DE19829527A1 - View-based object identification and data base addressing method - Google Patents

View-based object identification and data base addressing method

Info

Publication number
DE19829527A1
DE19829527A1 DE19829527A DE19829527A DE19829527A1 DE 19829527 A1 DE19829527 A1 DE 19829527A1 DE 19829527 A DE19829527 A DE 19829527A DE 19829527 A DE19829527 A DE 19829527A DE 19829527 A1 DE19829527 A1 DE 19829527A1
Authority
DE
Germany
Prior art keywords
objects
views
view
real
characteristic features
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Withdrawn
Application number
DE19829527A
Other languages
German (de)
Inventor
Ingo Dipl Ing Elsen
Dirk Dr Krumbiegel
Karl-Friedrich Dr Kraiss
Peter Dipl Ing Walter
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.)
KRAISS KARL FRIEDRICH PROF DR
Original Assignee
KRAISS KARL FRIEDRICH PROF DR
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 KRAISS KARL FRIEDRICH PROF DR filed Critical KRAISS KARL FRIEDRICH PROF DR
Priority to DE19829527A priority Critical patent/DE19829527A1/en
Publication of DE19829527A1 publication Critical patent/DE19829527A1/en
Withdrawn legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/28Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Library & Information Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Image Analysis (AREA)

Abstract

Automatically produces photorealistic views of large object quantities from existing CAD data, and detects characteristic features of these objects, e.g. edges, texture, color, etc., automatically through image processing. These characteristic features are used for the training of neural networks, which perform a copy of the object view on a data base key.- DETAILED DESCRIPTION - Views of the objects correspond to the actual appearance of the objects, and are produced from CAD data through computer graphic and virtual reality processes, e.g. real-time ray tracing, radiosity, texture mapping, etc. NB No drawing available

Description

Technisches GebietTechnical field

Objekterkennung, Data Mining, Künstliche Intelligenz, Computergrafik.Object recognition, data mining, artificial intelligence, computer graphics.

Stand der TechnikState of the art

Derzeitige Suchverfahren in elektronischen Katalogen basieren auf zusätzlichen Identifikatoren, wie Teilenummern oder Barcodes. Diese sind aber oftmals nicht vorhanden, defekt oder verschmutzt und lassen sich daher nicht zur Identifikation heranziehen.Current searches in electronic catalogs are based on additional identifiers, such as part numbers or barcodes. These are but often not present, defective or dirty and can be therefore do not use it for identification.

Teilenummern und Barcodes sind, wenn verfügbar, an bestimmten, von Objekt zu Objekt variierenden Stellen angebracht. Während der Identifikation müssen diese Stellen gesucht und die Information von Hand (Teilenummer) eingegeben, oder mit einer elektronischen Hilfe (Barcodeleser, Transponder) erfaßt werden.Part numbers and barcodes, if available, are available on certain of Locations that vary from object to object. During the Identification must look for these places and the information by hand (Part number) entered, or with an electronic help (Barcode reader, transponder) can be detected.

Sind diese Informationen nicht vorhanden, bleibt zur Identifikation nur der manuelle Vergleich mit Abbildungen in den Katalogen.If this information is not available, only that remains for identification manual comparison with illustrations in the catalogs.

Aufgabentasks

Die automatische Identifikation der Objekte anhand einer Ansicht des Objektes ohne Zuhilfenahme zusätzlicher objektfremder Identifikatoren wie Barcodes, Transpondern, etc.The automatic identification of the objects based on a view of the Property without the help of additional identifiers such as Barcodes, transponders, etc.

Gewerbliche NutzbarkeitCommercial usability

Gewerblich nutzbar ist die Erfindung (Abb. 2)
The invention can be used commercially ( Fig. 2)

  • - in der KFZ Industrie, wo die Produktkataloge mehrere hunderttausend Teile umfassen, deren Zusammensetzung sich häufig ändert- in the automotive industry, where the product catalogs are several hundred thousand Include parts whose composition changes frequently
  • - im Bereich der industriellen Bearbeitung von Retouren- in the field of industrial processing of returns
  • - bei der Identifikation von Teilen, die einem großen Produktspektrum (z. B. Kleinteile der Metallverarbeitung) angehören.- When identifying parts that have a wide range of products (e.g. small parts from metal processing).
Erreichte VorteileAchieved advantages

Identifikation ohne zusätzliche Information. Entlastung des Bedienpersonals.Identification without additional information. Relief for the operating personnel.

Ausführungexecution Verfahren zur ansichten-basierten Adressierung von DatenbankenProcedure for the view-based addressing of databases

Das System zur ansichten-basierten Abfrage von elektronischen Katalogen setzt sich aus folgenden Komponenten zusammen:The system for the view-based query of electronic catalogs consists of the following components:

In der Trainingsphase (Abb. 1):
In the training phase ( Fig. 1):

  • - Datenbanksystem- database system
  • - Objekterkennungssystem- Object detection system
  • - Echtzeit-Rendering Subsystem- Real time rendering subsystem
  • - Merkmalsextraktion-Subsystem- Feature extraction subsystem
  • - Neuronale Netze Subsystem- Neural networks subsystem
  • - Kataloggenerator- Catalog generator

In der Identifikationsphase (In the identification phase (

Abb.Fig.

2):
2):

  • - Objekterkennungsplatz (Kamera)- object recognition station (camera)
  • - Assistenzsystem- assistance system
  • - Objekterkennung-Subsytem- Object recognition subsystem
  • - Echtzeit Rendering Subsystem- Real time rendering subsystem
  • - Data Mining Subsystem- Data mining subsystem
  • - Datenbanksystem- database system
  • - Präsentationskomponente (Mensch-Maschine Schnittstelle).- Presentation component (human-machine interface).

In der Trainingsphase werden CAD Daten der einzelnen Objekte aus einer Datenbank entnommen und mit Methoden der Computergrafik und der Virtuellen Realität fotorealistisch aus verschiedenen Ansichten dargestellt. Durch Anwendung von bildverarbeitenden Verfahren entstehen hieraus charakteristische Merkmale der einzelnen Objekte mittels derer Neuronale Netze und Expertensysteme trainiert werden. Die Neuronalen Netze erzeugen aus den Ansichten der Objekte die Schlüssel mittels derer die Datenbanken adressierbar sind.In the training phase, CAD data of the individual objects from one Database extracted and using methods of computer graphics and Virtual reality shown in a photo-realistic way from different perspectives. This results from the application of image processing methods characteristic features of the individual objects by means of their neural Networks and expert systems are trained. The neural networks  generate the keys from the views of the objects by means of which the Databases are addressable.

In der Erkennungsphase wird eine Ansicht eines realen Objektes mit einer Kamera aufgenommen und denselben bildverarbeitenden Verfahren unterworfen wie die fotorealistisch erzeugten Bilder der CAD Modelle. Die in der Trainingsphase erzeugten Neuronalen Netze liefern jetzt als Teil des Assistenzsystemes eine Objektidentität die dem Data Mining Subsystem als Schlüssel zur Suche im elektronischen Katalog dient. Das Data Mining System enthält ein zusätzliches Expertensystems, das Mehrdeutigkeiten (Objekte können identische Ansichten haben) auflöst, indem der Benutzer interaktiv zu einer eindeutigen Identifikation geführt wird. Dies kann sowohl durch Darstellung der möglichen Objekte als auch durch gezieltes Eliminieren von Mehrdeutigkeiten erfolgen.In the recognition phase, a view of a real object with a Camera recorded and the same image processing method subject like the photorealistically generated images of the CAD models. The Neural networks generated in the training phase now deliver as part of the Assistance system an object identity that the data mining subsystem as Key used to search the electronic catalog. Data mining System contains an additional expert system, the ambiguities (Objects can have identical views) by the user interactively lead to a clear identification. This can both by displaying the possible objects as well as by targeted Eliminate ambiguities.

Claims (4)

Verfahren zur ansichten-basierten Adressierung von Datenbanken gekennzeichnet durch Method for the view-based addressing of databases characterized by 1. Automatische Erzeugung fotorealistischer Ansichten großer Objektmengen aus vorliegenden CAD Daten.
Mittels Verfahren der Computergrafik und der Virtuellen Realität (echzeitfähiges Raytracing, Radiosity, Texture Mapping, etc.) entstehen aus CAD Daten Ansichten der Objekte, die dem realen Erscheinungsbild der Objekte entsprechen.
1. Automatic generation of photo-realistic views of large quantities of objects from existing CAD data.
Using computer graphics and virtual reality (real-time ray tracing, radiosity, texture mapping, etc.), CAD data is used to create views of the objects that correspond to the real appearance of the objects.
2. Automatische Generierung charakteristischer Merkmale dieser Objekte anhand der erzeugten Ansichten.
Aus den Ansichten lassen sich mittels bildverarbeitender Verfahren Objektcharakteristika (Kanten-, Textur-, Farbmerkmale, etc.) bestimmen. Diese dienen dem Training Neuronaler Netze, welche eine Abbildung der Objektansicht auf einen Datanbankschlüssel (z. B. Produktnummer) durchführen.
2. Automatic generation of characteristic features of these objects based on the generated views.
Object characteristics (edge, texture, color features, etc.) can be determined from the views using image processing methods. These are used to train neural networks, which map the object view to a database key (e.g. product number).
3. Bestimmung der Objektidentität anhand einer Ansicht eines realen Objektes, welches sich im Aufnahmebereich einer Kamera befindet, durch automatischen Abgleich der charakteristischen Merkmale mit den in 1. und 2. erzeugten Datensätzen.
Hierzu kommen Verfahren der künstlichen Intelligenz und des Konnektionismus zum Einsatz.
3. Determination of the object identity on the basis of a view of a real object, which is located in the recording area of a camera, by automatically comparing the characteristic features with the data records generated in 1 and 2.
For this purpose, methods of artificial intelligence and connectionism are used.
DE19829527A 1998-07-02 1998-07-02 View-based object identification and data base addressing method Withdrawn DE19829527A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
DE19829527A DE19829527A1 (en) 1998-07-02 1998-07-02 View-based object identification and data base addressing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
DE19829527A DE19829527A1 (en) 1998-07-02 1998-07-02 View-based object identification and data base addressing method

Publications (1)

Publication Number Publication Date
DE19829527A1 true DE19829527A1 (en) 1999-02-25

Family

ID=7872721

Family Applications (1)

Application Number Title Priority Date Filing Date
DE19829527A Withdrawn DE19829527A1 (en) 1998-07-02 1998-07-02 View-based object identification and data base addressing method

Country Status (1)

Country Link
DE (1) DE19829527A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE10206871A1 (en) * 2002-02-18 2003-09-04 Vidair Ag Method and device for testing a fire detection device
WO2009127271A1 (en) * 2008-04-18 2009-10-22 Robert Bosch Gmbh Traffic object detection system, method for detecting a traffic object, and method for setting up a traffic object detection system
CN105809094A (en) * 2014-12-31 2016-07-27 研祥智能科技股份有限公司 Bar code identification method based on machine vision

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE10206871A1 (en) * 2002-02-18 2003-09-04 Vidair Ag Method and device for testing a fire detection device
WO2009127271A1 (en) * 2008-04-18 2009-10-22 Robert Bosch Gmbh Traffic object detection system, method for detecting a traffic object, and method for setting up a traffic object detection system
CN105809094A (en) * 2014-12-31 2016-07-27 研祥智能科技股份有限公司 Bar code identification method based on machine vision

Similar Documents

Publication Publication Date Title
Campbell et al. Interpreting image databases by region classification
DE69333431T2 (en) Method for recognizing handwritten symbols
EP0780002B1 (en) Process and apparatus for reconstructing raster-shaped line structures
Antoine et al. Shape characterization with the wavelet transform
DE19831413C2 (en) Image processing methods and devices for recognizing objects in traffic
EP0560779B1 (en) Method of detecting and estimating the position in space of objects from a two-dimensional image
DE102017220307B4 (en) Device and method for recognizing traffic signs
DE102015113434A1 (en) Method for object localization and pose estimation for an object of interest
EP2284795A2 (en) Quantitative analysis, visualization and motion correction in dynamic processes
DE112021003955T5 (en) Information processing device and information processing method
DE19829527A1 (en) View-based object identification and data base addressing method
EP0364614B1 (en) Method of recognising the spatial position and orientation of already known objects
EP0829809B1 (en) Method for process-visualization
DE112023000151T5 (en) INSPECTION SUPPORT SYSTEM, INSPECTION SUPPORT PROCEDURE AND INSPECTION SUPPORT PROGRAM
Wang et al. Using spatial relationships as features in object recognition
DE102008057979B4 (en) Learning unit for an object recognition system and object recognition system
EP3754544A1 (en) Detection system, work method and training method
EP0693200B1 (en) Object sorting process
EP1789909B1 (en) Method for assigning a digital image to a class of a classification system
DE102011075335A1 (en) Method for the computer-aided estimation of an object
DE4017430C2 (en) Device and method for recognizing patterns
CN117975310A (en) Unmanned aerial vehicle-based electric power constructor behavior recognition method
Fleischanderl Time Series Transformation into Images for Production State Recognition/submitted by Sabrina Fleischanderl
Arques et al. Applying distance histograms for robust object recognition
WO2022237930A1 (en) Component-classifying apparatus, method for classifying components, and method for training a component-classifying apparatus

Legal Events

Date Code Title Description
OAV Publication of unexamined application with consent of applicant
8122 Nonbinding interest in granting licences declared
8101 Request for examination as to novelty
8110 Request for examination paragraph 44
8105 Search report available
8139 Disposal/non-payment of the annual fee