JPH0576674B2 - - Google Patents

Info

Publication number
JPH0576674B2
JPH0576674B2 JP61103863A JP10386386A JPH0576674B2 JP H0576674 B2 JPH0576674 B2 JP H0576674B2 JP 61103863 A JP61103863 A JP 61103863A JP 10386386 A JP10386386 A JP 10386386A JP H0576674 B2 JPH0576674 B2 JP H0576674B2
Authority
JP
Japan
Prior art keywords
character
pattern
projection
width
isolated pattern
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.)
Expired - Lifetime
Application number
JP61103863A
Other languages
Japanese (ja)
Other versions
JPS62262194A (en
Inventor
Michio Terai
Shigeru Horii
Yoshikazu Kobayashi
Kazuo Ito
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.)
Oki Electric Industry Co Ltd
Original Assignee
Oki Electric Industry 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 Oki Electric Industry Co Ltd filed Critical Oki Electric Industry Co Ltd
Priority to JP61103863A priority Critical patent/JPS62262194A/en
Publication of JPS62262194A publication Critical patent/JPS62262194A/en
Publication of JPH0576674B2 publication Critical patent/JPH0576674B2/ja
Granted legal-status Critical Current

Links

Description

【発明の詳細な説明】 (産業上の利用分野) 本発明は光学式文字読取装置に関し、特に帳票
上に記載された印刷文字を認識する方法に関す
る。
DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an optical character reading device, and more particularly to a method for recognizing printed characters written on a form.

(従来の技術) 第2図は従来の光学式読取装置を示すブロツク
図である。同図において、21は帳票上の読取対
象を光学的に読取る読取部、22は読取部21に
より読取つたアナログ信号又は16階調のデイジタ
ル信号をある閾値に基づいて黒を“1”、白を
“0”として2値化して画像信号を得る2値化部、
23は2値化部22により得た画像信号を文字行
毎に切出す行切出部、24は行切出部23によつ
て切出された文字行に含まれる画像信号を行と直
交する方向に投影し、黒の有無を“1”、“0”で
表した射影を抽出する射影抽出部、25は射影抽
出部14により抽出した射影のデータを一時記憶
する射影記憶部、26は射影抽出部24により抽
出し、射影記憶部25に記憶されている射影にお
いて左右を“0”(白)に挟まれた連続する“1”
(黒)の射影部分に相当する画像信号を孤立パタ
ーンとして切出す孤立パターン切出部、27は孤
立パターン切出部26により切出された孤立パタ
ーンの幅を検出し、検出した幅を1文字とみなす
ことができる限度を示す最大文字幅と比較して検
出した幅が最大文字幅以内のときは孤立パターン
切出部26で切り出した孤立パターンを後述する
認識部29に供給し、大きいときは後述する強制
切出部28に供給するパターン幅判定部、28は
孤立パターンを最大文字幅で強制的に切出し、切
出した文字パターンを後述する認識部29に供給
する強制切出部、29は孤立パターン切出部26
又は強制切出部28からの文字パターンを辞書メ
モリに予め登録してある文字パターンと照合して
認識する認識部である。
(Prior Art) FIG. 2 is a block diagram showing a conventional optical reading device. In the figure, 21 is a reading unit that optically reads an object to be read on a form, and 22 is an analog signal or a 16-gradation digital signal read by the reading unit 21, and the black is set to “1” and the white is set to “1” based on a certain threshold value. a binarization unit that binarizes as “0” to obtain an image signal;
23 is a line cutting unit that cuts out the image signal obtained by the binarization unit 22 for each character line; 24 is a line cutting unit that cuts out the image signal included in the character line cut out by the line cutting unit 23 at right angles to the line; 25 is a projection storage unit that temporarily stores the projection data extracted by the projection extraction unit 14; 26 is a projection In the projection extracted by the extraction unit 24 and stored in the projection storage unit 25, continuous “1” sandwiched between “0” (white) on the left and right sides
27 detects the width of the isolated pattern cut out by the isolated pattern cutting section 26, and converts the detected width into one character. When the detected width is within the maximum character width, the isolated pattern cut out by the isolated pattern cutting section 26 is supplied to the recognition section 29, which will be described later. A pattern width determining section 28 forcibly cuts out an isolated pattern at the maximum character width and supplies the cut out character pattern to a recognition section 29 (described later); 29 is an isolated Pattern cutting section 26
Alternatively, it is a recognition unit that recognizes the character pattern from the forced extraction unit 28 by comparing it with character patterns registered in advance in the dictionary memory.

次に、第2図を用いて従来例の動作を説明す
る。先ず、読取部21では、帳票上の文字等の読
取対象を読取り、2値化部22で黒を“1”、白
を“0”として2値化し、行切出部23で文字行
毎に切り出す。射影抽出部24では、切出された
文字行について、行と垂直な方向に投影し、黒の
有無を“1”、“0”で表わす。この結果は孤立パ
ターンと射影の関係を示す第3図からわかるが、
ここで同図において、30〜32を射影、33〜35をイ
メージであり、射影を求める操作を射影の抽出と
呼ぶことにする。射影の抽出によつて得られた
“1”、“0”の射影の情報は、射影記憶部25に
記憶される。孤立パターン切出部26では、抽出
した射影について左右を“0”(白)に挟まれた
連続する“1”(黒)の射影部分に対応する文字
行上の第3図のイメージ33,34,35を孤立
パターンとして切出し、個々の孤立パターンを1
文字と考えて、認識部29の辞書メモリに予め登
録してある文字パターンと照合を行い、該当する
もとを探す。基本的には、この方法で文字の認識
が可能である。
Next, the operation of the conventional example will be explained using FIG. First, the reading section 21 reads the object to be read, such as characters on a form, and the binarization section 22 converts it into a binary value with black as "1" and white as "0". break the ice. The projection extraction unit 24 projects the extracted character line in a direction perpendicular to the line, and indicates the presence or absence of black as "1" or "0". This result can be seen from Figure 3, which shows the relationship between isolated patterns and projections.
Here, in the figure, 30 to 32 are projections, 33 to 35 are images, and the operation to obtain the projection will be referred to as projection extraction. Information on the projections of “1” and “0” obtained by extracting the projections is stored in the projection storage unit 25. The isolated pattern cutting unit 26 extracts images 33 and 34 of FIG. 3 on the character line corresponding to the projection part of continuous "1" (black) sandwiched between "0" (white) on the left and right sides of the extracted projection. , 35 are cut out as isolated patterns, and each isolated pattern is divided into 1
Considered as a character, the character pattern is compared with a character pattern registered in advance in the dictionary memory of the recognition unit 29 to find a matching character. Basically, characters can be recognized using this method.

しかし、印刷文字の場合には、複数個の文字が
連続することがあり、第3図bに示すように孤立
パターンが必ずしも1文字に対応するとは限らな
い。そこで、印刷文字の場合には、第2図のパタ
ーン幅判定部27で第3図aの孤立パターンの幅
Aを検出し、これが最大文字幅を越えるときに
は、第2図の強制切出部28において、最大文字
幅で切出しを行い、認識部29の中の辞書メモリ
内の文字パターンと照合し認識を行う。
However, in the case of printed characters, a plurality of characters may be consecutive, and the isolated pattern does not necessarily correspond to one character, as shown in FIG. 3b. Therefore, in the case of printed characters, the pattern width determining section 27 in FIG. 2 detects the width A of the isolated pattern in FIG. At , cutting is performed using the maximum character width, and recognition is performed by comparing the character pattern with the character pattern in the dictionary memory in the recognition unit 29 .

(発明が解決しようとする問題点) しかしながら、上記の方法では、連続文字があ
つた場合、最大文字幅で強制切出しを行うため、
文字を正しく切出せないことがある。例えば、第
4図aに示すような連続文字は、強制切出しの結
果、同図bのようにfとiの半分を1文字、iの
半分を別の1文字として、認識しようとする。し
たがつて、このように切出された文字は、リジエ
クトされたり、誤認識されたりしてしまうという
問題点があつた。
(Problem to be solved by the invention) However, in the above method, when there are consecutive characters, forced cutting is performed using the maximum character width.
Characters may not be cut out correctly. For example, as a result of forced segmentation of continuous characters as shown in FIG. 4a, half of f and i are recognized as one character, and half of i is recognized as another character, as shown in FIG. 4b. Therefore, there is a problem in that characters cut out in this manner may be rejected or misrecognized.

本発明はこれらの問題点を解決するためのもの
で、連続文字があつた場合でも、リジエクトや誤
認識を少なくし、より正しく印刷文字を認識する
ことのできる認識率の優れた光学式文字読取装置
を提供することを目的とする。
The present invention is intended to solve these problems, and is an optical character reader with an excellent recognition rate that can reduce rejects and misrecognitions and more accurately recognize printed characters even when there are consecutive characters. The purpose is to provide equipment.

(問題点を解決するための手段) 本発明は前記問題点を解決するために、帳標上
の読取対象を光学的に読取り、さらに2値化して
画像信号を得る読取部と、画像信号を文字行毎に
切出す行切出部と、この行切出部により切出され
た1文字行に含まれる画像信号を行と直交する方
向に投影して射影を抽出する射影抽出部と、この
射影抽出部により抽出した射影を一時格納する射
影記憶部と、この射影抽出部により抽出した射影
の黒の射影部分のみに相当する画像信号を孤立パ
ターンとして切出す孤立パターン切出部と、この
孤立パターン切出部により切出されれた孤立パタ
ーンの幅を所定の文字幅と比較し、所定の文字幅
以内のときは孤立パターンは1文字に相当するも
のとし、所定の文字幅より大きいときは孤立パタ
ーンは複数の文字が連続して構成する連続文字に
相当するものと判定するパターン幅判定部と、1
文字の文字パターンを予め登録してある単独文字
辞書メモリと、連続文字の文字パターンを予め登
録してある連続文字辞書メモリと、パターン幅判
定部の判定結果に基づいて、孤立パターンを単独
辞書メモリまたは連続文字辞書メモリに登録して
ある文字パターンと、孤立パターンから抽出した
特徴と照合して読取対象を認識する認識部とを具
備している。
(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention provides a reading section that optically reads an object to be read on a ledger and further binarizes it to obtain an image signal; a line cutting section that cuts out each character line; a projection extraction section that extracts a projection by projecting an image signal included in one character line cut out by the line cutting section in a direction perpendicular to the line; a projection storage section that temporarily stores the projection extracted by the projection extraction section; an isolated pattern cutting section that cuts out, as an isolated pattern, an image signal corresponding only to the black projection part of the projection extracted by the projection extraction section; The width of the isolated pattern cut out by the pattern cutting section is compared with a predetermined character width, and if it is within the predetermined character width, the isolated pattern is considered to be equivalent to one character, and if it is larger than the predetermined character width, the isolated pattern is considered to be equivalent to one character. a pattern width determination unit that determines that the isolated pattern corresponds to a continuous character composed of a plurality of consecutive characters;
A single character dictionary memory in which character patterns of characters are registered in advance, a continuous character dictionary memory in which character patterns of continuous characters are registered in advance, and an isolated pattern is stored in the individual dictionary memory based on the determination result of the pattern width determination unit. Alternatively, it includes a recognition unit that recognizes a reading target by comparing character patterns registered in a continuous character dictionary memory with features extracted from isolated patterns.

(作用) 以上のような構成を有する本発明によれば、読
取部は帳票上の読取対象を光学的に読取り、さら
に2値化して画像信号を得る。そして、行切出部
では画像信号を文字行毎に切出される。この切出
された1文字行に含まれる画像信号は射影抽出部
により行と直交する方向に投影されて射影が抽出
されて射影記憶部に一時格納される。そして、孤
立パターン切出部は抽出した射影の黒の射影部分
のみに相当する画像信号を孤立パターンとして切
出す。パターン幅判定部では切出された孤立パタ
ーンの幅を所定の文字幅と比較する。そして、比
較した結果、孤立パターンの幅が所定の文字幅以
内のときは、切出されれた孤立パターンは1文字
に相当するものと判定する。一方、孤立パターン
の幅が所定の文字幅より大きいときは、切出され
た孤立パターンは複数の文字が連続して構成する
連続文字に相当するものと判定する。そして、こ
の判定結果、認識部では孤立パターンから抽出し
た特徴と、所定の文字幅以内のときは孤立パター
ンを単独文字辞書メモリに予め登録してある1文
字の文字パターンとを照字し、所定の幅より大き
いときは孤立パターンを連続文字辞書メモリに予
め登録してある連続文字の文字パターンとを照合
して帳票上の読取対象を認識する。
(Function) According to the present invention having the above-described configuration, the reading section optically reads the object to be read on the form, and further binarizes the object to be read to obtain an image signal. Then, in the line cutting section, the image signal is cut out for each character line. The image signal included in this cut out one character line is projected in a direction perpendicular to the line by a projection extraction section, the projection is extracted, and the extracted image signal is temporarily stored in the projection storage section. Then, the isolated pattern cutting section cuts out the image signal corresponding only to the black projected portion of the extracted projection as an isolated pattern. The pattern width determination section compares the width of the cut out isolated pattern with a predetermined character width. Then, as a result of the comparison, if the width of the isolated pattern is within a predetermined character width, it is determined that the cut out isolated pattern corresponds to one character. On the other hand, when the width of the isolated pattern is larger than the predetermined character width, it is determined that the cut out isolated pattern corresponds to a continuous character composed of a plurality of consecutive characters. As a result of this determination, the recognition unit compares the features extracted from the isolated pattern and, if the character width is within a predetermined character width, the isolated pattern with a character pattern of one character previously registered in the single character dictionary memory, and If the width is larger than the width, the isolated pattern is compared with a character pattern of continuous characters registered in advance in a continuous character dictionary memory to recognize the object to be read on the form.

したがつて、本発明は前記問題点を解決するこ
とができ、作業効率の良好で、かつ認識率の優れ
た光学式文字読取装置を提供できる。
Therefore, the present invention can solve the above-mentioned problems, and can provide an optical character reading device with good working efficiency and excellent recognition rate.

(実施例) 以下、本発明の一実施例を図面に基づいて説明
する。
(Example) Hereinafter, an example of the present invention will be described based on the drawings.

第1図は本発明の一実施例を示すブロツク図で
ある。同図において、1は帳票上の読取対象を光
学的に読取る読取部、2は読取部1により読取つ
たアナログ信号又は16階調のデイジタル信号をあ
る閾値に基づいて黒を“1”、白を“0”として
2値化して画像信号を得る2値化部、3は2値化
部2により得た画像信号を文字行毎に切出す行切
出部、4は行切出部3によつて切出された文字行
に含まれる画像信号を行と直交する方向に投影
し、黒の有無を“1”、“0”で表した射影を抽出
する射影抽出部、5は射影抽出部4により抽出し
た射影のデータを一時記憶する射影記憶部、6は
射影抽出部4により抽出し射影記憶部5に記憶さ
れている射影において左右を“0”(白)に挟ま
れた連続する“1”(黒)の射影部分に相当する
画像信号を孤立パターンとして切出す孤立パター
ン切出部、7は孤立パターン切出部6により切出
された孤立パターンの幅を検出し、検出した幅を
1文字とみなすことができる限度を示す最大文字
幅と比較して比較判定の結果を後述する辞書切替
部13に供給するパターン幅判定部、8は、パタ
ーン幅判定部7を介した孤立パターンを後述する
各辞書メモリに予め登録してある文字パターンと
照合して認識を行う認識部、9は1文字の文字パ
ターンを予め登録してある単独文字辞書メモリ、
10は複数の文字が連続して構成する連続文字の
文字パターンを予め登録してある連続文字辞書メ
モリ、11は孤立パターン切出部6により切出さ
れた孤立パターンから特徴を抽出する特徴抽出
部、12は特徴抽出部11により抽出された特徴
と単独文字辞書メモリ9又は連続文字辞書メモリ
10に登録されている文字パターンとを照合して
文字判定を行う判定部、13はパターン幅判定部
7の判定結果に基づいて判定部2で照合するため
に単独文字辞書メモリ9又は連続文字辞書メモリ
10を切替える辞書切替部である。
FIG. 1 is a block diagram showing one embodiment of the present invention. In the figure, 1 is a reading unit that optically reads the object to be read on a form, and 2 is an analog signal or a 16-gradation digital signal read by the reading unit 1, and based on a certain threshold, black is set to 1 and white is set to 1. A binarization unit that binarizes as “0” to obtain an image signal, 3 a line cutting unit that cuts out the image signal obtained by the binarization unit 2 for each character line, and 4 a line cutting unit 3 that cuts out the image signal obtained by the binarization unit 2. 5 is a projection extraction unit 4 that projects the image signal included in the character line cut out in the direction perpendicular to the line and extracts a projection in which the presence or absence of black is expressed as “1” or “0”; A projection storage section 6 temporarily stores the data of the projection extracted by the projection extraction section 4, and a projection storage section 6 stores continuous "1" sandwiched between "0" (white) on the left and right sides in the projection extracted by the projection extraction section 4 and stored in the projection storage section 5. 7 detects the width of the isolated pattern cut out by the isolated pattern cutting unit 6, and converts the detected width into 1. A pattern width determination unit 8 compares the comparison with the maximum character width indicating the limit that can be considered as a character and supplies the comparison determination result to the dictionary switching unit 13, which will be described later. 9 is a recognition unit that performs recognition by comparing it with a character pattern registered in advance in each dictionary memory; 9 is a single character dictionary memory in which a character pattern of one character is registered in advance;
Reference numeral 10 denotes a continuous character dictionary memory in which character patterns of continuous characters constituted by a plurality of consecutive characters are registered in advance, and 11 a feature extraction unit that extracts features from the isolated pattern cut out by the isolated pattern cutout unit 6. , 12 is a determination unit that performs character determination by comparing the features extracted by the feature extraction unit 11 with character patterns registered in the single character dictionary memory 9 or the continuous character dictionary memory 10, and 13 is a pattern width determination unit 7. This is a dictionary switching unit that switches between the single character dictionary memory 9 and the continuous character dictionary memory 10 for comparison in the determination unit 2 based on the determination result.

次に、第1図を用いて本実施例の動作を説明す
る。
Next, the operation of this embodiment will be explained using FIG.

先ず、読取部1は、帳票上の文字等の読取対象
を読取り、黒白の濃淡の情報をアナログ信号又
は、16階調のデイジタル信号として出力する。2
値化部2では、黒白の濃淡情報をある閾値で2値
化し、黒を“1”、白を“0”として2値のデイ
ジタル信号として出力する。これを行切出部3で
文字行ごとに切出し、射影抽出部4では文字行上
のイメージを行と直交する方向に投影し、射影を
抽出して、射影記憶部5に格納する。さらに、孤
立パターン切出部6では、抽出した射影から
“0”(白)の部分を探し、“0”(白)に挟まれた
連続する“1”(黒)の射影部分に対応する文字
行上のイメージを孤立パターンとして切出す。
First, the reading section 1 reads an object to be read, such as characters on a form, and outputs black and white shading information as an analog signal or a 16-gradation digital signal. 2
The digitization unit 2 binarizes the black and white shading information using a certain threshold value, and outputs it as a binary digital signal with black as "1" and white as "0". The line cutting unit 3 cuts out each character line, and the projection extraction unit 4 projects the image on the character line in a direction perpendicular to the line, extracts the projection, and stores it in the projection storage unit 5. Furthermore, the isolated pattern cutting unit 6 searches for "0" (white) parts from the extracted projections, and searches for characters corresponding to consecutive "1" (black) projected parts sandwiched between "0" (white). Extract the image on the row as an isolated pattern.

さて、切出された個々の孤立パターンは、1文
字に対応するものもあれば、第3図aに示すよう
に2文字あるいは第3図bに示すようにそれ以上
の連続文字に対応するものもある。しかし、フオ
ント指定により決まる最大文字幅を越えるパター
ン幅を有する孤立パターンは、複数個の文字が連
続したものであると考えることができる。そこ
で、パターン幅判定部7で個々の孤立パターンの
幅を判定し、パターンデータと共に認識部8へパ
ターン幅の情報も送る。認識部8では、まず特徴
抽出部11で各孤立パターンから特徴を抽出しデ
ータを判定部12に送る。判定部12では、デー
タを各辞書メモリに登録してある文字パターンと
照合し、文字判定を行うが、単独文字か連続文字
かにによつて辞書メモリを切替える必要がある。
第1図に示した例では、辞書切替部13を設け、
パターン幅検出部7で得られたパターン幅の判定
結果に基づいて孤立パターンの幅が最大文字幅以
内であれば、その孤立パターンは1文字に対応す
るものとして単独文字辞書メモリ9に登録してあ
る文字パターンで照合を行い、孤立パターンの幅
が最大文字幅を越える場合には、複数個の文字が
連続したものとして、連続文字辞書メモリ10に
登録してある文字パターンで照合を行い、判定部
12の判定結果を認識結果として出力する。
Now, some of the isolated isolated patterns that have been cut out correspond to one character, while others correspond to two characters as shown in Figure 3a, or more than two consecutive characters as shown in Figure 3b. There is also. However, an isolated pattern having a pattern width exceeding the maximum character width determined by the font specification can be considered to be a plurality of consecutive characters. Therefore, the pattern width determining section 7 determines the width of each isolated pattern, and sends pattern width information to the recognizing section 8 along with the pattern data. In the recognition section 8 , first, the feature extraction section 11 extracts features from each isolated pattern and sends the data to the determination section 12 . The determination unit 12 performs character determination by comparing data with character patterns registered in each dictionary memory, but it is necessary to switch dictionary memories depending on whether the data is a single character or a continuous character.
In the example shown in FIG. 1, a dictionary switching section 13 is provided,
If the width of the isolated pattern is within the maximum character width based on the pattern width determination result obtained by the pattern width detection section 7, the isolated pattern is registered in the single character dictionary memory 9 as corresponding to one character. If a certain character pattern is compared and the width of the isolated pattern exceeds the maximum character width, it is assumed that multiple characters are consecutive, and the character pattern registered in the continuous character dictionary memory 10 is compared and determined. The determination result of the unit 12 is output as a recognition result.

ここで、連続文字辞書メモリに登録してある文
字パターンについて説明する。連続文字は、F,
f,r,Tなどのように、文字の上部が左右又
は、そのどちらか一方に広がつた文字と小文字、
特にiなどとの組合せが多く、ある程度組合せの
種類を特定できるので、全ての組合せを辞書とし
て持つ必要はない。また、文字の連続数は2個の
ものがほとんどで、3個のものは少なく、4個以
上は希である。辞書メモリに登録してある文字パ
ターンとしては、2文字連続のものと3文字連続
のものを用意しておけば、最大文字幅を越えるも
のの大部分には対応できる。したがつて、最大文
字幅を越える孤立パターンが切出された場合に
は、まず連続文字辞書メモリの中の2文字連続の
辞書メモリ内の文字パターンと照合を行い、該当
する組合せがあれば、それを認識結果として出力
し、該当する組合せがなければ、3文字連続の辞
書メモリ内の文字パターンと照合を行い、該当す
る組合せがあれば、それを認識結果として出力
し、それでも該当するものがなければリジエクト
する。
Here, character patterns registered in the continuous character dictionary memory will be explained. Consecutive characters are F,
Letters and lowercase letters where the upper part of the letter spreads to the left or right, such as f, r, T, etc.
In particular, there are many combinations such as i, and the types of combinations can be specified to some extent, so it is not necessary to have all combinations as a dictionary. Furthermore, the number of consecutive characters is mostly two, rarely three, and rarely four or more. By preparing two consecutive character patterns and three consecutive character patterns as character patterns registered in the dictionary memory, most characters exceeding the maximum character width can be handled. Therefore, when an isolated pattern exceeding the maximum character width is cut out, it is first compared with the character pattern in the dictionary memory of two consecutive characters in the continuous character dictionary memory, and if a matching combination is found, It is output as a recognition result, and if there is no matching combination, it is compared with the character pattern of three consecutive characters in the dictionary memory, and if there is a matching combination, it is output as a recognition result. If not, it will be rejected.

(発明の効果) 以上説明したように、本発明によれば、、連続
文字があつた場合でも、連続文字の文字パターン
を予め登録してある辞書メモリを設けて連続文字
のまま照合し、かつ認識することにより、リジエ
クトや誤認識を少ない、即ち作用効率が良く、さ
らに認識率の優れた光学式文字読取装置を提供す
ることができる。
(Effects of the Invention) As explained above, according to the present invention, even if there are consecutive characters, a dictionary memory in which the character patterns of the consecutive characters are registered in advance is provided to match the consecutive characters as they are, and By recognizing these characters, it is possible to provide an optical character reading device that has fewer rejects and erroneous recognitions, that is, has good operational efficiency and has an excellent recognition rate.

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

第1図は本発明の一実施例を示すブロツク図、
第2図は従来の光学式文字読取装置を示すブロツ
ク図、第3図は孤立パターンと射影の関係を示す
図、第4図は連続文字の強制切出しの一例を示す
図である。 1……読取部、2……2値化部、3……行切出
部、4……射影抽出部、5……射影記憶部、6…
…孤立パターン切出部、7……パターン幅判定
部、8……認識部、9……単独文字辞書メモリ、
10……連続文字辞書メモリ、11……特徴抽出
部、12……判定部、13……辞書切替部。
FIG. 1 is a block diagram showing one embodiment of the present invention;
FIG. 2 is a block diagram showing a conventional optical character reading device, FIG. 3 is a diagram showing the relationship between isolated patterns and projection, and FIG. 4 is a diagram showing an example of forced extraction of continuous characters. 1... Reading section, 2... Binarization section, 3... Line cutting section, 4... Projection extraction section, 5... Projection storage section, 6...
...Isolated pattern cutting section, 7... Pattern width determination section, 8... Recognition section, 9... Single character dictionary memory,
10... Continuous character dictionary memory, 11... Feature extraction section, 12... Judgment section, 13... Dictionary switching section.

Claims (1)

【特許請求の範囲】 1 帳票上の読取対象を光学的に読取り、さらに
2値化して画像信号を得る読取部と、 前記画像信号を文字行毎に切出す行切出部と、 該行切出部により切出された1文字行に含まれ
る前記画像信号を行と直交する方向に投影して射
影を抽出する射影抽出部と、 該射影抽出部により抽出した前記射影を一時格
納する射影記憶部と、 前記該射影抽出部により抽出した前記射影の黒
の射影部分のみに相当する画像信号を孤立パター
ンとして切出す孤立パターン切出部と、 該孤立パターン切出部により切出された孤立パ
ターンの幅を所定の文字幅と比較し、所定の文字
幅以内のときは前記孤立パターンは1文字に相当
するものとし、所定の文字幅より大きいときは前
記孤立パターンは複数の文字が連続して構成する
連続文字に相当するものと判定するパターン幅判
定部と、 1文字の文字パターンを予め登録してある単独
文字辞書メモリと、 前記連続文字の文字パターンを予め登録してあ
る連続文字辞書メモリと、 前記パターン幅判定部の判定結果に基づいて、
前記孤立パターンを前記単独辞書メモリまたは前
記連続文字辞書メモリに登録してある文字パター
ンと、前記孤立パターンから抽出した特徴と照合
して前記読取対象を認識する認識部とを具備する
ことを特徴とする光学式文字読取装置。
[Scope of Claims] 1. A reading unit that optically reads an object to be read on a form and further binarizes it to obtain an image signal; a line cutting unit that cuts out the image signal for each character line; and a line cutting unit that cuts out the image signal for each character line. a projection extraction unit that extracts a projection by projecting the image signal included in one character line extracted by the extraction unit in a direction perpendicular to the line; and a projection storage that temporarily stores the projection extracted by the projection extraction unit. an isolated pattern cutting section that cuts out, as an isolated pattern, an image signal corresponding only to the black projection portion of the projection extracted by the projection extraction section; and an isolated pattern cut out by the isolated pattern cutting section. The width of the isolated pattern is compared with a predetermined character width, and if it is within the predetermined character width, the isolated pattern is considered to be equivalent to one character, and if it is larger than the predetermined character width, the isolated pattern is considered to be a series of multiple characters. a pattern width determination unit that determines that the pattern corresponds to the constituent continuous characters; a single character dictionary memory in which character patterns of one character are registered in advance; and a continuous character dictionary memory in which character patterns of the continuous characters are registered in advance. and, based on the determination result of the pattern width determination section,
It is characterized by comprising a recognition unit that recognizes the reading target by comparing the isolated pattern with a character pattern registered in the individual dictionary memory or the continuous character dictionary memory and a feature extracted from the isolated pattern. optical character reader.
JP61103863A 1986-05-08 1986-05-08 Optical character reader Granted JPS62262194A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP61103863A JPS62262194A (en) 1986-05-08 1986-05-08 Optical character reader

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP61103863A JPS62262194A (en) 1986-05-08 1986-05-08 Optical character reader

Publications (2)

Publication Number Publication Date
JPS62262194A JPS62262194A (en) 1987-11-14
JPH0576674B2 true JPH0576674B2 (en) 1993-10-25

Family

ID=14365284

Family Applications (1)

Application Number Title Priority Date Filing Date
JP61103863A Granted JPS62262194A (en) 1986-05-08 1986-05-08 Optical character reader

Country Status (1)

Country Link
JP (1) JPS62262194A (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2630261B2 (en) * 1994-06-29 1997-07-16 日本電気株式会社 Character recognition device
JP2000353215A (en) 1999-06-11 2000-12-19 Nec Corp Character recognition device and recording medium where character recognizing program is recorded

Also Published As

Publication number Publication date
JPS62262194A (en) 1987-11-14

Similar Documents

Publication Publication Date Title
US4757551A (en) Character recognition method and system capable of recognizing slant characters
JPH0564834B2 (en)
JPS63216189A (en) Character recognition system
JPS58103075A (en) Character reader
US4769851A (en) Apparatus for recognizing characters
JPH0576674B2 (en)
JPS6160184A (en) Optical character reader
JPS6146573A (en) Character recognizing device
JP2877380B2 (en) Optical character reader
JPS59158482A (en) Character recognizing device
JP2973898B2 (en) Character recognition method and device
JPH0475557B2 (en)
JPH0816720A (en) Character recognition device
JPH0319589B2 (en)
JP3027232B2 (en) Character recognition device
JPS6111886A (en) Character recognition system
JPH01265378A (en) European character recognizing system
JPH0731711B2 (en) Optical character reader
JPH06231306A (en) Character recognition device
EP0114996A2 (en) Character recognition utilizing transition measurements
JPH0578068B2 (en)
JPH0969139A (en) Optical character reading method and its device
JPH0353392A (en) Character recognizing device
JPH0340186A (en) Character recognizer
JP2002230481A (en) Optical character reader