JPH0143351B2 - - Google Patents
Info
- Publication number
- JPH0143351B2 JPH0143351B2 JP57175222A JP17522282A JPH0143351B2 JP H0143351 B2 JPH0143351 B2 JP H0143351B2 JP 57175222 A JP57175222 A JP 57175222A JP 17522282 A JP17522282 A JP 17522282A JP H0143351 B2 JPH0143351 B2 JP H0143351B2
- Authority
- JP
- Japan
- Prior art keywords
- noise
- difficulty
- image
- difficulty level
- recognition area
- 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
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/24—Character recognition characterised by the processing or recognition method
- G06V30/248—Character recognition characterised by the processing or recognition method involving plural approaches, e.g. verification by template match; Resolving confusion among similar patterns, e.g. "O" versus "Q"
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Description
【発明の詳細な説明】
(1) 発明の技術分野
本発明はビデオカメラ等により得られる認識領
域内の画像像から対象物の形状、位置を認識する
画像認識方法に関するものである。DETAILED DESCRIPTION OF THE INVENTION (1) Technical Field of the Invention The present invention relates to an image recognition method for recognizing the shape and position of an object from an image within a recognition area obtained by a video camera or the like.
(2) 従来技術と問題点
従来、ビデオカメラ等により得られる画像から
対象物の形状位置を認識する場合、各種の方法が
考えられる。たとえば、認識領域内に各種のノイ
ズとともに存在する対象物を行方向に走査し、画
素の背景“０”に対する“１”情報の変化点を検
出し、この点を基準点として輪郭線を追跡し、こ
れがノイズか対象物かを１つ１つ識別判定する。
これは大小さまざまのノイズに対し、単一のアル
ゴリズムが適用される代りに非常に時間がかかる
という欠点があつた。(2) Prior Art and Problems Conventionally, various methods have been considered when recognizing the shape position of an object from an image obtained by a video camera or the like. For example, an object that exists with various types of noise in the recognition area is scanned in the row direction, the point where the pixel information changes from "1" to the background "0" is detected, and the outline is traced using this point as the reference point. , identify and judge whether this is noise or an object one by one.
This had the disadvantage that a single algorithm was applied to noises of various sizes, and it was very time consuming.
本発明者はあらかじめ大きさの判つている場
合、ノイズの大きさにより幾つかの認識の難易度
を設定し、それぞれに適応したアルゴリズムを適
用することにより、全体の処理時間を格段に短縮
できることに着目したものである。 The inventor found that when the noise size is known in advance, the overall processing time can be significantly reduced by setting several recognition difficulty levels depending on the size of the noise and applying algorithms adapted to each. This is what we focused on.
(3) 発明の目的
本発明の目的は認識領域内に存在するノイズ大
きさに応じ対象物の認識の難易度を設定し、対象
物を短時間にかつ高精度に認識できる画像認識方
法を提供することである。(3) Purpose of the Invention The purpose of the present invention is to provide an image recognition method that can recognize a target object in a short time and with high accuracy by setting the difficulty level of recognizing the target object according to the size of noise existing in the recognition area. It is to be.
(4) 発明の構成
前記目的を達成するため、本発明の画像認識方
法は認識領域内の画像から対象物の形状、位置を
認識する画像認識方法において、該認識領領域内
に存在するノイズの大きさを対象物の形状と関連
して設定された複数の難易度により判定し、その
判定された難易度に応じた最適の処理手段を選択
し、対象物を認識することを特徴とするものであ
る。(4) Structure of the Invention In order to achieve the above object, the image recognition method of the present invention recognizes the shape and position of an object from an image within the recognition area. The object is recognized by determining the size of the object based on a plurality of difficulty levels set in relation to the shape of the object, and selecting the optimal processing means according to the determined difficulty level. It is.
(5) 発明の実施例
本発明の原理を説明すると、まず、認識領域内
が“１”の情報により成る対象物と、“０”の情
報より成る背景だけであり、ノイズが存在しない
場合は簡単なアルゴリズムＡで対象物を精度よく
認識することができる。しかし、対象物と同じ情
報をもつノイズが存在すると、ノイズを考慮して
いないアルゴリズムＡでは誤認識の可能性がある
から、このような対象物にはノイズを考慮してア
ルゴリズムＢで認識する必要があり、当然アルゴ
リズムＡより処理時間は長くなる。しかしこれら
を組合せると、総合的に処理時間を短縮すること
ができるものである。(5) Embodiments of the Invention To explain the principle of the present invention, first, if the recognition area contains only an object consisting of information of “1” and a background consisting of information of “0”, and there is no noise, A simple algorithm A allows objects to be recognized with high accuracy. However, if there is noise that has the same information as the target object, Algorithm A, which does not take noise into account, may misrecognize it, so it is necessary to recognize such objects using Algorithm B, which takes noise into account. Therefore, the processing time is naturally longer than that of Algorithm A. However, if these are combined, the processing time can be reduced overall.
以下実施例による難易度の設定とそれぞれに対
応する処理手段の概略を説明する。 The settings of the difficulty level and the corresponding processing means according to the embodiment will be explained below.
第１図ａは本発明の実施例を適用する画像例を
示す。 FIG. 1a shows an example image to which an embodiment of the invention is applied.
同図において、認識領域１内で“１”画素より
成る対象物として、ボンデイングパツド２の下部
梯形部を示す。そして行方向の底辺ｌの中心
（ｘ、ｙ）を求めるものとし、“０”画素より成る
背景領域３内に“１”画素より成る長さｍのノイ
ズ４が存在する場合を考える。この対象物の梯形
部を認識する場合の難易度をｌとｍとの関係にお
いて次の３種類に分ける。 In the figure, the lower trapezoidal portion of the bonding pad 2 is shown as an object consisting of "1" pixel within the recognition area 1. The center (x, y) of the base l in the row direction is determined, and a case is considered in which a noise 4 having a length m and consisting of "1" pixels exists in the background region 3 consisting of "0" pixels. The degree of difficulty in recognizing the trapezoidal part of the object is divided into the following three types based on the relationship between l and m.
難易度；ｍ＝０の場合
難易度；０＜ｍ＜ｌの場合
難易度；ｍ≧ｌの場合
これより、難易度の場合は、ノイズを考慮せ
ずにｌと同じ長さの行をサーチし、そこをＰ点の
ｙ座標とし、求めた行の左端“１”の画素を調べ
てゆき、その点にｌ／２を加えたところをｘ座標
としてＰ点を求める。Difficulty: if m=0 Difficulty; if 0<m<l Difficulty; if m≧l From this, in the case of difficulty, search for a line with the same length as l without considering noise. Then, using this as the y-coordinate of point P, check the leftmost "1" pixel of the found row, and add l/2 to that point to find point P as the x-coordinate.
難易度の場合は、ｙ座標は難易度１の場合と
同じでよいが、ｘ座標を同じように検出すると孤
立したノイズにより誤認識の可能性があるので、
同図ｂのような梯形部のコーナ検出用テンプレー
ト５でマツチングをとつてＰ点を求める。 In the case of difficulty level, the y-coordinate may be the same as in the case of difficulty level 1, but if the x-coordinate is detected in the same way, there is a possibility of misrecognition due to isolated noise, so
Point P is obtained by performing matching using a corner detection template 5 of a trapezoidal portion as shown in FIG.
難易度になると、ｌと同じ長さのノイズが存
在することになるから、ｙ座標を求める場合に、
各行の画素数の変化率を調べて決定する。これは
ボンデイングパツドの場合は同図ｂに示すよう
に、45度の角度（Ａ−A′）であるから、これに
応じた変化率である。ノイズの場合は不規則であ
るから区別できる。 When the difficulty level is reached, there will be noise with the same length as l, so when calculating the y coordinate,
Determine by examining the rate of change in the number of pixels in each row. In the case of the bonding pad, this is an angle (A-A') of 45 degrees, as shown in FIG. Noise can be distinguished because it is irregular.
また、ｘ座標は同図ｃに示す梯形部に合せたボ
ンデイングパツド用のテンプレート６でマツチン
グを行なつて決定する。 The x-coordinate is determined by matching using a bonding pad template 6 that matches the trapezoidal portion shown in FIG.
第２図は上述の原理に基づく本発明の実施例の
構成説明図である。 FIG. 2 is an explanatory diagram of the configuration of an embodiment of the present invention based on the above-mentioned principle.
同図において、TVカメラ１１で得られた認識
領域内の画像情報を画像処理装置１２で処理した
後、画像メモリ１３に格納する。この画像メモリ
１３から走査方向に従い画素情報を読出し、本発
明の要部の認識制御部２０内の画像２値化回路１
４で所定閾値を設けて２値化し、難易度判定回路
１５で前述のノイズの走査方向の長さに応じ難易
度、、を判定する。この判定された難易度
に応じ処理回路（）１６1、（）１６2、（）
１６3の１つを選択し、前述のそれぞれのアルゴ
リズムに従つた処理を行なう。その結果ノイズを
識別、除去し対象物のＸ、Ｙ座標のみをXYテー
ブル１７に格納する。このようにして対象物を高
速、高精度に認識することができる。 In the figure, image information within a recognition area obtained by a TV camera 11 is processed by an image processing device 12 and then stored in an image memory 13. The image binarization circuit 1 in the recognition control unit 20, which is the main part of the present invention, reads out pixel information from this image memory 13 in accordance with the scanning direction.
4, a predetermined threshold value is set and binarized, and the difficulty level determining circuit 15 determines the level of difficulty according to the length of the noise in the scanning direction. Depending on the determined difficulty level, the processing circuits ()16 1 , ()16 2 , ()
16 3 is selected and processed according to the respective algorithms described above. As a result, noise is identified and removed, and only the X and Y coordinates of the object are stored in the XY table 17. In this way, objects can be recognized at high speed and with high precision.
第３図は第２図の実施例の動作を示す流れ図で
ある。 FIG. 3 is a flowchart showing the operation of the embodiment of FIG.
同図において、画像入力後、本発明の認識制御
部２０により画素情報の２値化を行ない、難易度
の判定を行なうことは前述のとおりである。 In the figure, after inputting the image, the recognition control unit 20 of the present invention binarizes the pixel information and determines the difficulty level, as described above.
難易度の場合には、ノイズがないからｌに等
しい行を見付け、その行の両端から“１”に変化
する位置を抽出し対象物を認識する。 In the case of the difficulty level, a line equal to l is found since there is no noise, positions where the value changes to "1" are extracted from both ends of the line, and the object is recognized.
難易度の場合には、ｌに等しい行を見付け、
その行の両端からコーナ用テンプレートでマツチ
ングをとり、対象物を認識する。 For difficulty, find the row equal to l,
Matching is performed using corner templates from both ends of the row, and the object is recognized.
難易度の場合には、ｌを見付けるのに行方向
の画素の変化率を調べｌを決定する。そしてパツ
ドコーナ用のテンプレートでマツチングをとつて
両端を決定し、対象物を認識する。 In the case of difficulty, l is determined by examining the rate of change of pixels in the row direction. Then, matching is performed using a template for padded corners to determine both ends and the object is recognized.
(6) 発明の効果
以上説明したように、本発明によれば、認識領
域内に存在するノイズの大きさにより難易度を判
定し、この難易度に従つてそれぞれ対応するアル
ゴリズムによる処理手段を選択し処理するもので
ある。これにより、従来の単一のアルゴリズムに
よる処理手段で一律に処理するのに対し、全体と
して処理を簡略化するとともに、高速化すること
ができる。さらに、難易度に最適の処理手段を適
用できるから高精度の認識が可能となる。(6) Effects of the Invention As explained above, according to the present invention, the level of difficulty is determined based on the magnitude of noise existing in the recognition area, and processing means using the corresponding algorithm is selected according to this level of difficulty. It is then processed. This makes it possible to simplify and speed up the processing as a whole, compared to uniform processing using conventional processing means using a single algorithm. Furthermore, since it is possible to apply a processing means that is optimal for the difficulty level, highly accurate recognition is possible.
第１図ａ〜ｃは本発明の実施例の概略と要部の
説明図、第２図は本発明の実施例の構成説明図、
第３図は第２図の実施例の動作を示す流れ図であ
り、図中は、１は認識領域、２は対象物、３は背
景、４はノイズ、５，６はテンプレート、１１は
TVカメラ、１２は画像処理装置、１３は画像メ
モリ、１４は画像２値化回路、１５は難易度判定
回路、１６1〜１６3は処理回路、１７はＸ−Ｙテ
ーブルを示す。
FIGS. 1 a to c are schematic diagrams of an embodiment of the present invention and explanatory diagrams of essential parts; FIG. 2 is an explanatory diagram of the configuration of an embodiment of the present invention;
FIG. 3 is a flowchart showing the operation of the embodiment shown in FIG.
12 is an image processing device, 13 is an image memory, 14 is an image binarization circuit, 15 is a difficulty level determination circuit, 16 1 to 16 3 are processing circuits, and 17 is an XY table.
Claims (1)
認識する画像認識方法において、該認識領域内に
存在するノイズの大きさを対象物の形状と関連し
て設定された複数の難易度により判定し、その判
定された難易度に応じた最適の処理手段を選択
し、対象物を認識することを特徴とする画像認識
方法。 ２ 前記難易度を前記認識領域内のノイズの有無
と、該ノイズと対象物につき行方向の長さの大小
関係とにより設定したことを特徴とする特許請求
の範囲第１項記載の画像認識方法。[Claims] 1. In an image recognition method for recognizing the shape and position of an object from an image within a recognition area, the size of noise existing within the recognition area is set in relation to the shape of the object. An image recognition method characterized in that a target object is recognized by making a determination based on a plurality of difficulty levels and selecting an optimal processing means according to the determined difficulty level. 2. The image recognition method according to claim 1, wherein the difficulty level is set based on the presence or absence of noise in the recognition area and the relationship between the noise and the length of the object in the row direction. .
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP57175222A JPS5971581A (en) | 1982-10-05 | 1982-10-05 | Picture recognizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP57175222A JPS5971581A (en) | 1982-10-05 | 1982-10-05 | Picture recognizing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPS5971581A JPS5971581A (en) | 1984-04-23 |
| JPH0143351B2 true JPH0143351B2 (en) | 1989-09-20 |
Family
ID=15992419
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP57175222A Granted JPS5971581A (en) | 1982-10-05 | 1982-10-05 | Picture recognizing method |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPS5971581A (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3021556B2 (en) * | 1990-06-20 | 2000-03-15 | ソニー株式会社 | Video information processing apparatus and method |
| US5901255A (en) * | 1992-02-07 | 1999-05-04 | Canon Kabushiki Kaisha | Pattern recognition method and apparatus capable of selecting another one of plural pattern recognition modes in response to a number of rejects of recognition-processed pattern segments |
-
1982
- 1982-10-05 JP JP57175222A patent/JPS5971581A/en active Granted
Also Published As
| Publication number | Publication date |
|---|---|
| JPS5971581A (en) | 1984-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JPH0215102B2 (en) | ||
| KR930002347B1 (en) | Pattern position recognition device | |
| JPH02306386A (en) | Character recognizing device | |
| JPH04198741A (en) | Shape defect detecting device | |
| JPH0765149A (en) | Person recognition device | |
| JPH0143351B2 (en) | ||
| JP2637591B2 (en) | Position recognition apparatus and method | |
| JPH11219435A (en) | Automotive white line detector | |
| JPH065545B2 (en) | Figure recognition device | |
| JP2000194861A (en) | Image recognition method and apparatus | |
| JP3031069B2 (en) | Appearance inspection method | |
| JPH09147121A (en) | Robot object recognition method and apparatus | |
| JPS58129888A (en) | Circular end face position detection device | |
| JPH10312460A (en) | Image processing method and high-precision image processing device | |
| JPS61286704A (en) | Image border detection method | |
| JPH0115910B2 (en) | ||
| JPS62233705A (en) | Position detecting method | |
| JP3348938B2 (en) | Position recognition device | |
| JPH07104137B2 (en) | Method of detecting the rotation angle of the target pattern | |
| JPH065544B2 (en) | Figure recognition device | |
| JPH05113315A (en) | Center position detection method for circular image data | |
| JPS6360432B2 (en) | ||
| JP3522383B2 (en) | Work image processing device | |
| JPH04286085A (en) | Pattern recognizing method | |
| JPH0412510B2 (en) |