JPH0117191B2 - - Google Patents
Info
- Publication number
- JPH0117191B2 JPH0117191B2 JP56088508A JP8850881A JPH0117191B2 JP H0117191 B2 JPH0117191 B2 JP H0117191B2 JP 56088508 A JP56088508 A JP 56088508A JP 8850881 A JP8850881 A JP 8850881A JP H0117191 B2 JPH0117191 B2 JP H0117191B2
- Authority
- JP
- Japan
- Prior art keywords
- character
- center
- character pattern
- pattern
- retry
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/98—Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Input (AREA)
- Character Discrimination (AREA)
Description
【発明の詳細な説明】
本発明はマスクマツチング法で認識不能となり
再試行を行なう場合、文字パターンの基準点を補
正して行なうようにした文字認識方式に関するも
のである。DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a character recognition method that corrects the reference point of a character pattern when the mask matching method fails to recognize the character and a retry is made.
従来、活字ドラムで印字された文字を光学文字
読取装置(OCR)により読取り認識するためマ
スクマツチング法が用いられる。マスクマツチン
グ法とは第1図aに示すような用紙に印字された
文字を認識する場合、1文字の領域(M×N)を
用紙を送りながらOCRのセンサで読取つた文字
パターンが同図bとすると、文字パターンの形状
から文字中心を求める。すなわち、左右幅x、上
下高さyのそれぞれ1/2の交点○イを求めるものと
する。この文字パターンの中心を、同図cに示す
ように、あらかじめメモリに格納されている標準
文字パターンの中心と合せて最も重複の程度の高
い文字が認識すべき文字となる。 Conventionally, a mask matching method is used to read and recognize characters printed on a type drum using an optical character reader (OCR). What is the mask matching method? When recognizing characters printed on paper as shown in Figure 1a, the character pattern read by the OCR sensor while feeding the paper in the area of one character (M x N) is the same as shown in Figure 1a. If b, the center of the character is found from the shape of the character pattern. That is, let us find the intersection point ○a of 1/2 each of the left and right width x and the top and bottom height y. The center of this character pattern is aligned with the center of the standard character pattern previously stored in the memory, as shown in FIG. 3c, and the character with the highest degree of overlap becomes the character to be recognized.
正常な文字の場合は上述の方法で認識できる
が、同図dのように文字の○ロに示すごみか汚れが
付着した場合や、同図eのように欠け○ハがある場
合には上下の長さがy′,y″のようになりその1/2
より求めた中心が正常な時の中心位置○イよりずれ
る。従つて前述のマスクマツチング法では認識不
能となる場合が起る。 Normal characters can be recognized using the method described above, but if there is dirt or dirt attached to the character as shown in the circle (○) in the figure d, or if there is a chip (○) as shown in the figure e, the upper and lower The length of becomes y′, y″ and 1/2 of that
The center determined by this method deviates from the normal center position ○a. Therefore, there may be cases where the mask matching method described above cannot be recognized.
このような場合は再試行を何回やつても結果は
同じである。むしろ正常な場合の中心を見付けれ
ば認識可能となる場合が考えられる。 In such cases, no matter how many times you try, the result is the same. Rather, it may be possible to recognize the center by finding the center of the normal case.
本発明の目的はマスクマツチング法で読取不能
となり再試行を行なう場合文字パターンの中心を
正常な場合の中心位置に補正して行なうようにし
た文字認識方式を提供することである。 SUMMARY OF THE INVENTION An object of the present invention is to provide a character recognition system that corrects the center of the character pattern to the normal center position when a retry is made due to unreadable characters using the mask matching method.
前記目的を達成するため、本発明の文字認識方
式は文字を光学文字読取手段で読取つた文字パタ
ーンの形状より文字基準点を求め別に格納した標
準文字パターンの中心と合せてマスクマツチング
法により認識する文字認識方式において、認識不
能の文字が発生し再試行を行なう場合該文字パタ
ーンの基準点をすでに認識した文字パターンの基
準点から類推した位置に補正してマスクマツチン
グ法により再試行することを特徴とするものであ
る。 In order to achieve the above object, the character recognition method of the present invention finds a character reference point from the shape of a character pattern read by an optical character reading means, aligns it with the center of a separately stored standard character pattern, and recognizes the character using a mask matching method. When an unrecognized character occurs in a character recognition method and a retry is performed, the reference point of the character pattern is corrected to a position analogous to the reference point of the already recognized character pattern and the retry is performed using the mask matching method. It is characterized by:
以下本発明を実施例につき詳述する。 The present invention will be described in detail below with reference to examples.
第2図は本発明の原理説明図である。 FIG. 2 is a diagram explaining the principle of the present invention.
同図において「5」に対する読取・認識の結
果、たとえば第1図dに示す欠陥により認識不能
となり再試行が行なわれるものとする。この場合
従来の方法で中心○イを決定したのでは依然認識不
能となるから、本発明ではすでに認識された直上
の文字「0」の中心位置から類推して決定する。
すなわち左右は同一とし、上下は文字領域間の標
準ピツチZだけ下方に移した点○イ′を中心として
マスクマツチング法を適用する。このようにして
認識すると、たとえゴミや汚れ○ロがあつたとして
も文字としてはマツチングが許容範囲に入り、認
識可能とすることができる。 In the same figure, it is assumed that as a result of reading and recognizing "5", recognition becomes impossible due to the defect shown in FIG. 1d, for example, and a retry is performed. In this case, if the center ○I is determined using the conventional method, it will still be unrecognizable, so in the present invention, it is determined by analogy from the center position of the character "0" immediately above, which has already been recognized.
That is, the left and right sides are the same, and the mask matching method is applied to the top and bottom with a point ○a' shifted downward by the standard pitch Z between character areas as the center. When recognized in this way, even if there is dust or dirt, the matching is within an acceptable range and the characters can be recognized.
第3図は上述の原理に従う本発明の実施例の構
成を示す説明図である。 FIG. 3 is an explanatory diagram showing the configuration of an embodiment of the present invention according to the above-described principle.
同図において、用紙に印字された文字をOCR
等の読取部11で読取り、AD変換部12、スラ
イス回路によりたとえば2値情報に変換し、一旦
データ収納部13に格納する。この内容を1領域
ずつ文字パターン中心検出回路14に読出し、文
字パターンおよび前述の手法に基づいて文字パタ
ーンの左右xと上下yのそれぞれ1/2の位置から
求めた中心座標を比較回路15に送り、標準パタ
ーンテーブル16の各文字パターンの中心座標と
合せてマスクマツチング法を適用する。検出され
た文字パターンが中心を合せた標準パターンによ
りマスクされる程度が最も大きくかつこれがある
値以上の時認識可能(OK)であり、これが満足
されない時認識不能(NG)となる。認識可能
(OK)ならばその情報をインタフエース17を
介してCPUに通知し、たとえば投票券払戻しの
場合には払戻処理が行なわれる。認識不能
(NG)の場合はNGになつた文字の番地と文字パ
ターンとを文字中心補正部21に送り、あらかじ
めデータ収納部13のデータと同じ文字データ群
が格納されている記憶部からNG文字の番地を探
索し、たとえば前述のようにその直上の領域に位
置する認識可能であつた文字の中心座標(x1,
y1)に対し、x1はそのままとし、y1は標準ピツチ
Zだけ下方に移した座標(x1,y1+Z)をNG文
字の中心とし、その文字パターンとともに、比較
回路15に送り標準パターンの中心と合せて再び
マスクマツチング法を適用して再試行を行なう。 In the same figure, the characters printed on the paper are OCRed.
The data is read by a reading unit 11 such as the above, converted into binary information by an AD conversion unit 12 and a slice circuit, and temporarily stored in a data storage unit 13. This content is read out to the character pattern center detection circuit 14 one area at a time, and the center coordinates obtained from the left and right x and 1/2 up and down y positions of the character pattern based on the character pattern and the above-mentioned method are sent to the comparison circuit 15. , and apply the mask matching method to the center coordinates of each character pattern in the standard pattern table 16. When the degree to which the detected character pattern is masked by the centered standard pattern is greatest and exceeds a certain value, it is recognized (OK), and when this is not satisfied, it is unrecognizable (NG). If recognition is possible (OK), the information is notified to the CPU via the interface 17, and, for example, in the case of a voting ticket refund, a refund process is performed. If it is unrecognizable (NG), send the address and character pattern of the character that has become NG to the character center correction unit 21, and extract the NG character from the storage unit that stores the same character data group as the data in the data storage unit 13. For example, as mentioned above, the center coordinates (x 1 ,
y 1 ), x 1 is left as is, y 1 is shifted downward by the standard pitch Z, the coordinates (x 1 , y 1 + Z) are set as the center of the NG character, and the character pattern is sent to the comparison circuit 15 along with the standard pitch. Apply the mask matching method again to match the center of the pattern and try again.
以上説明したように、本発明によれば、マスク
マツチング法で認識不能となり再試行を行なう場
合、文字パターンの基準点を認識可能であつた近
傍の文字パターンの基準点から類推した位置に補
正することにより、従来認識不能として棄ててい
たかなりの文字を認識可能として救済でき読取率
を格段に向上するとともに、読取時間の短縮を図
ることができるものである。 As explained above, according to the present invention, when a character pattern becomes unrecognizable in the mask matching method and a retry is performed, the reference point of a character pattern is corrected to a position that is inferred from the reference point of a nearby character pattern that was recognizable. By doing so, it is possible to salvage a large number of characters that were conventionally rejected as unrecognizable by making them recognizable, thereby significantly improving the reading rate and shortening the reading time.
なお、パターンの基準点はパターン中心に限ら
れるものではない。 Note that the reference point of the pattern is not limited to the center of the pattern.
第1図a〜eは従来例の概略説明図、第2図は
本発明の原理説明図、第3図は本発明の実施例の
構成を示す説明図であり、図中、11は読取部、
12はA/D変換部、13はデータ収納部、14
は文字パターン中心検出回路、15は比較回路、
16は標準パターンテーブル、17はインタフエ
ース、21は文字中心補正部を示す。
1A to 1E are schematic explanatory diagrams of a conventional example, FIG. 2 is an explanatory diagram of the principle of the present invention, and FIG. 3 is an explanatory diagram showing the configuration of an embodiment of the present invention. ,
12 is an A/D conversion section, 13 is a data storage section, 14
15 is a character pattern center detection circuit, 15 is a comparison circuit,
16 is a standard pattern table, 17 is an interface, and 21 is a character center correction section.
Claims (1)
ーンの形状より文字基準点を求め、別に格納した
標準文字パターンの基準点と合せてマスクマツチ
ング法により認識する文字認識方式において、認
識不能の文字が発生し再試行を行なう場合該文字
パターンの基準点をすでに認識した文字パターン
の基準点から類推した位置に補正してマスクマツ
チング法により再試行することを特徴とする文字
認識方式。1 Character reference points are determined from the shape of a character pattern read by an optical character reading means, and unrecognized characters are A character recognition method characterized in that when a character pattern occurs and a retry is performed, the reference point of the character pattern is corrected to a position analogous to the reference point of a character pattern that has already been recognized, and a retry is performed using a mask matching method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP56088508A JPS57203177A (en) | 1981-06-09 | 1981-06-09 | Character recognizing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP56088508A JPS57203177A (en) | 1981-06-09 | 1981-06-09 | Character recognizing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPS57203177A JPS57203177A (en) | 1982-12-13 |
| JPH0117191B2 true JPH0117191B2 (en) | 1989-03-29 |
Family
ID=13944755
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP56088508A Granted JPS57203177A (en) | 1981-06-09 | 1981-06-09 | Character recognizing system |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPS57203177A (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2881762B2 (en) * | 1987-12-11 | 1999-04-12 | 住友電気工業株式会社 | Optical character reading device |
| KR100221459B1 (en) | 1997-03-07 | 1999-09-15 | 이형도 | Recording liquid ejecting apparatus and method thereof of print head |
| US6123414A (en) * | 1997-03-12 | 2000-09-26 | Samsung Electro-Mechanics Co., Ltd. | Apparatus for injecting a recording solution of a print head using phase transformation of thin film shape memory alloy |
| US6130689A (en) * | 1997-11-19 | 2000-10-10 | Samsung Electro-Mechanics Co., Ltd. | Apparatus and actuator for injecting a recording solution of a print head and method for producing the apparatus |
| CN108073925B (en) * | 2016-11-17 | 2021-09-17 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
-
1981
- 1981-06-09 JP JP56088508A patent/JPS57203177A/en active Granted
Also Published As
| Publication number | Publication date |
|---|---|
| JPS57203177A (en) | 1982-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5164996A (en) | Optical character recognition by detecting geo features | |
| JPS63216189A (en) | Character recognition system | |
| US4797940A (en) | Optical character reader | |
| US6694065B2 (en) | Image collating apparatus and image collating method | |
| JPH0117191B2 (en) | ||
| US4769851A (en) | Apparatus for recognizing characters | |
| JPS63211082A (en) | Character reader | |
| US5272765A (en) | System for processing character images | |
| JP2606816B2 (en) | Character reader | |
| JPS5949671A (en) | Optical character reader | |
| JPH0247786B2 (en) | ||
| JPH04167193A (en) | Character recognizing method | |
| JP2925270B2 (en) | Character reader | |
| JPH036552B2 (en) | ||
| JPH06187450A (en) | Pattern recognition method and recognition device | |
| JPH04338889A (en) | Method and device for character recognition | |
| JP2004013188A (en) | Business form reading device, business form reading method and program therefor | |
| JP2578767B2 (en) | Image processing method | |
| JPH01259476A (en) | Character reader | |
| JPH1196290A (en) | Slip identification device, slip identification method and medium in which slip identification program is recorded | |
| JPH07111729B2 (en) | Reader | |
| JPS6041396B2 (en) | Address detection device | |
| JPS61289479A (en) | Character recognizing device for reading out connected characters | |
| JPS60159986A (en) | Optical character reading device | |
| JPS6134683A (en) | Optical character reader |