JPS6037082A - Character cutting system - Google Patents

Character cutting system

Info

Publication number
JPS6037082A
JPS6037082A JP58143427A JP14342783A JPS6037082A JP S6037082 A JPS6037082 A JP S6037082A JP 58143427 A JP58143427 A JP 58143427A JP 14342783 A JP14342783 A JP 14342783A JP S6037082 A JPS6037082 A JP S6037082A
Authority
JP
Japan
Prior art keywords
character
projection
image
line
cutting
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.)
Pending
Application number
JP58143427A
Other languages
Japanese (ja)
Inventor
Mamoru Maeda
護 前田
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ricoh Co Ltd
Original Assignee
Ricoh Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ricoh Co Ltd filed Critical Ricoh Co Ltd
Priority to JP58143427A priority Critical patent/JPS6037082A/en
Publication of JPS6037082A publication Critical patent/JPS6037082A/en
Pending legal-status Critical Current

Links

Landscapes

  • Character Input (AREA)

Abstract

PURPOSE:To cut exactly a character in the characters underlined extending over two characters deriving a projection which subtracted a value corresponding to an underline from a vertical projection of a character line image, and detecting a character cutting point by its projection. CONSTITUTION:As for an image data inputted from a scanner, noise is eliminated by a noise eliminating and line cutting part, and thereafter, a line is cut by a horizontal projecting method. Subsequently, the image data of a cut one character line is stored in an image memory part 11. On the other hand, a vertical projection detecting part 12 derives the image vertical projection of one character line, and inputs it to a character cutting part 13. At this time, a mask part 14 eliminates an underlined part from a character image, detects a character cutting point, and sends it out to a recognizing part.

Description

【発明の詳細な説明】 〔技術分野〕 本発明は、アンダーラインを含む文字行のイメージから
、個々の文字イメージを切り出す文字切出方式に関する
DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to a character cutting method for cutting out individual character images from an image of a character line including an underline.

〔従来技術〕[Prior art]

従来の文字切出方式は、文字行イメージの垂直射影の切
れ目を切出し点として1個々の文字イメージを切り出す
というものである。このような方式では、2文字以」二
に跨がる長いアンダーラインが(=Jされた文字列の場
合、各文字の垂直射影が連結さJしてしまうため、文字
イメージを切り出すことができないという欠点がある。
In the conventional character cutting method, each character image is cut out using a break in the vertical projection of a character line image as a cutting point. With this method, if a long underline spanning two or more characters is a character string, the vertical projections of each character are concatenated, making it impossible to extract the character image. There is a drawback.

〔目 的〕〔the purpose〕

本発明の目的は、上述のような欠点を除去した文字切出
方式を提供することにある。
An object of the present invention is to provide a character extraction method that eliminates the above-mentioned drawbacks.

〔実施例〕〔Example〕

まず第1図によって本発明の詳細な説明する。 First, the present invention will be explained in detail with reference to FIG.

図に示ずように長いアンダーラインlがイリされた文字
行2の、イメージに対しては、垂直射影3を4くめる。
As shown in the figure, for the image of the character line 2 with a long underline l, the vertical projection 3 is set by 4.

この凪直躬影3はアンダーライン1の影響で切、tt 
1」がなくなりでいるから、従来の方式(:’ 11文
字’ A J r B J r CJを切り出せない。
This Nagi Naomikage 3 is cut off due to the influence of Underline 1, tt
1" is missing, so the conventional method (:'11 characters' A J r B J r CJ cannot be extracted.

そこ゛(本発明は、垂直射影3からアンダーラインに相
当する値を差し引いた第2射影4をめ、この第2射影4
にしたがって切出し点を決め、各文字のイメージを切り
出す。
Therefore, the present invention obtains a second projection 4 obtained by subtracting a value corresponding to the underline from the vertical projection 3, and this second projection 4
Determine the cutting point according to the following and cut out the image of each character.

次に本発明の−・実施例を説明する。Next, embodiments of the present invention will be described.

第2図は本発明に係る文字切出装置のブロック図である
。図示されていないスキャナーから人力されるイメージ
データはノイズ除去・行切出部10によりノイズ除去の
後、公知の水平射影法により行切出しが行われる。切出
しされた1文字行(アンダーライン部も名む)のイメー
ジデータはイメージメモリ部11に記憶される1、垂直
射影検出部12は1文字fTのイメージの垂直列・1影
をぷめ、それを文字切出部13に人力する。14は文字
イメージがりアンダーライン部分を除去(マスク)Vろ
7ノ、ン部である。マスク部1・1から出力される叉′
5−イノーシデータは認識部(図示さJbt−いない)
八人力さJ[る、 第3目矧は文字切出部13の処理を示すフローザ!−1
・である。:の図によっで文字切出部13の作用を説明
する。
FIG. 2 is a block diagram of a character cutting device according to the present invention. Image data manually input from a scanner (not shown) is subjected to noise removal by a noise removal/line cutting section 10, and then line cutting is performed using a known horizontal projection method. The cut out image data of one character line (also called the underline part) is stored in the image memory unit 11, and the vertical projection detection unit 12 collects a vertical column/one shadow of the image of one character fT and calculates it. is manually input to the character cutting section 13. 14 is a section for removing (masking) the underlined portion of the character image. The fork output from the mask section 1.1
5- Wild boar data is recognized by the recognition unit (Jbt-not shown)
Yajin Rikisa J[ru, The third eye is Floza which shows the processing of the character cutting section 13! -1
・It is. The operation of the character cutout section 13 will be explained with reference to the figures.

ステップIO: ・続きの歪面1j影の長さく第・1図
参照)が2文字幅以上か判定する。N (−1ならばス
テップ(釦へ進む。
Step IO: - Determine whether the length of the shadow of the continuation distorted surface 1j (see Figure 1) is two characters or more wide. N (If -1, proceed to step (button).

ステップ(2+:垂直射影の始点(第・1図参照)より
前のイメージデータど終点(第・1図参照、)上り後の
イメージデータを調べ、垂直射影の直前と直後がスペー
スか判定する。Noの場合、第5図に示すような文字パ
ターンが連結している@自と判断し、ステップ(B+へ
進む。
Step (2+: Examine the image data before the starting point of the vertical projection (see Figure 1) and the image data after the ending point (see Figure 1), and determine whether the area immediately before and after the vertical projection is a space. In the case of No, it is determined that the character pattern shown in FIG. 5 is connected as @self, and the process proceeds to step (B+).

ステップ■:垂直射影からアンダーラインだけの場合の
垂直射影の値(垂直射影の高さ)を差り引いた第2射影
をめる。
Step ①: Find the second projection by subtracting the value of the vertical projection (height of the vertical projection) in the case of only an underline from the vertical projection.

ステップ■:各第2躬影の始点と萩点の近fシー切出し
点とする。1つの第2射影の終点と次の第2射影の始点
との間隔が1文′5・9幅より狭い喝i)、その中点を
切出し位置とする。
Step (2): The starting point of each second shadow and the Hagi point are set as near f-C cutting points. If the interval between the end point of one second projection and the start point of the next second projection is narrower than the width of one sentence'5.9, the midpoint is set as the cutting position.

ステップ[B・ニステップ中かlらこのステツノ′に進
んだ場合、垂直月影の始点と終点の近傍をLll li
 t。
If you proceed from step [B/Ni-step middle l to this steptsuno', the vicinity of the starting point and ending point of the vertical moon shadow will be Lll li
t.

点とし、イメージメモリ部11より1文字分の一1′メ
ージデータを読み出し、マスク部llを介しIj2識部
へ出力する。
11' image data corresponding to one character is read out from the image memory section 11 and output to the Ij2 recognition section via the mask section 11.

ステップ■からこのステップに進んた場白、垂直躬影を
文字幅で分割した点を切出し点と1.、各文字イメージ
をイメージメモリ部11より読出し、出力する。
When proceeding from step ① to this step, the point where the vertical shadow is divided by the character width is the cutting point 1. , each character image is read out from the image memory unit 11 and output.

ステップ+7(+からこのステップに進んだ@白、ステ
ップ(4)で検出された切出し7点て区別される各文字
イメージを、最下部から上部・・向って−1(下走査し
、水下ル1影を測定する。2−」以−1,の分離した4
く・[射影が検出された場合、最Fの本”F!11影に
含J4(る最−14ラインの番号を記g2tろ、文字と
アンダーラインが連結している場j>は、水’F fl
、I影の切れ1」を検出できない。この場自は、水・1
1J影内のj′ンターラ(ン位置の近傍で、水・P射影
値が急激に変化するラインの番号を記憶する。その後1
文へイメージデータを最−■4ラインよりパ旧−1メー
ジメモリ部11より読み出し、マスク部14を介し認識
部へ出力する。この際、記憶し、でいろう1′ン爪号以
降の名ライン読出し時にマスク1,1号をマツ、り部1
′1へljえ、当該名ラインのイメージデ−タを−(′
スキンクさせる。つまり、アンダーライン在マスA、ノ
グした文字イメージデータか認識部/、出力される。
Step +7 (proceeded to this step from + @white, each character image distinguished by the 7 cutout points detected in step (4) is scanned from the bottom to the top... -1 (downwards), Measure the shadow of Le 1.
- [If a projection is detected, record the number of the -14th line included in the most F book "F! 11 shadow. 'F fl
, I shadow cut 1'' cannot be detected. This place is water 1
1. Store the number of the line where the water/P projection value changes rapidly near the j'ntara (n position) in the J shadow.Then 1
The image data for the sentence is read from the last line 4 from the image memory section 11 and output to the recognition section via the mask section 14. At this time, memorize the mask 1 and 1 when reading out the name line after the 1' nail number, and the rim part 1.
'1 and save the image data of the name line -('
Make it skink. In other words, the underlined cell A is output as the recognized character image data.

文字rメージデータの後に、アンターラインコードをマ
スク部11Iを介し認識部へ送る。
After the character r image data, an underline code is sent to the recognition unit via the mask unit 11I.

なお上記のステップで41において、第21・J形量の
中点を切出し候補点とし、元の垂直月影の切出し、候補
点近傍の射影値を調べ、射影値が最小の点を切出し点と
するようにしてもよい。
In step 41 above, the midpoint of the 21st J-shaped quantity is set as a candidate point for cutting out, the original vertical moon shadow is cut out, the projection value near the candidate point is checked, and the point with the minimum projection value is set as the cutting point. You can do it like this.

〔効 果〕〔effect〕

以上説明し、たように、木光明は文字行(7I、−ジの
垂直月影からアンダーライン川当の値を差し引いた第2
射影をめ、その第2射影によっ゛(え宇切出し点を検出
するから、2文字以」−に跨が−)(アンダーラインが
引かれた文字行中の文字を正確に切り出すことができる
As explained above, Kikomei is the second value obtained by subtracting the value of the underline Kawato from the vertical moon shadow of the character line (7I, -ji).
By using the second projection, we can accurately cut out the characters in the underlined character line. .

【図面の簡単な説明】[Brief explanation of drawings]

第1図は長いアンダーラインをaむ文字行−rメージと
その垂直射影および第2射影の例を示l−図、第2図は
本発明に係る文字切出装置のプロツノ1,4、第3図は
文字切出し処理のフロー図、第4図は垂直射影の説明図
である。 10・・ノイズ除去・行切出部、II・・・イミラーー
;メモリ部、13・・・垂直射影検出部、l11 マス
ク部、15・・文字切出部。 代理人弁理上 鈴 木 誠 ′°゛。 第1図 第4図 第2図 第3図
FIG. 1 shows an example of a character line image with a long underline and its vertical projection and second projection. FIG. FIG. 3 is a flowchart of character extraction processing, and FIG. 4 is an explanatory diagram of vertical projection. 10... Noise removal/line cutting section, II... Mirror; memory section, 13... Vertical projection detection section, l11 mask section, 15... Character cutting section. Makoto Suzuki ′°゛ on behalf of attorney. Figure 1 Figure 4 Figure 2 Figure 3

Claims (1)

【特許請求の範囲】[Claims] (1)文字行イメージの垂直射影をめ、この垂直射影か
らアンダーラインに相当する射影値を差し引いた第2射
影をめ、この第2射影に基いて個々の文字イメージの切
出し点を決めることを特徴とする文字切出方式。
(1) Determine the vertical projection of the character line image, determine the second projection by subtracting the projection value corresponding to the underline from this vertical projection, and determine the cutting point of each character image based on this second projection. Characteristic character cutting method.
JP58143427A 1983-08-05 1983-08-05 Character cutting system Pending JPS6037082A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP58143427A JPS6037082A (en) 1983-08-05 1983-08-05 Character cutting system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP58143427A JPS6037082A (en) 1983-08-05 1983-08-05 Character cutting system

Publications (1)

Publication Number Publication Date
JPS6037082A true JPS6037082A (en) 1985-02-26

Family

ID=15338471

Family Applications (1)

Application Number Title Priority Date Filing Date
JP58143427A Pending JPS6037082A (en) 1983-08-05 1983-08-05 Character cutting system

Country Status (1)

Country Link
JP (1) JPS6037082A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS60160487A (en) * 1984-01-31 1985-08-22 Toshiba Corp Optical character reader
CN108108734A (en) * 2016-11-24 2018-06-01 杭州海康威视数字技术股份有限公司 A kind of licence plate recognition method and device

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS60160487A (en) * 1984-01-31 1985-08-22 Toshiba Corp Optical character reader
CN108108734A (en) * 2016-11-24 2018-06-01 杭州海康威视数字技术股份有限公司 A kind of licence plate recognition method and device
CN108108734B (en) * 2016-11-24 2021-09-24 杭州海康威视数字技术股份有限公司 License plate recognition method and device

Similar Documents

Publication Publication Date Title
JPH0546591B2 (en)
JPS6037082A (en) Character cutting system
JPS62254282A (en) Method and apparatus for separating overlapped pattern
JP2503208B2 (en) Business card image processing method
JP2975772B2 (en) Voice recognition device
JPH07230525A (en) Method for recognizing ruled line and method for processing table
JPH0136154B2 (en)
JP2909132B2 (en) Optical character reader
JP4254008B2 (en) Pattern detection apparatus and method
JPS6111886A (en) Character recognition system
JP3160458B2 (en) Character reading device and character reading method
JPH051510B2 (en)
JPS61105680A (en) Letter separating method for optical character reader
JP4580520B2 (en) Character recognition method and character recognition apparatus
JP2004110369A (en) Character recognizing method and character recognizing device
JP2913103B2 (en) Pattern recognition device
JPS60200375A (en) Template matching system
JP2925270B2 (en) Character reader
JPS581277A (en) Character characteristic pickup system
JP2964684B2 (en) Character data collection method for character reading
JP2894095B2 (en) Character recognition device
JPS63101983A (en) Character string extracting system
JPH0353392A (en) Character recognizing device
JPS5953977A (en) Character recognizing device
JPH0632074B2 (en) Normalization method