EP2047403A1 - Procédé de reconnaissance automatisée des objets 3d et détermination de position - Google Patents

Procédé de reconnaissance automatisée des objets 3d et détermination de position

Info

Publication number
EP2047403A1
EP2047403A1 EP07786429A EP07786429A EP2047403A1 EP 2047403 A1 EP2047403 A1 EP 2047403A1 EP 07786429 A EP07786429 A EP 07786429A EP 07786429 A EP07786429 A EP 07786429A EP 2047403 A1 EP2047403 A1 EP 2047403A1
Authority
EP
European Patent Office
Prior art keywords
geometric elements
sub
regular geometric
arrangement
rule
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.)
Ceased
Application number
EP07786429A
Other languages
German (de)
English (en)
Inventor
Martin Stotz
Ira Effenberger
Markus HÜTTEL
Thomas Dunker
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.)
Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV
Original Assignee
Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV
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 Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV filed Critical Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV
Publication of EP2047403A1 publication Critical patent/EP2047403A1/fr
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/64—Three-dimensional objects
    • G06V20/653—Three-dimensional objects by matching three-dimensional models, e.g. conformal mapping of Riemann surfaces
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/24—Aligning, centring, orientation detection or correction of the image
    • G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning

Definitions

  • the invention relates to a method for the automatic detection of three-dimensional objects and a device for the treatment of individual objects which are present in a complex environment.
  • the task is to determine the spatial position of any objects oriented in space, such as workpieces or components.
  • the position of the objects is unknown and their position and orientation in space must be determined, for example, to be able to determine suitable gripping positions for a robot.
  • Included in the problem is always the demand for an extensive automation of object recognition and orientation.
  • An aggravation of the problem arises from other objects in the vicinity of the objects to be gripped.
  • a common case in practice are, for example, workpieces stored in boxes which are present as completely disordered bulk material or in any packing order. Each individual object is surrounded by a multitude of other objects.
  • the object to be recognized must be segmented from the background, i. all points of the point cloud that do not belong to the object must be separated from the points that belong to the object.
  • photogrammetric methods such as "3D Robot Vision” from ISRA VISION SYSTEMS AG are used, although it is necessary for features to be present on the objects that make it possible to use photogrammetric methods
  • photogrammetric methods For example, you might want to use CAD models to generate a large number of images from different views that represent the model, but you need 4000 to 12,000 images be taught, which is associated with a great deal of computing power, time and storage space.
  • the object of the present invention is therefore to provide a method which allows the detection and orientation of objects and a treatment of the objects, without special features must be attached to the objects and without that a detailed CAD model of the recognizable Object must exist.
  • the invention relates to objects characterized by regular geometric elements, e.g. Layers, cylinders or cones describe or at least partially can be described by rule geometric elements and can be identified by this.
  • regular geometric elements e.g. Layers, cylinders or cones describe or at least partially can be described by rule geometric elements and can be identified by this.
  • the scene in which the objects must be recognized is digitized by means of an appropriate detection system.
  • a detection system for example, a 2.5D or 3D data acquisition unit, such as light section, fringe projection or time of flight (ToF) in question.
  • These systems provide depth images or point clouds.
  • 3D data In the following, such data will always be referred to as 3D data.
  • the corresponding evaluation can be done with a computer system.
  • geometric elements are understood to mean those elements whose shape or geometry is closed
  • Formulas is writable, i. for example, by parametric or implicit formulas.
  • 3D object recognition and position determination in space are realized on the basis of an automated recognition of individual rule geometric elements while observing certain geometric conditions, as far as the shape and position parameters of the individual objects are concerned.
  • the method according to the invention serves to manipulate and / or measure, more generally to treat, at least one object present in a limited area in an arrangement that can be described by one or more regular geometric elements that are related to one another, preferably in a specific and / or known relationship is and whose position and position can be described by arrangement parameters.
  • the arrangement parameters may include or be, for example, coordinates, rotation angles, tangency, orthogonality, parallelism, and / or angles to objects.
  • the dimensions of the objects can also be described therein.
  • the treatment may, for example, consist of gripping, moving, and even changing the object itself.
  • the object can also be reshaped.
  • the object is only too measured.
  • the limited area is essentially given by the area of influence of the treatment device. However, it can also consist, for example, in a box in which the objects to be treated, selected or to be measured are accommodated.
  • the object to be treated must first be selected.
  • the selection first means the determination of an object among the several, for which then subsequently data are determined, which describe the shape and the position of the object.
  • the recording of data describing the object and its location is also counted here for the selection of the object.
  • the first step it is necessary to record data that maps the arrangement of the objects.
  • data may be taken from the entire limited area or from a part of this area. They are preferably three-dimensional data describing the objects for each point of the room.
  • the 3D measurement data are recorded using a corresponding 3D sensor. Any sensor that generates dense 3D point clouds can be used here.
  • regular geometric elements can now be fitted into the three-dimensional data, by means of which the object to be selected can be described.
  • individual regular geometric elements are automatically fitted using a best-fit method, thereby providing a first segmentation. taken.
  • this substep one also contains the shape, position, rotation and / or arrangement parameters of the individual rule geometric elements and / or the respective 3D points associated with the element.
  • a regular geometric element can now be selected as the starting element in the regular geometric elements fitted in the first sub-step.
  • the arrangement parameters of this rule-geometric element can be determined automatically or are known. It is therefore possible to select a starting element in the regions of the geometrical single elements which have already been pre-segmented by the first sub-step.
  • the object to be selected is then combined as a whole into all or part of those regular geometric elements which describe the object to be selected and / or of which in the optional first to third
  • the arrangement parameters of the object to be selected are determined by the rule geometrical elements describing the object or the rule geometry elements selected in the third step, taking into account the relationships of the corresponding rule geometric elements in the object to be selected in the three-dimensional Data to be fitted.
  • the object to be found From the object to be found, it may be known or ascertained which rule geometric elements belong to the object and in which spatial relationship they stand with one another. Based on this information, the appropriate conditions for the combined fit can be generated.
  • the fitting of the entire object to be selected can thus be carried out taking into account the geometric constraints which result from the information about the geometric elements belonging to the object and their position relative to one another. In doing so, the object "combined" from the individual geometric elements is fitted as a whole selected object to be selected and achieved a better quality fitting result. As a result, a sufficiently accurate position and LübeStimmung of the workpiece can be achieved.
  • This treatment device may be a gripper, a suction device for aspirating an object or else another device for changing or detecting the object.
  • a device for the examination of objects is also possible.
  • the method can be varied or extended in various ways.
  • a start segment of the 3D measurement data is first set and divided become. This determination can be made, for example, using height information or, if present, pre-segmentation in the depth image. Height information and a depth image can, for example, be derived from the three-dimensional data or recorded with a suitable sensor.
  • the optional first sub-step of the method can be further accelerated by performing a curvature analysis in the 3D data before the best fit. At each point, the curvature is approximated and then subdivided into regions of different curvature as pre-segmentation. Now in the pre-segmented parts a Besteinpassung individual Regelgeometrischer elements performed.
  • individual rule-geometric elements are assigned to the object to be selected by using information about which rule-geometric elements belong to the object and in which spatial relationship they relate to one another.
  • these geometric conditions that is to say the arrangement parameters (for example with respect to the shape and position of the geometric elements present in the object to be found)
  • the selection of the starting element in the optional second sub-step can be supported by the use of additional information, which can result, for example, from preprocessing using depth images.
  • the method according to the invention can be carried out on a system which has a 2.5D or 3D data acquisition unit, such as light section, fringe projection or Time of Flight (ToF), which receives the parts from one or more views and digitally available.
  • the system may include a computer on which the evaluation method is implemented.
  • a system is described with which disordered parts, which consist of several regular geometric elements, can be identified in their position and with which for these gripping points can be determined.
  • Such parts may be, for example, castings, such as scaffold supports.
  • the following evaluation procedure is implemented.
  • the 3D data is divided into different regions and a start region selected.
  • the individual geometric objects are automatically adjusted, for example for a frame holder, cylinders and planes.
  • a found rule geometric element is selected as a starting element, for example the topmost, ie the most accessible one.
  • an environment searches for further geometric objects belonging to the object, that is to say in the case of a framework holder for surrounding cylinders and planes.
  • the type and number of the control geometry near the starting element are initially used as cutting criteria. individual elements.
  • geometric constraints are determined in advance based on the geometry of the object to be detected so that the position of the object can be determined from different views.
  • these geometric conditions may refer, for example, to three planes that must be perpendicular to each other or to cylinders of the same radius and parallel axes that are perpendicular to the ground plane.
  • the determination of the geometric constraints is automatically based on an existing CAD model or by an object is scanned in advance and using this scan data, an automatic determination of the geometric conditions between individual rule geometric elements. It is successively determined which and how many elements are necessary in order to be able to unambiguously determine the position of the object to be recognized.
  • the starting element is selected using depth image information.
  • depth image information e.g. Pre-segmentation with a region-wide method in combination with height information will be helpful.
  • the 3D measurement a curvature analysis. For each point, the curvature is calculated approximately and then areas with similar curvature are summarized and absegmentiert. Only in the pre-segmented data in this way, the automatic Besteinpassung individual Cannabisgeometrischer elements. This speeds up the process and prevents excessive expansion of the individual control geometric elements.
  • the further procedure corresponds to that described in the first example.
  • the application possibilities of the method according to the invention range from gripping disordered objects to 3D scene analysis to metrological applications (for example, shape and position testing), which are summarized here among other applications under the term of treatment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Image Processing (AREA)
  • Processing Or Creating Images (AREA)

Abstract

L'invention concerne un procédé de reconnaissance automatique des objets tridimensionnels, qui peuvent être décrits par des éléments géométriques de réglage, et un dispositif de traitement de ces objets, les objets pouvant se situer dans un environnement complexe. Le procédé se caractérise en ce que les paramètres de disposition de l'objet à reconnaître sont définis de telle sorte qu'au moins une partie des éléments géométriques de réglage décrivant l'objet à reconnaître, en tenant compte des relations des éléments géométriques de réglage dans l'objet, est intégrée en commun dans des données captées par un capteur 3D.
EP07786429A 2006-08-03 2007-07-30 Procédé de reconnaissance automatisée des objets 3d et détermination de position Ceased EP2047403A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102006036346A DE102006036346B4 (de) 2006-08-03 2006-08-03 Verfahren zur automatisierten 3-D-Objekterkennung und Lagebestimmung
PCT/EP2007/006728 WO2008014960A1 (fr) 2006-08-03 2007-07-30 Procédé de reconnaissance automatisée des objets 3d et détermination de position

Publications (1)

Publication Number Publication Date
EP2047403A1 true EP2047403A1 (fr) 2009-04-15

Family

ID=38606438

Family Applications (1)

Application Number Title Priority Date Filing Date
EP07786429A Ceased EP2047403A1 (fr) 2006-08-03 2007-07-30 Procédé de reconnaissance automatisée des objets 3d et détermination de position

Country Status (3)

Country Link
EP (1) EP2047403A1 (fr)
DE (1) DE102006036346B4 (fr)
WO (1) WO2008014960A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2385483A1 (fr) 2010-05-07 2011-11-09 MVTec Software GmbH Reconnaissance et détermination de la pose d'objets en 3D dans des scènes en 3D en utilisant des descripteurs de paires des points et de la transformée généralisée de Hough
EP2720171A1 (fr) 2012-10-12 2014-04-16 MVTec Software GmbH Reconnaissance et détermination de la pose d'objets en 3D dans des scènes multimodales

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102008020579B4 (de) 2008-04-24 2014-07-31 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Verfahren zur automatischen Objektlageerkennung und Bewegung einer Vorrichtung relativ zu einem Objekt
DE102009009569B4 (de) * 2009-02-19 2019-12-19 Daimler Ag Verfahren zum Ermitteln einer Teilfläche eines Bauteils
DE102011100919A1 (de) * 2011-05-09 2012-11-15 Lufthansa Technik Ag Verfahren zur automatisierten Detektion von Einzelteilen einer komplexen differenziellen Struktur
US9233469B2 (en) 2014-02-13 2016-01-12 GM Global Technology Operations LLC Robotic system with 3D box location functionality
DE102014005181A1 (de) 2014-04-03 2015-10-08 Astrium Gmbh Positions- und Lagebestimmung von Objekten
DE102015100983A1 (de) 2015-01-23 2016-07-28 Sick Ag Verfahren zur Lokalisierung von Greifpunkten von Objekten
DE102021210903A1 (de) 2021-09-29 2023-03-30 Robert Bosch Gesellschaft mit beschränkter Haftung Verfahren zum Aufnehmen eines Objekts mittels einer Robotervorrichtung

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5988862A (en) * 1996-04-24 1999-11-23 Cyra Technologies, Inc. Integrated system for quickly and accurately imaging and modeling three dimensional objects

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
FAUGERAS O D ET AL: "THE REPRESENTATION, RECOGNITION, AND LOCATING OF 3-D OBJECTS", INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH, SAGE SCIENCE PRESS, THOUSAND OAKS, US, vol. 5, no. 3, 1 September 1986 (1986-09-01), pages 27 - 52, XP008018683, ISSN: 0278-3649 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2385483A1 (fr) 2010-05-07 2011-11-09 MVTec Software GmbH Reconnaissance et détermination de la pose d'objets en 3D dans des scènes en 3D en utilisant des descripteurs de paires des points et de la transformée généralisée de Hough
EP2720171A1 (fr) 2012-10-12 2014-04-16 MVTec Software GmbH Reconnaissance et détermination de la pose d'objets en 3D dans des scènes multimodales

Also Published As

Publication number Publication date
DE102006036346B4 (de) 2010-12-30
WO2008014960A1 (fr) 2008-02-07
DE102006036346A1 (de) 2008-02-07

Similar Documents

Publication Publication Date Title
WO2008014960A1 (fr) Procédé de reconnaissance automatisée des objets 3d et détermination de position
DE102014102943B4 (de) Robotersystem mit Funktionalität zur Ortsbestimmung einer 3D- Kiste
DE112011103794B4 (de) Aufnehmervorrichtung für Werkstücke
DE102016100101A1 (de) System zum maschinellen sehen und analytisches verfahren zur segmentierung planarer oberflächen
DE19936364A1 (de) Verfahren zur Identifizierung und Lokalisierung von Marken in einem 3D-Volumendatensatz
EP2019283A2 (fr) Procédé et dispositif de mesure des données de mesure réelles d'un composant
DE19633693C1 (de) Verfahren und Vorrichtung zur Erfassung von Targetmustern in einer Textur
DE69829091T2 (de) Messungsgerät
EP2058765A1 (fr) Procédé et dispositif destinés à texturer un objet d'un modèle géométrique tridimensionnel virtuel
EP3794311B1 (fr) Procédé, dispositif et produit-programme informatique pour générer un plan de mesure pour mesurer un objet, procédé et dispositif pour mesurer un objet
DE102021103726A1 (de) Messparameter-Optimierungsverfahren und -Vorrichtung sowie Computersteuerprogramm
WO2009046781A1 (fr) Procédé et dispositif pour enregistrer des informations d'un outil
DE19951146A1 (de) Verfahren zum Reduzieren des Rauschens in einem durch Abbildung erhaltenen Signal
DE102015220031A1 (de) Verfahren zur Konfidenzabschätzung für optisch-visuelle Posenbestimmung
EP1098268A2 (fr) Méthode pour la mésure optique tridimensionelle de surfaces d'objets
DE112015006181B4 (de) Ladungsträgerstrahlvorrichtung, Ausrichtungsverfahren für die Ladungsträgerstrahlvorrichtung, Ausrichtungsprogramm und Speichermedium
DE102006036345B4 (de) Verfahren und Vorrichtung zur Lagebestimmung von Objekten im dreidimensionalen Raum
EP0364614B1 (fr) Méthode de reconnaissance de la position et de l'orientation spatiale d'objets déjà connus
DE102008020579B4 (de) Verfahren zur automatischen Objektlageerkennung und Bewegung einer Vorrichtung relativ zu einem Objekt
DE102022201279B3 (de) Verfahren zum Erfassen einer Umgebung eines Fahrzeugs, Kameravorrichtung und Fahrzeug
DE102006005990A1 (de) Werkstückvermessung für 3-D Lageerkennung in mehreren Multi-Roboter-Stationen
DE102009007024A1 (de) Verfahren und Vorrichtung zum Vereinzeln von Bauteilen
DE102017122627B4 (de) Optisches Messsystem und Messverfahren
EP2345004B1 (fr) Procédé et système de contrôle pour réaliser un contrôle optique du contour d'un objet
EP1988506A2 (fr) Procédé de détermination automatique de la zone de contrôle, procédé de contrôle et système de contrôle

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20090102

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC MT NL PL PT RO SE SI SK TR

AX Request for extension of the european patent

Extension state: AL BA HR MK RS

RIN1 Information on inventor provided before grant (corrected)

Inventor name: DUNKER, THOMAS

Inventor name: HUETTEL, MARKUS

Inventor name: EFFENBERGER, IRA

Inventor name: STOTZ, MARTIN

17Q First examination report despatched

Effective date: 20090921

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED

18R Application refused

Effective date: 20110111