straight line comprises in the described first step: extract the position and the content parameters of literal and straight line with Different CAD primitive types, all dissimilar literal and straight line respectively merger be the literal and the straight line of same type; Identical direction, join end to end or straight line that center section repeats is merged into a single straight line; The straight line adjustment of slight inclination in the error range is become the straight line that angle is 0 ° or 90 °.
The method that sorts in the described first step is that all straight lines and literal are sorted according to coordinate position respectively.
The present invention is in described the 3rd step, according to the setting of described template, seek a key word in the described electronic drawings and archives, seek its key element lattice according to described key word again, the method of seeking the key element lattice according to described key word comprises, position according to the key word place is determined 1 P arbitrarily, again by last, down, a left side, right, find all straight line L according to following rule: draw vertical line to a straight line from described some P, intersection point drops on promptly to be thought on the described straight line and obtains a described straight line L; Choose the straight line on the four direction more in turn, all sort from the near to the remote on each direction, if horizontal line is sought two straight lines that intersect on the vertical direction according to distance; If perpendicular line is sought two straight lines that intersect on the horizontal direction,, be the four edges of described key element lattice until two horizontal linears and two vertical line of obtaining surrounding this point and intersecting in twos.
Described steps A comprises the cell that is adjacent at certain direction finding according to the described key element lattice that found, the method of seeking the adjacent cell in the left side comprises the steps: to seek all straight lines of the left side bearing that intersects at the described key element lattice rectangle that has found and obtains intersection point respectively, described intersection point is separated into several line segments A to the left side bearing of the described key element lattice that found, find and intersect at each described line segment A respectively, on, two horizontal linear B of lower extreme point and C, seek the left side that is positioned at described line segment A again, and be parallel to described line segment A, also intersect at described horizontal linear B and C, the nearest straight line D of while distance A, described line segment A, straight line B and C, and the cell that straight line D forms is exactly the adjacent cell in the left side.
Seek the method for described value cell lattice information in described the 4th step, comprise the steps: basis and the key element lattice that described value cell lattice match, utilize the described topological relation of stipulating in the described template, find corresponding value cell lattice.If not with the key element lattice of described value cell lattice pairing, then in whole cells of described electronic drawings and archives, seek the immediate cell in position with the described value cell lattice of described template definition.
Seek the method for described mixed cell lattice information in described the 4th step, comprise the steps: that search earlier comprises the cell of mixed cell lattice key word, removes keyword fragment then in the content of cell.
Seek the method for described group of cell information in described the 4th step, comprise the steps: at first to search for the key element lattice that match with described group of cell, on the direction of regulation, search for and the described group of cell that the cell width is identical repeatedly then.
The present invention passes through rule definition flexibly, has improved the method for template matches, realizes the accurate location to cells dissimilar in the form, thus the influence of having avoided the title block format change that identification is brought.Set of rule can adapt to the close multiple title block of layout, thereby solves the problem that repeats to discern.The present invention is by to the extraction of title block general character, according to each category feature, and the topological relation between the grid particularly, to the cell in the title block, comprise the cell that does not have key word to guide, can both discern and locate, thus the title block content in extraction and the analysis chart shelves.
Description of drawings
Fig. 1 is the title block synoptic diagram in the drawing.
Fig. 2 is a cell synoptic diagram dissimilar in the title block.
Fig. 3 has topological relation with the key element lattice but non-conterminous cell synoptic diagram.
Fig. 4 is a plurality of cells whiles and a synoptic diagram that cell is adjacent.
Fig. 5 is the general flow chart of cell identifying of the present invention.
Fig. 6 is the present invention comprises this point according to some searchings a cell search procedure process flow diagram.
Fig. 7 be the present invention according to a cell, the cell process flow diagram that is adjacent at certain direction finding.
Embodiment is described in further details the present invention below in conjunction with the drawings and specific embodiments.
The present invention is the automatic processing procedure of a computing machine, is used for analyzing and the title block content of extraction figure shelves.This process is formulated the rule of a cover coupling by the extraction to the title block general character, and its core is according to each category feature, and particularly the topological relation between the grid is accurately located the cell in the title block (being a grid that can not segment again in the form).Find after the cell that needs, the literal in the reading unit lattice extracts the needed title block information that becomes.
This process need computing machine, described computing machine have a CPU (central processing unit), and this CPU (central processing unit) is connected to a memory device, a controller, an input equipment and an output device in operation.Use memory device to come straight line and the literal of using in the stores processor process, and the matched rule that is applicable to dissimilar electronic drawings and archives.CPU (central processing unit) is carried out electronic drawings and archives after the pre-service, obtains available matched rule from storer, attempts one by one, then matching result is screened and handles, and obtains final extraction result.
The present invention passes through rule definition flexibly, has improved the method for template matches, realizes the accurate location to cells dissimilar in the form, thus the influence of having avoided the title block format change that identification is brought.Set of rule can adapt to the close multiple title block of layout, thereby solves the problem that repeats to discern.
Cell in the title block form is divided into four classes, is cell synoptic diagram dissimilar in the title block as shown in Figure 2, is defined as follows:
(1) the key element lattice 20
Have only feature prompt text or pattern in the grid, do not have other content
(2) the value cell lattice 30
Have only the information content in the grid, do not have prompt text.
The value cell lattice can be divided into two classes again:
(a) the key element lattice of existence and its pairing
(b) there are not key element lattice with its pairing, as shown in Figure 3.
(3) the mixed cell lattice 40
Existing feature literal in the grid also has the information content.
(4) group cell 50
Be detailed cell, the corresponding a plurality of cells of feature prompt text, as shown in Figure 4.
The rule of title block form identification institute foundation comprises:
(1) be basic point with certain key element lattice, the approximate range of definition title block search.
(2) definition of keywords cell, and the topological relation between a plurality of key element lattice.Can judge according to this topological relation whether coupling is successful.If there is unique feature key word in the title block, then can only define key element lattice.It is a plurality of alternative that the key word of same thresholding can provide, as: figure number, drawing number, Drawing No. can be considered as same key word.
(3) for the value cell lattice, can select whether to define with it the key word of pairing, and with corresponding key element lattice between topological relation.
(4) for the mixed cell lattice, definition key word wherein
(5) for group cell, definition of keywords and the direction of search.
Be the general flow chart of cell identifying of the present invention as shown in Figure 5, be summarized as follows based on the identifying of cell topological relation:
(1) reads the content of literal in the electronic drawings and archives and straight line and accurate position.
(2) pre-service
A) all dissimilar literal and straight line respectively merger be the literal and the straight line of same type.Primitive types about straight line and literal among the CAD is a lot.As straight line simple straight line, polyline, ray etc. are just arranged.Need be decomposed into the simplest pel to the primitive types of complexity, reduce the difficulty of subsequent treatment.
B) identical direction, join end to end or straight line that center section repeats is merged into a single straight line.
C) the straight line adjustment of slight inclination in the error range is become the straight line that angle is 0 ° or 90 °.When judging straight-line intersection, the calculated amount of mutually perpendicular straight line is very little, can significantly improve speed.
D) all straight lines and literal are sorted according to coordinate position.
(3) take out a template.
(4) in electronic drawings and archives, find arbitrary key word of title block, determine to surround the cell of this key word.According to a bit, seek the algorithm of the cell that comprises this point and met journey A.
(5) be the center with these key element lattice,, determine its all cells on every side according to the approximate range that template is guided.With first key element lattice that obtain the location is starting point, the cell that distinguish up and down, the searching of left and right four direction is adjacent, and until finding new cell again, perhaps the hunting zone exceeds the scope of template definition.Write down these cells topological relation each other.
Wherein use an algorithm repeatedly: according to a cell, the cell that is adjacent at certain direction finding.Met journey B.
(6), can determine whether current title block mates according to the topological relation between the different key element lattice in the template.If only defined key element lattice, and this cell has found, thinks directly that then the match is successful.If not only having defined key element lattice, the relative position relation between the key element lattice of template definition, consistent with among the actual figure thinks that then the match is successful.
(7) other value cell lattice in the scope of the current template definition of search, mixed cell lattice, group cell.
A) searching method of value cell lattice:
If the key element lattice of definition pairing then in all cells, are sought and immediate one of the cell of template definition;
If defined the key element lattice of pairing, then adopt step (4), find this key element lattice, then according to topological relation, find corresponding value cell lattice.This method is specially adapted to the situation that cell is stretched and is out of shape.Some cell in title block is stretched under the situation of distortion, the position of real value cell lattice often with template in different, if but find the key element lattice of pairing earlier, look for according to topological relation again, then can accurately locate.Especially, if this topological relation is to have crossed over some cells, rather than direct neighbor, then can solve " identification and the location that do not have the cell of key word guide " problem.As the blank cell among Fig. 2, can be described as: with the right adjacent secondary series cell of " approval " key element lattice.Topological relation has been described the position relation between mutual related two cells, the interference of effectively having avoided the cell change of shape to bring.The cell adjacent with certain direction of a cell may more than one (Fig. 3) in the actual conditions, but this can come quantitative description with direction+sequence number.
B) searching method of mixed cell lattice
Search comprises the cell of mixed cell lattice key word, removes keyword fragment then in the content of cell, and remaining is exactly the information that needs identification.
C) searching method of group cell
The at first search and the key element lattice of its pairing, the method that adopts of applying step (4), the search cell identical on the direction of regulation repeatedly then with its width.What all searched all is suitable group cell.
(8) comprehensively all recognition result merges the information that identified region repeats
Process A: according to a bit, seek the cell that comprises this point, the search procedure process flow diagram as shown in Figure 6.
(1) by last, down, a left side, right, find all straight lines according to following rule: draw vertical line to this straight line from this point, intersection point drops on the straight line
(2) choose straight line on the four direction (on each direction all according to distance ordering from the near to the remote) in turn, judge in turn according to following rule: if horizontal line must guarantee that two straight lines on this line and the vertical direction intersect; If perpendicular line must guarantee that two straight lines on this line and the horizontal direction intersect
(3) obtain surrounding this point and crossing in twos two horizontal linears and two vertical line at last, form the minimum unit lattice that surround this point
Process B: according to a cell, seek the cell that be adjacent in certain direction (to be example) left, the process flow diagram of its search procedure as shown in Figure 7.
(1) all crossing straight lines of the left side bearing of searching and current cell rectangle are obtained intersection point
(2) according to intersection point the left side bearing of cell is separated into some line segments
(3) to wherein each line segment A, two horizontal linear B that find respectively that upper and lower end points with this line segment intersects and C
(4) left side at A is parallel with A and B, C all intersect, the nearest straight line D of distance A to seek one.The cell that A, B, C, D form is exactly one of them cell that needs searching.
Claims (9)
1. title block identification method reads title block information in the electronic drawings and archives with the computing machine Automatic Program, it is characterized in that, comprises the steps:
The first step reads literal and straight line in the described electronic drawings and archives, more described literal of merger and straight line, ordering then;
In second step, take out a template;
In the 3rd step,, seek key word and key element lattice thereof in the described electronic drawings and archives according to the setting of described template; Repeat following steps A again, till can't in described electronic drawings and archives, finding new cell or hunting zone to exceed the scope of described template definition;
Steps A: with the key element lattice that found is the center, seeks the described key element lattice that found all cells on every side according to the title block hunting zone that described template is set in described electronic drawings and archives;
Write down described all cells topological relation each other that has found, and with described template in topological relation between the different key element lattice mate, if the match is successful then enter next step, otherwise got back to for second step, in second step, take out next template;
In the 4th step,, seek described value cell lattice, mixed cell lattice and group cell and extraction information wherein according to the key element lattice of described template definition and the topological relation between other value cell lattice, mixed cell lattice and the group cell;
The 5th step, if also have other templates in the described computer program, then got back to for second step, from described other templates, take out next template, otherwise merge the recognition result of handling each template, eliminate the part that wherein repeats.
2. title block identification method according to claim 1, it is characterized in that, the method of described literal of merger and straight line comprises in the described first step: extract the position and the content parameters of literal and straight line with Different CAD primitive types, all dissimilar literal and straight line respectively merger be the literal and the straight line of same type; Identical direction, join end to end or straight line that center section repeats is merged into a single straight line; The straight line adjustment of slight inclination in the error range is become the straight line that angle is 0 ° or 90 °.
3. title block identification method according to claim 2 is characterized in that, the method that sorts in the described first step is that all straight lines and literal are sorted according to coordinate position respectively.
4. title block identification method according to claim 1, it is characterized in that, in described the 3rd step, setting according to described template, seek a key word in the described electronic drawings and archives, seek its key element lattice according to described key word again, the method of seeking the key element lattice according to described key word comprises, position according to the key word place is determined 1 P arbitrarily, again by last, down, a left side, right, find all straight line L according to following rule: draw vertical line to a straight line from described some P, intersection point drops on promptly to be thought on the described straight line and obtains a described straight line L; Choose the straight line on the four direction more in turn, all sort from the near to the remote on each direction, if horizontal line is sought two straight lines that intersect on the vertical direction according to distance; If perpendicular line is sought two straight lines that intersect on the horizontal direction,, be the four edges of described key element lattice until two horizontal linears and two vertical line of obtaining surrounding this point and intersecting in twos.
5. title block identification method according to claim 1, it is characterized in that, described steps A comprises: the cell that is adjacent at certain direction finding according to the described key element lattice that found, the method of seeking the adjacent cell in the left side comprises the steps: to seek all straight lines of the left side bearing that intersects at the described key element lattice rectangle that has found and obtains intersection point respectively, described intersection point is separated into several line segments A to the left side bearing of the described key element lattice that found, find intersect at respectively each described line segment A on, two horizontal linear B of lower extreme point and C, seek the left side that is positioned at described line segment A again, and be parallel to described line segment A, also intersect at described horizontal linear B and C, the nearest straight line D of while distance A, described line segment A, straight line B and C, and the cell that straight line D forms is exactly the adjacent cell in the left side.
6. title block identification method according to claim 1, it is characterized in that, seek the method for described value cell lattice information in described the 4th step, comprise the steps: the key element lattice that basis and described value cell lattice match, utilize the topological relation of stipulating in the described template, find corresponding value cell lattice.
7. title block identification method according to claim 1, it is characterized in that, seek the method for described mixed cell lattice information in described the 4th step, comprise the steps: that search earlier comprises the cell of mixed cell lattice key word, removes keyword fragment then in the content of cell.
8. title block identification method according to claim 1, it is characterized in that, seek the method for described group of cell information in described the 4th step, comprise the steps: at first to search for the key element lattice that match with described group of cell, on the direction of regulation, search for and the described group of cell that the cell width is identical repeatedly then.
9. title block identification method according to claim 6, it is characterized in that, if not with the key element lattice of described value cell lattice pairing, then in whole cells of described electronic drawings and archives, seek the immediate cell in position with the described value cell lattice of described template definition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008102167954A CN101388079B (en) | 2008-10-17 | 2008-10-17 | Title block identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008102167954A CN101388079B (en) | 2008-10-17 | 2008-10-17 | Title block identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101388079A CN101388079A (en) | 2009-03-18 |
| CN101388079B true CN101388079B (en) | 2011-06-08 |
Family
ID=40477490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008102167954A Active CN101388079B (en) | 2008-10-17 | 2008-10-17 | Title block identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101388079B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102592108B (en) * | 2011-01-04 | 2014-09-10 | 中国移动通信集团设计院有限公司 | Picture label information extracting method and device |
| CN107908843B (en) * | 2017-11-07 | 2022-05-27 | 杭州格原信息技术有限公司 | Management system based on drawing of map label information |
| CN109299675B (en) * | 2018-09-07 | 2021-05-07 | 广州算易软件科技有限公司 | Electrical drawing based distribution box identification method and device |
| CN109784236B (en) * | 2018-12-29 | 2023-09-19 | 河南辉煌科技股份有限公司 | Method for identifying table contents in railway drawing |
| CN112949596A (en) * | 2021-04-02 | 2021-06-11 | 上海肇观电子科技有限公司 | Identification method and broadcasting method of table pictures, electronic equipment and storage medium |
| CN113837068A (en) * | 2021-09-23 | 2021-12-24 | 纬衡浩建科技（深圳）有限公司 | PDF table character recognition method and device |
| JP7268115B1 (en) | 2021-11-09 | 2023-05-02 | 西松建設株式会社 | Rebar arrangement list reader, list reader, bar arrangement list reading method and program |
| US11752639B2 (en) | 2022-01-21 | 2023-09-12 | Saudi Arabian Oil Company | Engineering drawing review using robotic process automation |
-
2008
- 2008-10-17 CN CN2008102167954A patent/CN101388079B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN101388079A (en) | 2009-03-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101388079B (en) | Title block identification method | |
| CN112651373B (en) | Method and device for identifying text information of building drawing | |
| CN104517112A (en) | Table recognition method and system | |
| CN101826099B (en) | Method and system for identifying similar documents and determining document diffusance | |
| CN106156082A (en) | A kind of body alignment schemes and device | |
| CN101853299A (en) | Image searching result ordering method based on perceptual cognition | |
| CN102542067A (en) | Automatic image semantic annotation method based on scale learning and correlated label dissemination | |
| CN111611766A (en) | Method, apparatus and storage medium for determining circuit layout constraints | |
| CN104298700A (en) | Method for generating control-code by a control-code-diagram | |
| CN110362596A (en) | A kind of control method and device of text Extracting Information structural data processing | |
| CN109857957B (en) | Method for establishing label library, electronic equipment and computer storage medium | |
| CN106599064A (en) | Method for automatically classifying, obtaining and storing complex knowledge of high-end device | |
| CN117648093A (en) | RPA flow automatic generation method based on large model and self-customized demand template | |
| CN109871743B (en) | Text data positioning method and device, storage medium and terminal | |
| CN109062921B (en) | Method and system for extracting ship tray management information | |
| CN101673347B (en) | Spitting method of electronic drawing file | |
| Jabal et al. | A comparative study on extraction and recognition method of CAD data from CAD drawings | |
| CN106484788A (en) | Patent search system based on industry keyword | |
| CN102592108B (en) | Picture label information extracting method and device | |
| CN110377768B (en) | Intelligent graph recognition system and method | |
| CN103309888A (en) | Method and device for verifying data of electronic map | |
| Cheng et al. | Research on recognition method of interface elements based on machine learning | |
| KR100835307B1 (en) | Agent system and method for crawling schedule on the web | |
| Adefowoke Ojokoh et al. | Automated document metadata extraction | |
| CN113837068A (en) | PDF table character recognition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |