JPH10235584A - Article recognizing method and device therefor - Google Patents
Article recognizing method and device thereforInfo
- Publication number
- JPH10235584A JPH10235584A JP9057158A JP5715897A JPH10235584A JP H10235584 A JPH10235584 A JP H10235584A JP 9057158 A JP9057158 A JP 9057158A JP 5715897 A JP5715897 A JP 5715897A JP H10235584 A JPH10235584 A JP H10235584A
- Authority
- JP
- Japan
- Prior art keywords
- article
- articles
- image
- line
- adjacent
- 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.)
- Withdrawn
Links
Landscapes
- Closed-Circuit Television Systems (AREA)
- Image Analysis (AREA)
- Manipulator (AREA)
- Image Processing (AREA)
Abstract
Description
【０００１】[0001]
【発明の属する技術分野】本発明は、ボトル等の物品認
識方法及び装置に関する。The present invention relates to a method and an apparatus for recognizing articles such as bottles.
【０００２】[0002]
【従来の技術】従来、物品同士が接触していると、撮像
手段により撮像された物品の２値化画像が単一物品の輪
郭を表わすものと異なるものとなり、物品の認識ができ
ない。従って、接触している物品は当該物品であること
の認識がなされずに見逃され、認識効率が低下する。ま
た、上述の物品の接触による認識不能を回避するために
前工程に物品の接触を解く物品切り離し機構を設ける必
要がある。或いは、上述の物品の接触による認識不能に
よって見逃された物品を回収して、再度物品の認識に供
するために物品回収コンベヤを設けることも必要にな
る。それを解決する方法として、ボトル等の物品の重心
位置や向きを認識する物品認識方法として、物品の内部
の基本画像（絵柄）のパターンマッチによるものがある
（図８）（従来技術１）。2. Description of the Related Art Conventionally, when articles are in contact with each other, the binarized image of the article taken by the image pickup means differs from that representing the outline of a single article, and the article cannot be recognized. Therefore, the contacting article is overlooked without being recognized as the article, and the recognition efficiency is reduced. Further, in order to avoid the above-mentioned recognition failure due to the contact of the article, it is necessary to provide an article separating mechanism for releasing the contact of the article in the previous process. Alternatively, it is also necessary to provide an article collection conveyor to collect the articles missed due to the inability to recognize due to the above-mentioned article contact and to use the articles again for recognition. As a method for solving this, there is an article recognition method for recognizing the position and orientation of the center of gravity of an article such as a bottle by pattern matching of a basic image (picture) inside the article (FIG. 8) (prior art 1).
【０００３】また、特開平6-309459号公報には、各種形
状パターンが互いに接触又は重なり合うことにより、外
部輪郭線と内部輪郭線とが存在する場合であっても、外
部輪郭線から幾何学計算方法によって、内部輪郭線を抽
出し、外部輪郭線及び内部輪郭線から円及び四角形の幾
何学的特徴によって図形を認識し、被認識形状パターン
から各種形状パターンを迅速に認識するものもある（従
来技術２）。[0003] Japanese Patent Application Laid-Open No. 6-309459 discloses that even when an external contour and an internal contour are present, various geometric patterns come into contact with each other or overlap with each other. According to some methods, an internal contour is extracted, a figure is recognized based on a geometric feature of a circle and a rectangle from the external contour and the internal contour, and various shape patterns are quickly recognized from a recognized shape pattern. Technology 2).
【０００４】[0004]
【発明が解決しようとする課題】然しながら、従来技術
１では、下記、の問題点がある。 物品の内部特徴（絵柄等）によるパターンマッチなの
で、内部特徴の無い物品は認識できない。However, the prior art 1 has the following problems. Since there is a pattern match based on the internal features (pictures, etc.) of the article, an article having no internal features cannot be recognized.
【０００５】全方向に対するマッチングが必要であっ
たため、予め記憶しておく情報も多く、処理時間も掛か
っていた（例えば約1.5 秒）。[0005] Since matching in all directions is necessary, much information is stored in advance and processing time is required (for example, about 1.5 seconds).
【０００６】また、従来技術２では、外部輪郭線及び内
部輪郭線から円及び四角形の幾何学的特徴によって図形
を認識するものであるため、円形もしくは四角形の認識
に限定される。Further, in the prior art 2, since a figure is recognized from the outer contour and the inner contour by geometrical features of a circle and a rectangle, the recognition is limited to the recognition of a circle or a rectangle.
【０００７】本発明の課題は、物品同士が接触している
場合にも、任意の形状の物品を、高速で認識できるよう
にすることにある。An object of the present invention is to enable an article of an arbitrary shape to be recognized at a high speed even when the articles are in contact with each other.
【０００８】[0008]
【課題を解決するための手段】請求項１に記載の本発明
は、 2個以上の互いに接触している物品をそれぞれ独立
の物品として認識する物品認識方法において、物品の幾
何学的特徴量を予め登録しておき、撮像手段により物品
群の画像を取り込み、物品の画像の外部輪郭線を抽出
し、この外部輪郭線が急激に折れる変曲点を求め、 2個
の変曲点を結ぶ直線の両側で、上記外部輪郭線に囲まれ
る面積を求め、それらの面積が物品の幾何学的特徴量と
して予め定められている物品の面積の整数倍であれば、
その直線を分離線と判定し、隣接する物品を上記分離線
により切り離し、それらの物品を独立させて認識するよ
うにしたものである。According to a first aspect of the present invention, there is provided an article recognition method for recognizing two or more articles in contact with each other as independent articles. Registered in advance, the image of the article group is taken in by the imaging means, the outer contour of the article image is extracted, the inflection point at which this outer contour sharply breaks is found, and a straight line connecting the two inflection points On both sides, determine the area surrounded by the outer contour, if those areas are integral multiples of the area of the article predetermined as the geometric feature of the article,
The straight line is determined as a separation line, adjacent articles are separated by the separation line, and those articles are recognized independently.
【０００９】請求項２に記載の本発明は、請求項１に記
載の本発明において更に、前記隣接する物品の分離処理
に先立ち、物品の画像の面積値から隣接する物品の有無
と、その物品数を求め、上記隣接する物品数だけ物品の
分離処理を行なうようにしたものである。According to a second aspect of the present invention, in addition to the first aspect of the present invention, prior to the separation processing of the adjacent article, the presence or absence of the adjacent article is determined based on the area value of the image of the article. The number of the articles is obtained, and the articles are separated by the number of adjacent articles.
【００１０】請求項３に記載の本発明は、請求項１又は
２に記載の本発明において更に、前記独立させて認識し
た物品の外部輪郭線から物品の重心と慣性主軸を求め、
この物品の重心と慣性主軸を用いる幾何学的演算から、
物品の位置と向きを認識するようにしたものである。According to a third aspect of the present invention, in the first or second aspect of the present invention, the center of gravity and the principal axis of inertia of the article are obtained from the external contour of the article independently recognized,
From the geometric operation using the center of gravity of this article and the principal axis of inertia,
The position and orientation of the article are recognized.
【００１１】請求項４に記載の本発明は、請求項３に記
載の本発明において更に、前記物品認識情報を、後工程
の物品ハンドリング装置に転送するようにしたものであ
る。According to a fourth aspect of the present invention, in the third aspect of the present invention, the article recognition information is further transferred to an article handling apparatus in a subsequent process.
【００１２】請求項５に記載の本発明は、 2個以上の互
いに接触している物品をそれぞれ独立の物品として認識
する物品認識装置において、物品の幾何学的特徴量を予
め登録するデータ記憶部と、撮像手段により取り込んだ
物品群の画像を記録する画像メモリと、画像メモリの記
録データとデータ記憶部の登録データとを得て、物品の
画像の外部輪郭線を抽出し、この外部輪郭線が急激に折
れる変曲点を求め、 2個の変曲点を結ぶ直線の両側で、
上記外部輪郭線に囲まれる面積を求め、それらの面積が
物品の幾何学的特徴量として予め定められている物品の
面積の整数倍であれば、その直線を分離線と判定し、隣
接する物品を上記分離線により切り離し、それらの物品
を独立させて認識する演算部とを有してなるようにした
ものである。According to a fifth aspect of the present invention, there is provided an article recognizing apparatus for recognizing two or more articles in contact with each other as independent articles, wherein a data storage unit for pre-registering a geometric feature of the article. And an image memory for recording an image of the article group captured by the imaging means, and recording data of the image memory and registration data of the data storage unit, and extracting an outer contour of the article image. Finds the inflection point at which it breaks sharply, and on both sides of the straight line connecting the two inflection points,
The area surrounded by the outer contour line is obtained, and if those areas are integral multiples of the area of the article predetermined as the geometric feature of the article, the straight line is determined as a separation line, and the adjacent article is determined. Are separated by the above-mentioned separation line, and an arithmetic unit for recognizing those articles independently is provided.
【００１３】請求項６に記載の本発明は、請求項５に記
載の本発明において更に、前記演算部が、前記隣接する
物品の分離処理に先立ち、物品の画像の面積値から隣接
する物品の有無と、その物品数を求め、上記隣接する物
品数だけ物品の分離処理を行なうようにしたものであ
る。According to a sixth aspect of the present invention, in the fifth aspect of the present invention, the arithmetic unit determines the adjacent article based on the area value of the article image prior to the separation processing of the adjacent article. The presence / absence and the number of articles are obtained, and the articles are separated by the number of adjacent articles.
【００１４】請求項７に記載の本発明は、請求項５又は
６に記載の本発明において更に、前記演算部が、前記独
立させて認識した物品の外部輪郭線から物品の重心と慣
性主軸を求め、この物品の重心と慣性主軸を用いる幾何
学的演算から、物品の位置と向きを認識するようにした
ものである。According to a seventh aspect of the present invention, in the fifth or sixth aspect of the present invention, the calculation unit further calculates the center of gravity and the principal axis of inertia of the article from the external contour of the article independently recognized. Then, the position and orientation of the article are recognized from the geometric operation using the center of gravity of the article and the principal axis of inertia.
【００１５】請求項８に記載の本発明は、請求項７に記
載の本発明において更に、前記演算部が、前記物品認識
情報を、後工程の物品ハンドリング装置に転送するよう
にしたものである。According to an eighth aspect of the present invention, in the seventh aspect of the present invention, the operation unit transfers the article recognition information to an article handling apparatus in a subsequent process. .
【００１６】本発明によれば、下記〜の作用があ
る。 絵柄パターンのマッチングに限らず、幾何学的特徴量
による演算で認識できるものであり、情報量が少なく、
処理を高速化（例えば約0.03秒）できる。According to the present invention, the following effects are obtained. It can be recognized not only by the pattern pattern matching but also by the calculation based on the geometric feature amount, the information amount is small,
Processing can be accelerated (for example, about 0.03 seconds).
【００１７】同一形状物品の接触であれば、任意の形
状の物品を認識できる。An article of an arbitrary shape can be recognized as long as the articles have the same shape.
【００１８】隣接する物品の分離処理に先立ち、物品
の画像の面積値から隣接する物品の有無と、その物品数
を求めることにより、物品の認識の確実を図ることがで
きる。Prior to the separation process of the adjacent articles, the presence or absence of the adjacent articles and the number of the articles are obtained from the area value of the image of the articles, so that the recognition of the articles can be ensured.
【００１９】分離処理した物品の外部輪郭線から物品
の重心と慣性主軸を求め、この物品の重心と慣性主軸を
用いる幾何学的演算から、物品の位置と向きを認識する
ことにより、物品同士が接触していた場合にも、それら
物品の位置と向きを認識できる。The center of gravity and the principal axis of inertia of the article are obtained from the outer contour line of the separated article, and the positions and orientations of the articles are recognized from the geometrical operation using the center of gravity and the principal axis of inertia of the article. Even in the case of contact, the position and orientation of those articles can be recognized.
【００２０】[0020]
【発明の実施の形態】図１は本発明の一実施形態を示す
模式図、図２は画像処理装置の画像処理フローを示す模
式図、図３は隣接物品の分離処理原理を示す模式図、図
４は隣接物品の分離処理手順を示す模式図、図５は物品
の方向・位置検出原理を示す模式図、図６は物品の認識
手順を示す流れ図、図７は隣接物品の分離処理手順を示
す流れ図、図８は従来方法を示す模式図である。DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a schematic diagram showing an embodiment of the present invention, FIG. 2 is a schematic diagram showing an image processing flow of an image processing apparatus, FIG. FIG. 4 is a schematic diagram showing a procedure for separating adjacent articles, FIG. 5 is a schematic diagram showing the principle of detecting the direction and position of an article, FIG. 6 is a flowchart showing a procedure for recognizing articles, and FIG. FIG. 8 is a schematic diagram showing a conventional method.
【００２１】物品取扱い装置１０は、図１に示す如く、
供給コンベヤ１１と排出コンベヤ１２を有する。そし
て、物品取扱い装置１０は、供給コンベヤ１１が供給す
るボトル１を一定の向き（底部を排出方向に向ける方
向）に整列させた状態にて排出コンベヤ１２に供給す
る。The article handling apparatus 10 is, as shown in FIG.
It has a supply conveyor 11 and a discharge conveyor 12. Then, the article handling apparatus 10 supplies the bottles 1 supplied by the supply conveyor 11 to the discharge conveyor 12 in a state where the bottles 1 are aligned in a fixed direction (a direction in which the bottom is directed to the discharge direction).
【００２２】このとき、物品取扱い装置１０は、ハンド
リング装置２０の前工程に物品認識装置３０を有してお
り、供給コンベヤ１１にて供給されるボトル１の向きと
重心位置を認識し、この認識情報を後工程のハンドリン
グ装置２０に転送する。ハンドリング装置２０は、物品
認識装置３０の認識情報を得て、後述するハンド２２に
より各ボトル１の重心位置を吸着し、各ボトル１のボト
ル底部を排出方向に向けるようにハンド２２を旋回し、
それらのボトル１を排出コンベヤ１２に供給する。At this time, the article handling apparatus 10 has an article recognition apparatus 30 in a process preceding the handling apparatus 20, and recognizes the direction and the center of gravity of the bottle 1 supplied by the supply conveyor 11, and recognizes this. The information is transferred to the handling device 20 in the subsequent process. The handling device 20 obtains the recognition information of the article recognition device 30, sucks the position of the center of gravity of each bottle 1 with the hand 22 described later, and turns the hand 22 so that the bottle bottom of each bottle 1 faces the discharge direction.
The bottles 1 are supplied to a discharge conveyor 12.
【００２３】以下、ハンドリング装置２０と物品認識装
置３０について説明する。 （ハンドリング装置２０）（図１〜図３）ハンドリング
装置２０は、図１に示す如く、ロボット２１のハンド２
２をＸＹ方向に移動可能とするとともに、旋回角度を制
御可能とし、ボトル１を吸着するための吸着パッドを昇
降自在に備える。Hereinafter, the handling device 20 and the article recognition device 30 will be described. (Handling Device 20) (FIGS. 1 to 3) As shown in FIG.
2 can be moved in the XY directions, the turning angle can be controlled, and a suction pad for sucking the bottle 1 is provided so as to be movable up and down.
【００２４】即ち、ハンドリング装置２０は、ロボット
２１の移動動作により、供給コンベヤ１１上の各ボトル
１をハンド２２に吸着した後、ハンド２２の旋回により
ボトル１の向きを整えるとともに、ロボット２１の移動
動作によりそれらボトル１を排出コンベヤ１２の排出部
のいずれかの所定位置に移送して供給する。That is, the handling device 20 adsorbs each bottle 1 on the supply conveyor 11 to the hand 22 by the movement operation of the robot 21, and then adjusts the direction of the bottle 1 by turning the hand 22, and moves the robot 21. By operation, these bottles 1 are transferred to any predetermined position of the discharge section of the discharge conveyor 12 and supplied.
【００２５】（物品認識装置３０）（図１〜図７）物品
認識装置３０は、図１に示す如く、カメラ３１（撮像手
段）と、画像処理装置３２とを有して構成される。そし
て、画像処理装置３２は、画像メモリ３３と、データ記
憶部３５と、演算部３６とを有する。演算部３６は、隣
接ボトル分離処理部３７を有する。(Article Recognition Apparatus 30) (FIGS. 1 to 7) As shown in FIG. 1, the article recognition apparatus 30 includes a camera 31 (imaging means) and an image processing apparatus 32. The image processing device 32 includes an image memory 33, a data storage unit 35, and a calculation unit 36. The calculation unit 36 includes an adjacent bottle separation processing unit 37.
【００２６】カメラ３１は供給コンベヤ１１の全幅を視
野とし、供給コンベヤ１１を搬送されてくるボトル１群
の画像を取り込む。The camera 31 views the entire width of the supply conveyor 11 and captures an image of a group of bottles conveyed on the supply conveyor 11.
【００２７】画像メモリ３３はカメラ３１から取り込ん
だボトル１群の画像（図２（Ａ））を記録する。The image memory 33 records an image of the group of bottles 1 taken from the camera 31 (FIG. 2A).
【００２８】データ記憶部３５は、ボトル１の幾何学的
特徴量（独立ボトル１の面積Ｓ、周囲長Ｌ等）を予め登
録する。The data storage unit 35 registers in advance the geometric features of the bottle 1 (the area S of the independent bottle 1, the perimeter L, etc.).
【００２９】演算部３６は、画像メモリ３３の記録デー
タと、データ記憶部３５の登録データとを得て、隣接ボ
トル分離処理部３７により接触しているボトル１、１の
分離処理を行なう。ボトル１の画像の面積値から隣接す
るボトル１の有無（接触の有無）と、そのボトル数Ｎを
求める（図２（Ｂ））。そして、演算部３６は、ボトル
１の画像の外部輪郭線を抽出し、この外部輪郭線が外角
θA で急激に折れる変曲点Ａを求める。このとき、図３
において、θA ＝tan-1 ＡＣ／ＡＢであり、θA ＜180
度（好ましくはθA ＜135 度）となる点を変曲点とす
る。更に、演算部３６は、ボトル１の外部輪郭線上の 2
個の変曲点Ａ1 、Ａ2 を結ぶ直線の両側で、上記外部輪
郭線に囲まれる面積Ｓ1 、Ｓ2 を求め、それらの面積Ｓ
1 、Ｓ2 のボトル１の幾何学的特徴量として予め定めら
れているボトル１の面積Ｓの整数倍であれば、その直線
Ａ1 Ａ2 を分離線Ｋと判定とする（図３）。これによ
り、演算部３６は、隣接するボトル１、１を上記分離線
Ｋにより切り離し、それらのボトル１、１を独立させて
認識する（図２（Ｃ））。尚、演算部３６は、上述した
隣接するボトル数Ｎだけ、ボトル１の分離処理を繰り返
して行なう。The arithmetic unit 36 obtains the recorded data in the image memory 33 and the registered data in the data storage unit 35, and performs the separation processing of the bottles 1 and 1 in contact by the adjacent bottle separation processing unit 37. The presence or absence of the adjacent bottle 1 (the presence or absence of contact) and the number N of the bottles are obtained from the area value of the image of the bottle 1 (FIG. 2B). The arithmetic unit 36 extracts the external contour line of the image bottle 1, obtaining the inflection point A rapidly break in the external outline external angle theta A. At this time, FIG.
Where θ A = tan −1 AC / AB and θ A <180
A point where the degree (preferably θ A <135 degrees) is taken as an inflection point. Further, the calculation unit 36 calculates the 2
On both sides of the straight line connecting the inflection points A 1 and A 2 , the areas S 1 and S 2 surrounded by the external contour are obtained, and the areas S 1 and S 2 are obtained.
If it is an integral multiple of the area S of the bottle 1 which is predetermined as the geometric feature of the bottle 1 of S 1 and S 2 , the straight line A 1 A 2 is determined to be the separation line K (FIG. 3). As a result, the arithmetic unit 36 separates the adjacent bottles 1 and 1 by the separation line K, and recognizes the bottles 1 and 1 independently (FIG. 2C). The arithmetic unit 36 repeats the bottle 1 separation process for the number N of adjacent bottles described above.
【００３０】また、演算部３６は、隣接ボトル分離処理
部３７の分離処理により独立させて認識したボトル１の
外部輪郭線（外形パターン）からボトル１の重心Ｇと慣
性主軸Ｉs を求め、このボトル１の重心Ｇと慣性主軸Ｉ
s を用いる幾何学的演算から、ボトル１の位置と向きを
認識する。ボトル１の位置は重心Ｇの位置そのものであ
る。ボトル１の向きは、ボトル１の全高をｈとすると
き、重心Ｇから慣性主軸Ｉs 方向にｈ／2 の点Ｐa が外
部輪郭線内にある方が口部側、外部輪郭線外にある点Ｐ
b が底部側とする（図５）。[0030] The calculation unit 36 obtains the center of gravity G and the principal axes of inertia I s of the bottle 1 from the outer contour line of the bottle 1 which is recognized by independently by separate processing of the adjacent bottle separation processing unit 37 (outline pattern), this Center of gravity G of bottle 1 and principal axis of inertia I
The position and orientation of the bottle 1 are recognized from the geometric operation using s . The position of the bottle 1 is the position of the center of gravity G itself. Bottle 1 orientation, when the total height of the bottle 1 is h, it is the mouth portion of the principal axis of inertia I s direction to a point P a of h / 2 from the center of gravity G is in the outer contour line, outside the outer contour line A point P
Let b be the bottom side (FIG. 5).
【００３１】また、演算部３６は、ボトル１の位置と向
きの認識情報を、後工程のハンドリング装置２０に転送
する。The arithmetic unit 36 transfers the recognition information of the position and the orientation of the bottle 1 to the handling device 20 in the subsequent process.
【００３２】以下、物品認識装置３０による物品認識手
順について説明する（図２、図６）。 (1) 供給コンベヤ１１の一定の送りに対応するタイミン
グ信号により、カメラ３１から画像を取り込み、これを
画像メモリ３３に記録する（図２（Ａ））。The article recognition procedure by the article recognition apparatus 30 will be described below (FIGS. 2 and 6). (1) An image is fetched from the camera 31 by a timing signal corresponding to a fixed feed of the supply conveyor 11, and is recorded in the image memory 33 (FIG. 2A).
【００３３】(2) 画像メモリ３３の記録データとデータ
記憶部３５の登録データを得て、演算部３６の隣接ボト
ル分離処理部３７により、隣接しているボトル１の分離
処理を行なう（図２（Ｂ））。隣接するボトル１、１は
分離線Ｋにより切り離され、それらのボトル１、１は互
いに独立して認識される。(2) The recording data of the image memory 33 and the registration data of the data storage unit 35 are obtained, and the adjacent bottle separation processing unit 37 of the calculation unit 36 performs the separation processing of the adjacent bottles 1 (FIG. 2). (B)). Adjacent bottles 1,1 are separated by separation lines K, and these bottles 1,1 are recognized independently of each other.
【００３４】(3) 独立して認識されたボトル１の重心Ｇ
と慣性主軸Ｉs 等の幾何学的演算から、ボトル１の位置
（重心Ｇの位置）と向きを求める（図２（Ｃ））。(3) Center of gravity G of bottle 1 independently recognized
Then, the position (position of the center of gravity G) and the orientation of the bottle 1 are obtained from the geometrical operation of the inertia main axis Is and the like (FIG. 2C).
【００３５】(4) 上記(3) で求めた認識情報（向き、重
心位置）を後工程のハンドリング装置２０に転送し、そ
のボトル１に対し前述のハンドリング操作を加え、これ
を排出コンベヤ１２のいずれかの排出部に供給する。(4) The recognition information (direction and center of gravity) obtained in the above (3) is transferred to the handling device 20 in the post-process, and the above-described handling operation is performed on the bottle 1, and this is transferred to the discharge conveyor 12. Supply to any of the outlets.
【００３６】(5) 上記(1) 〜(4) を繰り返す。(5) The above (1) to (4) are repeated.
【００３７】次に、演算部３６の隣接ボトル分離処理部
３７による隣接ボトルの分離処理手順について説明する
（図４、図７）。 (1) ボトル１の画像の面積値から隣接するボトル１の有
無と、そのボトル数Ｎを求める。Next, a procedure for separating adjacent bottles by the adjacent bottle separation processing unit 37 of the arithmetic unit 36 will be described (FIGS. 4 and 7). (1) The presence or absence of the adjacent bottle 1 and the number N of the bottles are obtained from the area value of the image of the bottle 1.
【００３８】(2) ボトル１の画像の外部輪郭線を抽出す
る。(2) The outer contour line of the image of the bottle 1 is extracted.
【００３９】(3) 上記(2) の外部輪郭線上の各点での外
角θを算出し、この外角θA が180度（好ましくは135
度）より小さい点を変曲点Ａ（図４（Ａ）の点１〜点１
０）を求める。(3) The outer angle θ at each point on the outer contour line in the above (2) is calculated, and the outer angle θ A is 180 degrees (preferably 135 degrees).
Point) (point 1 to point 1 in FIG. 4A).
0).
【００４０】(4) 上記(3) で求めた変曲点Ａの組み合わ
せから、隣接するボトル１、１を切り離すための分離候
補線を算出する。ここで、平面図形の幾何学的特徴から
2つの図形が 0.5Ｌ以上の接触面を持たないことより、
変曲点間の外部輪郭線に沿う長さは 0.5Ｌ以下にはなり
えない。即ち、変曲点間の外部輪郭線に沿う長さが0.5
Ｌ（Ｌ：ボトル１の周囲長）以上ある変曲点の組み合わ
せの中で、 2個の変曲点を結ぶ直線的距離を算出し、こ
の直線的距離の小さい順に、図４（Ｂ）に示す如くに並
べ、これらの組み合わせに係る変曲点間直線を分離候補
線とする。(4) From the combination of the inflection points A obtained in the above (3), a separation candidate line for separating the adjacent bottles 1 and 1 is calculated. Here, from the geometric features of the planar figure
Since the two figures do not have a contact surface of 0.5L or more,
The length along the outer contour line between the inflection points cannot be less than 0.5L. That is, the length along the external contour line between the inflection points is 0.5
Among the combinations of inflection points that are equal to or greater than L (L: perimeter of the bottle 1), the linear distance connecting the two inflection points is calculated, and FIG. The lines are arranged as shown, and the straight line between inflection points according to these combinations is set as a separation candidate line.
【００４１】(5) 上記(4) の分離候補線の並びの中で、
最先の変曲点間直線の両側で、外部輪郭線に囲まれる面
積Ｓ1 、Ｓ2 を求め、それらの面積Ｓ1 、Ｓ2 がボトル
１の幾何学的特徴量として予め定められているボトル１
の面積Ｓの整数倍であれば、その分離候補線を正しい分
離線Ｋとして判定する（図４（Ｃ））。(5) In the arrangement of the candidate separation lines in the above (4),
On both sides of the first inflection point straight line, the areas S 1 and S 2 surrounded by the outer contour are obtained, and the areas S 1 and S 2 are predetermined as the geometric feature of the bottle 1. Bottle 1
If the area S is an integral multiple of the area S, the separation candidate line is determined as a correct separation line K (FIG. 4C).
【００４２】尚、上述の算出面積Ｓ1 （又はＳ2 ）がボ
トル１の面積Ｓと同等（ボトル１の1本分の面積）のと
き、その面積Ｓ1 （又はＳ2 ）の周囲長を求め、この算
出周囲長が前述のＬであることを上述の分離線Ｋの判定
条件として付加しても良い。これは分離線Ｋの判定精度
を向上するものである。When the calculated area S 1 (or S 2 ) is equal to the area S of the bottle 1 (the area of one bottle 1), the peripheral length of the area S 1 (or S 2 ) The calculated perimeter may be added to the above-described L as a determination condition of the above-described separation line K. This improves the determination accuracy of the separation line K.
【００４３】(6) 上記分離線の判定結果が最先の分離候
補線（4-7)を正しい分離線Ｋと認識しないもの（ＮＧ）
であれば、前記(4) の分離候補線の並びの中で次の分離
候補線(4-6) について上記(5) を行なう（図４
（Ｄ））。(6) When the separation line determination result does not recognize the earliest separation candidate line (4-7) as a correct separation line K (NG)
If so, the above (5) is performed for the next separation candidate line (4-6) in the arrangement of the separation candidate lines (4) (FIG. 4).
(D)).
【００４４】上記分離線の判定結果が最先の分離候補線
(4-7) を正しい分離線Ｋと認識するもの（ＯＫ）であれ
ば、前記(4) の分離候補線の並びの中でその分離線Ｋの
両端変曲点Ａ1 、Ａ2 (4、7) を含まない次の分離候補線
(3-10 ) について上記(5) を行なう（図４（Ｅ））。The separation line determination result is the earliest separation candidate line.
If (4-7) is recognized as a correct separation line K (OK), the inflection points A 1 , A 2 (4 , 7) next candidate line
The above (5) is performed for (3-10) (FIG. 4E).
【００４５】(7) 前記(1) で求めてある隣接ボトル１の
数Ｎだけの分離線の抽出（Ｎ− 1個の分離線の抽出）
を、上記(5) 、(6) の繰り返しにより行なう。(7) Extraction of separation lines for the number N of adjacent bottles 1 obtained in (1) above (extraction of N-1 separation lines)
Is performed by repeating the above (5) and (6).
【００４６】従って、本実施形態によれば、下記〜
の作用がある。 絵柄パターンのマッチングに限らず、幾何学的特徴量
による演算で認識できるものであり、情報量が少なく、
処理を高速化（例えば約0.03秒）できる。Therefore, according to the present embodiment, the following:
Has the effect of It can be recognized not only by the pattern pattern matching but also by the calculation based on the geometric feature amount, the information amount is small,
Processing can be accelerated (for example, about 0.03 seconds).
【００４７】同一形状物品の接触であれば、任意の形
状のボトル１を認識できる。As long as the articles have the same shape, the bottle 1 having an arbitrary shape can be recognized.
【００４８】隣接するボトル１の分離処理に先立ち、
ボトル１の画像の面積値から隣接するボトル１の有無
と、その物品数を求めることにより、ボトル１の認識の
確実を図ることができる。Prior to the separation of adjacent bottles 1,
By determining the presence / absence of an adjacent bottle 1 and the number of articles of the adjacent bottle 1 from the area value of the image of the bottle 1, the recognition of the bottle 1 can be assured.
【００４９】分離処理したボトル１の外部輪郭線から
ボトル１の重心と慣性主軸を求め、このボトル１の重心
と慣性主軸を用いる幾何学的演算から、ボトル１の位置
と向きを認識することにより、ボトル１同士が接触して
いた場合にも、それらボトル１の位置と向きを認識でき
る。The center of gravity and the principal axis of inertia of the bottle 1 are obtained from the outer contour of the bottle 1 subjected to the separation processing, and the position and orientation of the bottle 1 are recognized from the geometrical operation using the center of gravity of the bottle 1 and the principal axis of inertia. Also, even when the bottles 1 are in contact with each other, the position and orientation of those bottles 1 can be recognized.
【００５０】以上、本発明の実施の形態を図面により詳
述したが、本発明の具体的な構成はこの実施の形態に限
られるものではなく、本発明の要旨を逸脱しない範囲の
設計の変更等があっても本発明に含まれる。The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration of the present invention is not limited to this embodiment, and the design may be changed without departing from the scope of the present invention. The present invention is also included in the present invention.
【００５１】[0051]
【発明の効果】以上のように本発明によれば、物品同士
が接触している場合にも、任意の形状の物品を、高速で
認識できる。As described above, according to the present invention, an article having an arbitrary shape can be recognized at a high speed even when the articles are in contact with each other.
【図１】図１は本発明の一実施形態を示す模式図であ
る。FIG. 1 is a schematic diagram showing one embodiment of the present invention.
【図２】図２は画像処理装置の画像処理フローを示す模
式図である。FIG. 2 is a schematic diagram illustrating an image processing flow of the image processing apparatus.
【図３】図３は隣接物品の分離処理原理を示す模式図で
ある。FIG. 3 is a schematic diagram illustrating a principle of a separation process of adjacent articles.
【図４】図４は隣接物品の分離処理手順を示す模式図で
ある。FIG. 4 is a schematic diagram showing a procedure for separating adjacent articles.
【図５】図５は物品の方向・位置検出原理を示す模式図
である。FIG. 5 is a schematic diagram illustrating the principle of detecting the direction and position of an article.
【図６】図６は物品の認識手順を示す流れ図である。FIG. 6 is a flowchart showing an article recognition procedure.
【図７】図７は隣接物品の分離処理手順を示す流れ図で
ある。FIG. 7 is a flow chart showing a procedure for separating adjacent articles.
【図８】図８は従来方法を示す模式図である。FIG. 8 is a schematic diagram showing a conventional method.
１ ボトル（物品） ２０ ハンドリング装置 ３０ 物品認識装置 ３１ カメラ（撮像手段） ３３ 画像メモリ ３５ データ記憶部 ３６ 演算部 ３７ 隣接ボトル分離処理部 DESCRIPTION OF SYMBOLS 1 Bottle (article) 20 Handling device 30 Article recognition device 31 Camera (imaging means) 33 Image memory 35 Data storage unit 36 Operation unit 37 Adjacent bottle separation processing unit
Claims (8)
れぞれ独立の物品として認識する物品認識方法におい
て、 物品の幾何学的特徴量を予め登録しておき、 撮像手段により物品群の画像を取り込み、 物品の画像の外部輪郭線を抽出し、この外部輪郭線が急
激に折れる変曲点を求め、 2個の変曲点を結ぶ直線の両側で、上記外部輪郭線に囲
まれる面積を求め、それらの面積が物品の幾何学的特徴
量として予め定められている物品の面積の整数倍であれ
ば、その直線を分離線と判定し、 隣接する物品を上記分離線により切り離し、それらの物
品を独立させて認識することを特徴とする物品認識方
法。1. An article recognition method for recognizing two or more articles in contact with each other as independent articles, wherein a geometric feature of the article is registered in advance, and an image of the article group is registered by an imaging means. Import, extract the outer contour line of the image of the article, find the inflection point where this outer contour line breaks sharply, and find the area enclosed by the outer contour line on both sides of the straight line connecting the two inflection points If their area is an integral multiple of the area of the article predetermined as the geometric feature of the article, the straight line is determined to be a separation line, adjacent articles are separated by the separation line, and those articles are separated. An article recognition method characterized by recognizing items independently.
物品の画像の面積値から隣接する物品の有無と、その物
品数を求め、 上記隣接する物品数だけ物品の分離処理を行なう請求項
１記載の物品認識方法。2. Prior to the separation process of the adjacent article,
2. The article recognition method according to claim 1, wherein the presence / absence of adjacent articles and the number of articles are obtained from the area value of the article image, and the article is separated by the number of adjacent articles.
線から物品の重心と慣性主軸を求め、この物品の重心と
慣性主軸を用いる幾何学的演算から、物品の位置と向き
を認識する請求項１又は２記載の物品認識方法。3. The method according to claim 1, wherein the center of gravity and the principal axis of inertia of the article are obtained from the outer contour of the article independently recognized, and the position and orientation of the article are recognized from a geometric operation using the center of gravity and the principal axis of inertia of the article. Item 1. The article recognition method according to Item 1 or 2.
ドリング装置に転送する請求項３記載の物品認識方法。4. The article recognition method according to claim 3, wherein said article recognition information is transferred to an article handling apparatus in a subsequent process.
れぞれ独立の物品として認識する物品認識装置におい
て、 物品の幾何学的特徴量を予め登録するデータ記憶部と、 撮像手段により取り込んだ物品群の画像を記録する画像
メモリと、 画像メモリの記録データとデータ記憶部の登録データと
を得て、物品の画像の外部輪郭線を抽出し、この外部輪
郭線が急激に折れる変曲点を求め、 2個の変曲点を結ぶ
直線の両側で、上記外部輪郭線に囲まれる面積を求め、
それらの面積が物品の幾何学的特徴量として予め定めら
れている物品の面積の整数倍であれば、その直線を分離
線と判定し、隣接する物品を上記分離線により切り離
し、それらの物品を独立させて認識する演算部とを有し
てなることを特徴とする物品認識装置。5. An article recognition apparatus for recognizing two or more articles in contact with each other as independent articles, a data storage unit for pre-registering a geometric feature of the article, and an article captured by an imaging unit. An image memory for recording the image of the group, the recording data of the image memory and the registration data of the data storage unit are obtained, and the outer contour of the article image is extracted. The area surrounded by the outer contour line on both sides of the straight line connecting the two inflection points is calculated,
If their area is an integer multiple of the area of the article predetermined as the geometric feature of the article, the straight line is determined as a separation line, adjacent articles are separated by the separation line, and those articles are separated. An article recognition device, comprising: an operation unit for recognizing independently.
処理に先立ち、物品の画像の面積値から隣接する物品の
有無と、その物品数を求め、上記隣接する物品数だけ物
品の分離処理を行なう請求項５記載の物品認識装置。6. The arithmetic unit determines the presence or absence of an adjacent article and the number of the articles from the area value of the image of the article prior to the separation processing of the adjacent article, and separates the articles by the number of the adjacent articles. The article recognition device according to claim 5, wherein
物品の外部輪郭線から物品の重心と慣性主軸を求め、こ
の物品の重心と慣性主軸を用いる幾何学的演算から、物
品の位置と向きを認識する請求項５又は６記載の物品認
識装置。7. The arithmetic unit obtains the center of gravity and the principal axis of inertia of the article from the external contour of the article independently recognized, and calculates the position of the article and the position of the article from a geometric calculation using the center of gravity and the principal axis of inertia of the article. The article recognition device according to claim 5, wherein the direction is recognized.
工程の物品ハンドリング装置に転送する請求項７記載の
物品認識装置。8. The article recognition apparatus according to claim 7, wherein said calculation section transfers said article recognition information to an article handling apparatus in a subsequent process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP9057158A JPH10235584A (en) | 1997-02-26 | 1997-02-26 | Article recognizing method and device therefor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP9057158A JPH10235584A (en) | 1997-02-26 | 1997-02-26 | Article recognizing method and device therefor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPH10235584A true JPH10235584A (en) | 1998-09-08 |
Family
ID=13047771
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP9057158A Withdrawn JPH10235584A (en) | 1997-02-26 | 1997-02-26 | Article recognizing method and device therefor |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPH10235584A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006089235A (en) * | 2004-09-24 | 2006-04-06 | Kao Corp | Article moving device |
| JP2009536731A (en) * | 2006-05-11 | 2009-10-15 | プライム・センス・リミテッド | Modeling of humanoid shape by depth map |
| JP2012158406A (en) * | 2011-01-31 | 2012-08-23 | Shibuya Kogyo Co Ltd | Container treatment apparatus |
| US8565479B2 (en) | 2009-08-13 | 2013-10-22 | Primesense Ltd. | Extraction of skeletons from 3D maps |
| US8582867B2 (en) | 2010-09-16 | 2013-11-12 | Primesense Ltd | Learning-based pose estimation from depth maps |
| US8594425B2 (en) | 2010-05-31 | 2013-11-26 | Primesense Ltd. | Analysis of three-dimensional scenes |
| US8787663B2 (en) | 2010-03-01 | 2014-07-22 | Primesense Ltd. | Tracking body parts by combined color image and depth processing |
| JP2014229153A (en) * | 2013-05-24 | 2014-12-08 | 株式会社ブレイン | Article identification system and program for the same |
| US9002099B2 (en) | 2011-09-11 | 2015-04-07 | Apple Inc. | Learning-based estimation of hand and finger pose |
| US9019267B2 (en) | 2012-10-30 | 2015-04-28 | Apple Inc. | Depth mapping with enhanced resolution |
| US9047507B2 (en) | 2012-05-02 | 2015-06-02 | Apple Inc. | Upper-body skeleton extraction from depth maps |
| US10043279B1 (en) | 2015-12-07 | 2018-08-07 | Apple Inc. | Robust detection and classification of body parts in a depth map |
| US10366278B2 (en) | 2016-09-20 | 2019-07-30 | Apple Inc. | Curvature-based face detector |
| JPWO2021053750A1 (en) * | 2019-09-18 | 2021-03-25 |
-
1997
- 1997-02-26 JP JP9057158A patent/JPH10235584A/en not_active Withdrawn
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006089235A (en) * | 2004-09-24 | 2006-04-06 | Kao Corp | Article moving device |
| JP2009536731A (en) * | 2006-05-11 | 2009-10-15 | プライム・センス・リミテッド | Modeling of humanoid shape by depth map |
| US8249334B2 (en) | 2006-05-11 | 2012-08-21 | Primesense Ltd. | Modeling of humanoid forms from depth maps |
| US8565479B2 (en) | 2009-08-13 | 2013-10-22 | Primesense Ltd. | Extraction of skeletons from 3D maps |
| US8787663B2 (en) | 2010-03-01 | 2014-07-22 | Primesense Ltd. | Tracking body parts by combined color image and depth processing |
| US8594425B2 (en) | 2010-05-31 | 2013-11-26 | Primesense Ltd. | Analysis of three-dimensional scenes |
| US8781217B2 (en) | 2010-05-31 | 2014-07-15 | Primesense Ltd. | Analysis of three-dimensional scenes with a surface model |
| US8824737B2 (en) | 2010-05-31 | 2014-09-02 | Primesense Ltd. | Identifying components of a humanoid form in three-dimensional scenes |
| US8582867B2 (en) | 2010-09-16 | 2013-11-12 | Primesense Ltd | Learning-based pose estimation from depth maps |
| JP2012158406A (en) * | 2011-01-31 | 2012-08-23 | Shibuya Kogyo Co Ltd | Container treatment apparatus |
| US9002099B2 (en) | 2011-09-11 | 2015-04-07 | Apple Inc. | Learning-based estimation of hand and finger pose |
| US9047507B2 (en) | 2012-05-02 | 2015-06-02 | Apple Inc. | Upper-body skeleton extraction from depth maps |
| US9019267B2 (en) | 2012-10-30 | 2015-04-28 | Apple Inc. | Depth mapping with enhanced resolution |
| JP2014229153A (en) * | 2013-05-24 | 2014-12-08 | 株式会社ブレイン | Article identification system and program for the same |
| US10043279B1 (en) | 2015-12-07 | 2018-08-07 | Apple Inc. | Robust detection and classification of body parts in a depth map |
| US10366278B2 (en) | 2016-09-20 | 2019-07-30 | Apple Inc. | Curvature-based face detector |
| JPWO2021053750A1 (en) * | 2019-09-18 | 2021-03-25 | ||
| WO2021053750A1 (en) * | 2019-09-18 | 2021-03-25 | 株式会社Ｆｕｊｉ | Work robot and work system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JPH10235584A (en) | Article recognizing method and device therefor | |
| JP3768174B2 (en) | Work take-out device | |
| JP4039505B2 (en) | Appearance inspection device | |
| JP3464990B2 (en) | Small object inspection equipment | |
| JP4558467B2 (en) | Method and apparatus for aligning the direction of an article | |
| JP2000263482A (en) | Attitude searching method and attitude searching device of work, and work grasping method and work grasping device by robot | |
| JP3223108B2 (en) | Article recognition method and apparatus | |
| JPH09290392A (en) | Collision preventing method and device for article handling device | |
| JP4337418B2 (en) | Chip type electronic component handling apparatus and chip type electronic component handling method | |
| JPS63109923A (en) | Component position correction recognition mechanism | |
| US20030048941A1 (en) | Method for recognizing position of honeycomb structure | |
| JPH06246236A (en) | Parts feeder | |
| JPH0797059A (en) | Object takeout device | |
| JP2611935B2 (en) | Image processing device for parts feeder | |
| JP3336123B2 (en) | Pellet position detection method for pellet bonding equipment | |
| JP3379909B2 (en) | Appearance inspection equipment for minute objects | |
| JPH073990U (en) | Mobile robot | |
| JP4424798B2 (en) | Object recognition method, recognition apparatus, and object selection apparatus | |
| JPH06115669A (en) | Electronic parts aligning device | |
| JP7470926B2 (en) | IMAGE PROCESSING METHOD, PROGRAM, AND IMAGE PROCESSING SYSTEM | |
| JPH08247740A (en) | Visual inspection device of chip component | |
| JPH06246237A (en) | Parts feeder | |
| JPH08201041A (en) | Method for recognizing three-dimensional shape | |
| JPH1158187A (en) | Transfer supply control method of part in production line and device therefor | |
| JPH08130171A (en) | Semiconductor manufacturing apparatus and recognition method of id mark for that |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A300 | Withdrawal of application because of no request for examination |
Free format text: JAPANESE INTERMEDIATE CODE: A300 Effective date: 20040511 |